Heh Coe

nted from ANNAES OF THE Missouri Borsnrcar
¢ 7315-376.” February-Apmil, 1915.

Ate 7 ale 4
¢

feet) porte ah tho Porat
iversary Celebration of the Missouri
Garden, October 16,1914-0 0

REC

ve
y
i
Iv, »
\ \ey
\ I
N oP
$4

~
Cay
0

: S
oD —
|
re ae
Cc) Ii —————EO
Lm

PHYLOGENY AND RELATIONSHIPS IN THE
ASCOMYCETES?

GEO. F. ATKINSON
Cornell University

Part I. ARGUMENT

Perhaps there is no other large group of plants whose origin
and phylogeny have given rise to such diametrically opposed
hypotheses as the fungi. The presence of chlorophyll and the
synthesis of carbohydrates from inorganic materials are such
general and dominant characteristics of plants, that many
students regard them as the fundamental traits which pri-
marily marked the divergence of plant from animal life. Ac-
cording to this hypothesis all plants possess chlorophyll or
were derived from chlorophyll-bearing ancestors.

No one questions the origin of the chlorophylless seed plants
from chlorophyll bearing ones by the loss of chlorophyll and
reduction of photosynthetic organs.2, What is more natural
then, than the hypothesis that the fungi have been derived
from chlorophyll-bearing ancestors? It is not my purpose
to discuss the question as to whether or not the Phycomycetes,
or lower fungi, had an independent origin, or were derived
from one or several different groups of the green algae. I
wish to consider some of the evidence which points to the
origin of the Ascomycetes from fungus ancestry, rather than
from the red algae.

1 The first part of this paper is the abstract or argument as read at the anni-
versary proceedings. Because of the brief character of the abstract which renders
many of the statements more or less categorical, while some therefore will appear
dogmatic, the subject is further elaborated, and illumined by examples in a series
of Notes which follow as an appendix in Part II.

*The chlorophylless seed plants constitute comparatively small, isolated
groups of separate origin from different families or orders of the spermatophytes.
They do not constitute a phylum. The situation is quite different with the
Ascomycetes, which make up a great phylum with ascending and diverging lines,
as well as descending branches. They do not give evidence of many isolated
groups derived by degeneration from many separate families of the red algae.

ANN. Mo. Bot. GARD., VOL. 2, 1915 (315)

[VoL. 2
316 ANNALS OF THE MISSOURI BOTANICAL GARDEN

In this abstract the statements must be more or less cate-
gorical, and some will therefore appear rather dogmatic.

1. The phylogenetic relation of the odblastema filaments of
the red algae, and the ascogenous threads of the sac fungi.—
The nuclear history in the two structures is very different. In
the red algae there is a single fusion of one pair of sex nuclei
in the egg, forming a true diploid nucleus which multiplies by
division in the odblastema filament providing the primary
nucleus for each cystocarp. The odblastema filament fuses
with vegetative auxiliary cells to furnish attachment and base
for food supply of the cystocarp, but the diploid and haploid
nuclei of the fusion cell repel each other. The attempt to
show a phyletic relation between the copulation of short
odblastema filaments with cells of the procarp, or the fusion
of the procarp cells, after the union of haploid gametic nuclei,
in some groups of red algae, and the communication of func-
tional archicarp cells of certain sac fungi, as well as entertain-
ing the notion that fusions of approximate cells of the asco-
genous hyphae are phyletically related to the fusion of oodblas-
tema filaments and auxiliary vegetative cells, introduces
additional confusion into a doctrine already overburdened
with questionable hypotheses. The odblastema filaments and
ascogenous threads are parallel developments. They present
an example of morphological homology or analogy, not of
phylogenetic affinity.

2. The phylogenetic relation of the ascus and carpospore,
or tetrasporangium (see Part 11, Notes m and m1).—There are
two horns to the dilemma here, and either one requires several
additional supporting hypotheses. The origin of the ascus
from a coenocytic zygote, in some cases by reduction, in others
terminating a progressive splitting of the same, is far more
comprehensible. The nuclear fusion in the ascus is not vege-
tative (see Note m). It takes place in all forms thus far in-
vestigated and is to be considered the final stage of the sexual
act, however modified this may be. Were it merely vegetative
fusion there would be no need of conjugate division in the
ascus hook to avoid the union of sister nuclei. The nucleo-
cytoplasmic relation, or balance, would be just as easily at-

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 317

tained by fusion of sister nuclei, or even by contemporaneous
growth of nucleus and cytoplasm, such as is well known to
occur in many other cases, for example in sexual cells, gonoto-
konts, ete.

8 The phylogenetic relation of the ascocarp and cystocarp.
—If this principle of the resemblance between different types
of cystocarp and ascocarp has any force, it would mean that
the sac fungi had as many points of origin from the red algae
as there are points of resemblance between their fruit struc-
tures. JI presume no one at the present time holds any such
view of the polyphyletic origin of the Ascomycetes.

4, The phylogenetic relation of the trichogyne and sexual
apparatus of the Ascomycetes and those of the red algae.—
The sexual apparatus of some of the Ascomycetes, particu-
larly the trichogyne, and the so-called spermatia, is generally
conceded to be the strongest evidence in support of their
phyletic relation to the red algae. This theory, however, re-
quires a jump from the simple trichogyne, a continuous pro-
longation of the egg of the red algae, to the complex, multi-
septate one of the Ascomycetes. It requires further the re-
duction of this trichogyne to a unicellular one, and then to the
simple gamete. It also requires the transition from free an-
theridia, or spermatia, to fixed ones, and from this specialized
condition to the simple gamete, thus finally attaining the gen-
eralized condition of the copulation of simple gametangia.
This appears to me to be a rather strained backward reading
of the evidence.

ORIGIN OF THE ASCOMYCETES FROM FUNGUS ANCESTRY

Although Sachs’ suggestion of the relation of the Ascomy-
cetes to the red algae was received with favor by many stu-
dents at that time, and the doctrine has received a fresh im-
petus in recent years, it was not accepted by some of the
foremost students of the fungi at that time (Winter, ’79;
deBary, ’84). DeBary plead for the application of the
theory of descent which had come to be used as the basis
of classification for the higher plants. As a result of his ex-
tensive studies of development in the Phycomycetes and As-

[Vou, 2
318 ANNALS OF THE MISSOURI BOTANICAL GARDEN

comycetes he was led to the conclusion that the Ascomycetes
were derived from the Phycomycetes. This doctrine is based
chiefly on the evidence of a phyletic relation between the sex-
ual organs of the two groups. In spite of the persistence of
the belief in the origin of the sac fungi from the red algae,
deBary’s doctrine of their descent from the Phycomycetes has
had many adherents. Nowhere in deBary’s writings have I
been able to find any statement which can be construed as
favoring the origin of the sac fungi from the red algae. The
esteem in which his judgment is held, even at the present day,
has led to the republication of a rumor of an ante mortem
statement by deBary to the effect that he was inclined to the
view that the procarps of the two groups pointed to the origin
of the Ascomycetes from the Rhodophyceae!

Our present knowledge of the cytology of the ascus would
not perhaps favor such close contact between the Ascomycetes
and Phycomycetes as would appear from the knowledge pos-
sessed in deBary’s time. Unfortunately we are not yet in
possession of any cytological knowledge of spore production
in the zygote of the Phycomycetes which we can use for com-
parison. But at any rate, the difficulties in this relation are no
greater than are met with in attempting to derive the ascus
from the carpospore or tetrasporangium of the red algae.

Origin of the ascogenous threads—The ascogenous threads
are outgrowths of the zygote or o6gonium and represent one
method of splitting up and proliferation of the same in accord-
ance with recognized principles of progression in the same
direction of increase in the output of spores following the
sexual process, or its equivalent, and terminating the diploid
phase.

One of the most instructive forms suggesting a mode of
transition from the Phycomycetes to the Ascomycetes, is Dipo-
dascus. Its sexual organs are strikingly like those of certain
Mucorales or Peronosporales in their young stages. The
sexual organs, which can be recognized as antheridium and
oogonium, arise either from adjacent cells of the same thread,
or from different threads. After resorption of the wall at
the point of contact, the fertilized odgonium (or zygote) grows

1915]

ATKINSON—PHYLOGENY IN THE ASCOMYCETES 319

out into an elongate stout ‘‘ascus’’ or zygogametangium with
the production of numerous spores. While all phases of the
nuclear phenomena have not yet been made clear, the gametes
are multinucleate, and multiplication either of the sex nuclei,
or of the fusion nucleus,
takes place in the gen-
eralized ‘‘ascus.’’ This
so-called ascus is an
outgrowth of the undif-
ferentiated o6gonium or
ascogonium. The split-
ime up of such a
generalized ascus by fil-
amentous outgrowths,
the ascogenous threads,
which branch and pro-

Fig. 1. A, copulation

Dipodascus albidus:
of gametangia; B, communication established

duce terminal asci con-
taining fewer spores,
would be a very natural
course in progressive
evolution, specialization,
and increase in spore

between antheridium and odgonium; OC, the two
sex nuclei approaching each other; D, fusion

nucleus large, vegetative nuclei small; BE,
growth of generalized ascus from odgonium side
ef copulating gametes, early stages of, in C and
D; F, generalized ascus with numerous spores;
G, spore mass crowded out of end of ascus. a,
antheridium; 0, odgonium.—A-E, after Juel;

F and G, after Lagerheim.
output.

Origin of the ascus in the Endomycetaceae.—The tendency
of generalized forms to split up in different directions, often
giving rise to divergent lines or series, is a well founded prin-
ciple in the doctrine of descent. These series are often of
different character in respect to numbers and diversity of
forms, as well as to progression or reduction in one or more
structures. One of the directions in which descent from such
a generalized, coenocytic, germinating zygote (or ascus) as
represented by Dipodascus has taken place is that of reduction
in size of the generalized ascus and in the number of spores.
Evidence of this reduction is furnished by Dipodascus itself ;
for, as the culture ages the asci become smaller and smaller
and the spores fewer in number. In this way by reduction in
number of spores to 8 and 4, just permitting the meiotic
nuclear divisions, forms like Eremascus and Endomyces have

[Vou 3
320 ANNALS OF THE MISSOURI BOTANICAL GARDEN

arisen. Further reduction of one of the gametes, or of the
vegetative stages, would result in apogamous forms of En-
domyces, the Exoasceae,' the Saccharomycetes, or yeasts, ete.
By reduction and loss of one of the gametes without reduction
in size of the generalized ‘‘ascus,’’ such forms as Ascoidea,
Protomyces, Taphridium, ete., may have arisen.

Origin, progression and sterilization of the so-called tricho-
gyne.—There is no well developed trichogyne-like structure in
any of the known Phycomycetes. But there is evidence in a
few of the forms, like certain species of
Cystopus, of a tendency of the odgonium,
probably under chemotactic stimulation and
a softening of the wall, to develop a short
process directed toward the antheridium.
This has been suggested by a number of stu-
dents (Lotsy, ’07, p. 468) to be an indication
of the origin of the trichogyne in the As-
comycetes. It does not mean that Cystopus?
is to be regarded as an ancestral form of the
Ascomycetes, though certain species do pos-

Ses ee eae sess a number of peculiarities which may be
matie representation attributed to such a hypothetical form. This
of the archicarp of peculiar feature of the odgonium of some
lichens and many . . S
silves Ancoinjecices) eerie of Cystopus is, however, of impor-
The fertile part is tance as it indicates one probable method
ai ae eof origin of the trichogyne in the Ascomy-
the so-called “tricho- cetes. The trichogyne is not a character
Bus 3 “cho possessed by all Ascomycetes, even of those

which still retain two functional gametangia.
This, I believe, is strong evidence of the independent origin of
the trichogyne in the Ascomycetes.

It arose as a copulating process or beak from the odgonium

Archi car

Asc ogoni um

Such an origin for the Hxoasceae is more comprehensible than the theory
that their mycelium may represent ascogenous hyphae which have migrated from
the condition of parasitism in the vegetative portion of a former ascocarp, to
parasitism on their present hosts, as suggested by Harper (’00, p. 392).

One of these features is the generalized character of the sexual organs,
which are polyenergid, but particularly the great variation in number of func-
tional egg nuclei in different species as described by Stevens (’99, ’01).

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 321

under chemotactic stimulation, combined with a transverse
splitting of the odgonium or archicarp.

The failure of the antheridium to perform its function in
the sexual process, its reduction or loss, are well known fea-
tures in the life history of a number of Ascomycetes. In
many cases where the antheridium or its supposed equivalent,
the spermatiun, is to all appearance potentially functional, its
failure to function appears to be due to the sterilization of the
terminal portion of the archicarp.!

Analogous situations are known in the seed plants. I need
only cite the case of Elatostoma acuminatum (see Strasburger,
709). The nucleus of the embryo sac mother cell enters the
preliminary phases of the heterotypic division. After synap-
sis the further stages of the heterotypic division are inhibited,
and by typic or ‘‘vegetative’’ division the eight-nucleated
embryo sac is formed. The egg, therefore, ripens with a dip-
loid nucleus, and, without fertilization, develops the embryo.
The walls of the inner integument grow together at the micro-
pylar end of the ovule and harden, thus forming an effectual
barrier to the entrance of the pollen tube (Treub, 05; Stras-
burger, 09). While great disturbances occur in pollen de-
velopment and most of the pollen grains are empty or un-
developed, some pollen is formed which appears normal. In
some cases the mother cell, which usually forms the diploid
embryo sac, undergoes a true reduction division forming a
row of four cells, the lower one of which forms a normal em-
bryo sac with a haploid egg. The few male plants of this
species, Strasburger thinks, result from fertilization of such

* While the “trichogyne” or terminal portion of the archicarp assumed vegeta-
tive characters in an increasing degree, it seems that it did not in every case
lose all of the features appropriate to a receptive organ. It appears in a few
eases at least to still respond to chemotactic or analogous stimuli, seeking the
fixed spermatia as in Collema pulposum (according to Bachmann, 713) and
Zodiomyces vorticellarius (Thaxter, ’96). In a number of cases there seem to
be receptive areas on the trichogyne where the free sperms become fixed, where
fusion of sperm and trichogyne takes place. The perforation of the transverse
walls of the trichogyne, which is said to occur after fusion with the sperm, also
appears to be another example of the retention of an ancestral character of the
archicarp which primarily permitted the passage of sperm nuclei through the
terminal segment, or the association of nuclei of different segments as partheno-
genesis or apogamy was introduced.

[Vou. 2
322 ANNALS OF THE MISSOURI BOTANICAL GARDEN

haploid eggs by sperms from the normal pollen.

This sterility of the archicarp, I believe, has been brought
about by its assumption more and more of a vegetative char-
acter. The formation of septa at the base of the ‘‘trichogyne”’
in such forms as Pyronema and Monascus, which primarily
may have been the beginning of a transverse splitting of the
oogonium, would make more difficult the fertilization of the
basal portion of the archicarp. In Aspergillus repens the so-
called ‘‘trichogyne,’’ or terminal cell of the archicarp, some-
times gives rise to ascogenous hyphae! (according to Miss
Dale, 09). The basal portion of the two-celled archicarp, or
the basal or central portions of the several-celled archicarp,
seem to be the portions which have retained the function of
ascogenic cells where that function still resides in the archi-
carp. As the archicarp becomes longer, the sterile portion,
which is non-ascogenic, becomes longer and more septate.
This only increases the difficulties of the passage of the sperm
nuclei.

The increasing vegetative character of the terminal portion
of the archicarp has given rise to the long, simple, multisep-
tate ‘‘trichogyne’’ of the lichens and many Pyremomycetes
and Discomycetes, as well as to the profusely branched multi-
septate trichogyne of certain Laboulbemales.? It is an inter-
esting fact that in many of the cases of the extraordinary
vegetative development of the terminal portion of the archi-
carp (the ‘‘trichogyne’’), antheridia and spermatia are en-
tirely wanting.®

The degeneration changes of the sterile portion of the archi-
carp (multiseptate and often also much branched ‘‘tricho-
gyne’’) which are described as taking place after connection of
the spermatium with the receptive terminal cell (for lichens see

1It is worthy of note in this connection that Olive’s studies (’05) of Mon-
ascus led him to regard the “trichogyne,”’ or terminal cell of the archicarp, as
the ascogonium, and the second cell, or ascogonium according to others, as a nurse
cell.

2Thaxter (’96) says that when the spermatia do not become attached to the
receptive cell of the trichogyne the vegetative growth of the trichogyne is greatly
increased.

® (Lachnea cretea, according to Fraser, 13; in Teratomyces actobu, Thaxter,
796, was not able to find antheridia.)

1915]
ATKINSON—-PHYLOGENY IN THE ASCOMYCETES cages

Stahl, ’77, Baur, ’98, Bachmann, ’13; for the Laboulbeniales,
Thaxter, ’96, p. 225), may be classed as secondary or accom-
panying sexual phenomena. It does not necessarily follow
that the sperm nucleus reaches the egg or fertile portion of
the archicarp. The trichogyne changes taking place after
the entrance of the sperm into, or its connection with the re-
ceptive terminal cell, are not dependent on the final fate of the
sperm, i. e., whether it reaches the egg or not. They are ante-
cedent phenomena and in no sense a proof that fertilization
has taken place. These disintegration changes, initiated, it
would seem, by the influence of the sperm on the receptive cell
of the archicarp, terminate the vegetative growth of the archi-
carp and thus the reflex upon the fertile portion at the middle
or base releases the ascogenic cells from the inhibiting influ-
ence of the vegetative phenomena, and they then proceed with
the modified sexual process among the ascogonial nuclei which
may be now associated in sexual pairs, or this pairing be post-
poned to some period in the development of the ascogenous
hyphae.

Origin of spermatia in the Ascomycetes——The presence of
the so-called spermatia in many lichens and other Ascomy-
cetes, associated at the same time in numerous instances with
the trichogyne-like termination of the archicaryp, is one of the
major pieces of evidence brought forward in supporting the
doctrine of the red algal origin of the sac fungi. If we accept
this doctrine, then in the Ascomycetes we must read the his-
tory of the antheridia in the following order: They appeared
first as free structures, spermatia, abjointed from spermatio-
phores, large numbers of which were crowded in highly spe-
cialized receptacles.

At the next step there were few, imbedded, isolated anther-
idiophores to which a few spermatia remained attached, until
finally the stage was reached where spermatium and anther-
idiophore were merged into the simple antheridium. This
doctrine also requires that along with the change from free
spermatia to the simple antheridium, there was a transition
from the condition in which the spermatia do not function to

[Vou, 2
324 ANNALS OF THE MISSOURI BOTANICAL GARDEN

that where the sperm nuclei of the simple antheridium are
functional.

Notwithstanding this interesting course of evolution of the
antheridium and of sexuality which we trace if the red algae
are accepted as the source of the Ascomycetes, I believe, just
as in the case of the archicarp and trichogyne, the evidence
warrants us rather in reading it in just the opposite direction;
and that in the last stages of progressive development of the
sexual apparatus in the Ascomycetes, the resemblances to the
sexual apparatus of the red algae are merely those of mor-
phological homology and analogy, not phylogenetic homology
and affinity.

According to this view, then, the ancestral forms of the
Ascomycetes were fungi with well developed, simple but gen-
eralized gametangia. This condition is retained in a number
of existing Ascomycetes, in many of which true sexuality
exists.1

In connection with the specialization of the antheridium and
the origin of the spermatia of the Ascomycetes, Monascus is
an extremely interesting form. The antheridium is an
elongate terminal cell of a hypha. The archicarp arises as a
branch below the septum. It curves closely against the an-
theridium, bending it over more or less at right angles, and
copulates at any point along the side of the antheridium, there
being no portion of the latter especially selected as a copula-
tion place. The conidia in Monascus are formed in chains by
constriction and septation of terminal portions of hyphae
similar in diameter to the antheridium. The archicarp some-
times copulates with a conidium of the chain before their final
separation (Barker, ’03). A chain of conidia is thus homol-
ogous with the antheridium, and a conidium with any section
of the antheridium. It would be but a step from this condi-

1Examples of generalized, simple (non-septate) gametangia are found in
Dipodascus and Gymnoaseus. Examples of simple specialized gametangia, i. e.,
uninucleate gametangia, are found in the powdery mildews (Hrysiphaceae) and
Eremascus. A second stage is presented in forms where the antheridium remains
simple and generalized, but there is a beginning of specialization in the archicarp
where it is split transversely into two cells, the terminal one (trichogyne) func-
tioning as a copulating organ and migration tube for the sperm nuclei. Examples
are found in Pyronema and Monascus.

1915]

ATKINSON—PHYLOGENY IN THE ASCOMYCETES 325

tion to the copulation of the archicarp with free conidia. The
situation in Collema pulposum (Bachmann, 713), Ascobolus
carbonarius (Dodge,

ilar where the tricho-
gyne copulates with
spermatia (conidia) still
attached to the sperma-
tiophore. ‘These cases
are very strong evidence
suggesting the homology
of conidia (or pycno-
spores as the case may
be) and spermatia! in
the Ascomycetes.
Progression in the di-
rection of multiplication
of antheridia, or sper-
matiophores, and their
association in groups
followed from the sim-

Fig. 3.
sexual organs and fruit.

Monascus, showing development of
an, antheridium; ar,
archicarp; tr, trichogyne; asc, ascogonium;
con, conidium with which trichogyne is cop-
ulating; A.h, ascus hooks or croziers; B, young

ple and more or less
isolated situation, pro-
gressing along the same
course which is recog-

fruit showing ascogenous hyphae within, at left
is a very young fruit body showing ascogonium
becoming surrounded by the enveloping fila-
ments; ©, mature fruit body with asci and
ascospores.—Upper row of figures after Barker;
lower group after Schikorra.

nized in the association
and massing of conidiophores into bundles, cushions, or
pycnidia. It is the same course which is universally recog-
nized as a striking indication of progression in other groups
of plants, a cephalization of fruiting or reproductive struc-
tures, as in the bryophytes, lycopods, conifers, and angio-
sperms. In the latter it has given us the flower, and further
cephalization of the flower has resulted in the head of the com-
1 Their function in the ancestral or early forms may have been generalized
enough to permit of their performing as conidia or sperms, as in the case of
Ectocarpus, Prostosiphon, Ulothria, ete. Strasburger (’05, p. 25) has expressed
the idea that the pycnospores of the Ascomycetes might have been spermatia,

and that the process of fructification now presented by these fungi is a secondary
adaptation in place of the erstwhile fertilization by spermatia.

[Vou. 2
326 ANNALS OF THE MISSOURI BOTANICAL GARDEN

posites, the highest stage of phyletic evolution in the plant
world.

In conclusion, the Ascomycetes present a very rich variety
of form, structure, and adaptation with very marked diverging
series. Some of these series present evidences of progres-
sion from simple, generalized forms to highly specialized
forms, while others indicate descent by reduction. The evi-
dences of progression are of the same kind and value as are
generally yen in other groups of plants.

: Sachs, in his later writings,
agreed with deBary in recog-
nizing the Ascomycetes as a
distinct phylum, with an as-
cending series from simple
and generalized forms to com-
plex and specialized ones. He

Fig. 4. Gymnoascus Reessii: A-D, never mentioned the tricho-
formation of sexual organs, fusing at C; qd -
E, sexual organs in uninucleate condi- Syne as evidence of their phy-

tion; F, fusing sexual organs in multi- letic relation to the red algae.
nucleate stage.—After Dale. But his theory was based on
the presence of a procarp whether with or without a tricho- .
gyne. He selected Gymnoascus, where the sexual apparatus
consists of simple copulating gametangia, as the simplest
ascomycete known at that time. It is only in recent years
that the trichogyne has been seized upon as evidence of the
phyletic relation of the two groups and has forced this
anomalous backward reading of the history.

Part II. Exuciation
NOTE I

The red algae are remarkable for the great constancy in the
form of the procarp (procarpic branch, carpogonial branch,
etc.) and the very great divergence in the processes subse-
quent to the fertilization of the egg (terminal cell of the pro-
carp, carpogonium) and ending in the production of the carpo-
spores. The general character of this divergence may be
shown by a brief presentation of several types, as follows:

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES S21

1. The simplest type of cystocarp development occurs in
the Nemalionales where the carpogonium, or egg cell, after
fertilization, gives rise to several branched sporogenous
threads in a compact cluster, bearing terminally the carpo-
spores (Nemalion, Lem-
anea, etc.), or in some
species the sporogenous
threads are more widely
extended in the thallus,
the branches producing

separated clusters of
D Fig. 5. A and B, Lemanea; C, Batrachosper-
carpospores ( ermo- mum: epbr, procarp or carpogonial branch;

nema dichotomum, see cpg, carpogonium or egg; tr, trichogyne; sp,
Schmitz and Haupt- °4"tnd B after Atkinon; ©, after Davis.
fleisch, °97). Fertiliza-

tion by the fusion of a sperm nucleus with the egg nucleus
after entrance into the trichogyne and migration down into
the carpogonium has been described in Nemalion (Wolfe, ’04)
and in Batrachospermum (Schmidle, ’99; Osterhout, ’00).

2. In Polysiphonia (Rhodomeniales) the procarp branch
of four cells is curved around so that the carpogonium is in
contact with an auxiliary cell lying between the carpogonium
and the pericentral cell which gave rise to the procarp. After
fusion of the sperm and egg nucleus in the carpogonium, the
fusion nucleus divides once. The carpogonium now connects
with the auxiliary cell mentioned, which fuses with the peri-
central cell. The two diploid nuclei migrate into the peri-
central cell, the carpogonium separates from the auxiliary cell,
while it and the remaining cells of the procarp degenerate. The
pericentral cell now fuses with several other auxiliary cells,
which arose from it as a branch, forming the central cell. The
diploid nuclei remain in the upper part of the central cell,
while the haploid nuclei from the auxiliary cells, some having
divided, now degenerate (Yamanouchi, ’06).

3. A somewhat different situation exists in FHrythro-
phyllum delesseroides (Gigartinales). The odblastema fila-
ment from the fertilized egg connects with the auxiliary cell
which is the basal cell of the seven or eight-celled pro-

[Vou. 2
328 ANNALS OF THE MISSOURI BOTANICAL GARDEN

carp. This in turn fuses with the two other large cells of the
basal portion of the procarp, thus forming the large fusion
cell from which the
gonimoblasts, or sporog-
enous threads arise
(Twiss, 711).

4. In Harveyella mi-
rabilis,, a large cell
which gives rise to the
four-celled procarp is
the auxiliary cell. A
short odblastema fila-

Fig. 6. A, Harveyella mirabilis; B and CO, ment from the egg con-

Erythrophyllum delesseroides; D, H, F, and G, nects with the latter,
Callithamnion corymbosum: cpbr, carpogonial :
branch; cpg, carpogonium; tr, trichogyne; of, which becomes the con

odblastema filament; ac, auxiliary cell; g, tral cell.
gonimoblast ; fe, fusion cell. 1, 2, and 3 are the T : Ss
three large basal cells of the procarp in Ery- D. n Callithamnion

throphyllum which fuse with the odblastema Ceramial :
filament to form the fusion cell. Shaded por- ( a es) the fusion

tions are diploid; note that in the fusion cell (diploid) nucleus in the

of Callithamnion the vegetative nucleus (hap- ral 7
loid) remains at a distamee from the diplona °88 divides into two.

nucleus.—A, after Sturch; B and OC, after Two short odblastema
Twiss; D, #, F, and G, after Oltmanns. filaments proceed from
the carpogonium, each containing a diploid nucleus, and fusing
with an auxiliary cell at the side of the base of the procarp.
Each of the two auxiliary cells now contains two nuclei. A
wall divides each cell into two. The upper daughter cell con-
tains the diploid nucleus and becomes the central cell, giving
rise to the sporogenous threads, while the haploid nucleus in
the lower cell degenerates (Oltmanns, ’04).

6. The most complicated type may be represented by
Dudresnaya purpurifera (Cryptonemiales) where several
odblastema filaments arise from the sterilized egg cell. These
fuse with auxiliary cells which are either certain cells of the
procarp branch, or terminal cells of its branched system, or
of more distant ‘‘secondary procarp branches.’’ An odbla-

1H. mirabilis is parasitic on certain species of Polysiphonia, and is devoid
of chlorophyll. For this reason it is regarded by some as indicating a step in the
direction of an ascomycete.

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 329

stema filament after fusing with one auxiliary cell may grow
forward and fuse with another and soon. The diploid nucleus
formed in the egg multiplies by division in the odblastema
filaments. In the fusion cell, resulting from the union of the
filament and auxiliary cell, the diploid and haploid nuclei repel
each other so that the former lies on the filament side while
the latter lies in the base of the auxiliary cell. An outgrowth

Fig. 7. Dudresnaya purpurifera: A, odblastema filaments fusing with
auxiliary cells; B, C and D, outgrowth from the fusion cell to form the
central cell; C, diploid nucleus dividing; D, central cell of cystocarp separated
by a wall. Note that the nucleus of the auxiliary cell remains distant from
the diploid nucleus of the odblastema filament. Shaded portions are diploid.
epbr, carpogenic branch; cpg, carpogonium; tr, trichogyne; of, odblastema
filament; fc, fusion cell; ac, auxiliary cell; an, auxiliary cell nucleus; cy,
central cell of cystocarp.—After Oltmanns.

arises from the odblastema filament at the point where the
diploid nucleus lies. The latter divides, one nucleus migrat-
ing into the outgrowth, while a wall separates it from the
fusion cell. This new cell with its diploid nucleus becomes
the central cell (Oltmanns, ’04).

7. In Cruoriopsis cruciata the situation is similar. The
odblastema filament by coursing widely through the thallus,
fuses with the terminal cell (auxiliary cell) of ‘‘secondary
procarp branches.’’ Each of these fusion cells, or auxiliary
cells, then gives rise to one or two simple rows of 2-4 spores
(Schmitz, ’79, ’83), or a single 24-celled spore chain (Olt-

manns, 704).

[VoL, 2
330 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Relation between the fusions of procarp and auailiary cells,
and those of archicarp cells —Several persons have made the
interesting suggestion that certain similarities between the
events which take place in the fusion of one or more of the
middle or basal cells of the procarp with an outgrowth from
the carpogonium, either direct, or through the medium of an
auxiliary cell, as represented in Erythrophyllum, Harveyella,
Callithamnion, ete. (third, fourth and fifth types mentioned
above), and those occurring in the fusion among themselves of
the middle or basal cells of the archicarp prior to the forma-
tion of the ascogenous threads, may be evidence of a phylo-
genetic relationship between the red algae and Ascomycetes.
Thus Baur (’98) suggests that the first fertile cell of the sev-
eral-celled ascogone of Collema crispum may be the egg cell,
that this may be fertilized by the entrance of the sperm nu-
cleus and its fusion with the egg nucleus. This fusion nucleus
may now divide. The other cells of the ascogone below the
egg are conceived of as auxiliary cells into each one of which
a nucleus resulting from the division of the fertilized egg
nucleus migrates after pore formation in the intervening
walls.

In an interesting paper on the morphological relationships
of the Florideae and Ascomycetes, Dodge (’14) emphasizes
this theory by pointing to a number of cases in the lichens
and other Ascomycetes where fusion, or pore connections, are
known to occur between the ascogenous cells of the archicarp
where more than one cell gives rise to ascogenous hyphae.
Examples among the lichens are Collema crispum (Baur, ’98),
Physcia pulverulenta (Darbishire, ’00), Anaptychia ciliaris
(Baur, ’04), and Collema pulposum (Bachmann, 713), while
among the other Ascomycetes may be mentioned the follow-
ing: Ascobolus (Harper, ’96. Here there is but one asco-
genous cell which gives rise to the ascogenous hyphae, but pore
formation in intervening walls permits intercommunication
between several adjacent cells in the middle of the archicarp.
The species is not given), Ascophanus carneus (Cutting, ’09),
Lachnea cretea (Fraser, 713), Polystigma rubrum (Nienburg,
714).

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 331

Now as to the suggested relationship between the phenom-
enon of broad or narrow pore formation in the walls of
certain cells near the middle or base of the archicarp in
certain lichens and other Ascomycetes, and that shown in
the communications taking place between the carpogonium
and auxiliary cells (often including one or more of the
other procarp cells), it may be said (1) that in the red algae
this communication of the carpogonium (terminal procarp
cell) with other procarp cells when it does take place is not
direct, but by a roundabout method, either through a distinct
outgrowth from the carpogonium, or through the medium of
one or more auxiliary cells, or by a combination of both, to
form the central cell; (2) no evidence of any similar round-
about method has been observed in the archicarp of the sac
fungi. The intercommunication between the middle or basal
cells of the archicarp is always direct, and no communication
in the multicellular archicarp occurs by means of which either
a fertilized nucleus, or a sperm nucleus has been observed to
migrate from the terminal cell to the middle or basal cells;
(3) that in a number of the fungi where pore formation occurs
between cells of the fertile portion of the archicarp, the
‘‘trichogyne’’ is either absent, or admittedly degenerate, or the
antheridium is absent. Examples are: Ascobolus, studied by
Harper (’96), antheridium and trichogyne absent; Asco-
phanus carneus, antheridium absent, trichogyne doubtful or
degenerate; Lachnea cretea, no antheridium observed, tri-
chogyne not functional; Polystigma rubrum,! trichogyne not
functional, from a multicellular cell at base of archicarp one
nucleus migrates into the adjacent uninucleate archicarp cell,
which is regarded as the ascogonium (Nienburg, 714). In
none of the lichens has a sperm or other nucleus been observed
to move down into the fertile part of the archicarp. Pore
formation in the archicarp of the Ascomycetes has no phyletic
relation to the fusions of auxiliary cells among themselves or
with a short odblastema thread or the egg cell. It occurs in-

Blackman and Welsford (712), who earlier investigated the cytology of
Polystigma rubrum, are of the opinion that the “spermatia” as well as the archi-
carps degenerate, and that certain vegetative cells become transformed into as-
cogones.

[VOL. 2
332 ANNALS OF THE MISSOURI BOTANICAL GARDEN

dependently in different groups of the fungi as a means of
permitting the association of nuclei, often in conjunction with
the association of sex nuclei or their equivalent modified sex
nuclei (see the situation in Basidiobolus, Hidam, ’86; Raci-
borski, ’96; Fairchild, ’97; Olive, 07; Woycicki, ’04).

Relation of odblastema filaments and ascogenous hyphae.—
In the Ascomycetes the processes in the growth of the zygote
or ascogenic cell present to a certain extent a somewhat analo-
gous course of progression to that of the carpogenic cell of the
red algae. In the less complicated process, as shown in the
Laboulbemiales, the carpogenic cell may undergo a few divi-
sions, the subterminal cell of the series forming the as-
cogonium. The ascogonium then usually divides to form two
or four ascogenic cells, or without division forms the single
ascogenic cell (Thaxter, 96; Faull, 712). The ascogenic cells
give rise directly, by budding, to the asci. They are, there-
fore, somewhat comparable or analogous to the gonimoblasts
of the red algae. In Sphaerotheca (Harper, ’95*, p. 475)
there is a single short ascogenous thread of a few cells (arising
from the one-celled o6gonium or ascogonium) forming a single
ascus from the subterminal cell. Where the process is more
complex, as in Pyronema (Harper, ’00; Claussen, 712), several
long ascogenous hyphae arise from the large single-celled
zygote or ascogonium, giving rise ultimately to numerous
terminal asci. In other forms the ascogonium is several-
celled, a number of the cells developing ascogenous hyphae
(Collema, Stahl, ’77; Baur, 98; Bachmann, 712, 713; Anap-
tychia ciliaris, Baur, ’04; Physcia pulverulenta, Darbishire,
700; Ascophanus carneus, Cutting, ’09; Lachnea cretea,
Fraser, ’13; etc.).

Some of the chief objections in the way of accepting the
theory of a phylogenetic relation between the odblastema fila-
ments of the red algae and the ascogenous threads of the sae
fungi are as follows:

1. The fusion of a free sperm and the egg nucleus in the
single uninucleate odgonium or carpogenic cell. So far as we
know this is universal in the red algae. In the Ascomycetes
the odgonium is usually multinucleate or multiseptate. In no

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 333

cease has fertilization by a free sperm been determined, and in
forms with a multiseptate ‘‘trichogyne,’’ or odgonium, the so-
called spermatia, or antheridia, do not, so far as we know, play
the usual réle in fertilization, not even a modified role by asso-
ciation with the odgonial nuclei.

2. The individual nuclei of the odblastema filaments are of
the usual diploid character, and there is no fusion of these
nuclei prior to the formation of the carpospores. The indi-
vidual nuclei of the ascogenous threads, or ascogenie cells, are
probably haploid in character, and sooner or later form the
so-called synkarion, an association of two nuclei, together
equivalent to a diploid nucleus. Fusion of the paired nuclei
takes place before the formation of the ascospores.

3. It has been suggested that the complex processes in the
extensive migration, branching and fusions of the odblastema
filaments with auxiliary cells as is known to occur in the
Cryptonemiales (as in Dudresnaya, Cruoriopsis, Gloeosi-
phonia, etce.), may furnish still more important evidence of
the ancestry of the Ascomycetes than that suggested in the
fusions of procarp and auxiliary cells on the one hand, and
archicarp cells on the other (Dodge, ’14). The fusions of the
odblastema filaments with auxiliary cells and the production
of sporogenous threads from the central cells thus formed,
are supposed to be represented by the fusions which are known
to occur between the ultimate and antepenult cells of the ascus
hook prior to the formation of additional asci. The processes
in both groups result in the multiplication of spore origins
and consequently in an increase in spore output. Perhaps the
nearest analogue to the process in the Ascomycetes which re-
sults in the formation of the ascus with its four to eight spores,
is found in Cruoriopsis, where one or two spore chains of two
to four spores each are produced as a result (Schmitz, ’79, ’83;
Oltmanns, ’04). The theory of ‘‘second sexual fusions’’ in
the red algae was founded on the discovery of these fusions of
the odblastema filaments with auxiliary cells, since it was sup-
posed that a fusion occurred between the nucleus of the
odblastema filament (derived from the diploid nucleus of the
fertilized egg) and the nucleus of the vegetative auxiliary cell

[VoL, 2
334 ANNALS OF THE MISSOURI BOTANICAL GARDEN

(Schmitz, ’83). Recent cytological work on the red algae has
not confirmed this theory, but, on the other hand, has discred-
ited it, since in the cases examined the diploid nucleus of the
odblastema filament and the haploid nucleus of the auxiliary
cell are said to repel each other and no fusion between them
occurs. It should be emphasized that the fusion of the odbla-
stema filament and the auxiliary cell is a fusion of a diploid
structure with a haploid one, that it is probably of a nutritive,
or parasitic, nature comparable to the fusion of the moss
sporogonium with the tissue of the gametophyte, a physiolog-
ical, nutritive requirement in the absence of other means of
nourishing the moss sporogonium. The fusions occurring be-
tween cells of the same ascogenous hypha are fusions between
cells of the same phase and serve to bring into association
nuclei of more or less remote ancestry, but each endowed with
the same number of chromosomes (probably the 1x number).

Thus, while there are somewhat analogous variations in the
splitting up of the ascogonium in the sac fungi, and of the car-
pogonium in the red algae, with progress in the direction of
increasing the output of spores, it seems fair to conclude, that,
so far as the evidence at present in hand is concerned, the rela-
tion between the fusion of odblastema filaments and auxiliary
cells in the red algae, and those between the ultimate and an-
tepenult cells of the ascus hook (of the ascogenous hyphae),
however interesting it may be, has no phylogenetic signifi-
cance, and is at best a rather strained parallel.

Ascogenous hyphae, gonimoblasts, ooblastema filaments, the
several fertile cells of certain ascogonia which communicate
by resorption of the intervening septa, the fused procarp, may
be considered as morphological equivalents, as suggested by
Dodge (14, p. 174), but there is no evidence of a phyletic rela-
tion between the ascogenous hyphae and fusing ascogonial
cells, and their morphological equivalents in the red algae.
They illustrate different modes of increase of spore output by
splitting up of the odgonium.

NOTE II

The fundamental difference in the method of development of

ascospores and carpospores is one of the great barriers in the

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 335

way of the descent of the Ascomycetes from the Florideae.
Some (Bessey, E., 713, p. 151) have attempted to overcome this
difficulty by suggesting the homology of the ascus and tetra-
sporangium. But this effort leads to so many suppositions
and supporting hypotheses because of the fundamental dif-
ference between the process of spore formation in the ascus,
and the processes of carpospore or tetraspore formation, that
the descent of the ascus fungi from the red algae would re-
quire a far more labyrinthian course than would be necessary
in deriving them from the Phycomycetes.

NOTE III
IS NUCLEAR FUSION IN THE ASCUS OF A VEGETATIVE OR SEXUAL
NATURE?

It is unfortunate that there is such great divergence of
opinion in the interpretation of the nuclear phenomena in the
archicarp and ascogenous threads. These conflicting results
are probably, in a large measure, due to the difficulties pre-
sented in the minute size of the nuclei. The divergence of
opinion relates primarily to the question as to whether the
fusion nucleus of the ascus is the result of two successive
nuclear fusions, the first taking place in the ascogonium and
the second in the ascus, or whether the nuclear fusion in the
ascus is the only one.

The principle of a single nuclear fusion, that in the ascus,
interprets this act as the final stage in the process of fertiliza-
tion, by the fusion of two nuclei of more or less remote an-
cestry. At some time prior to ascus formation these two
nuclei may possibly become associated in pairs into a syn-
karion and multiply in the ascogenous threads by conjugate
division, or the synkarion and conjugate division may be post-
poned to the ascus hook and the complicated series of fusions
between the ultimate and antepenult cells of the crozier, or
proliferations of the young ascus with accompanying con-
jugate divisions of the synkarion.

Dangeard (’94) first described the presence of two nuclei in
the young ascus, and their fusion, in several species (Borrera
ciliaris, Peziza vesiculosa, Helvella ephippium, Geoglossum

(Vou. 2
336 ANNALS OF THE MISSOURI BOTANICAL GARDEN

hirsutum, Acetabula calyx, Exoascus deformans, and some
lichens). The origin of the ascus was correctly described in
a number of cases, but in the majority of cases at that time he
thought the young ascus arose by the copulation of two unicel-
lular gametes according to a method similar to the formation

|
m7

{i
"}

Fig. 8. Pyronema confluens: A, section of mature discocarp; B, group
of archicarps copulating with antheridia by means of the slender prolonga-
tion (trichogyne) of the ascogonium which is separated as a distinct cell;
O, pair of sexual organs copulating by means of the trichogyne cell, asco-
gonium at left, antheridium at right; D, showing multinucleate condition of
sexual organs and communication of antheridium and trichogyne. a,
antheridium; 0, trichogyne; c, ascogonium. JZ, older stage of a similar
group of sexual organs after the antheridial nuclei have entered the asco-
gonium and the trichogyne nuclei have degenerated; also showing early stage
of growth of ascogenous hyphae from the ascogonium; F’, showing relation
of ascogonia, ascogenous hyphae, asci, and paraphyses in mature fruit
body.—After Harper.

1915]
ATKINSON—-PHYLOGENY IN THE ASCOMYCETES oor

of the ascus in Hremascus, so that the ascus appeared to be
supported on two stalks. Frequently, however, in Peziza
vesiculosa and Helvella ephippium he observed the origin of
the ascus from a single hypha curving at the end in the form
of a hook or crozier. The four nuclei resulting from the divi-
sion of two were so situated in the crozier that after the forma-
tion of two cross walls the ultimate and antepenult cells each
contained one nucleus, while the penult cell contained two
nuclei. The association of two nuclei in the young ascus and
their fusion he interpreted as a sexual act, and the young ascus
was looked upon as an odgonium. Later, Dangeard found that
the crozier method of ascus formation was the usual one in the
forms studied and that in no case in these higher forms did
the ascus arise immediately from the conjugation of two dif-
ferent hyphae.

This important pioneer work by Dangeard was a great
stimulus to further studies which has led to a more or less
clear knowledge of the history of the nuclei from the archicarp
through the ascogenous hyphae to the ascus, while the origin
of the ascogenous hyphae from the fertile cells of the archi-
carp was first described by Janczewski (’71) in Ascobolus,
and later by Kihlman (’83) in Pyronema confluens. Harper
first demonstrated the origin of the ancestral ascus nuclei in
the archicarp of Sphaerotheca castagnet (’95°) and Pyronema
confluens (’00) and their migration in the ascogenous hyphae,
though he does not give the nuclear history in the ascogenous
hyphae, except the later stages at the time of formation of the
ascus. Their archicarp origin has been abundantly confirmed
by several investigators in a number of different forms, both
among the lichens and other Ascomycetes.

The different opinions in regard to the significance of
nuclear fusion in the ascus rest upon the interpretation by
different investigators of the behavior of the nuclei in the
archicarp, or ascogenous cells, before they begin to move into
the ascogenous hyphae. Some maintain that there is a fusion,
in pairs, of the sex nuclei (1) in the archicarp when fertilized

[Vou, 2
338 ANNALS OF THE MISSOURI BOTANICAL GARDEN

by an antheridium (Harper in Sphaerotheca castagnei,' 795°,
96; Erysiphe, 96; Pyronema confluens, ’?00; Phyllactima, ’05;
Blackman and Fraser in Sphaerotheca, ’?05; Claussen in Bou-
diera [=Ascodesmis], ’05); or (2) in the archicarp where the
antheridium is functionless or absent (Blackman and Fraser,
06, in Humaria granulata; Fraser, ’07, in Lachnea stercorea).
In Aspergillus herbariorum Miss Fraser (07, p. 420) finds that
the antheridium often degenerates and did not observe disap-
pearance of the intervening wall when fusion with the tricho-
gyne took place. She nowhere describes or figures fusion in
pairs of the ascogonial nuclei. She merely assumes it, for, in
the summary (’07, p. 428) she says: ‘‘It seems probable that
normal fertilization occurs in some cases, and that in others it
is replaced by a fusion of ascogonial nuclei in pairs’’; Wels-
ford (’07) in Ascobolus furfuraceus; Dale (’09) in Aspergillus
repens; Cutting (’09) in Ascophanus carneus believe in the
fusion of archicarp nuclei in pairs; or (3) of nuclei in vegeta-
tive cells where the archicarp is wanting or functionless
(Fraser, ’07, ’08, in Humaria rutians, fusion of the nuclei said
to take place soon after entering the ascogenous hyphae; Car-
ruthers, 711, in Helvella crispa; Blackman and Welsford, 712,
merely found evidence of nuclear fusion in vegetative cells of
Polystigma rubrum).

1Dangeard (’97) claims that the antheridium is functionless and that the
single nucleus in the odgonium divides into two. After his study of Pyronema
Claussen (712) is inclined to question the fusion of the two sex nuclei in the
oégonium of Sphaerotheca, Erysiphe, Phyllactinia, and Pyronema as described
by Harper (95a, 796, ’00, 705), and by Blackman and Fraser (’05) in Sphaerotheca
as well as in the case of Boudiera (= Ascodesmis) studied by him in 1905. In
respect to his work on Boudiera he now says: “My own statements upon the nuclear
fusion in the ascogone of Boudiera (Ascodesmis) are clearly wrong.” He points
out that in none of these cases is the history of the nuclei in the ascogenous hyphae
known, and thinks that a reinvestigation will show paired nuclei here. A question
to be considered, says Strasburger (’05, p. 24), is whether the chromosomes of the
nuclei united in the odgonium do not remain in separated groups in the ascogenous
hyphae, in order to fuse as individual nuclei in the ascus. Lotsy (’07) has ex-
pressed a somewhat similar view in an attempt to harmonize the situation in the
Ascomycetes and Basidiomycetes. The fusion nucleus in the odgonium remains
for a time a 2x nucleus but some time prior to ascus formation the 2x nucleus
separates into two individual 1x nuclei in the ascogenous hypha, forming a syn-

karion. Conjugate division now takes place with ascus formation occurring
immediately or after several successive conjugate divisions.

1915]

ATKINSON—PHYLOGENY IN THE ASCOMYCETES 339

Others maintain with equal assurance that there is no fusion
of the sexual nuclei in the archicarp. There is merely an asso-
ciation of sex nuclei.

(1). In forms with a func
tional antheridium and archicarp
may be mentioned Monascus1
(Schikorra, ’09) and Pyro-
nema confluens (Claussen, ’07,
712).

(2). In forms where the an-
theridium is absent or function-
less may be mentioned Pyronema
confluens (Brown, W. H., ’09, an-
theridium functionless), Lachnea
scutellata (Brown, W. H., 711,
antheridium absent). In both of
these examples, cases of division
of the nuclei in the ascogonium
were observed which might be
mistaken for fusion. Since no

Fig. 9. Pyronema confluens: A,
B, and C, conjugate division of
nuclear pairs in the ascogenous
hyphae; D, conjugate division in
ascus hook; H, tips of branched

ascogenous hyphae with ascus
hooks, young asci, and beginning of

divisions of nuclei in the as-
cogonium have been described by
authors in the forms where they
believe sexual fusions of nuclei
to take place, W. H. Brown (711)
suggests that they may have had
before them division stages. In

conjugation of the ultimate and
antepenult cells of the ascus hooks;
F, completed conjugation of the
ultimate and antepenult cells of
the hook and association of their
nuclei as a pair. Ascog, ascogo-
nium; asc. h, ascogenous hyphae
with paired “sexual” nuclei.—After
Claussen.

Ascophanus carneus and Ascobolus immersus the anther-
idium is absent, but association of the nuclei in several of the
multinucleate ascogonial cells occurs after pore formation in
the walls. Most of these nuclei become paired and remain
paired as they migrate in the ascogenous hyphae to the ascus
hooks, where conjugate division takes place. The only fusion
of nuclei is that in the ascus, except in badly fixed prepara-
tions or in degenerating nuclei in the ascogonium (Ramlow,
14). In Leotia (Brown, W. H., 710) the ascogenous hyphae

1 Barker (’03) ascribed his failure to find a fusion of nuclei in the ascogonium
of Monascus to the absence of proper stages in his material,

[Vou 2
340 ANNALS OF THE MISSOURI BOTANICAL GARDEN

are supposed to arise from an ascogonium in the base of the
ascocarp, but the nuclei are believed to arise from a haploid
nucleus. Conjugate division occurs in the ascus hooks, the
majority of which are formed by proliferation of the binu-
cleate penult cell and from fusions of the ultimate and ante-
penult cells of croziers, so that many conjugate divisions of
the haploid nuclei take place, and the first nuclear fusion is
in the ascus. In Laboulbenia chaetophora and L. Gyrimdarum
(Faull, ’11, 712) fusion of nuclei does not occur in the as-
cogonium, the mature binucleate ascogenic cell develops the
asci by budding, each ascus bud being preceded by a con-
jugate division of the nuclear pair. In Polystigma rubrum
(Nienburg, 714) no fusion in the ascogonium occurs. In
Collema pulposum (Bachmann, ’13) the nuclei in the ascogenic
cells were often found in pairs, but no cases of fusion were
observed.

(3). Forms in which an archicarp is absent or functionless,
and certain vegetative cells take on the function of ascogenic
cells, in which the authors believe nuclear fusion does not take
place except in the ascus: Gnomoma erythrostoma (Brooks,
10); Helvella elastica (McCubbin, 710) in which the ‘‘as-
cogenous hyphae’’ form an intricately interwoven subhy-
menial layer of threads each with two nuclei in the end. The
ends of these hyphae form croziers with conjugate division of
the two nuclei followed by about six repeated proliferations
of the young ascus and crozier formations, accompanied by
fusions of the ultimate and antepenult cells and crozier
formation, resulting in many successive conjugate divisions
of the haploid nuclei, with fusion first in the ascus. In
Xylaria tentaculata (Brown, H. B., ’13) the ascogenic cells
which appear to be derived by the separation of the cells of
‘‘Woronin’s hypha’’ are uninucleate and soon become multi-
nucleate by nuclear division. The nuclei multiply also in the
ascogenous hyphae.

The theory of a vegetative fusion in the ascus arose from
the belief on the part of some students that sexual fusion of
the nuclei occurred in the ascogonium, that the nuclear fusion
in the ascus must be a second fusion with no relation to the

1915]
ATKINSON—-PHYLOGENY IN THE ASCOMYCETES 341

sexual process, and, therefore, it must be of a vegetative na-
ture. If a second fusion of the nuclei occurred it would call
for a triple division of the fusion nucleus in order that the
haploid condition should be again reached.

The universal occurrence of the triple division in the ascus
in the formation of the spores is by some ascribed to a ‘‘quad-
rivalent character’’ of the chromosomes in the fusion nucleus,
and rendered necessary in the return to the univalent condi-
tion (Harper, ’05; Overton, ’06), and Overton states, ‘‘that
all these divisions persist, no matter how many spores are to
be produced, which shows their necessity in the process of
reduction.’’*! Hremascus controverts this statement since
there is certainly but one fusion (Stoppel, 07; Guilliermond,
709) and yet triple division occurs in the ascus.

The results of cytological investigations by different stu-
dents in connection with the triple division show considerable
variation. Thus Harper (’00, ’05) finds the same number of
chromosomes in all three divisions (10 in Pyronema, 8 in
Phyllactinia). The two ascus nuclei ‘‘fuse with all their cor-
responding parts’’ (Harper, ’05, p. 67), so that the quadriva-
lent nature of the chromosomes in the fusion nucleus is not to
be seen, though he conceives it to exist. Synapsis occurs in
the first division.

Miss Fraser (’07, ’08) describes Humaria rutilans as having
16 chromosomes in the first division where synapsis occurs
(heterotypic) which split transversely and the daughter nuclei
have each 16 chromosomes which appear on the nuclear plate
in the second division. In the second division the chromo-
somes split longitudinally (homéotypic) and 16 chromosomes
pass to each daughter nucleus. In the third division the 16
chromosomes are supposed to separate at the nuclear plate
without division, 8 going to each daughter nucleus. This divi-
sion she terms ‘‘brachymeiotic’’. A similar situation is de-
scribed by Fraser and Welsford (’08), Fraser and Brooks
(709), and Carruthers (711). Faull (’05) finds the same num-

1 Polysporous asci resulting from several to many nuclear divisions may be

the retention of an ancestral character, the number of divisions being reduced
to three in most forms.

[Vou. 2
342 ANNALS OF THE MISSOURI BOTANICAL GARDEN

ber of chromosomes in all three divisions, in some species 4
or 5 (Hydnobolites), in others 8 (Neotiella).

More recently Claussen (712) after a very thorough study
of Pyronema confluens finds the same number of chromosomes
(about 12) in all three divisions. The first division is hetero-
typic accompanied by synapsis, diakinesis and a splitting of
the chromosomes. The second is homéotypic, while the third
is typic. Faull (’12) in a recent study on Laboulbenia also
finds that the two first divisions in the ascus agree with the
usual phenomena accompanying reduction in spore mother
cells, the first being heterotypic, while the second follows
‘‘very swiftly on the heels of the first.’’ He concludes that
‘*probably the only nuclear fusion in the life cycle is that in
the ascus,’’ and that conjugate divisions of nuclei are an im-
portant phase in the sexual phenomena of the Ascomycetes.

The evidence from recent investigations, therefore, supports
more and more the interpretation of nuclear fusion in the
ascus as a process of exactly the same significance as the
nuclear fusion in the basidium of the Basidiomycetes, and in
the teleutospore of the Uredinales, i. e., it is the fusion of a
pair of nuclei of a longer or shorter history of conjugate divi-
sions from a pair of ancestral nuclei of more or less remote
association. This association of nuclei arises in a variety of
ways and at different periods in the ontogeny just as it does
in the Basidiomycetes (Maire, ’02; Ruhland, ’01; Harper, ’02;
Nichols, S. P., ’04; Kniep, °13), and Uredinales (Sappin-
Trouffy, ’96; Maire, 99, ’01; Blackman, ’04; Christman, ’05,
07; Blackman and Fraser, ’06; Olive, ’08; Hoffmann, 712;
Werth und Ludwigs, 712). The association is accomplished
in some eases through the copulation of two gametangia
(Pyronema, Monascus, Gymnoascaceae, and the Erysipheae).
Such an association represents nearly, if not quite exactly,
the true type of sexuality. The other methods of association
represent a variety of modified types of sexuality (see Note 1)
where the archicarp is present and the antheridium absent, or
functionless, or where the archicarp is absent and vegetative
cells, either with or without the migration into them of nuclei

1915]
ATKINSON—-PHYLOGENY IN THE ASCOMYCETES 343

from adjacent vegetative cells, give rise to the ascogenous
threads.

The results of recent work tend more and more to show
that there is no fusion of the associated nuclei in the as-
cogonium, or ascogenic cells, whether certain of the nuclei
have been derived from an antheridium (Pyronema, Claussen,
12; Monascus, Schikorra, ’09), or not. Conjugate division
in the ascogenous threads has been abundantly proven, though
in some cases it may occur only one or a few divisions prior
to the formation of the ascus.

What the peculiar features of nuclear fusion in the ascus
are which characterize it as vegetative, seem to rest more on
an ex parte judgment of a fusion of nuclei in the ascogonium
than upon any well established idea of the nature of vegeta-
tive nuclear fusion. Thus, Miss Fraser (’08, p. 37) states
that in Humaria rutilans the two nuclei in the ascus enter in-
dependently upon the prophases of the first division, fusing
in the spirem stage. This she regards as evidence in disproof
of the sexual nature of the fusion of nuclei in the ascus
(708, p. 44). Harper (’05) raises a similar objection. On
the other hand, it seems to me that it is excellent evidence
that it is not of a vegetative nature. It is well known in a
number of cases that the egg and sperm nuclei, lying side by
side in the egg, undergo the prophase stages of division up
to the formation of the chromosomes before fusion of the two
takes place. I cite certain examples in the Abietineae: Pinus
sylvestris (Blackman, ’98); P. strobus (Miss Ferguson, ’01,
04); Tsuga canadensis (Murrill, ’00).

In support also of the supposed vegetative nature of the
fusions in the ascus Miss Fraser (’13, p. 559) cites ‘‘vegeta-
tive nuclear fusions’”’ in the quadrinucleate ascus of Humaria
rutilans and her work on this plant in 1908. But she no-
where describes or figures the fusion of the. four nuclei in
such asci. She says (’08, p. 41) ‘‘trinucleate (Fig. 50) and
quadrinucleate (Fig. 51) asci are sometimes formed; their
fate could not be determined.’’ It is very likely that such
tetranucleate young asci found by Miss Fraser in Humaria
result from further conjugate division prior to the prolifera-

[VoL. 2
344 ANNALS OF THE MISSOURI BOTANICAL GARDEN

tion of the young ascus to form branches and further croziers
resulting in an increase of asci as shown to take place in
Pyronema confluens by Claussen (712, p. 25, fig. 6, IIT).

It has been suggested by some who regard the fusion in
the ascus as a second fusion of nuclei (Harper, ’05;
Overton, ’06) that if the synkaryophytic condition of the
terminal portion of the ascogenous hyphae in Pyronema,
and far back in those of Galactinia succosa (Maire, ’03, ’05),
could ‘‘work back until the egg cell was reached,’’ an
apogamous condition might result similar to that in the
Hymenomycetes. Certainly those who have suggested this
theory have not thought far enough ahead, for how would
the univalent condition of the spore nucleus pass to the
bivalent condition of each nucleus prior to the paired (= quad-
rivalent) condition in the ascogenous hyphae of the next
generation unless this were preceded by a nuclear fusion.
Such a condition would not be apogamy. The quadrivalent
character of the fusion nucleus of the ascus, or of the syn-
karion in the ascogenous threads, demands two successive
nuclear fusions, if the triple division in the ascus brings
about the reduction of a quadrivalent nucleus to a univalent
one as maintained by the adherents of this theory. As to
such an apogamous condition being similar to that in the
Hymenomycetes it must be remembered that there are only
two divisions in the reduction process in the Hymenomycetes,
so that when two univalent nuclei become associated in cells
of the mycelium or basidiocarp the bivalent condition of these
cells is attained.

In a very interesting and scholarly argument Harper (’05)
has attempted to explain the inclusion and fusion of two
nuclei in the young ascus on the basis of the nucleo-cytoplasmic
relation or balance in the cell. ‘The abundance of food ma-
terial in the tips of the ascogenous hyphae inhibits cell wall
formation so that two nuclei are enclosed in one cell. Rapid
growth of the ascus and cytoplasm follows in order to balance
the relation of the latter with the nuclear mass. The fusion
of the nuclei and growth of the fusion nucleus again over-
balance the cytoplasm, which then by growth increases again

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 345

in mass. The process is thus a reversible one, and by a sort
of see-saw growth of nucleus and cytoplasm the ascus cell is
pushed up to the large size characteristic of spore mother
cells.

It is very true that the ‘‘regulative function is a reversible
one,’’ that an active cell with a large amount of cytoplasm
demands a correlative amount of nuclear substance, that the
increase in one may result in the increase of the other. Also
it is very true that the ascus belongs to the category of spore
mother cells, which are characterized by relatively large nuclei
and cytoplasmic mass compared with most vegetative cells, but
this does not explain why, when ascus or spore mother cell
formation is about to take place, cell division does not occur
at a period when the food relation would permit the forma-
tion of young uninucleate asci if these nuclei are bivalent in
nature. The regulative functions accompanying growth and
maturity of such a young gonotokont would assure sufficient
size, sufficient food material, and the necessary equilibrium.
The fact that asci in different species and groups vary so
greatly in size shows this, and also that there is no general
standard of mass in relation to surface area which would
demand two nuclei at the origin of the ascus.

In fact it is very clear, from the morphological processes
which take place in the tip of the ascogenous hyphae of most
of the forms studied, that cell division, or cell wall formation,
is more likely governed by the last division of the two nuclei
so that the cell walls are laid down between the daughter
nuclei. If the inclusion and fusion of two nuclei in the young
ascus were controlled entirely by nutritive and cyto-regula-
tive processes, why are not sister nuclei included? Surely the
purely cyto-regulative functions would be just as well satis-
fied. It appears that in rare cases sister nuclei may be in-
cluded in the ascus (Brown, W. H., 710, in Leotia chloro-
cephala).

Of the four nuclei resulting from the two successive divi-
sions of the zygote nucleus in Spirogyra, Chmielewski (’90)
states that two fuse to form the nucleus of the single germling
which is usually formed in the Zygnemaceae. Harper inter-

[Vou 2
346 ANNALS OF THE MISSOURI BOTANICAL GARDEN

prets this as a vegetative fusion in support of his interpreta-
tion of vegetative fusion in the ascus. Karsten (’08) de-
scribes the divisions of the zygote nucleus into four nuclei in
Spirogyra jugalis, but does not state the relation of the nuclei
to the germling (second division sometimes omitted). Tr6n-
dle (’07) interprets the process in Spirogyra Spréeiana as
presenting but a single division of the zygote nucleus. Re-
sults of this nature, so divergent from expectations based on
the normal history in many other organisms in widely sepa-
rated groups, are usually received with considerable reserve,
particularly where they are pioneer investigations in a group
not yet studied. Recently Kurssanow (’11) in a thorough
study of nuclear division and germination of the zygote in
two species of Zygnema (Z. cruciatum and Z. stellinum) has
shown that the process is normal, there being two successive
divisions, three of the nuclei usually degenerating, while one
becomes the nucleus for the single germling characteristic of
the Zygnemaceae. Occasionally only two of the nuclei de-
generated, but then two germlings were formed, an interest-
ing case showing a tendency to retain what is believed to
be the ancestral condition where four germlings are formed
as in the Mesotaeniaceae, while in the desmids two germlings
are regularly formed.

Other cases cited as examples of vegetative nuclear fusion
and classed with nuclear fusion in the ascus, are those of the
endosperm nucleus with the second sperm nucleus in seed
plants (Harper, ’05), and (Fraser, 713) nuclear fusions in
paraphyses and in hairs of the excipulum of certain discomy-
cetes. Such cases, however, cannot be legitimately compared
to fusion in the ascus, since those nuclei are shut off from
further participation in the line of successive ontogenies.

The example cited by Harper of Boveri’s (’88) experiment
in shaking sea urchin’s eggs after fertilization, resulting in
the production of an abnormally large larva with 72 instead
of 36 chromosomes, is in a different class from most of the
other examples of vegetative fusion given. This is equiva-
lent to a true double fertilization and it is quite within the
bounds of possibility that among many such larvae some

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 347

might under favorable conditions be the starting point of a
new ontogeny which would be similar to certain mutants. The
case of Oenothera gigas (see De Vries, ’03, 713) a mutant from
Oe. Lamarckiana with double the number of chromosomes is
similar.t Other tetraploid mutants are known (see Gates,
13), the diploid gametophyte and tetraploid sporophyte of the
mosses produced experimentally by Marchal (’09, 711) is in-
teresting in this connection.

Now, the possibility of a similar double fertilization in an as-
comycete is not, a priori, excluded. There might be an isolated
example. But the normal expectation is that it would have
afterward a nuclear history in its ontogeny similar to others
with one nuclear fusion and one reduction from 2x tolx. But
it is not likely that the entire group of sac fungi is founded
on such a mutation, followed by a double reduction with triple
division and then double fertilization again and so on. The
several cases where it has been quite well established that
there is no nuclear fusion prior to the ascus, together with
the great uniformity of the ascus nuclear phenomena in the
group, controverts the idea of any such origin for the sac
fungi.

All of these facts go to prove that the inclusion and fusion
of two nuclei in the young ascus is of a very different and far
greater significance than a vegetative one. The process of
nuclear fusion in the ascus does not comprise in itself the
entire series of events generally accepted as belonging to the
process of fertilization, for in most organisms nuclear fusion
occurs in the same cell where nuclear association takes
place. It is generally conceded that before the haploid con-
dition of the nucleus is again established important pro-
cesses take place which we call reduction phenomena, the
full significance of which we perhaps are as yet ignorant
of. These processes, including synapsis, cannot take place
unless nuclear fusion has occurred, and some students see in

1 Just how the doubling arose in this instance is of course difficult to
determine. Stomps (12) suggested that it arose through the union of two
unreduced diploid gametes, while Gates (’09, ’13) thinks it arose through “sus-
pended mitosis of a megaspore mother cell” having (4x) 28 chromosomes, and
its apogamous development.

[Vou, 2
348 ANNALS OF THE MISSOURI BOTANICAL GARDEN

them the real act of fertilization (Strasburger, ’00, ’04, ’05).
Remarks on the origin of the specialized ascus—In the
direction of progression from the generalized ascus by split-
ting up of the zygote, the diploid phase has been prolonged
and the number of spores multiplied. The filamentous out-
growths of the zygote, or its equivalent, provide numerous
terminal cells of restricted size suitable for the production of
a small number of spores in each, following the meiotie divi-
sions of the fusion nucleus which terminate the diploid phase.
The situation in species with polysporic asci, where the
spores result from numerous divisions of the fusion nucleus,
is interpreted by some as a germination phenomenon (Over-
ton, 06), but it seems to me more comprehensible to regard it
as a retention of a primitive feature existing in certain phyco-
mycetous ancestors, and characteristic also of primitive As-
comycetes like Dipodascus. )

The formation of internal non-motile spores through free
cell formation in the zygote, under conditions adapted for dis-
persion by ejection from either the generalized or specialized
ascus, may be sufficient to account for the distinctive processes
of spore formation in the sac fungi. In the odgonium of
Saprolegnia, functional nuclei in the odgonium are very simi-
lar to the nuclei of the ascus preceding ascospore formation.
The nucleus is provided with a prominent central body at its
pointed end from which kinoplasmic radiations extend (Har-
tog, ’95; Claussen, ’08; Miicke, ’08).

In most of the Ascomycetes the cytoplasm in the ascus is
differentiated into epiplasm and spore plasm, the former as-
sisting in the ejection of the ascospores. This separation of
the plasm may have been one of the direct causes of the
peculiar method of ascospore formation.

NOTE IV
THE PHYLOGENETIC RELATION OF THE TRICHOGYNE AND SEXUAL APPARATUS OF
THE ASCOMYCETES AND THOSE OF THE RED ALGAE
The sexual apparatus of the Ascomycetes, particularly the
trichogyne and the so-called spermatia, is generally conceded
to be the strongest evidence in support of their phylogenetic

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 349

relation to the red algae. The analogy at least between the
trichogyne of the red algae and that of the Ascomycetes is
very striking. The evidence brought forward by Stahl (’77)
and others of the relation of the trichogyne to the ascogo-
nium in the lichens, together with the fusion of spermatia to
the trichogyne, followed by the gradual and peculiar degenera-
tion of the latter and the subsequent development of the as-
cogenous threads, was generally accepted as proof of fertil-
ization in the ascogonium by a spermatium. Also the early
studies of Polystigma rubrum (Fisch, ’82; Frank, ’83) and
Gnomona erythrostoma (Frank, ’86) in which similar struc-
tures and phenomena were observed at that time, were gen-
erally accepted as indicating a well developed condition of
sexuality. These studies gave a great impetus to the theory
suggested by Sachs (’96) that the Ascomycetes had their
origin from the red algae, or that the two groups had an-
cestors in common. This theory has taken very deep root
and probably is accepted by a majority of botanists even
at the present time, especially by those who are not special
students of the fungi. It should be stated also that a number
of our foremost students of the fungi, perhaps a majority of
them, are firm disciples of this theory.

Recent investigation, however, including a cytological study
of several of the now classic types, including Collema (Bach-
mann, Miss F. M., 712, °13), Polystigma rubrum (Blackman
and Welsford, 712; Nienburg, 714), Gnomonia erythrostoma
(Brooks, 710) have failed to furnish any evidence of a real
sexual function on the part of either the trichogyne or sper-
matia in any of the species of fungi possessing these two
structures. Pairing of nuclei in the odgonium, or the pairing
of these with nuclei from adjacent cells of the ascogonial
branch or archicarp, furnish the synkaria, or the synkaria are
organized at different stages in the development of the as-
cogenous hyphae (see Note 11). In some quarters these re-
sults have led to a loss of confidence in the sexual significance
of the trichogyne and spermatia of the Ascomycetes. Some
have therefore attributed to the trichogyne a physiological
significance of another kind, that of a respiratory organ for

[VoL. 2
350 ANNALS OF THE MISSOURI BOTANICAL GARDEN

example (Brooks, 710), or a boring organ, a terebrator
(Lindau, 799). Zukal (’89) interpreted the trichogyne of
Pyronema confluens as a haustorium to provide food for the
large ascogonium with its numerous ascogenous threads.

Recent investigations on Collema pulposum (Bachmann,
F. M., 713) have revealed an interesting departure in the rela-
tion of the trichogyne and spermatia from that thus far found
in other lichens, and is in strong contrast with the condition
found by Stahl in Collema. The ‘‘spermatia’’ are not free
and are not formed in large numbers in superficial receptacles,
but are imbedded in the thallus and remain attached to the
supporting hypha. The trichogyne does not extend to the
surface but migrates through the interior of the thallus, seeks
the spermatia and fuses with one. Then the trichogyne un-
dergoes the usual deterioration, but no evidence was obtained
of the migration of the nucleus of a spermatium to the as-
cogonium, although a nucleus supposed to be the sperm
nucleus appears to have been observed in the terminal cell of
the trichogyne.

In the red algae the only variations and progression in the
trichogyne is in variations in length to meet the requirements
of thin or thick cortex, some more or less sinuous or spirally
wound, and a few stout and blunt. It is universally a con-
tinuous, enucleate,! prolongation of the odgone, 1. e., not sep-
tate nor a separate cell. So far as we know the sperm always
functions in the red algae. In the sac fungi, there is great
variation and marked morphological progression from an
oogone without a trichogyne through short one-septate
trichogynes to long, simple, several-celled ones, and also to
profusely branched, multi-septate trichogynes. It is more
comprehensible to regard this progression and variation in the
light of evolution from the simple to the complex, in the as-
comycete phylum, independent of the red algae, than to con-

*Davis (96) describes the trichogyne of Batrachospermum as having a
nucleus of its own, but it is not separated from the egg nucleus by a wall until
just prior to the development of the gonimoblasts from the egg. He also states
that the sperm nucleus never passes out of the trichogyne into the egg. How-

ever, Schmidle (’99) and Osterhout (’00) find no trichogyne nucleus and describe
a real fertilization by fusion of sperm and egg nucleus.

1915]
ATKINSON—-PHYLOGENY IN THE ASCOMYCETES 55

ceive the long septate trichogyne of the highly specialized
Collema to be derived directly from the simple trichogyne of
the red algae, and then degenerate to the simple gamete of
lower more generalized Ascomycetes.

NOTE V
MODIFICATION OF SEXUAL PROCESS ALONG WITH STERILITY OR LOSS OF THE
ANTHERIDIUM AND STERILIZATION OF THE ARCHICARP

Sterility or loss of the antheridtum.—Several species are
known in which the antheridium, though present, does not
function. In such cases sexuality is modified in such a way
that sex differentiation occurs among the nuclei in the as-
cogonium or in the ascogenous hyphae. Several examples
may be cited as follows: In Pyronema confluens (Brown,
W. H., ’09) the antheridium sometimes fuses with the
trichogyne but there is no migration of its nuclei; in other
cases it may not connect with the trichogyne. The antheridial
nuclei degenerate. In still other cases the antheridium is ab-
sent. In Lachnea stercorea the antheridium fuses with the
terminal cell of the archicarp but its nuclei degenerate
(Fraser, ’07). In Aspergillus herbariorum (Fraser and
Chambers, ’07) and A. repens (Dale, ’09) a similar situation
exists. In those numerous examples where spermatia (mostly
free ‘‘antheridia’’) are present it is very likely that the sperm
nuclei no longer play a réle in fecundation due to such exten-
sive sterilization of the terminal segments of the archicarp,
but the cytology of only a few species has been determined.
They no longer perform the function of fecundation in Poly-
stigma rubrum (Blackman and Welsford, 712; Nienburg, 714),
Gnomonia erythrostoma (Brooks, 710), and in Collema pul-
posum (Bachmann, 713) the sperm nucleus has not been
traced through the long succession of sterile segments of the
archicarp, and it is very probable that it does not reach the
ascogonial cells. The spermatia are entirely absent in a num-
ber of species where archicarps are present, as in Laboul-
benia chaetophora (Thaxter, ’96; Faull, ’12).

Sterilization of the terminal portion of the archicarp and
differentiation of sex nuclei in the ascogomum or ascogenous

[VoL, 2
352 ANNALS OF THE MISSOURI BOTANICAL GARDEN

hyphae.—A moderately large number of species, in which
more or less extensive sterilization of the terminal portion of
the archicarp has occurred, have been examined by cytological
methods and in most cases a reduced or modified sexual con-
dition has been found.

In Pyronema confluens great variations occur in the sexual
nature of the ascogonium. In what may be called normal
cases, antheridial nuclei enter and become associated with the
ascogonial nuclei (Harper, ’00; Claussen, ’07, 712). Under
cultural conditions the antheridium may be normal, rudi-
mentary or absent, but the ascogonium develops in a normal
manner (van Tieghem, ’84). Different strains may also be-
have differently. In some the antheridium does not fuse with
the trichogyne, while in others it does (Brown, W. H., ’09).
In some cases even when the antheridium fuses with the
trichogyne, its nuclei do not pass into the ascogonium (Dan-
geard, ’07), but degenerate in situ (Brown, W. H., 709). In
these cases where the antheridium does not function the sex-
uality of the ascogonium is modified in as much as its nuclei
are differentiated sooner or later so that in pairs they per-
form the function of sperm and egg nuclei. According to
W. H. Brown (’09) in cases where the origin of the pair of
nuclei in the ascus hook could be determined, they were sisters.
After the one conjugate division in the hook the two nuclei
in the ascus, or penult cell, are ‘‘cousin’’ nuclei.

The archicarp of Lachnea scutellata (Woronin, ’66; Brown,
W. #H., ’11) consists of about nine cells. No antheridial struc-
ture has been observed. The penultimate cell functions as
the ascogonium (Brown, W. H., 711). It is multinucleate and
no fusion of nuclei in pairs takes place here. The nuclei are
increased in numbers by division, not only in the ascogenous
threads where they do not appear to be paired or show con-
jugate division, but also in the ascus hook where conjugate
division takes place. The numerous fusions of the terminal
and basal cells of the ascus hook result in numerous succes-
sive conjugate divisions. In Leotia, although the archicarp.
has not been clearly observed, it would appear from the ac-
count (Brown, W. H., ’10) that the antheridium is absent (or

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 353

if present, functionless) and that the ascogonium consists of
a single coenocytic cell. Conjugate division takes place in
the ascus hook, and the subsequently fusing cells, so that in
most cases rather distantly related pairs of nuclei form the
fusion nucleus in the ascus. In L. chlorocephala (Brown,
W. H., 710), it appears that the pair of ascus nuclei are some-
times sisters. This would indicate an extreme case in the
modification of sexuality, the distance of relationship between
the sex nuclei being reduced to the minimum. It recalls the
very close relationship of the sex nuclei in many of the lower
algae, particularly in certain diatoms! (Oltmanns, ’04), and
in the species of Spirogyra having buckle-joint conjugation
(Chodat, 710). In the case of Spirogyra it is not known
whether the pair of sex nuclei in this type of conjugation
are cousins or sisters, or whether now one and then another
of these possibilities exists. Such species of Spirogyra in
which certain threads present scalariform as well as buckle-
joint conjugation offer an interesting parallel to the variation
in distant relationship of the fusing nuclei in the young ascus.

In some other species where the antheridium is function-
less or wanting, sex differentiation is said to take place among
the nuclei in the ascogonium. ‘This indicates a sex differentia-
tion much earlier than that which is supposed to occur in the
species just cited. This differentiation in sex nuclei has been
described in Humaria granulata (Blackman and Fraser, ’06).

Another species in which similar phenomena are described
is Lachnea stercorea (Fraser, ’07). Here the archicarp con-
sists of several coenocytic large cells and the terminal tricho-
gyne of 4-6 smaller coenocytic cells. The unicellular coenocytic
antheridium fuses with the terminal cell of the trichogyne, but
its nuclei do not reach the single-celled ascogonium, among
whose nuclei sex differentiation is said to take place.

For a number of years Polystigma rubrum, a parasite on
cherry leaves, as the result of studies by Fisch (’82) was re-
garded as an example of fertilization of an ascogone coil by

1In Achnanthes subsessilis, the protoplast divides into two parts along with

nuclear division. The two uninucleate protoplasts now immediately unite in
auxospore formation,

[Vou, 2
354 ANNALS OF THE MISSOURI BOTANICAL GARDEN

sperm nuclei from spermatia after passing through a long
succession of cells constituting the trichogyne or sterile por-
tion of the archicarp. The trichogyne, or sterile portion of
the archicarp, is very long and branches into two portions,
one extending to either surface of the leaf. But according to
Nienburg (’14) sex differentiation has occurred between the
basal cells of the archicarp and a nucleus from the basal cell
migrates into the adjacent cell, which becomes the ascogonium
or ascogenic cell, but nuclear fusion does not take place here.

Loss of function by the archicarp or its disappearance.—
A number of examples are known in which the archicarp has
either lost its function as a sexual organ or ascogone, or has
disappeared. In such cases differentiation of sex occurs in
special vegetative cells, sometimes by the migration of a nu-
cleus from certain cells into adjacent ones. In Gnomona
erythrostoma, although Frank (’86) described coiled ascogone-
like structures with trichogynes, and believed that the coils
were fertilized through the agency of the spermatia, recent
cytological work (Brooks, ’10) on this species appears to
show that the tufts of hair-like structures emerging through
the stomates of cherry leaves, on which this species of Gno-
monia is parasitic, are not now connected with the coiled
hyphae deeper in the tissue. It appears also from the same
work that the ascogenous hyphae do not arise from the coils,
but from one or more slightly differentiated hyphae in the
center of each coil.

A similar example is found in Xylaria polymorpha (Fisch,
82), where an extensively coiled hypha (‘‘ Woronin’s hypha’’)
occurs in the early stages of the formation of the ascocarp,
but later disappears and certain vegetative cells give rise to
the ascogenous hyphae.

In Humaria rutilans (Fraser, 08) no archicarp or ascogone
coil is discernible, but certain vegetative cells function as as-
cogenic cells following the migration into them of nuclei from
adjacent cells.

MORPHOLOGY OF THE ARCHICARP

If the history of the Ascomycetes is correctly read from

the simpler and more generalized forms to the complex and

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 355

highly specialized ones as Sachs (’74, ’96), de Bary (’81,
84), and many other students have advocated, the female
organ or archicarp first appeared as a ‘‘unicellular’’ or con-
tinuous organ, not differentiated into an odgonium or fertile
portion, and a trichogyne. The presence of a ‘‘procarp,’’
whether consisting of one or several cells, which ultimately
gave rise to the asci or ascogenous threads was the predomi-
nant character which led Sachs in 1896 to believe in the phy-
letic relation of the sac fungi and red algae, although earlier
he had regarded the morphology of the ascocarp and cysto-
carp of greater importance in showing relationship. No
known red alga possesses a procarp simple enough to repre-
sent the prototype of the two groups. Gymnoascus was se-
lected by Sachs as representing the simplest Ascomycetes.
The archicarp of Gymnoascus is a continuous structure more
or less coiled around the antheridium from which it copulates
directly without the intervention of a trichogyne.

After copulation the ascogonium divides into several cells
which give rise to the ascogenous hyphae. In some forms the
splitting up of the ascogonium by transverse division occurs
at an earlier period, before copulation. There is some evi-
dence which indicates that the ‘‘trichogyne’’ in the Ascomy-
cetes primarily was a prolongation of the ‘‘unicellular’’
oogone (or carpogone), and that when it was first separated
as a distinct cell it was still a fertile part of the archicarp. In
Aspergillus repens the terminal cell, or ‘‘trichogyne,’’ some-
times gives rise to ascogenous hyphae (Fraser, ’08).

The terminal cell became merely a trichogyne when it ceased
to give rise to ascogenous hyphae, and acted as a transport
tube for the sperm nuclei from the antheridium to the as-
cogonium, as in Pyronema and Monascus. The septum be-
tween the terminal cell and the functional ascogonium was
an impediment to the passage of the sperm nuclei, as well as
the fact that when they entered the terminal cell of the archi-
carp they did not meet with functional egg nuclei. This situa-
tion very likely favored the assumption of sperm and egg
functions by the nuclei of the functional ascogonial cell. The
variations in Pyronema where the antheridium may or may

[Vou, 2
356 ANNALS OF THE MISSOURI BOTANICAL GARDEN

not be present, and often when present and fused with the
trichogyne its nuclei degenerate and the ascogonium is still
functional producing ascogenous hyphae and asci, is in sup-
port of this interpretation.

Further sterilization of the terminal portion of the archi-
carp proceeds as it becomes longer and more septate, the fer-
tile ascogonial cell or cells being near the center or base. All
of the sterile portion of the archicarp distal to the ascogonial
cells is usually interpreted as the trichogyne. I believe it
would be more in harmony with the historical origin of the
archicarp, and with the real homologies, if only the terminal
sterile receptive cell of the archicarp were called the trich-
ogyne, the other portions to be regarded as sterile portions of
the archicarp or ascogonium. This would be in harmony also
with Thaxter’s (’96) interpretation of the archicarp of the
Laboulbeniales.1. In this group the inferior and superior sup-
porting cells are sterile cells of the archicarp derived by a
transverse splitting of the ascogonium. Even with this inter-
pretation of the trichogyne of the Ascomycetes, it would be
a different structure from that of all the red algae where it
is merely a continuous prolongation of the egg cell.

NOTE VI

The coenocytic character of the mycelium of the Phycomy-
cetes has been presented as an obstacle to the derivation of
the sac fungi from the sporangium fungi (Bessey, E. A., 713);
this character can, however, have very little or no significance,
for many of the Ascomycetes are coenocytic. As in most of
the fungi, cell wall formation is delayed so that new portions
of filaments are often multinucleate, the cell walls being laid
down subsequently, sometimes enclosing one nucleus, some-
times several in a cell. There are the monoenergid and poly-
energid species of sac fungi. In the Phycomycetes cell wall
formation is usually longer delayed or does not occur except
where reproductive cells are formed. In the Mucorales old
mycelium frequently becomes multiseptate. It should be
noted that in Basidiobolus (Kidam, ’86; Raciborski, ’96; Fair-

1 Except in the case of the multiseptate branched trichogynes.

1915]
ATKINSON—-PHYLOGENY IN THE ASCOMYCETES 357

child, ’97, and others) the cells are uninucleate. The varia-
tion in coenocytic character of mycelium probably is due in
some measure to the usually fundamental difference between
cross wall formation in dividing cells, in the thallophytes and
the higher groups of plants, where the fibers of the inner
spindle play a part and the cell wall development is centrif-
ugal, while in most thallophytes the spindle fibers do not
play such a part, wall formation being centripetal, like a clos-
ing iris diaphragm.

The strong plasma connections between the protoplasts of
the Laboulbemales (Thaxter, 96) present a very striking re-
semblance to those in the red algae. This feature is regarded
by some as very strong evidence of a phylogenetic relation
between the Laboulbeniales and the red algae. But intercellu-
lar plasma connections are a common feature in all groups of
plants, though in many plants these connections are very
minute. The single central pore in the wall of the Laboul-
beniales is perhaps the result of incomplete closing of the ring-
forming wall, and in the Laboulbeniales would seem to be of
physiological rather than of phylogenetic significance. The
firm cell walls which are characteristic of the members of this
group bear a very definite relation to their habit as external
parasites of insects. Standing out free from their bodies and
thus having no other means of support than their own rigidity,
thick cross walls would interfere with transport of food ma-
terial, while the prominent plasma connections permit easy
passage of nutrients.

NOTE VII

BRIEF OUTLINE OF SOME OF THE THEORIES AS TO THE PHYLOGENY OF THE
ASCOMYCETES

I. Descent from the Rhodophyceae.—Sachs (’74, p. 287)
regarded the resemblances between cystocarp and ascocarp
as the most important character indicating a relationship be-
tween the red algae and sac fungi, although the form of the
sexual organs, particularly the carpogonial branch, was also
believed to point in the same direction. In his ‘Lehrbuch der
Botanik’ he did not even suggest that the Ascomycetes were
derived from the Florideae. The relationships were based

[Vou. 2
358 ANNALS OF THE MISSOURI BOTANICAL GARDEN

on the principle of morphological homology, which he believed
were great enough to justify their inclusion in the same class.
To justify his arrangement in one large group of plants with
such diverse aspects and habitats, he cites the inclusion of the
Lemnaceae and palms in the great group of the monocots. We
could not then interpret his inclusion of the sac fungi and red
algae in one class, the Carposporeae, as indicating that the
former were derived from the latter.

Sachs says (’74, p. 288) that in order to find the relation-
ships between plant divisions one must compare the simplest,
not the highest forms. By this method he finds that the
Coleochaetaceae and Characeae are linked, on one hand to
the simplest Florideae, and on the other to the simplest As-
comycetes. Each of these series, he says, has developed in
its own peculiar manner to higher forms, so that if one com-
pared the most complete Ascomycetes with the coleochaetes
only very slight resemblances are to be found. From this it
is very clear that Sachs, at that time, had no thought of the
derivation of the Ascomycetes from the Florideae. There is
nothing to indicate that he believed the Ascomycetes descended
from the charas and simplest coleochaetes, to which he says
the simplest Ascomycetes are most closely related. Nor
would his theory require a common ancestor for the two
groups. Because of the morphological resemblance between
eystocarp and ascocarp, he would have united the Ascomy-
cetes and Florideae into a higher group even had he believed
that the former were derived from the Phycomycetes.

It has been said by Sachs (’96, p. 204) that the fungi as a
whole cannot be valued as an architype because, as apochlo-
rates, they must be descended from green plants. The bacteria
he would derive from the Cyanophyceae, the Phycomycetes
from the Siphoneae, and the Ascomycetes (or at least the
Discomycetes) from the Rhodophyceae. The predominant
feature indicating the descent of the sac fungi from the red
algae he now sees in the procarp of both groups (96, p. 205).

The chlorophylless seed plants have only a slight form-pro-
ducing power or motive, as Sachs has pointed out (96, p. 205),
since they occur mostly as small plant groups within certain

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 359

green leaved families and show very plainly the morpholog-
ical characters of their antecedents. But he says it is quite
otherwise with the fungi. The simplest primitive forms of
the Ascomycetes, Phycomycetes and Basidiomycetes have
given rise independently to an enormously high state of dif-
ferentiation. Now Sachs in 1896 (and earlier, ’74, p. 310)
recognized Gymnoascus as belonging to the simplest Ascomy-
cetes, the sexual organs of which are a simple carpogone and
pollinode. It is very clear then that Sachs would not derive
the Ascomycetes from any primitive form at all like any
known red algae, much less through such forms as the highly
specialized Collema or Polystigma. This warrants us in con-
eluding that Sachs had in mind a primitive hypothetical an-
cestor of the sac fungi and red algae, which possessed simple
copulating gametes. With the knowledge we possess to-day
of such forms as Dipodascus, Eremascus, etc., where the
zygote becomes the ascus (generalized or simple) I believe he
would have recognized in the Phycomycetes, as we know them
to-day, a situation very closely approximating an ‘‘Urform”’
for the Ascomycetes, particularly in view of the fundamental
difference in the cytology of the red algae and sac fungi.

But whether the fungi represent one or several architypes
it by no means follows that, because of the absence of chloro-
phyll, they must be derived from green plants, or that each
great series must be derived separately from different groups
of algae.

The appearance of the higher fungi (Humycetes) was, in
the opinion of Vuillemin (’12, p. 223), contemporaneous with
the emergence of sea-shore, which abandoned certain red algae
to a terrestrial life. This new environment introduced the
change, which, accompanied by loss of chlorophyll, gave rise
first to the Pyrenomycetes, from which the other higher fungi
(Uredinales, Basidiomycetes) have originated. The sapro-
phytic forms represent the productive and progressive stock.
Parasitic groups, like the Uredinales, Laboulbeniales, lichens,
etc., are composed of highly specialized and uniform members,
their progressive potentialities being suppressed, but they re-
tain their hold on existence because of their specialized hab-

[Vou, 2
360 ANNALS OF THE MISSOURI BOTANICAL GARDEN

itat. The first Pyrenomycetes, according to his view, were
some of these depatriated red algae, losing their pigments
while preserving the structure, the sexual organs and the gen-
eral evolution. But he recognized no known member of the
red algae as a prototype of the Pyrenomycetes. Primitive
trichogyne-bearing algae gave rise to the red algae on one
hand, and to the Pyrenomycetes on the other, the now known
colorless red algae (like Harveyella mirabilis, Choreocolax
alba) being recently reduced forms having no significance in
the origin of the sac fungi. But the Pyrenomycetes with well
developed trichogyne and spermatia are chosen as the primi-
tive forms, the simplest represented by Polystigma (in his
““Polystigmales’’) the higher ones (his ‘‘Pyreniales’’) giving
rise successively to the Hysteriales and Phacidiales. From
the Polystigmales three other lines arose, their simplest forms
being represented by first, Gymnoascus; second, Pyronema;
and the third line represented by the Laboulbemales (see
Vuillemin, 712, pp. 338-341).

Bessey (’14) regards the Discolichenes as the most primi-
tive Ascomycetes. This theory is based on the supposed
phyletic relation of the multiseptate trichogyne of the lichens
(Collema, for example) to the trichogyne (a mere tubular,
continuous, prolongation of the egg) of the red algae. Cer-
tain of the red algae became parasitic on blue-green algae and
on simple members of the green algae, forming a lichen
thallus. It is supposed that this parasitism may have had
its origin while both kinds of organisms still lived in the water,
but finally the lichen assumed the land habit. The improba-
bility of such a derivation of the sac fungi as suggested in the
above theories has been fully discussed in the preceding pages.

II. Descent from the Phycomycetes——De Bary (’81, ’84,
87), as already stated in the first part of this paper, be-
lieved the Ascomycetes were derived from the Phycomycetes,
particularly through such forms as the Peronosporales. The
criterion for the relationship is the close homology and mor-
phological resemblance of the sexual organs, though he sug-
gested that Hremascus might have been derived from the
Mucorales through some such form as Piptocephalus where

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 361

the zygote is the outgrowth from the fusion point of two equal
gametangia.

Brefeld (’89, ’91) also derived the Ascomycetes from the
Phycomycetes but interpreted the ascus as the phyletic hom-
ologue of the sporangium, the ascus representing a special-
ized structure derived from the generalized sporangium in one
direction, while the conidia were regarded as reduced one-
spored sporangia. But the nuclear fusion and reduction
phenomena in the ascus are so fundamentally different from
any known cytological processes in the sporangium, that its
phyletic relation to the sporangium is doubtful.t The con-
jugation of the gametangia he interpreted as ordinary fusion
of hyphae which occurs in numerous instances devoid of all
sexual significance. Protomyces, Ascoidea and Thelebolus,
with numerous spores in the ascus, were interpreted as rep-
resenting an intermediate condition between the generalized
sporangium of the Mucorales and the specialized ascus. In
Thelebolus it has been found that the development of the ascus
follows the type with crozier formation and that it is closely
related to Ascobolus and Rhyparobius (see Ramlow, ’06;
Dangeard, ’07). As for Protomyces and Ascoidea they prob-
ably represent forms with reduced sexuality while retaining
the ancestral character of many divisions of nuclei to form
numerous spores.

Zukal (’89), influenced by Brefeld, derived the hymenial
Ascomycetes (like Ascobolus, Pezizales, etc.) through Thele-
bolus and Monascus; the stromatic Ascomycetes (whether
Pyrenomycetes or Discomycetes) from the Uredinales; the
Gymnoascales and others with asci arising directly from the
mycelium, from another ancestral type.

Lotsy (’07, p. 469) sees no difficulty in deriving the polyener-
gid forms like Pyronema from the Phycomycetes. The forms
with spermatia, which are usually monoenergid, it would seem
rational, he thinks, to derive from the red algae, and this
raises the question as to whether the Ascomycetes are of poly-
phyletic (or biphyletic) origin. The great uniformity of the

1The nuclear phenomena in the “germ” sporangium (from the zygote) are
not known. .

[Vou. 2
362 ANNALS OF THE MISSOURI BOTANICAL GARDEN

ascus in the entire group is a great obstacle in the way of
accepting a polyphyletic origin for the group. All things con-
sidered he is inclined to accept de Bary’s view of their phy-
comycetous origin.

The origin of the Ascomycetes from the Phycomycetes is
recognized by Dangeard (’07) through such forms in which
there is still a union of gametangia. Dipodascus and Ere-
mascus represent such forms in his scheme. The generalized
ascus resulting from the union of the gametangia of Dipo-
dascus he terms a ‘‘sporogone.’’ From Eremascus, by re-
duction, forms like Endomyces arose, while the Ascomycetes
with ascogenous hyphae were derived from such forms as
Dipodascus by delayed nuclear fusion and the proliferation of
the gametangium into what he terms ‘‘gametophores’’ (= as-
cogenous hyphae). The gametes then are formed in the
nuclear pair which fuses in the ascus. This terminology
arises from his persistent belief that the ascus is the egg.
Shorn of the change in terminology and his, perhaps, unfor-
tunate insistence on homologizing the ascus with the egg, his
interpretation of the relation which such a form as Dipodascus
bears to the Ascomycetes, has much merit.

Nienburg (’14) suggests the origin of the Ascomycetes from
the Phycomycetes through some such form as Monoblepharis.
He would find the evidence for this in the homology of the
archicarp of Polystigma rubrum with such forms of Monoble-
pharis in which the stalk cell of the odgonium is an anther-
idium, and where the odgonium is terminated by one or more
sterile cells. The archicarp of Polystigma he interprets as
having two fertile cells at the base and prolonged into a long
sterile septate portion (so-called trichogyne) which forks,
sending a branch to either surface of the leaf. The basal
multinucleate cell is the antheridium. After pore formation
one nucleus migrates into the unicellular egg. Interesting as
this suggestion is, forms of Pythwwm (see de Bary, ’81, ’84;
Atkinson, ’95) with intercalary odgonia and stalk antheridia
present a closer analogy to the archicarp of Polystigma as
described by Nienburg, but it is extremely doubtful if the
point of contact is to be sought through such structures.

1915]
ATKINSON—-PHYLOGENY IN THE ASCOMYCETES 363

Brief comparative summary of the above views on the
phylogeny of the Ascomycetes—The adherents to the doc-
trine of the red algal origin of the Ascomycetes interpret the
point of contact in three different ways: first, sac fungi with
highly developed ‘‘trichogyne”’ (sterilized archicarp) of the
Collema type with red algae like certain of the existing forms,
Nemalion, or some of the higher forms in the vicinity of Har-
veyella, ete.; second, sac fungi with highly developed
‘*‘trichogyne’’ (= sterilized archicarp) of the Polystigma
type with hypothetical trichogyne algae representing the com-
mon stock for the origin of both groups; third, sac fungi with
simple generalized copulating gametes of the Gymnoascus
type with hypothetical algae having a simple procarp repre-
senting the stock from which both groups originated.

According to the two first interpretations the sac fungi
have been derived through highly developed and specialized
forms from either quite highly developed and specialized red
algae, or both groups from a common trichogyne algal stock,
and then by degeneration have slid backward from complex
and specialized structures to simple, generalized and primi-
tive ones. The third view which recognizes a simple procarp,
without regard to a trichogyne, as the important character of
the hypothetical stock, is far more comprehensible.

But if we must go back to some hypothetical ancestor, which
cannot be represented by any known red alga, for the source
of the sac fungi it is far more reasonable to search for one
in another fungus line, where, in the light of present-day
knowledge, there are known forms with sexual organs very
much like the sexual organs of simple, known forms of the
Ascomycetes. But we are not yet in a position to name any
known phycomycete! as a probable ancestor of the Ascomy-
cetes, though it appears very likely that the ancestral stock
possessed phycomycetous characters.

. + Lotsy (’07) suggests Cystopus; Miss Dale (’03) in her study of Gymnoascus
suggests Basidiobolus; Nienburg (’14), Monoblepharis; while Dangeard (’07)
suggests Myzocytiwm vermicolum as the prototype of the higher fungi.

[VoL, 2
364 ANNALS OF THE MISSOURI BOTANICAL GARDEN

PROVISIONAL ARRANGEMENT OF MAIN LINES OF DEVELOPMENT IN
ASCOMYCETES

For those who are interested in the suggestions as to the
phylogeny and relationships of the Ascomycetes presented in
this paper, a diagrammatic arrangement of the. principal
series or lines which will illustrate the relationships tenta-
tively held by the writer may be acceptable. It is with con-
siderable hesitation that this arrangement is presented. The
writer trusts that it will be accepted as provisional and in
the nature of a working hypothesis which he hopes will fur-
ther stimulate investigation, suggestions and criticisms on
the ideas embodied in this paper, all of which, for or against,
will be gladly welcomed.

Dipodascus, a primitive form, cells of mycelium polyen-
ergid, gametogenous branches large, unequal, polyenergid.
Ascus is elongated, broadened zygospore, zygote germinating
immediately forming a broad germ tube in which spores are
formed. Since the process does not go on to the formation
of a sporangium, a different mode of internal free cell-forma-
tion then arose in connection with the precocious formation of
spores in the zygote and retention of epiplasm which assists
in discharge of spores. Dzpodascus retains tendency of
gamogenic branches to copulate early before they become
strongly differentiated as gametangia, just as in Mucorales.

J. ProroascoMycetss are derived by descent and degenera-
tion from some such primitive ascomycete form as Dipodascus.
The ascus when of sexual origin is the zygote, except in Nad-
sonia.

Endomyces Magnusii is the nearest known form to the gen-
eralized condition seen in Dipodascus. Cells of mycelium
usually polyenergid, those of stout mycelium are polyenergid.
Formation of ascus in Hndomyces Magnusii repeats formation
of zygospore in Zygorhynchus. Gamete branches in both are
multinucleate, but when cell wall is laid down delimiting the
gametangia all but one nucleus in each gametangium of E£.
Magnusu are excluded. After contact of the two sexual
branches the male gametangium is formed by enlargement of
its tip, into which protoplasm and the one nucleus migrates,

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 365

exactly as male gamete of Zygorhynchus is formed, except the
latter is multinucleate. By disappearance of the separating
wall, ascus is formed of the two gametes.

Endomyces series, then, derived from Dipodascus-like an-
cestors, with Endomyces Magnusu the lowest and most gen-
eralized.

Developmental tendencies from here in four, five, or six
different directions:

1. Eremascus, both gamogenic branches uninucleate, ascus more
definite and specialized in shape. Loss of conidial formation.
Endomyces fibuliger indicates step toward Hremascus (E. fer-
tilts) in small size of gametes.

2. Endomyces diverging into the two series, one chiefly with sprout
conidia, the other chiefly with oidia; the latter preserves the L.
Magnusii character, the former takes on sprout conidia in addi-
tion to oidia (EF. fibuliger and E. capsularis form both oidia and
sprout conidia) ; oidia formation the more primitive and gener-
alized condition in Ascomycetes.

3. Saccharomycetes. Still more specialized and reduced than in
Endomyces fibuliger and in this same line. Schizosaccharomyces
may have come from same line with dropping of sprout conidia,
or may be descended from form near Endomyces Magnusit.

4. Exoascaceae. From Endomyces-like ancestors. Nuclear phe-
nomena not well known. Diploid young ascus may have arisen
in connection with cell wall formation, two nuclei being retained
in ascogone instead of one as in HL. Magnus, where all but one
are excluded at time of wall formation, i. e., ascus fundament
may have retained the polyenergid character of the most primi-
tive forms like FE. Magnusu. Tendency to form hymenia may
be controlled by host since asci in all, except Taphrina laurencia,
come to surface to mature.

5. Ascocorticium, saprophytic on wood where food is not so rich,
tendency to drop conidial formation (?), association of asci in
hymenium, highest development of the Endomyces series, or of
the Protoascomycetes. Series is terminated early, tendency in
Endomyces line to specialization of zygote into one ascus with
reduced number of spores, and line soon terminated.

6. Ascoidea, Protomyces, Taphridium, etc., probably represent
forms derived by reduction and loss of distinct sexual organs but
preserving primitive feature of many divisions of nucleus in the
generalized ascus.

II. HEvascomycetses. Lowest forms with generalized archi-
carp. Similar to Monascus, Gymnoascus, ete.

[Vou. 2
366 ANNALS OF THE MISSOURI BOTANICAL GARDEN

1. Tendency to late copulation of gamogenic branches, so that
archicarp becomes large and many-nucleate, or tendency to
elongate, or both.

2. As it elongates tendency to septation, first a single terminal cell
(“trichogyne”’), and later longer and multiseptate “trichogyne,”
or rather sterilization of terminal portion of archicarp. One of
the early tendencies in connection with elongation of the archi-
carp may have been the origin of a receptive terminal] portion
under chemotactic or similar stimulation; such a condition sug-
gested in Cystopus.

3. This made the passage of antheridial nuclei increasingly dif_i-
cult, and resulted in early tendency to sterilization of anther-
idium or failure to function because of functionless condition of
“trichogyne.” Led in many cases to modified sexuality by dif-
ferentiation of sex among nuclei in ascogonium, vegetative cells,
or ascogenous threads.

4. Progressive tendency to multiplication of spores by postpone-
ment of nuclear fusion and spore formation; conjugate division
of sex nuclei, and multiplication of the specialized structures
(asci) in which spores are formed, so that spore formation and
distribution is extended over greater period of time. This most
advantageously attained by sprouting of zygote (ascogone),
branching of threads, and terminal formation of specialized asci.

Diverging lines from Gymnoascus and Monascus-like an-
cestors or related prototypes in which asci are irregularly
arranged but associated in groups with imperfect envelope.

1. A line with interwoven asci, Plectascales as a highly specialized
lateral group, with Gymnoascaceae at base. Aspergillaceae a
progressive line, with Perisporiales an offshoot, or Perisporiales
direct from Monascus-like ancestors.

2. Hlaphomycetaceae, asci interwoven in groups but separated by
sterile walls.

3. Pezizales, asci remaining in groups not interwoven in mycelium,

' but spaced by sterile threads (paraphyses). Pyronema repre-
sents one of the generalized, lower forms. The Helvellales, etc.,
are probably derived from the Pezizales.

4.The Microthyriales! have usually been placed among the Peri-
sporiales with which they have little in common. I believe they

1 Recent studies by several authors, particularly by von Héhnel (710) and
by Theissen (’12, ’13, 714) have greatly increased our knowledge of these interest-
ing fungi, partly by the discovery of new forms but especially by uncovering
many forms from the clouded situation in which they have been placed for lack
of an adequate study of their structure.

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 367

represent reduced forms derived on the one hand from the Pha-
cidiales and perhaps on the other from the Sphaeriales and pos-
sibly some from the Perisporiales. The formation of the char-
acteristic shield has rendered superfluous the perithecial wall as
a protective structure. The genus Diplocarpon, the structure
and development of which was investigated by one of my former
students (see Wolf, ’12), I believe is an excellent illustration of
a form on the way (by reduction of the perithecial wall in con-
junction with the formation of the shield) from the Phacidiales
to the condition presented by many members of the Micro-
thyriales.

The above provisionally suggested relationships may be
represented by the following five or six series, or lines of
development, with the accompanying diagram (fig. 10) :

1. Apocarp line from Dipodascus-like forms and by reduction.

2. Plectocarp line from Dzpodascus-like forms, perhaps similar to
Monascus.

3. Perispore line arising from Monascus-like prototype, before
splitting of archicarp, or from Aspergillaceae.

4. Pyrenocarp line arising near Monascus-like prototype. Laboul-
beniales side line near base, and some of the
Mycrothyriales as reduced from Sphaeriales.

5. Discocarp line from Dzpodascus-like forms near Monascus, but
lower (it is not improbable that some of the
members of the stock of primitive Ewascomy-
cetes showed considerable variation in the
strength of the ascocarp envelope, also in its
presence or absence in forms where it is more or
less rudimentary!) ; and some of the Microthy-
riales as reduced forms from Phacidiales.

Or a 6th line also, Laboulbeniales from Monascus-like ancestor.

1 This variation sometimes occurs in existing forms. Zukal (’89) describes
an abnormal case in Hurotium (Aspergillus) herbariorum where the antheridial
branch and envelope are wanting, the mass of asci being exposed. In this con-
nection it is worthy of note that Fraser and Chambers (’07) regard Aspergillus
“as representing a primitive ascomycetous type from which most others can be
derived.” This suggestion was based on the assumption that the red algae were
the ancestors of the sac fungi. On the basis of the counter theory (phyco-
mycetous origin) Gymnoascus and Monascus-like forms are more comprehensible
as primitive Huascomycetes.

[VoL. 2
368 ANNALS OF THE MISSOURI BOTANICAL GARDEN

a
8
3
>
S
£ 3
s § of es
are § & é yi, )
\ iS </ & o/
g \ a 4 & S/yt Hysteriales <S
2, <x Sj S/—?> Microthyriales o/
2 \ ef,
Reh cet
% 2\

aoe _-~ Tuberales
eek Helvellales

ote
=. of

y Se
MONASCU q

<2 Ascocorticium
witA—— Exoascaceae

eyo Saccharomucetes

Taphridium
Protomyces

DIPODASCUS ___<7)Ne

STOCK APOCAR

PROTOASCOMYCETES
ASCOMYGERES

PHYCOMYCETES

Fig. 10. Chart showing suggested phylogeny of the Ascomycetes.

LITERATURE CITED

Atkinson, Geo. F. (’95). Damping off. Cornell Univ. Agr. Exp. Sta., Bull. 94;
231-272. pl. 1-6. 1895.
Bachmann, Miss F. M. (’12). A new type of spermogonium and fertilization in
Collema. Ann. Bot. 26: 747-760. pl. 69. 1912.
, (13). The origin and development of the apothecium in Collema pul-
posum (Bernh.) Ach. Archiv f. Zellforsch, 10; 369-430. pl. 30-36. 1913.

Atkinson, Geo. F. Monograph of the Lemaneaceae of the United States. Ann. Bot. 4: 177-277;

pls. 7-9, 1890.

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 369

Barker, B, T. P. (’03). The morphology and development of the ascocarp in
Monascus. Ann, Bot. 17: 167-236. pl. 12-13. 1908.

, (704). Further observations on the ascocarp of Rhyparobius. British
Assoc. Adv. Sci., Cambridge, Rept. 1904; 825-826. 1905.

de Bary, A. (’81). Untersuchungen iiber die Peronosporeen und Saprolegnieen
und die Grundlagen eines natiirlichen Systems der Pilze. In de Bary und
Woronin, Beitr. z. Morph. u. Physiol. d. Pilze 4; 1-145. pl. 1-6. 1881.

, (784). Vergleichende Morphologie und Biologie der Pilze, usw.
Leipzig, 1884.

, (87). Comparative Morphology and Biology of the Fungi, Mycetozoa,
and Bacteria. 1887.

Baur, E. (’98). Zur Frage nach der Sexualitiit der Collemaceen. Ber. d. deut.
bot. Ges. 16: 363-367. pl, 23. 1898.

, (04). Untersuchungen iiber die Entwickelungsgeschichte der Flechten-
apothecien. Bot. Zeit. 62: 21-44. pl. 1-2. f. 1. 1904.

Bessey, C, E. (’14). Revisions of some plant phyla. Univ. Neb. Stud. 14: 37-109.
1914.

Bessey, E. A, (713). Some suggestions as to the phylogeny of the Ascomycetes.
Mye. Centralbl. 3: 149-153. 1913.

Blackman, V. H. (’98),. On the cytological features of fertilization and related
phenomena in Pinus sylvestris L. Roy. Soc. London, Bot., Phil. Trans. 190:
395-427. pl. 138-14. 1898.

(704). On the fertilization, alternation of generations and general
cytology of the Uredineae. New Phytol. 3; 23-28. 1904,

, (04). On the fertilization, alternation of generations and general
cytology of the Uredineae. Ann. Bot. 18; 323-373. pl. 21-24. 1904.

, and Fraser, H. C. I. (’05). Fertilization in Sphaerotheca. Ann. Bot.
19: 567-569. 1905.

» ———, (06). On the sexuality and development of the ascocarp of
Humaria granulata Quel. Roy. Soc. London, Bot., Proc. 77: 354-368. pl.
13-15. 1906.

—_—,, —————,, (’06). Further studies on the sexuality of the Uredineae.
Ibid. 20: 35-48. pl. 3-4. 1906.

, and Welsford, E, J. (712). The development of the perithecium of Poly-
stigma rubrum DC, Ann. Bot. 26: 761-767. pl. 70-71. 1912.

Boveri, Th. (788). Zellen Studien II. Die Befruchtung und Zellteilung des Eies
von Ascaris megalocephala. Jena, 1888.

Brefeld, O. (’88). Basidiomyceten II. Protobasidiomyceten. Untersuchungen aus
dem Gesammtgebiete der Mykologie 7: I-X and 1-178. pl. 1-11. 1888.

, (789). Basidiomyceten III. Autobasidiomyceten und die Begriindung
des natiirlichen Systemes der Pilze. Ibid. 8; 1-274. pl. 1-11. 1889.

, (91). Die Hemiasci und die Ascomyceten. Jbid. 9: 1-156. pl. 1-3B.

, (91). Ascomyceten II. Ibid. 10: 157-378. pl. 4-13, 1891.

Brooks, F. T. (710). The development of Gnomonia erythrostoma Pers, Ann. Bot.
24: 585-605. pl. 48-49. 1910.

<*%

[Vou 2
370 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Brown, H. B. (713). Studies in the development of Xylaria. Ann. Mye, 11; 1-13.
pl. 1-2. 1913.

Brown, W. H. (’09). Nuclear phenomena in Pyronema confluens. Preliminary
note. Johns Hopkins Univ. Cire. N. 8S. 28°: 42-45 (1-6). f. 1-3. 1909.

, (10). The development of the ascocarp of Leotia. Bot. Gaz. 50; 443—
459. f. 1-47. 1910.

, (11). The development of the ascocarp of Lachnea scutellata. Ibid.
52: 273-305. pl, 9. f. 1-51. 1911.

Carruthers, C. (711). Contributions to the cytology of Helvella crispa. Ann.
Bot. 251: 243-252. pl. 18-19. 1911.

Chmielewski, W. F. (90). Matériaux pour servir 4 la morphologie et physiologie
des procés sexuels chez les plantes inférieurs. 1890.

Chodat, R. (710). Etudes, sur les Conjugées I. Sur la copulation d’un Spirogyra.
Soc. Bot. Genéve, Bull. II. 2: 158-167. f. a-g. 1910.

Christman, A. H. (05). Sexual reproduction in the rusts. Bot. Gaz, 39; 267-275.
pl. 8. 1905.

, (07). The nature and development of the primary uredospore. Wis.
Acad, Sci., Trans. 15: 517-526. pl. 29. 1907.

Claussen, P. (705). Zur Entwicklungsgeschichte der Ascomyceten. Boudiera. Bot.
Zeit. 63: 1-28, pl. 1-3. f. 1-6. 1905.

, (07). Zur Kenntnis der Kernverhiiltnisse von Pyronema confiuens. —
Ber, d. deut. bot. Ges. 25: 586-590. f. 1, 1907.

, (708). Ueber Hientwicklung und Befruchtung bei Saprolegnia. Ibid.
26: 144-161. pl. 6-7. 1908.

, (12). Zur Entwickelungsgeschichte der Ascomyceten. Pyronema con-
fluens. Zeitschr. f. Bot. 4; 1-64. pl. 1-6. f. 1-10. 1912.

Cutting, E, M. (’09). On the sexuality and development of the ascocarp in Asco-
phanus carneus Pers. Ann. Bot. 23: 399-417, pl. 28. 1909.

Dale, Miss E. (’03). Observations on the Gymnoasceae. Ann. Bot. 17: 571-596.
pl. 27-28. 1903.

, (09), On the morphology and cytology of Aspergillus repens. Ann.
Mye. 7: 215-225. pl. 2-3. 1909.

Dangeard, P. A. (’92). Recherches sur la reproduction sexuelle des champignons.
Le Botaniste 3: 222-281. pl. 20-23. 1892.
, (94). La reproduction sexuelle des Ascomycetes. Ibid. 4; 21-58. f.
1-10. 1894.
, (97). La reproduction sexuelle des Ascomycetes. Ibid, 5; 245-284.
f. 1-17. 1897.
, (07). Recherches sur le développement du périthéce chez les Ascomy-
eetes. Ibid. 10: 1-385. pl. 1-91. 1907.
Darbishire, O. V. (700). Uber die Apothecienentwickelung der Flechte Physcia
pulverulenta (Schreb.) Nyl. Jahrb. f. wiss. Bot. 34; 329-345. pl. 11. 1900.
Davis, B. M. (’96). The fertilization of Batrachospermum. Ann. Bot. 10: 49-76.
pl. 6-7, 1896.

, (03). Odgenesis in Saprolegnia. Bot. Gaz. 35: 233-249, 320-349.
pl. 9-10. 1903.

1915]
ATKINSON——-PHYLOGENY IN THE ASCOMYCETES 371

Dodge, B. O. (712). Artificial cultures of Ascobolus and Aleuria. Mycologia
4: 218-222. pl. 72-78. 1912.

——, (7122). Methods of culture and the morphology of the archicarp in
certain species of the Ascobolaceae. Bull. Torr. Bot. Club 39: 139-197. pl.
10-15. f. 1-2. 1912.

——., (714). The morphological relationships of the Florideae and the
Ascomycetes. Ibid. 41; 157-202, f. 1-13. 1914.

Hidam, E. (’80). Beitrag zur Kenntniss der Gymnoasceen. Beitr. z. Biol, d. Pfl.
3: 267-305. pl. 12-15. 1880.

, (83). Zur Kenntniss der Entwicklung bei den Ascomyceten. Ibid.
3: 376-433. pl. 19-23. 1883.

» (86). Basidiobolus, eine neue Gattung der Entomophthoraceen. Ibid.
4: 181-251. pl. 9-12, 1886.

Fairchild, D. G. (’97). Ueber Kerntheilung und Befruchtung bei Basidiobolus
ranarum Hidam. Jahrb. f. wiss. Bot. 30: 285-296. pl. 13-14. 1897.

Faull, J. H. (705). Development of ascus and spore formation in Ascomycetes.
Boston Soc. Nat, Hist., Proc, 32: 77-113. pl. 7-11. 1905.

——, (711). The cytology of the Laboulbeniales. Ann. Bot. 25: 649-654.
1911.

, (12). The cytology of Laboulbenia chaetophora and L. Gyrinidarum.
Ann. Bot. 26: 325-353. pl. 37-40. 1912.

Ferguson, Margaret C, (’01). The development of the egg and fertilization in
Pinus Strobus. Ann. Bot. 15: 435-479. pl. 22-25. 1901.

, (04). Contributions to the knowledge of the life history of Pinus with
special reference to sporogenesis, the development of the gametophytes and
fertilization. Washington Acad. Sci., Proc. 6; 1-202. pl. 1-24. 1904.

Fisch, C. (’82). Beitrige zur Entwickelungsgeschichte einiger Ascomyceten. Bot.
Zeit. 40: 851-905. pl. 10-11. 1882.

Frank, A. B. (’83). Ueber einige neue und weniger bekannte Pflanzenkrankheiten,
II, Polystigma rubrum. Ber. d. deut. bot, Ges. 1:58-62. 1883.

, (86), Ueber Gnomonia erythrostoma, die Ursache einer jetzt herr-
schenden Blattkrankheit der Siisskirschen im Altenlande, nebst Bemerkungen
liber Infection bei blattbewohnenden Ascomyceten der Biiume iiberhaupt.
(Vorlaiufige Mittheilung.) Ibid. 4; 200-205. 1886.

Fraser, Miss H. C. I. (707). On the sexuality and development of the ascocarp in
Lachnea stercorea, Ann. Bot. 21: 349-360, 1907.

, (08). Contributions to the cytology of Humaria rutilans Fries. Ann.
Bot. 22: 35-55. pl. 4-5. 1908.

, (13). The development of the ascocarp in Lachnea cretea, Ibid. 27:
553-563. pl, 42-43. 1913.

————, and Brooks, W. E. St. John (’09). Further studies on the cytology of
the ascus. Ibid, 23: 537-549. pl. 34-40. f. 1. 1909.

— , and Chambers, H.S. (707). The morphology of Aspergillus herbariorum.
Ann. Mye. 5: 419-431. pl. 11-12. 1907.

, and Wéisford, E. J. (708). Further contributions to the cytology of
the Ascomycetes. Ann. Bot. 22: 465-477, pl. 26-27. 1908.

[Vou, 2
372 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Gates, R. R. (709). The stature and chromosomes of Oenothera gigas De Vries.
Archiv. f. Zellforsch. 3:525-552. 1909.

, (13). Tetraploid mutants and chromosome mechanisms. Biol. Cen-
tralbl. 33: 92-150. f. 1-7. 1913.

Guilliermond, A. (08). La question de la sexualité chez les Ascomycetes. Rey.
Gen. Bot. 20: 32-39, 85-89, 111-120, 178-182, 298-305, 333-334, 364-377.
f. 1-86. 1908.

————., (09). Recherches cytologiques et taxonomiques sur les Endomycetées.
Ibid. 21: 354-391, 401-419. pl. 13-19. 1909.

» (712). Les levures. 1-565. f. 1-163. Paris, 1912.

Harper, R. A. (’95). Beitrag zur Kenntniss der Kerntheilung und Sporenbildung
im Ascus, Ber. d. deut. bot. Ges. 13:(67)—(68). pl. 27. 1895.

———,, (95a). Die Entwickelung des Peritheciums bei Sphaerotheca Castagnei.
Ibid. 13: 475-481. pl. 89. 1895.

, (796). Ueber das Verhalten der Kerne bei der Fruchtentwickelung
einiger Ascomyceten, Jahrb. f. wiss. Bot. 29: 655-685. pl. 11-12. 1896.

, (99). Cell-division in sporangia and asci. Ann. Bot. 13; 467-525, pl.
24-26, 1899.

, (00). Sexual reproduction in Pyronema confluens and the morphology
of the ascocarp. Ann. Bot. 14; 321-400. pl. 19-21. 1900.

, (02). Binucleate cells in certain Hymenomycetes, Bot. Gaz. 33: 1-35.
pl. 1. 1902.

, (05). Sexual reproduction and the organization of the nucleus in
certain mildews. Carnegie Inst. Washington, Publ. 37: 1-104. pl. 1-7. 1905.

Hartog, M. M. (’95). On the cytology of the vegetative and reproductive organs
of the Saprolegnieae. Roy. Irish Acad., Trans. 30: 649-708. pl. 28-29. 1895.

Hoffmann, A. W. H. (712). Zur Entwicklungsgeschichte von Endophyllum sem-
pervivi. Centralbl. f. Bakt. IT. 32: 137-158. pl. 1-2. f. 1-14. 1912.

von Héhnel, F. (710). Fragmente zur Mykologie. X. Mitteilung. K. Akad. Wiss.
Wien., Math.- naturw. Kl., Sitzungsber. 119: 393-473 (1-81). f. 1. 1910.

Janczewski, E. (’71). Morphologische Untersuchungen iiber Ascobolus furfur-
aceus. Bot. Zeit. 29: 257-262, 271-278. pl. 4. 1871.

Juel, H. O. (702). Taphridium Lagerh. & Juel. Eine neue Gattung der Protomy-
cetaceen. Bihang K. Sv. Vet.- Akad. Handl. 27**: Afd. III. 1-29. pl. 1. 1902.

, (02). Uber Zellinhalt, Befruchtung und Sporenbildung bei Dipodascus.
Flora 91: 47-55. pl. 7-8. 1902.

Karsten, G. (’08). Die Entwicklung der Zygoten von Spirogyra jugalis Kizg.
Flora 99; 1-11. pl. 1. 1908,

Kihlman, O. (°83). Zur Entwickelungsgeschichte der Ascomyceten. Soc. Sci. Fen-
nicae, Acta 13; 1-43, pl. 1-2. 1883,

Kniep, H. (’13). Beitriige zur Kenntnis der Hymenomyceten, I, II. Zeitschr. f.
Bot. 5: 593-637. pl, 2-5. f. 1. 1913.

Kurssanow, L. (’11). Ueber Befruchtung, Reifung und Keimung bei Zygnema.
Flora 104: 65-84. pl, 1-4. 1911.

Lagerheim, G. de (92). Dipodascus albidus, eine neue, geschlechtliche Hemiascee,
Jahrb. f. wiss. Bot. 24; 549-565. pl. 24-26. 1892.

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 373

Lindau, G. (’88). Ueber die Anlage und Entwicklung einiger Flechtenapothecien.
Flora 71: 451-489. pl. 10. 1888.

, (99). Beitrige zur Kenntniss der Gattung Gyrophora, Festschrift
fiir Schwendener. Berlin, 1899.

Lotsy, J. P. (’07). Vortrige iiber botanische Stammesgeschichte 1: I-IV and
1-828. f. 1-430. 1907.

Maire, R. (’99). Sur les phénoménes cytologiques précédant et accompagnant la
formation de la téleutospore chez le Puccinia Liliacearum Duby. Compt.
rend. acad. Paris 129; 839-841. 1899.

, (01). L’evolution nucléaire chez les Urédinées et la sexualité. Bull.
Soe, Myc. 17: 88-96. 1901.

» (02). Recherches cytologiques & taxonomiques sur les Basidiomycetes.
Ibid. 18; 1-209. pl. 1-8. 1902.

» (03). Recherches cylotogiques sur le Galactinia succosa. Compt.
rend. acad. Paris 137: 769-771. 1903.

, (05). Recherches cytologiques sur quelques Ascomycetes. Ann. Myc.
3: 123-154. pl. 3-5. 1905.

Marchal, El. et Em. (’09). Aposporie et sexualité chez les Mousses, II. Bull.
acad, Belg. (classes des Sciences) 1909: 1249-1288. 1909.

——,, (711). Ibid. III. Ibid. 1911: 750-778. f. 1-19. 1911.

McCubbin, W. A. (710). Development of the Helvellineae. I. Helvella elastica.
Bot. Gaz. 49: 195-206. pl. 14-16. 1910.

Miicke, M. (708). Zur Kenntnis der Hientwicklung und Befruchtung von Achlya
polyandra de Bary. Ber. d. deut. bot. Ges. 268: 367-378. pl. 6. 1908.

Murrill, W. A. (’00). The development of the archegonium and fertilization in

the hemlock spruce (Tsuga canadensis Carr.). Ann. Bot. 14: 583-607. pl. 21-
22. 1900.

Nichols, M. A. (96). The morphology and development of certain pyrenomycetous
fungi. Bot. Gaz, 22: 301-328. pl. 14-16. 1896.

Nichols, S. P. (704). The nature and origin of the binucleated cells in some
Basidiomycetes. Wis. Acad. Sci., Trans. 15: 30-70. pl. 4-6, 1904.

Nienburg, W. (707). Beitrige zur Entwicklungsgeschichte einiger Flechtena-
pothecien. Flora 98: 1-40. pl. 1-7. 1907.

, (714). Zur Entwicklungsgeschichte von Polystigma rubrum DC.
Zeitschr, f, Bot. 6; 369-400. f. 1-17. 1914.

Olive, E. W. (05). The morphology of Monascus pupureus. Bot. Gaz. 39: 59-60.
1905.

——,, (’07). Cell and nuclear division in Basidiobolus. Ann. Myce. 5; 404—
418. pl. 10. 1907.

, (08). Sexual cell fusions and vegetative nuclear divisions in the
rusts. Ann. Bot. 22: 331-360. pl. 22. 1908.

Oltmanns, F. (’98). Zur Entwicklungsgeschichte der Florideen. Bot, Zeit. 56:
99-140. pl. 4-7. 1898.

————, (04). Morphologie und Biologie der Algen 1; 1-733. f. 1-467. Jena,
1904. :

Osterhout, W. J. V. (’00). Befruchtung bei Batrachospermum. Flora 87; 109-
115. pl. 5. 1900.

[Vou. 2
374 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Overton, J. B. (06). The morphology of the ascocarp and spore-formation in the
many-spored asci of Thecotheus Pelletieri. Bot. Gaz. 42: 450-492. pl, 29-30.
1906.

Raciborski, M. (’96). Studya Mykologiezne (Mycologische Studien I. Karyo-
kinese bei Basidiobolus ranarum, Absidia robusta nov. sp., Penicillium Poir-
aultii nov. sp., Entyloma Nymphaeae Cunningham). Akad. d. Wiss., Krakau,
Anz. 1896: 377-386. 1 pl. 19 f. 1896.

Ramlow, G. (’06). Zur Entwicklungsgeschichte von Thelebolus stercoreus Tode.
Bot, Zeit. 64: 85-99. 1906.

, (14). Beitriige zur Entwicklungsgeschichte der Ascoboleen. Myce.
Centralbl. 5: 177-198. pl. 1-2, f. 1-20. 1914.

Ruhland, W. (’01). Zur Kenntnis der intracellularen Karyogamie bei den Basi-
diomyceten, Bot. Zeit. 59; 187-206. pl. 7. 1901.

Sachs, J. (768). Lehrbuch der Botanik. 1-632. f. 1-465. Leipzig, 1868.
————., (74). Ibid. 1874.

, (796). Physiologische Notizen X, Phylogenetische Aphorismen und
iiber innere Gestaltungsursachen oder Automorphen. Flora 82; 173-223.
1896.

Sappin-Trouffy, M. (’96). Recherches histologiques sur la famille des Urédinées.
Le Botaniste 5; 59-244. f. 1-70. 1896.

Schikorra, W. (’09). Ueber die Entwicklungsgeschichte von Monascus. Zeitschr.
f. Bot. 1:379-410. pl. 2. f. 1-3. 1909.

Schmidle, W. (’99). Einiges iiber die Befruchtung, Keimung und Haarinsertion
von Batrachospermum. Bot. Zeit. 57: 125-135. pl. F. 1899.

Schmitz, F. (’79). Ueber die Fruchtbildung der Squamarieen. WNiederrhein.
Ges. f. Nat.- u. Heilkunde, Bonn, Sitzungsber. 36: 376-377. 1879.

————, (’80). Ueber die Zellkerne der Thallophyten. Ibid. 37; 122-132. 1880.

, (83). Untersuchungen iiber die Befruchtung der Florideen. K. Preuss.
Akad. Wiss., Berlin, Sitzungsber. 1883: 215-258. pl. 5. 1883.

————., und Hauptfleisch, P. (’97). Rhodophyceae. In Engler & Prantl,
Nat. Pflanzenfam. 1°: 298-544. f. 192-288. Leipzig, 1897.

Stahl, E. (’77). Beitriige zur Entwickelungsgeschichte der Flechten. 1-55. pl.
1-4. Leipzig, 1877.

Stevens, F. L. (99). The compound odsphere of Albugo bliti. Bot. Gaz. 28: 149-
176, 225-245. pl. 11-15. 1899.

, (01). Gametogenesis and fertilization in Albugo. Bot. Gaz. 32: 77-98,
157-169, 238-261. pl. 1-4. f. 1. 1901.

Stomps, T. J. (712). Die Entstehung von Oenothera gigas deVries. Ber. d. deut.
bot. Ges. 30: 406-416. 1912.

(707). Eremascus fertilis nov. spec. Flora 97; 333-346. pl. 11-12. f.

907.

Strasburger, E. (700). Uber Reduktionsteilung, Spindelbildung, Centrosomen und
Cilienbildner im Pflanzenreich. Histolog. Beitr. 6: 125. 1900.

Stoppel, R.
1-6. 1

——— , (04). Uher Reduktionstheilung. K. preuss. Akad. Wiss. Berlin, phys.-
math. K1., Sitzungsber. 18: 587-615. f. 1-9. 1904.

1915]
ATKINSON—PHYLOGENY IN THE ASCOMYCETES 375

» (05). Typische und allotypische Kernteilung, Ergebnisse und Erdér-
terungen. Jahrb, f. wiss. Bot. 42: 1-71. 1905.

» (09). Sexuelle und apogame Fortpflanzung bei Urticaceen. Jahrb. f.
wiss. Bot. 47; 245-288. pl, 7-10. 1909.

Thaxter, R. (96). Contribution toward a monograph of the Laboulbeniaceae.
Am. Acad., Mem. 12: 189-429. pl. 1-26. 1896,

» (08). Contribution toward a monograph of the Laboulbeniaceae. II.
Ibid. 13: 219-469, pl. 28-71, 1908.

Theissen, F. (’12). Die Gattung Clypeolela v. Héhn. Centralbl. f. Bakt. II. 34:
229-235. 1912,

» (712). Fragmenta brasilica IV nebst Bemerkungen iiber einige andere
Asterina-Arten, Ann. Myc. 10: 1-32. f. 1-5. 1912.

—, (712). Fragmenta brasilica V nebst Besprechungen einiger palaeo-
tropischer Microthyriaceen. Ann. Myce. 10: 159-204. 1912.

» (13). Lembosia-Studien. Ann. Mye, 11: 425-467. pl. 20. 1913.

» (13). Hemisphaeriales. (Vorliugfige Mitteilung.) Ann. Mye, 11:
468-469, 1913.

——, (’13). Uber einige Mikrothyriaceen. Ann. Myce. 11; 493-511. pl. 21.
fale LOS.

» (718). Die Gattung Asterina in systematischer Darstellung. K.K.
zool.-bot. Ges., Wien, Abhandl. III. 7: 1-130. pl. 1-8. 1913.

=—, (13). Zur Revision der Gattungen Mycrothyrium und Seynesia.
Osterr. bot. Zeitschr. 63: 121-131. 1913.

, (14). Trichopeltaceae n, fam. Hemisphaerialium. Centralbl. f. Bakt.
II. 39; 625-640. pl. 1. f. 1-7. 1914.

———,, (’14). Uber Polystomella, Microcyclus, u. a. Ann. Mye. 12: 63-75.
pl. 6-7. 1914.

Treub, M. (’05). L’apogamie de l’Elatostema acuminatum Brogn. Ann. Jard. Bot.
Buitenzorg II. 5: 141-152. pl. 4-11. 1905.

Tréndle, A. (’°07). Ueber die Kopulation und Keimung von Spirogyra. Bot. Zeit.
65*: 187-210. pl. 5. f. 1-13. 1907.

Twiss, W. C. (711). Erythrophyllum delesserioides J. Ag. Univ. Calif. Publ.
Bot. 4: 159-176. pl. 21-24. 1911.

van Tieghem, Ph. (’84). Culture et développement du Pyronema confluens. Soe.
Bot. France, Bull, 31: 355-360. 1884.

De Vries, H. (’03). Die Mutations-Theorie 1;I-XIV and 1-752. pl. 1-2. f. 1-159.
1901; 2: 1-XII and 1-648. pl. 1-8. f. 1-181. 1903.

» (13). Gruppenweise Artbildung unter spezieller Beriicksichtigung der
Gattung Oenothera I-VII and 1-365. pl. 1-22. f. 1-121. 1913.

Vuillemin, P. (’12). Les champignons. Essai de classification, 1-425. Paris,
1912.

Welsford, E. J. (07). Fertilization in Ascobolus furfuraceus, New Phytol. 6:
156-161. pl. 4. 1907.

Werth, E., and Ludwigs, K. (712). Zur Sporenbildung bei Rost- und Brandpilzen.
Ber. d, deut. bot. Ges. 30: 522-528. pl. 15. 1912,

[Vou. 2, 1915]
376 ANNALS OF THE MISSOURI BOTANICAL GARDEN

Wolf, F. A. (712). The perfect stage of Actinonema Rosae. Bot. Gaz. 54; 218-234.
pl. 13. 1912.

Wolfe, J. J. (704). Cytological studies on Nemalion. Ann. Bot. 18; 607-630.
pl. 40-41. f. 51. 1904,

Woronin, M. (66). Zur Entwickelungsgeschichte des Ascobolus pulcherrimus Cr.
und einiger Pezizen. In deBary und Woronin, Beitr. z. Morph. u. Physiol. d.
Pilze 2: 1-11. pl. 1-4. 1866.

, (70). Sphaeria Lemaneae, Sordaria coprophila, fimiseda, Arthrobotrys
oligospora. Ibid. 3: 1-36. pl. 1-6, 1870.

Woycicki, Z. (’04). Einige neue Beitriige zur Entwicklungsgeschichte von Basi-
diobolus ranarum. Flora 93; 87-97. pl. 4. f. 1. 1904.

Yamanouchi, 8. (’06). The life history of Polysiphonia. Bot. Gaz. 42: 401-449.
pl, 19-28. 1906.

Zukal, H. (89). Entwicklungsgeschichtliche Untersuchungen aus dem Gebiete der
Ascomyceten. K. Akad. Wiss., Wien, Math.- naturw. K1., Sitzungsber. 98:
520-603. pl. 1-4. 1889.

ad f

Ti

3 5185 00064