Reads
i

ha

weet

Mut)

:

° 4 Miva
Ree a bie
UEP) f i) ny)

ne

=

ue

or

srs

Senn

MURRAY PRINTING COMPANY,

TORONTO.

COMMITTEE OF MANAGEMENT,

Chairman: JAMES Loupon, M.A., LL.D., President of the
University.

Proressor. W. ALEXANDER, M.A., Ph.D.
PRINCIPAL J. GALBRAITH, M.A.

Proressor A. H. Reynar, M.A., LL.D.
Proressor R. Ramsay Wricut, M.A., B.Sc.
PROFESSOR GEORGE M. Wrona, M.A.

General Hditor: H. H. LANGTON, B.A., Librarian of the University.

THE GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.*

( Reprinted by permission from the Transactions of the
Canadian Institute, 1896-97 ).

Ie

ON account of their subterranean and inconspicuous prothallus and
the slow germination of their spores, the literature on the subject of the
sexual generation of the Ophzoglossacee is somewhat scanty.

Hofmeister’ was the first to give an account of the gametophyte in
this group. His friend Irmisch sent him specimens of the very young
sporophyte of Botrychtum Lunarza in 1854. On visiting the spot where

ERRATA,

Instead of ‘‘latter”’ in 19th. line of page 6. read ‘‘ former.”

Instead of “‘phlegmaria”’ in 29th. line of page 6. read ‘‘ Phlegmarza.”

Instead of ‘‘ Rhabenhorst” in foot-note of page 12. read ‘‘ Rabenhorst.”’
Instead of “‘phlegmaria”’ in 17th. line of page 14. read ‘‘ Phlegmaria.”
Instead of ‘‘ hypobasal” in 3rd. line of page 28. read ‘‘ epibasal.”

Figures 7, 8, 9, 10, 11, 12, 13 and 14 are all lithographed from photographs.

tem, green tips. 1ne Tourtm Irona conrormed to the usual type, and
probably made its appearance in the next period of vegetation. From
the situation of the embryo on the lower surface of the prothallus, the

*Most of the material for this investigation was secured by means of a grant from the Elizabeth
Thompson Scientific Fund.

r, Abhand, d k. Sachs. Gesellschaft d. Wissch. Bd. ii., pp. 657-662.

THE GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.*

( Reprinted by permission from the Transactions of the
Canadian Institute, 1896-97 ).

He

ON account of their subterranean and inconspicuous prothallus and
the slow germination of their spores, the literature on the subject of the
sexual generation of the Ophzoglossacee is somewhat scanty.

Hofmeister’ was the first to give an account of the gametophyte in
this group. His friend Irmisch sent him specimens of the very young
sporophyte of Botrychtum Lunarza in 1854. On visiting the spot where
the young plants had been discovered, he found other examples, some
of which were still attached to the maternal prothallus. The latter, he
describes as being oval in shape and about a millimetre in length, of
light brown colour externally, and yellowish white in section. The
cells were filled with clumps of material not of a starchy nature.
Antheridia were found mainly on the upper surface, the archegonia
being situated below. Root-hairs were sparingly interspersed among
the sexual organs. The antherozoids resembled those of the other
Filicinee, but were about one-half larger in size. The archegonia were
sunk almost level with the surface of the gametophyte. One prothallus
was found still attached to its spore, but attempts to germinate other
spores, under observation, were unsuccessful. No young embryos were
obtained, nor was it possible to study the development of the sexual
organs. As a result of the inferior position of the archegonza, the young
sporophyte appeared on the lower surface of the prothallus. The root
grew out first, indeed two roots often made their appearance, before the
first leaf became visible. The latter was bract-like and colourless. The
two following leaves resembled it, but they had, either one or both of
them, green tips. The fourth frond conformed to the usual type, and
probably made its appearance in the next period of vegetation. From
the situation of the embryo on the lower surface of the prothallus, the

* Most of the material for this investigation was secured by means of a grant from the Elizabeth
Thompson Scientific Fund.

1, Abhand, d k. Sachs. Gesellschaft d. Wissch. Bd. ii., Pp. 657-662.

4 JEFFREY: GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.

growing shoot was forced to make a half turn to assume its normal,
negatively geotropic position.

In 1856, Mettenius? published an account of the sexual phase of
Ophioglossum pedunculosum, which he found in considerable quantities,
in the earth of the pots containing the adult spore-plants. Attempts to
germinate the spores, under observation, failed also in this case. The
youngest prothallia were tuber-like in shape, and one to three milli-
metres in thickness. Out of the tuber grew subsequently a conical
process which elongated considerably (four to fifty millimetres), and
sometimes branched. At the tip of the outgrowth, or of its ramifications,
was found an apical cell, sometimes at least, of triangular pyramidal
shape. The cylindrical portion of the prothallus grew upwards towards
the surface of the soil, but, on reaching the light, became green and
died away at the apex, or divided into two or three lobes which
flattened out on the earth and developed no further. The tuber was
composed of starch-laden parenchyma. In the process some textural
differentiation was found, there being an axial, elongated, starch-free
strand, surrounded by short starch-bearing cells. Both kinds of sexual
organs were found in the same plant and not arranged in any definite
order, but generally situated on the cylindrical process. The anthertdia
were large in size and their wall was generally two layers of cells in
thickness. The antherozoids were large also, and composed of one
and a-half to two spiral turns. The anthertditum opened by a pore
produced by the breaking away of two superimposed cells in its wall.
The aperture was generally situated in that part of the wall nearest the
apex of the prothallium. The spermatozoids swarmed out of the
mother cells and about in the cavity of the antherzdium before making
their way out. The archegonza originated from two superficial cells, the
upper of which gave rise by repeated divisions to a neck of three to five
tiers of cells; the lower formed the axial row, which were not, however,
made out individually by this writer. On account of the small number
of embryos found, it was impossible to follow stage by stage, their
development. Nothing was noted in regard to the formation of the first
dividing walls. The youngest embryo was oval in shape and already
segmented into a number of cells. The older ones were similar in
configuration, but of larger size. The anterior end of the elliptical
embryo grew through the tissues of the prothallium towards its apex,
and bursting forth sooner or later, became the cotyledon, green in
colour, and lanceolate in outline. The root developed more slowly and
bored its way directly outwards. A rounded protuberance at the

2. Filices Horti Botanici Lipsiensis, pp. 119-120.

JEFFREY: GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM. 5

junction of the cotyledon and root, probably the foot, fastened the
young sporophyte to the base of the archegonium. The apical bud
appeared sometimes at the point of union of root and leaf, and some-
times further down on the root, thus simulating the adventitious buds
arising from the roots of the adult plant.

The most recent contributions to our knowledge of this group is due
to the discovery of the gametophyte of Botrychium virginianum by
Professor Douglas Campbells at Grosse Isle, Michigan, in 1893. The
prothallia were unfortunately, like those of Hofmeister’s Botrychium
Lunaria, which they resembled in appearance, although larger in size,
too old for the study of the development of the sexual organs and
embryo. They are described as being flattened tubers with folded
upper margins, covered with root-hairs and bearing the reproductive
organs on the superior surface. Brown externally, white in section, the
lower part of the gametophyte harboured an endophytic fungus. The
archegonia had rather long and straight necks, while the antheridia were
quite endogenous like those of Hgu¢setum and Marattia. No young
embryos were found, but only advanced young sporophytes, bearing
already the first or a subsequent leaf.

Professor Campbell was the first to bring about the germination of
the spores in this group. The process is exceedingly slow, requiring,
even in the warm climate of California, for Botrychium virgintanum,
eighteen months or more, and for Ophdoglossum pendulum, somewhat
less than that time. The most advanced stages yet obtained by him,
had only undergone two or three divisions. Chlorophyll was found in
the young prothallium of Botrychium virginianum, and a suspicion of
chlorophyll in that of Ophzoglossum pendulum. This may have been
due merely to the fact that germination took place in the light.

As there has been a tendency in recent years to associate the
Ophioglossee with the isosporous Lycopodinee, it is necessary to state
briefly what is at present known concerning the gametophyte in the
latter group. Fankhauser+ discovered in 1872 the brown subterranean
prothallus of Lycopodium annotinum. The examples found by him were
lobed, tuber-like, and marked by numerous ridges and depressions.
Antheridia and fully formed sporophytes were found on them, hence the
prothallia must have been moncecious. In 1884, Bruchmanns found
some much younger prothallia. These were of oval and flattened form,

3. Trans. British Association, Oxford Meeting, 1894. Structure of Mosses and Ferns, 1895, pp. 224-228
4. Bot. Zeitung. 1873. No.1.
5. Bot. Centralblatt. Bd. i., 1885, pp. 23-28.

6 JEFFREY : GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.

the superior margin being raised so as to produce a depression in the
centre. The antheridia occupied ridges in the bottom of this basin.
No archegonia were present, nor did the plants show a definite apical
meristem. The same observer remarked that the inferior cells of the
prothallus were occupied by an apparently symbiotic fungus, the
mycelium of which communicated with the outside by means of the
root-hairs with which the plants was provided. He referred the
symbiont to the genus Pythium. More recently Treub® has published
a description of the prothallium of Lycopodium cernuum. Here the
gametophyte, as in Ophzoglossum pedunculosum, starts from a primary
tubercule, and divides subsequently into green lobes. The sexual
organs have no definite arrangement and are moncecious. The arche-
gonta possess a single uninucleate canal-cell. The large antheridia have
a single-layered outer wall and produce biciliate moss-like antherozoids.
The embryo is peculiar in the possession of a rudimentary suspensor.
The stem in the young sporophyte is at first represented by a paren-
chymatous mass which has been designated the primary tubercule.
The first division in the embryo is transverse and gives rise to the
epibasal and hypobasal cells. The latter originates first the cotyledon ;
the stem-apex apparently not developing till after several leaves have
grown out. The first root also is derived from this segment, but only
after a number of foliar organs have unfolded. The prothallus in this
case was likewise occupied by a symbiotic fungus, which was considered
by the author to be a species of Pythzwm.

Goebel? about the same time described the sexual phase of another
species, Lycopodium inundatum. It closely resembled Lycopodium
cernuum in structure, and also harboured a fungus resembling Pythzum.
Treub® has also published an account of another form, viz: Lycopodium
phlegmaria, which is slender, much branched, and entirely subterranean.
It is especially interesting on account of the occurrence of a number of
canal-cells in the avchegontum and from the presence of paraphysis-like
growths among the antherzdza.

II.

In 1895, the writer came upon a large number of prothallia of Botry-
chium virginianum in a Sphagnum-swamp behind the village of Little
Metis, in the Province of Quebec. The presence of these plants was
revealed by the greenish-yellow cotyledons appearing above the surface

6. Etudes sur les Lycopodiacées. Annales du Jardin botanique de Buitenzorg. Tome iv., v., vii.
Viii., 1884-1890.

7. Bot. Zeitung. 1887. No. 11-12.

8. Op. Cit:

JEFFREY : GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM, 7

of a slight depression in the moss. On removing some of the overlying
vegetation, numbers of the larger prothallia were easily obtained. It
required, however, careful sorting of the peaty soil with the fingers to
secure the younger and more interesting stages. Nearly a week was
spent in working over about half the bed, the result being several
hundred examples in all stages of development, of the gametophyte and
attached sporophyte. Subsequently, in another season, a week was
spent on the spot, and all the plants which careful sifting of the soil
would yield, were removed. The second harvest amounted to over six
hundred specimens, by far the larger number of which, however, were
much too old for study. During the same summer, other and older
plants were found in rich woods about two miles back of Metis. In
the spring of 1896, additional discoveries were made in Foster’s Flats,
below the Whirpool, on the Niagara River, and on the east branch of
the river Don, a few miles from Toronto. The last mentioned spot
proved rich in interesting examples of older stages of the attached
sporophyte. Most of these were removed last autumn (1897).

IIT.

One of the greatest difficulties in the way of the present research,
was the proper preservation of the prothallia. They are singularly
impermeable to fixing reagents on account of the thick external cuticle,
and must be cut at intervals with a razor, to allow the preserving
medium to penetrate. The presence of oil in large quantities in the
tissues, also renders aqueous fluids useless, as they scarcely make their
way in at all. A saturated solution of picric acid in thirty per cent.
alcohol, gave fairly good results ; but the best fixation was obtained by
using a mixture of three parts of a saturated solution of corrosive sub-
limate in ninety per cent. alcohol, and one part of saturated solution of
picric acid in the same menstruum, diluted with distilled water to
reduce the alcohol to thirty per cent. strength. The same reasons
which rendered the material hard to preserve, made it difficult to
embed. Paraffine was mainly used, and the most satisfying results
were obtained by infiltrating with benzole, in a vertical tubular dialyzer
with a chamois leather diaphragm, revolved slowly by means of clock-
work. It was found that the ordinary type of stationary dialyzer was
quite unsuitable for these very delicate objects. When the prothallia in
alcohol were placed in the top compartment, and the benzole below, the
osmosis was exceedingly slow; and, if the position of the media was
reversed, the weight of the benzole carried it through too rapidly, and
injurious shrinkage was the result. The continued reversing of the

8 JEFFREY: GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.

relative positions of the two liquids by the clock movement, and the
accompanying agitation, were found to overcome these inconveniences.
Unfortunately, this device was hit upon only after numerous experi-
ments, and when the investigation was almost completed. The trans-
ference from benzole to paraffine was effected in a stationary dialyzer,
or by evaporating off the benzole in a water-bath, from a ten per cent.
solution of paraffine in benzole. Celloidin embedding has also great
advantages, but as the material has to be cut into slices not thicker than
two millimetres at most, and as the prothallia were often nearly twenty
millimeters in length, it was only employed for sections through certain
regions of the gametophyte, and for the much less impenetrable young
sporophyte. The stains chiefly used were either a combination of
alum-cochineal and eosin, or aqueous saffranin, made by dropping a
small amount of saturated alcoholic solution of equal parts of Griibler’s
alcohol and water soluble saffranins. This last method seems worthy
of a wider application.

IV.

The youngest prothallia obtained were already two millimetres in
length by one and a-half in breadth. As may be seen from figure 1,
they are of flattened oval shape, and covered with hairs. The growing
point is at the narrow thin end, and the prothallium thickens and
widens from thence backwards. Axthertdia alone are found at this
stage, and are entirely confined to the upper surface of the gametophyte.
They form a cluster at the older end, but thin out into a narrow median
row as they extend forward towards the growing point, figure 1, av.
In somewhat larger and older plants, the median row of antheridia is
raised on the crest of a distinct ridge, and the archegonia begin to make
their appearance upon its sides, figure 2, The antheridial ridge is a
marked feature of most of the older prothallia, and must have the same
significance in the process of fertilization as the inferior archegonial
prominence possesses in the leptosporangiate /’zdzcznee. In more
mature individuals the ridge is obliterated, especially in the posterior
region of the prothallus, by the more rapid growth of the sides of the
latter, which seems to bea provision for the nourishment of the fertilized
archegonia. This phenomenon probably is the cause of the antheridial
ridge not being noticed by Campbell % Figure 3 shows a plant in
which an embryo, em., has already reached a considerable size. The
antheridial prominence is still very marked; the root-hairs, however,
have largely disappeared. In figure 5, we have a somewhat younger
stage with the rhizoids still abundantly present, especially in the

g. Op. Cit.

ma Oo}

JEFFREY : GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM. 9

younger anterior region of the prothallus. Figure 4 is of a lobed game-
tophyte ; figure 6 shows a similar condition in which two embryos,
em. 1, em. 2, are to be seen. The depression of the antheridial ridge in
the posterior region by marginal growth is particularly well-marked.
These lobed forms are quite abundant among the Metis specimens, but
the Toronto plants did not manifest this peculiarity. I am inclined to
believe that the conditions of life in the two cases may have been the
cause of this difference. The Metis specimens were found in wet, peaty
soil. The Toronto plants, on the contrary, grew in rich, yet rather dry,
forest mould. Older lobed prothallia have almost invariably two
sporophytes attached to them. In figure 7, is represented an example
in which the first root of the young sporophyte has reached a consider-
able size. At this stage the axis of the young sporophyte, which, in
earlier phases, is nearly always at right angles to that of the prothallus,
becomes often more or less oblique, as in the example figured. This
rotation of the axis is probably due to the continued growth of the
prothallium after the formation of the embryo. Figure 8 shows a
prothallium in which two roots of the attached sporophyte have grown
to a considerable length, although the cotyledon is short and still
unfolded. In figure 9, we have a small gametophyte with only one
root, and yet having the cotyledon fully expanded. The first leaf may
expand either after one, two, or three roots have been formed, according
to the vigor of the plant, and may always be recognized by its seeming
to grow out of the proximal end of the first and stoutest root. Figure
10, is of a strong plant with three precotyledonary roots. The lamina
of the cotyledon is not bilaterally symmetrical, as in most of the
Filicinee, but of the palmate type represented by Ophzoglossum pedun-
culosum. As may be seen from figures 9 and 10, the first leaf varies
considerably in complexity in accordance with the greater or less
robustness of the plant from which it originates. In the next drawing,
figure II, is represented a lobed prothallium, on which are two older
sporeplants, deprived of the leaves of the year of their collection.
Figure 12 shows a Toronto specimen, bearing two well-advanced
sporophytes. Figure 13 is a representation of a bifurcated sporeplant,
two examples of which have been found. Figure 14 is interesting, for
it represents a sporophyte which has already developed the fertile
ventral segment, and is yet still attached to the mother prothallium.
The sporeplant in this case is eight years old, as indicated by the
number of foliar lacunz in the fibro-vascular cylinder. There seems to
be little danger of error in drawing this inference, for a considerable
acquaintance with the young sporophyte enables me to state positively,
that never more than one leaf is developed at a time, and in all

10 JEFFREY : GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.

probability, only one in a year. Attached sporophytes, five or six
years old, are sufficiently common, as has been already stated in the
preliminary notice.'°

The prothallia described in the foregoing account were from two
to twenty millimetres in length, and from one and a-half to fifteen
millimetres in breadth. The gametophyte of &. vzrginzanum is thus
considerably larger than any geophilous prothallus which has yet been
described. Attempts have been made to germinate the spores of this
species, but although these are still undecayed, no signs of growth have
yet made their appearance after eighteen months. Professor Douglas
Campbell got them to sprout in less time than this, but doubtless the
warmer climate of California had some influence in hastening the process.
He found a few large chloroplasts in the young plants; but it seems
probable that the presence of chlorophyll here is accidental, and depends
on the spores being sown contrary to the natural conditions, in the light-
An analogous phenomenon occurs when potato tubers are grown under
conditions of illumination. Most of the prothallia collected by the
writer were found ten centimetres or more below the surface of the soil.
Mature sporophytes have been dug up, with the foot-tubercle still intact,
and buried often thirty centimetres in the ground. These facts make it
very difficult to imagine that the tubercular, deeply subterranean,
gametophyte of 2. uvzrginzanum can have been preceded by a green
aerial phase as are the quite superficial, colorless, gametophytic buds of
Vittaria, Trichomanes and Hymenophyllum described by Goebel, or the
larger tuberlike, resting phase of the liverwort Geothallus recently studied
by Campbell. It is perhaps worth while to suggest that the slow
germination of the spores in the case of Pteridophyta, with subterranean
prothallia is an adaptation to enable the former to reach a favorable
depth in the substratum, before beginning their growth.

V.

A cross-section of the prothallus, such as is represented in figure 15,
reveals a number of important features. The antheridial ridge, x. is
seen above, containing several axtherzdia. On its sloping sides are the
archegonia, y. Multicellular hairs are often found attached to the ridge,
to its flanks and to the base of the prothallium. The position of several
of these is indicated in the figure at Z The intergal cells, a, 3atine
upper part of the plant appear light in color, and contain protoplasm
and small quantities of starch. The lower cells, 4., both in fresh and
stained sections, are dark-colored, and in their natural condition, filled

ro. Can. Inst. Proceed. Vol.i. Pt. 1, p.10. Annals of Botany. Vol. xi., p. 485.

JEFFREY : GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM. II

with a heavy oil which is not readily soluble in alcohol. They are
likewise occupied by a filamentous fungus which is presently to be
described. Figure 16 illustrates a median long-section of the prothallus.
At x. is seen the antheridial ridge cut lengthwise, and showing the
antheridia in various stages of development. The younger ones are
found nearer the anterior, sloping, apical region, ap. The distribution
of the fungiferous tissue is represented in this figure. It is to be noted
that it extends forward gradually, as the prothallus increases in length,
by the activity of the apical meristem. The fungus never occupies all
the cells on the lower side of the prothallus, but leaves free always a few
of the lower tiers. Above, as has been already stated, there is a
considerable mass of cellular tissue underneath the reproductive organs,
quite free from infection and containing a small amount of starch. The
symbiont is always present, as it has never been missed in the four or
five hundred plants which have been minutely studied. It is not possible
to state whether it is indifferent or beneficial to its host ; it certainly does
not seem to be injurious. The infected cells do not apparently suffer,
and perhaps the presence of oil in them, may be interpreted as an
indication of improved nutrition, Only experimental cultures can settle
this important question.

The growing region of the prothallus is always on the upper side,
figure 16, af. It is marked by the presence of a superficial layer of high
columnar cells like those found at the base of the apical incision of the
leptosporangiate gametophyte. These are represented in figure 17. One
of the columnar cells, @., is in all probability, the initial cell. It is very
difficult to secure exactly horizontal sections of the apical region except
in very young plants, of which my supply was somewhat limited. These
were all used up for longitudinal and transverse series, and I am accord-
ingly unable to describe the horizontal configuration of the initial cell.

The root-hairs are from one to four millimetres in length and are often
multicellular, especially when they arise from the crest or flanks of the
prothallium. Those which originate from the base are unicellular and
longer than the others. These rhizoids are generally about twenty
micra in width and are more or less completely cutinized. It is chiefly
through them that the symbiotic fungus makes its way into the prothal-
lium. The passage of the fungal ype through the cutinized wall of the
root-hair, is marked by the formation of thick sheaths which surround
the Zyphe for ten or more micra of their course. These sheaths are
apparently only formed where the fungus has to penetrate an already
cutinized wall, and one does not find the phenomenon repeated as the
hyphe pass successively through the walls of the internal cells of the

12 JEFFREY : GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.

host-plant. Figure 21 represents a broken root-hair, the basal wall of
which has become cutinized and consequently forms a sheath where the
hypha is passing through. The penetration of the next cell-wall inwards
is unaccompanied by this phenomenon. In figure 18 can be seen part
of a root-hair, c., on the lateral walls of which are two sheaths, and the
hair in this case being intact, sheaths are not formed in the uncutinized
basal wall. In the same figure sheaths can be seen at 4 and d@, where the
fungus has passed in through ordinary superficial cells of the prothallus.
This is apparently of rare occurrence.

After penetrating about two or three layers of cells, y, the symbiotic
filaments, which are from two to four micra in diameter, begin to grow
luxuriantly, and fill the succeeding strata of cells, 4, with a much-coiled
mycelium. If this be examined with a good apochromatic objective, it
is possible to discover that it is by no means always filamentous, but
that in many cases, the ype expand into large thin-walled vesicles,
which are often so abundant that they fill the cells with a botryose mass
resembling a Completorza, figure 19,0 and c. In other cells the filaments
prevail, zbzd. a. It is not difficult to satisfy oneself that the Zyphe and
vesicles belong to one and the same mycelium, figure 19, 6. Frequently
some of the vesicular structures become ruptured and shrivel up, zdzd.
Figure 20 shows a freshly infected cell of the prothallus, highly magni-
fied, in which the vesicular structures have just begun to form. Often
the advance of the symbiont through the prothallus is marked by the
penetration of filaments or by a mixed growth of Ayphe@ and vesicles into
new cells. Another kind of organ is also found in the mycelzum, viz.,
contdia. These are thick-walled and from fifteen to twenty micra in
diameter. They are generally formed at the end, but sometimes, though
rarely, in the course, of a Zypha, and are filled with a dense, coarsely
granular protoplasm. The contents of the conzdzum are not separated
from the filament by a septum and thus resemble the conzdza of the sub-
form Aphragmium™ of the genus Pythzum. The conzdium germinates 2m
situ, forming a tube which often makes its way into the adjoining cells of
the host-plant. I have never been able to detect the formation of
zoospores from these conzdia, and indeed it is difficult to imagine how
they could serve as a means of distribution for so completely endopara-
sitic a fungus. The stages of formation and germination of the conzdza
are shown in figure 22, a, 6 and «.

It will be seen from the above account that the symbiont of Botrychium

virgintanum presents several rather remarkable characteristics. In its
mode of penetration it resembles Completorta complens, as described by

11. Rhabenhorst, Krypt, Flora. Fischer, Phycomyceten, p. 397-

JEFFREY : GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM, 13

Leitgeb™ in the prothallia of Péerzs cretica, Aspidium falcatum, and other
ferns ; the formation of the dark brown sheath from the cell-wall of the
host-plant being very characteristic. Atkinson's has described a similar
phenomenon for a Comp/letoria found in the same species of prothallia
in America. In the filamentous portion of the undivided mycelzum as
well as in the formation of its conzdza it markedly resembles a Pythzum.
In the botryose vesicular masses completely filling the cells of the host,
it again strikingly simulates Comp/letoria. It may perhaps fairly be
considered as a form uniting the genera Pythium and Completoria. If,
on further investigation, the above view proves to be correct, it may
possibly be necessary to remove Completorta from the vicinity of the
Entomopthorace, where it has been placed on account of its ejaculatory
conidia by Nowakowski and Thaxter, and toreplace it with the Perono-
sporacee where Leitgeb, as a result of his careful investigation, con-
sidered it to belong.

The endophyte of the prothallium of Botrychium virginianum, unlike
that of Lycopodium cernuum, described by Treub,* and that of Z.
annotinum, described by Bruchmann,‘s is always intracellular and never
becomes intercellular, in the deeper layers of the host-plant. Treub’s
description is somewhat brief, but from the fuller account of Bruchmann,
the structure of the szycelzum in the symbiont of Lycopodium seems to
be quite different from that of the form found in Botrychium virginta-
num.

Only further study of the fungus can settle whether it is a distinct
species of Completoria or Pythium, or, on the other hand, an intercalary
species. Before leaving this subject, there is one more interesting fact
to record. In older prothallia bearing well-advanced sporophytes, the
symbiont is shrunken and dead. Whether this state of affairs is rightly
comparable to the similar phenomena observed by Frank in the
mycorhize and mycodomatia of various Phanerogamia, at the time of
flowering or seeding, and is to be considered as a digestion of the sym-
biont by its host, must for the present be left in suspense. The
prothallia often continue to live long after the death of the endophyte.
Nothing of the nature of an cogonium has yet been observed in any stage
of development of the fungus.

VI.

The antherzdia arise, after the first basal cluster has been formed, figure

12. Sitzungsberichte d. Akad. d. Wissch. Wien. Math.—Natwissch. Classe. Bd. 84. Abth. i., 1881,
p- 291 and p. 307.

13- Bull 94. Cornell Experimental Station, p. 52, 53-

14. Op. Cit. i., p. 124.

15. Op. Cit. pp. 310-313.

14 JEFFREY : GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.

I, always on the crest of the antheridial ridge, figure 23. The older
antheridia are found generally higher on the ridge than the younger ones,
figure 23, a’, a?, a. The first indication of the male organ is a richly
protoplasmic superficial cell, which divides transversely, giving rise to a
shallow outer cell and a deep inner one, figure 23 a. The former
becomes transformed into the outer wall of the antheridium, and the
latter originates by repeated divisions, the mother-cells of the anthero-
zoids. In figure 24 is represented a young stage in which both the inner
and outer cells have already undergone several divisions. When the
anthertdium attains about a third of its ultimate size, its outer wall is
doubled by periclinal divisions. In figure 25 these are represented as
just beginning. Subsequently, the mass of spermatocytes is shut off
internally from the prothallium cells by further periclinal divisions, figure
23, a7, @. Often the antheridia are accompanied by short multicellular
hairs, resembling those found on the rest of the surface of the prothallus
and comparable to the paraphyses described by Treub in Lycopodium
phlegmaria, figure 26, par. The more primitive mother-cells of the
antherozoids possess large nuclei with numerous nucleoli, figure 27, a.
After a number of simultaneous divisions of the spermatogenic tissue,
the definite spermatocytes are formed. In these the reserve chromatin
in the form of nucleolihas disappeared. The filar chromatin is arranged
in what appears to be a true reticulum. When the formation of the
antherozoids begins, the nucleus contracts somewhat and the bars of the
chromatic veczculum become thickened, figure 27 6. The nucleus then
assumes a lateral position, and begins to flatten out, figure 27 ¢c. This
process is continued, and by the lengthening out of the nucleus, the
condensation of its chromatin, and the curvature produced byits position
in the cell, the antherozoid is formed, figure 27d. The interesting
structure to which Webber’ in his recent studies on the antherozoids of
the Cycadee, has applied the name dlepharoplast, and which he compares
with the cilia-forming body lately discovered by Belajeff'7 in the
filicinee and Eguisetinee has been looked for in the developing anthero-
zoids of Botrychium virginianum, but has not been made out. This
is probably due to the fact that osmic acid fluids could not be used as
fixing reagents on account of the oil in the tissues, and because the stains
employed were not those used by Belajeff, but either a combination of
alum-cochineal and eosin, or aqueous saffranin alone. The material
illustrative of spermatogenesis was somewhat limited in amount, and it
was not thought advisable to risk the series by removing their covers

16. Bot. Gazette. Vol. xxiv., p. 233.

17, Ueber Nebenkern in Spermatog. Zellen u. d. Spermatogenese d. Farnkrautern. Berichte d.
deutsch. Bot. Gesell. Bd. xv., pp. 337-339. Idem —Die Spermatogenese d. Schachtelhalm. Ibid. Bd. xv.
PP- 339-342.

JEFFREY : GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM, 15

and re-staining with the reagents employed by Belajeff. The writer
hopes to secure more young prothallia in the coming summer, in which
event it will be possible to come to a decision on this important point.

The fully developed antherozoid forms a spiral of one and a-half turns
and has the structure usual in the F7dzccenee. The cilia come off from
the attenuated, anterior end of the spiral. I could not decide, from the
preserved examples which were the only ones I had the opportunity of
examining under high magnification, the exact length of the ciliary
region. The antherozoids, like those of Ophzoglossum pedunculosum
described by Mettenius"’, escape from the mother-cells while still within
the antheridium. They swim about freely in its cavity, figure 28, a and
6: sometimes still retaining their protoplasmic vesicles and in other
instances being already freed from them, figure 27, e7and e?. The
spermatozoids make their way out by means of an aperture formed by
the disappearance of two superimposed cells of the outer wall of the
antheridium. They do not escape all at once, as is quite generally the
case, but seem to be voided in several swarms, at intervals, under undis-
covered conditions. The cavity of the axtherzdium is filled with a thin
gelatinous matrix, resulting, probably, from the disintegration of the
spermatocytic walls, figure 28, a and 8.

Vil.

As has already been stated, the avchegonza originate on the flanks of
the median ridge of the prothallia, figure 15, v7. The youngest stage of
the archegonium is a single, richly protoplasmic, superficial cell, which,
as in the anztheridium, divides subsequently into an outer shallow cell
and an inner deeper one, figure 29. The former gives rise to the neck
of the archegonzum, and the latter to its axial row of cells. The next
stage is the horizontal division of the inner rudiment which separates
from it the large basal cell, figure 30. The superficial rudiment sub-
sequently begins to divide, first, by anticlinal walls, figure 31; and then
by periclinal ones, figure 32; thus forming the neck. The richly
protoplasmic basal cell divides, figure 32 ; and then the upper axial cell
undergoes a division, which results in the formation of the cervical
canal-cell and the ventral cell; figure 33 and figure 34. In the latter
figure is seen a paraphysis, a, which is in reality, only one of the multi-
cellular hairs common over the whole surface of the younger parts of
the prothallium. In figure 35, the nucleus of the cervical canal-cell has
divided, and as may be seen in the next figure 36, the nuclear division

18. Op. Cit.

16 JEFFREY : GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.

is not followed by the formation of a cell-wall, such as has been
described by Farmer and Campbell in Angiopteris, Marattia, and
Osmunda. From the study of many hundred archegonia in this stage
of development, the statement is made with some confidence that such
a wall is never present in Botrychium virginianum. In figure 37, is
represented an archegonium in which the ventral canal-cell has made its
appearance. One very rarely finds this canal-cell intact, as it quickly
disintegrates and in preserved material, at any rate, is represented by
an indistinct mass thrust against the wide base of the cervical canal-
cell. In figure 38, is seen a ripe archegonium which has ejected its
canal-cells. The apical cells of the neck are, as is usual in the Pteri-
dophyta, thrust outwards. At the same time one frequently notices
chromatolysis in the nuclei of the upper cells of the archegonial neck,
figure 37, although this phenomenon is by no means invariably present.

The mature egg is large and possesses a very dense protoplasm,
which however, generally encloses a hydroplastid. The free surface of
the oosphere rises into a median elevation, the receptive prominence.
Figure 38, was drawn from a preparation in which a single spermatozoid
had entered the canal of the archegonium. It has not been possible to
follow the stages of union of the sexual nuclei. After fertilization, the
canal is generally occluded by the closing together of the neck cells,
figure 39, although this is by no means invariably the case, figure 40.
The oospore grows to many times its original size before the first
division takes place. Figures 39 and 40, represent two stages of the yet
undivided oospore. In figure 41, the first segmentation has occurred,
and the basal wall is horizontal, as in the other eusporangiate Pteri-
dophyta. In figure 42, the embryo has become divided into quadrants
by the median wall, which is the next to appear, and which, in the
majority of cases at least, is parallel to the long axis of the prothallium.
The transverse wall next makes its appearance at right angles to the
other two. In figure 43, is represented an embryo which has already
undergone further divisions. The upper octants have been sub-divided
before any similar activity has appeared in the lower segments. There
is no indication of a suspensor, and as the lower part of the embryo is
not loaded with food materials, it seems probable that the earlier
divisions in the upper octants, are for the purpose of thrusting the
young sporophyte deep into the prothallium, that it may be more easily
nourished and attain its characteristically large size without exposure to
injury. The divisions are not always so regular, as in the case of the
embryo represented in figure 43. In some instances, the basal wall is
rather oblique, and corresponding differences exist in the orientation of

JEFFREY: GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM. 17

the ensuing divisions, figure 44. Quite often, too, no regular course of
segmentation can be made out at all, as in figure 45. When the embryo
is only a little larger than those figured in 43, 44 and 45, the basal,
median, and transverse walls are quite obscured by subsequent divisions.
It is not possible to detect any indication of apical initials such as
commonly occur in the early phases of the leptosporangiate sporophyte,
and such as have also been described in some, at least, of the eusporan-
giate Pteridophyta. The next phase which is chosen for representation,
is that in figure 46. Although no apical cells could be made out in
this preparation and others of the same age, there is in the example
figured, a very considerable formation of periclinal walls in the upper
internal region of the embryo. The whole lower portion of the young
sporophyte forms the foot, figure 46 f In figure 47, is shown an
embryo in which the root and shoot have already become differentiated.
The periclinal activity already referred to, has led to the formation of a
large amount of tissue in the upper portion of the embryo, and this is
supported on the broad basis furnished by the foot. A high merismatic
epidermis has already become differentiated at x, the cells of which are
very rich in protoplasm and have the elongated columnar configuration
of the shoot meristemeta of most of the Pteridophyta. Among these,
the one marked a seems to be the initial cell. At y, is a protuberance
which is the outward indication of the first root. Within this, at 0, is
the apical cell of the root, distinguished by its darkly-stained proto-
plasm, and by the fact that it has just undergone its first periclinal
division. The condition of the embryo of Botrychium virginianum at
this stage, is remarkable in that the stem-apex appears before the first
leaf. The cotyledon is consequently derived from the shoot meristem,
just as the later leaves are, but as in the case of the latter, it is not
possible to follow the changes in the meristem leading to the formation
of the foliar rudiment. The difficulty is greater in the case of the
cotyledon, on account of the comparative paucity of younger embryos
which have been cut exactly axially. For this investigation nearly
three hundred series o prothalli, from two to twenty millimetres in
length, have been sectioned. In spite of this not inconsiderable labor,
less than twenty per cent. proved to be of value, either because no
embryos were present, which is very commonly the case; or being
present, they were not cut in a truly median plane. The surface of the
gametophyte presents such irregularities that the proper orientation of
the younger phases of the embryo is entirely a matter of chance. So far
as I am aware the embryo of the Eguzsetacee presents the only other
case yet described, in which the primitive foliar organ is secondarily

18 JEFFREY: GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.

derived from the shoot-apex. Sadebeck makes the following state-
ment concerning the equisetaceous embryo :—“ Nach meinen Unter-
suchungen bin ich vielmehr zu dem Resultat gekommen, dass die obere
Halfte des noch zweizelligen Embryo ganz unmittelbar die primare
Axe darstellt, aus welcher sich in gleicher Weise, wie spater bei der
erwachsenen Stammknospe die Blatter erzeugen.”

The embryo of Jsvetes echinospora, as described by Campbell, also
resembles in a measure that of 2. virgintanum. It has a large foot
originating from doth the hypobasal quadrants, which by its position and
size, at least, somewhat strikingly resembles that of Botrychium. In the
case of the latter, it is quite impossible to state from which of the primi-
tive divisions of the fertilized egg, the foot takes its origin. A resem-
blance also exists in the formation of the root and shoot from the upper
part of the embryo. In J/. echinospora, however, the cotyledon is the
first shoot-organ to appear, and the stem-meristem does not definitely
develop until later, although there is an indication of its existence from
the first.

It is not to be supposed, however, that these resemblances are in any
way to be considered as indicative of relationship, for the development
of the embryo may vary greatly in the same natural group. In the
Marattiacee, for example, both Angiopteris and Marattia, as described
by Farmer*t and Campbell” are distinguished by the precocious
development of the cotyledon. In Danea, 3 on the other hand, it is the
root which first shows considerable development. A somewhat similar
state of affairs has been observed by the writer in the Eguzsetacee.
Eiquisetum arvense and E. hiemale have a precocious root, whilst Z.
Limosum and E. palustre develop first the shoot-organs. Among the
Ophtioglossacee themselves, in Ophioglossum pedunculosum, the cotyledon
is the first organ to rupture the calyptra. In Botrychium virginianum
and 4. Lunaria, the root is prior in appearance.

In figure 48, is represented an embryo, which, although larger, is yet
younger than that in figure 47. At a@and 6 are probably the root and
shoot initials. Figure 49 is an older stage than figure 47. The root, 7, is
already well advanced and its apical region is fully developed. Behind

19. Die Entwick, d. Keimes d. Schachtelhalme. Pringsheim. Jahrbucher t. Wiss. Botanik.
Bd. xi., p. 582.

20. Annals of Botany, vol. v., p. 244.
21. Annals of Botany, vol. vi., p. 265.
22. Annals of Botany, vol. viii.

23. Brebner, G. On the Prothallus and Embryo ot Danea simplicifolia. Annals of Botany, vol. *.,
Pp: 199.

JEFFREY : GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM, 19

its terminal meristem are elongated cells which, later, give rise to fibro-
vascular tissues. The cotyledon, ¢, is also for the first time visible, and
beside it is the stem-meristem, s. Below is the very massive foot, *
Figure 50, lithographed from a photomicrograph, represents a still later
stage of development. Here the root is almost ready to burst the
calyptra, cal. The cotyledon is distinctly seen, and at this stage, for the
first time, covers over the stem-apex, which now lies on the side of a
transverse fissure. No vascular tissue appears till the root has grown to
a length varying from five to twenty millimetres, and has burst the
calyptra. he first tracheides arise in the proximal region of the root
after ‘t has emerged from the prothallium. Subsequently they make
their appearance in the cotyledon and the stem-axis.

Before referring to the further developmental changes in the nascent
sporophyte, it will be well to consider an interesting abnormality. In
figure 51 is represented part of a prothallus in which tracheides are
present, near a region of superficial decay. The decayed spot probably
marks the position of an embryo which has been injured and in con-
sequence has rotted away, So far as I have been able to learn, by
reference to the literature on the subject, such prothallial tracheides are
the invariable accompaniment of apogamy. Their presence was first
described in connection with this phenomenon by Farlow*+ in the
apogamus prothallia of Pterzs cretica. ‘hey have since been seen by
many observers under similar conditions. Lang’s has recently found
them in the interesting reduced, apogamous, sporangiferous sporophytes
of Lastrea dilatata, Presl, var. Cristata gracilis, Roberts and Scolopendrium
vulgare, L., var. ramulosissimum, Woll. According to Bower, tracheides
also occur in the prothallia [endosperm] of certain Cycads. In view of
the recent discoveries of antherozoids in the pollen-tubes of this group,
it would be interesting to know if the Cycads also manifest the phenom-
enon of apogamy.

The example figured is the only occurrence of prothallial tracheides
which has come under my notice in examining a large number of
gametophytes. In this case both antheridia and archegonia were
present. Recently an example of apogamy in Péerzs aquilina has come
under my observation in which an apogamous and a normal embryo
were produced side by side on the same archegonial pad. The former
was accompanied by a single prothallial tracheid. The apparent rarity
of the phenomenon in Botrychtum virgintanum may be due to the con-
ditions under which the Metis specimens, which I have almost exclusively

24. Quarterly Journal of Microscopical Science, vol. xiv., N.S., p. 266.

25. Annals of Botany, vol. xi., pp. 157-168 ; also, Proc. of Royal Society of London.

20 JEFFREY : GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.

investigated, grew. They were found as has been already stated,
virtually submerged in a peat-bog, and as a consequence, absence of
proper water supply which has been noticed as a predisposing cause of
apogamy, would not make itself felt. Possibly prothallia from the rich,
rather dry soil of the Don valley might yield a greater number of
examples. If we may infer apogamy from the presence of prothallial
tracheides, the gametophyte of Botrychium virginianum is unique among
the eusporangiate vascular Zoidogama, in this respect ; unless the phe-
nomenon is shown to be present in the tracheid-bearing Cycad endos-
perms described by Bower, and apogamy can no longer be considered
as peculiar to the leptosporangiate Fzlicznea.

Returning to the young sporophyte, the shoot-organs and the root
possess fairly well marked apical cells, as is shown by Campbell** to be
true also of the mature spore-plant. Figure 52 represents the terminal
meristem of the young stem in vertical section. At @ is probably the
apical cell. In figure 53 the same region is shown in horizontal section.
In figure 54 is the apex of the cotyledon in longitudinal section. Figure
55 represents a long section of the apex of the first root in an embryo
which has not yet broken through the calyptra. A large primary
segment is found on the side of the jpzleorhiza,a state of affairs rarely
seen in later stages of the root, as subsequently the small cells of the
inner part of the root cap abut immediately on the apical cell. This is
possibly to be explained by the comparatively slight development of the
bileorhiza which consequently requires only very occasional contributions
from the apical initial. The root of Botrychium virgintanum is an endo-
trophic sycorhiza and, as has been shown by Frank, there is a tendency
to degeneracy in the root-cap of roots of this type. The apical cell is
much more active on its flanks although even here it divides slowly,
compared with the apical initial of the leptosporangiate F72/zcznee@. In
figure 56 the root-apex is seen in transverse section, and unlike that of
the stem, its initial cell is triangular in this plane.

Figure 57 shows an interesting case of polyembryony corresponding
to that described by Treub?7 in Lycopodium cernuum. It was first
noticed after a series had been made of what appeared externally to be
a bifurcated embryo. The central cylinders of two plants, a and 4, are
shown; a@ is larger and much more abundantly supplied with reserve
food-materials, which cause it to stain more intensely ; 0 is smaller, less
developed, and in a condition of malnutrition as is indicated by a cor-
responding paleness of hue; a? is the second root of embryo a, and is

26. Campbell. Mosses and Ferns; pp. 232, 235.

27. Etudes sur les Lycopodacées ; Extrait vi., p. 11.

JEFFREY : GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM, 21

quite fully matured ; a3 is the foliar trace of the cotyledon, which is just
béing separated by a layer of decidual periderm; 4 is the central
cylinder of a, with the trace of the second leaf just making its appear-
ance; 4? is the still embryonic second root of the smaller embryo 6;
63 is the young cotyledon and y is the central cylinder. Figure 58
represents a lower section in the same series with the same lettering as
before ; a‘ is the primary root of the better developed embryo, and 0" is
that of the smaller embryo. At a? is a prominence indicating the point
of origin of the second root of the larger embryo. Figure 59 is of a
section still lower down and passes through the common foot of the
geminal sporophytes. The staining alone indicates the boundary
between the two plants. Their central cylinders are separate throughout,
but the fundamental tissues appear to be in textural continuity. A quite
sharp demarcation, however, is produced by the different condition of
nutrition of their cells; those on the side of a being loaded with starch ;
those of 4,on the other hand, containing only a very small amount.
Unwillingness to sacrifice the series prevented the use of the ordinary
methods of demonstrating protoplasmic continuity for the purpose of
discovering whether the protoplasm of the two was in reality continuous.
The phenomena of nutrition would seem to negative such a supposition.
Figures 57, 58 and 59 have been lithographed from photomicrographs.

The first root of the young sporophyte is sometimes diarchous, but
just as often triarchous. There seems to be no relation between the
vigor of the root and the number of protoxlyem-strands ; as depauperate
plants sometimes have three strands, and, on the other hand, robust
individuals often have only two. I have not found a single example of
a monarchous root in the large number of specimens which I have
examined. Figure 60 is a drawing of a section of a diarchous primary
root in aqueous analinsulphate. The endodermis a is quite distinct, and
shows plainly the characteristic radial lignified zones. Between it and
the vascular tissue are one or more layers of pericycle cells. The
protoxylem tracheides, x, are reticulate in their sculpture and not ringed
or spiral as is generally the case. The metaxylem elements almost
always meet in the centre. The bast, y, is made up of thick-walled
elements, some of which are sieve-tubes and the rest elongated
parenchyma cells. Between the bast and the vessels, is a considerable
amount of wood parenchyma. Often two or three diarchous roots are
formed, but sooner or later triarchous, and finally tetrarchous ones are
produced.

The central cylinder of the stem becomes fully differentiated below
the point of origin of the cotyledon. From the very first it has a well-

22 JEFFREY: GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.

marked pith, figure 61 #. The pith communicates with the external
fundamental tissue through a gap caused by the exit of the cotyledonary
trace, as has been described by Van Tieghem*. The internal endodermis
discovered in the younger portion of the stem of Lotrychium Lunaria
and others of the Ophioglossacee by Van Tieghem*” and Poiraults, is
not present in this species, although the external endodermis is well-
marked, only disappearing opposite the foliar gaps. The bast-tissue
originates first in the young central cylinder and seems never to have
any secondary additions from the activity of the camdzum. Graf zu
Solms3* has thrown doubt on the existence of secondary wood in the
Ophioglossacee@, but in this species there can be no uncertainty as to its
presence ; in fact, the wood is practically all secondary, as may be
learned from the radial arrangement of its matured elements and by
following the course of its development, figure 63 x, and figure 64 ~.
The first-formed wood-elements are reticulately sculptured and are never
of the ringed or spiral type. In this respect they resemble those of the
stem of the Warattiaceg, and, in fact, also those of the Osmundacee ; for
the groups of typical protoxylem elements found in the upper region of
the bundles of the latter, really belong to the leaf-traces. It ismore than
probable that the absence of typical primitive tracheary tissue in all
these cases, is due to the very slow growth of the stem,a phenomenon
which renders their presence unnecessary. The writer has noticed the
absence of these elements in the slowly growing stems of species of
so-called polystelic Primule, viz :—P. Auricula and P. farinosa.

During this investigation, the rather interesting observation has been
made, that the periderm-tissue first described in the Ophzoglossacee by
Russow?? and Holle33, is formed in Botrychium virgintanum at the bases
of defunct leaves, and thus is merely an abciss-layer. Figure 65, froma
photomicrograph, shows a young sporophyte still attached to its pro-
thallium ; 7 is the first root and x the base of the cotyledon ; 7? and # are
developing leaves. -As may be seen from the figure, the course of the
cotyledonary bundle +, has been interrupted by the intercalation of a
layer of periderm. Figure 66 shows the tissues in question under a
sufficiently high magnification to make clear the details of periderm
formation. By the continued growth of the latter the distal part of the

28. Remarques sur la structure de la tige des Ophioglossées. Journal de Botanique, iv., Année; p. 407
2g. Op. Cit.
30. Recherches sur les Cryptogames vasculaires. Annales de Sci. Nat. Bot. Tome xviii. ; p. r70.

31. Fossil Botany, p. 223.
3z. Mémdel'Acad. Imp. des Sciences de St. Petersbourg. vii.Serie. Tome xix., No. 1, p. 117.

33. Bot. Zeit. 1875. Ueber Bau u. Entwicklung der Ophioglosseen, p. 12.

JEFFREY: GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM, 23

leafstalk is forced continually outwards and eventually decays, leaving
no trace of its existence. This is the reason that, in transverse sections
of older stems, the foliar bundles of fallen leaves apparently disappear
before reaching the external cortex. The periderm formation of B.
virginianum is thus connected with the occlusion of the leafstalks, and
is probably to be explained as an adaptation for protecting the subter-
ranean stem from infection by the fungi of the soil.

In a transverse section through the older region of the stem, the
periderm is never found to forma continuous investiture as in the higher
plants, but is strictly localized in areas representing the points of origin
of former leaves. The writer has not yet had an opportunity of inves-
tigating whether the mode of cork formation obtaining in 4. vergenzanum
is common to the whole group, but it seems probable that this may
prove tobe the case. Periderm is also often formed both in the sporo-
phyte and in the gametophyte where surface injuries have occurred: a
striking case of correspondence between the two generations.

The cotyledonary trace originates from the central cylinder asa single
strand, figure 61, co¢.; but separates shortly after reaching the petiole into
two approximately collateral bundles. These pass upwards through the
long leafstalk into the lateral lobes of the lamina, one of them giving off
a bundle for the median lobe, exactly as in the postcotyledonary leaves
of many /7zlécone@. The endodermis is never quite continuous on the
inner side of the cotyledonary trace, and in subsequent leaves becomes
less and less marked, till at the stage in which there are four petiolar
bundles, it is entirely absent. Figure 67 represents the laminar portion
of the ninth leaf of a sporophyte which was still attached to its prothal-
lium. The fertile segment, f s., of the lamina is already present. This
plant was at the same time the oldest sporophyte still in connection with
the gametophyte, and the youngest already producing spores, which has
come under my notice during the present investigation.

In figure 68 is a still attached young sporophyte. Its prothallium is
infected with the already defunct symbiont, a. The spore-plant still
bears its cotyledon /;' and two younger leaves, /? and # are in the process
of formation. In the primitive root, 7,can be seen at x and y, certain
dark spots which are cells occupied by the sporophytic endophyte.
There is no resemblance between the latter and that of the gametophyte
as its mycelial filaments are much larger, being generally about eight
micra in diameter. There are no vesicles nor conzdza present, and in fact
the sterile mycelium is uniformly filamentous in character. These features
are reproduced in figure 69. The occurrence of a symbiont in the roots

24 JEFFREY: GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.

of the Ophioglossacee has long been known, and is mentioned by Russow
and Holle in the works already cited. The latter refers to its presence
or absence, the varying number of protoxylem groups in the larger and
smaller roots of Botrychium matricariefolium. In Lb. virgintanum this
explanation cannot be accepted, as, although the first formed roots vary
greatly in the number of archixyles, it is only in rare cases like that
figured in 68 that the fungus is present.

Vill.

The results of this investigation may be summarized as follows :—

(1). The gametophyte of B. virgznzanum is entirely subterranean,
without chlorophyll and probably symbiotic. It is from two to twenty
millimetres in length by one and a-half to fifteen millimetres in breadth,
and oval in outline, whether viewed from above or from the side.

(2). The whole surface of the plant is beset with rhizoids, which are
generally multicellular. The upper part of the gametophyte is occupied
in most prothallia, which have not yet produced embryos, by a median
ridge. The reproductive organs are found exclusively on the superior
surface, the antherzdia being situated on the crest of the ridge, and the
archegonia on its flanks.

(3). The gametophyte grows by a well-marked apical meristem which
is situated on the upper side, anteriorly, and apparently originates from 4
single initial cell.

(4). There is present in the lower part of the prothallus, an endophy-
tic fungus, possessing characteristics which will perhaps, on further
study, justify its recognition as a form intermediate between the genera
Pythium and Completoria. The symbiont is accompanied by a large
amount of oil, and probably advantageously affects the nutrition of the
prothallus. The fungus dies after one or more embryos have reached a
considerable size.

(5). The axthercdium originates from a single superficial cell and is
characterized by possessing a double outer wall. The antherozoids are
of the ordinary filicineous type and are rather large in size.

(6). The archegonium likewise takes its origin from a single super-
ficial cell. The neck consists of seven or eight tiers of cells. The
cervical canal-cell is binucleate, but is never represented by two cells.
A stratum of basal cells is present.

(7). The first division of the fertilized egg is transverse, as in the other
eusporangiate Pteridophyta. The identity of the octant walls which are

JEFFREY: GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM., 25

formed in the usual way, is early lost, and the embryo grows to a
relatively large size before the organs make their appearance. The root
and shoot originate from the upper part of the embryo ; and it may per-
haps be inferred that, like those of /svetes echinospora, they are derived
from the upper octants. The foot is formed from the whole of the lower
region of the embryo. The cotyledon is apparently derived secondarily
from the shoot meristem.

(8). The root, the stem, and the cotyledon grow by the segmentation
of a single apical cell, as in the adult plant. The root develops more
rapidly than the other organs; and the second or third root may make its
appearance before the cotyledon unfolds. The latter is green and cap-
able of assimilation, as in Ophzoglossum pedunculosum.

(9). The root-system ot the young sporophyte is soon occupied by a
symbiotic fungus, which differs in the size of its filaments and in several
other respects, from that found in the gametophyte.

(10). Evidence of apogamy has been found in the form of prothallial
tracheides.

(11). One example of polyembryony was observed.

(12). The sporophyte remains for a long time attached to the gameto-
phyte. It is an open question whether this is a primitive characteristic,
or merely an adaptation. The fact that the young sporophyte of the
much less robust &. Lunaria, according to Hofmeister’s account remains
for a very short period attached to its gametophyte, would seem to
justify the latter assumption.

IX,

In coming to any conclusions as to the bearing of this research on the
phylogenetic position of the Ophzoglossace@, due weight should be given
to the fact that the present species is the only one which has been some-
what fully investigated ; and the results of recent observations on the
Marattiacee, Lycopodiace@, and Egutsetacee show that a very consider-
able variety of development may exist even within the same natural
group. Moreover the saprophytic habit of the gametophyte of B. vzr-
gintanum has in all probability more or less profoundly modified its
structure.

It will be convenient to consider first the position of B. vzxgenzanum
in regard to the other representatives of the Ophzoglossacee which have
been studied. Its prothallus resembles very closely that of B. Lunarza,
and shows indications of being only a more specialized type. That this

26 JEFFREY : GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.

is the case is rendered probable by the strict localization of the antheridia
on the antheridial ridge, and by the occurrence of the reproductive
organs on the upper surface of the gametophyte. It is interesting in
this connection to note the scattered disposition of the axtheridza in the
very young prothallus; for this is probably to be regarded as a primitive
feature. An embryological comparison between the two forms is not
possible, as the embryology of B. Lunarza is at present unknown. The
young sporophyte of B. virg7nianum, in that it is attached to the upper
surface of the prothallus, and has a completely developed and assimila-
tory cotyledon, differs from the sporophyte of 4. Lunaria. The young
spore-plant also remains much longer attached to the gametophyte than
is the case in the latter species. B. virginzanum seems, of all the
representatives of the genus in Canada at least, to be the most com-
pletely adapted to modern conditions ; for it is everywhere abundant in
rich woods, and always outnumbers the other species.

The prothallus of Ophzoglossum pedunculosum does not very closely
resemble that of B. virg7nzanum. The presence of a primary tubercle
and the formation of green prothallial lobes are its characteristic
features. It should be remembered, however, that within the single
genus Lycopodium, L. annotinum resembles in its prothallus BL. vergen-
zanum and B. Lunaria, whilst L. cernuum and L. inundatum have a
gametophyte like that of Ophzoglossum pedunculosum. It is possible
that a species of Lotrychtum may yet be found in which the prothallus
is like that of Ophzoglossum pedunculosum. The antheridia and anthero-
zoids of the present species quite exactly resemble Mettenius’ description
of those of Ophioglossum pedunculosum. The archegonia correspond,
too, in so far as the earlier description offers points of comparison. In
the development of the embryo, the account of Mettenius is rather too
meagre to allow of any exact inferences in regard to points of likeness
in the successive phases of segmentation. The young sporophyte of
Ophioglossum pedunculosum develops its cotyledon early, and the
primary root is slow in pushing its way out, which exactly reverses the
course of events in 4. vzxg¢ntanum and probably also in &. Lunaria.

Bowers has recently fully discussed the relationships of the Ophzo-
glossacee to the other groups of the Pteridophyta. He comes to the
conclusion that the ventral fertile leaf-segment of the Ophzoglossace@ is
the morphological equivalent of the single ventral sporangium of the
homosporous Lycopodinee, and derives it from the former by a process
of septation and branching. He also compares the two groups in

34. Studies in the morphology ot spore-producing members. Part 2. Ophioglossacee, p. 56, et seq.

JEFFERY: GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM. 27

regard to the structure of the vegetative organs of the mature sporo-
phyte, and finds that in this respect they also show a marked resem-
blance to one another. Lastly, the organization of the gametophyte
and the development of the sporophyte, are discussed in the same con-
nection with a like conclusion.

It is only necessary in considering the results of the present investi-
gation, to examine the latter features. In regard to the structure of
the prothalli, the two groups certainly do present marked likenesses; eg.,
the gametophyte of Ophzoglossum pedunculosum to those of Lycopodium
cernuum and L. znundatum, and the gametophytes of &. Lunxaria and
b. virginianum to that of L. annotinum. It is quite possible, however,
that the resemblance in these cases is due to a similarity in environment.

The male organs of the two groups are in some important features
quite different. The anztherzdium has a double outer wall in the
Ophioglossacee and the antherzoids are spiral and multiciliate. In the
homosporous Lycopodinee, the antheridium has a simple outer wall, and
the antherozoids have the general configuration and the two cilia of the
antherozoids of the Bryophyta.

The archegonia of B. virgintanum at least, resemble those of the
Filicinee, (excluding /soetes, which probably does not belong here), in
having a basal cell and a single binucleate canal-cell, or at most two
neck canal-cells. On the other hand the Lycopodinee and Egutsetacee
are without the basal cell and have a decided tendency to increase the
number of cervical canal-cells. Too much importance should not, how-
ever, be attached to these structural features of the archegonza.

The embryo of &. virgintanum and apparently that also of O.
pedunculosum, lacks the suspensor and primary sporophytic tubercle
which are so characteristic of most of the isosporous Lycopodineag, and in
these defects resembles the /7zl/zccxee. So far as the facts in the case of
B. virginianum go, it seems probable that the Ophzoglossacee are much
more closely allied to the eusporangiate /zdzc7nee than to the isosporous
Lycopodinee, although they may be possibly the nearest of the mega-
phyllous Pteridophyta to that group. In all probability, the Ophzoglos-
sace@ are more primitive than the Marattzacee which they in some
respects resemble.

As a result of the fuller knowledge in recent years of the segmenta-
tation of the embryo ofthe Pteridophyta, it is scarcely possible to retain
any longer the conception of octants propounded by Leitgeb and others
when the leptosporangiate /z/zcznee were practically the only ferns inwhich

28 JEFFREY: GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.

anything of the embryology was known. Inthe homosporous Lycopodinee
the apex of the stem, the cotyledon, and the root, are all according to
Treub’s description, derived from the hypobasal half of the embryo. In
Isoetes, echinospora, the same three organs, according to Campbell’s
account, originate from the epibasal octants, the foot being formed
from all the hypobasal octants. No recent complete investigation
of the embryology of the Se/aginellee is available, but the phases
of development described by Pfeffer can only be harmonized with the
octant theory by something like a tour de force. In the Eguzsetacee,
according to Sadebeck, the shoot originates from the upper octants, and
the root and foot from the lower octants, the primitive leaves being
derived secondarily from the shoot meristem. The Ophzoglossacee, as
represented by &. wzrg7nianum resemble embryologically Jsoetes echino-
spora. The segmentation of the Marattiacee alone, agrees fairly well
with the stages of development found in the leptosporangiate Fz/zcnea,
and it is not very difficult in this group to refer the organs to definite
pairs of octants. But of all the eusporangiate forms, the Warattzacee
come closest to the leptosporangiates, and this probably is the explana-
tion of their embryological agreement.

If we are to accept the hypothesis that the eusporangiate Pteridophyta
are primitive, and if we follow Bower in deriving their sporophytic
phase from the progressive sterilization of the potential sporogenous
tissue of intercalary sporogonium-like forms, the axis is certainly to be
regarded as primitive, and the leaves and roots must be considered as
secondary outgrowths from the axis; either by eruption as Bower sur-
mises, or by some other undiscovered process. According to this con-
ception, foot and shoot are the primitive organs, and leaf and root are
subsequently derived from the latter. This view of the matter har-
monizes with what is known of the embryology of the lower eusporan-
giates. In the highly specialized leptosporangiates on the other hand,
a process of acceleration and rearrangement has been carried out and
the organs appear precociously, in definite relation, to the earlier
segmentations of the embryo.

In conclusion, the writer wishes to express his special obligations to
Professor G. L. Goodale of Harvard University for very kindly putting
at his disposal the books of the Gray Herbarium.

JEFFREY: GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.,

EXPLANATION OF PLATES.
PEATE is

Fic. 1.—Youngest prothallium found, av. antheridial ridge. x 8.

Fic. 2.—An older stage in which the antheridial ridge has become more marked.
x 16.

Fic. 3.—A considerably older gametophyte on which is a developing embryo, em.
The antheridial ridge, av, is particularly prominent. This prothallium is lithographed
from a photomicrograph. x 7.

Fic. 4.—A lobed prothallus from a photomicrograph. x 4.

Fic. 5.—From a photomicrograph ; represents a younger phase in which the root-
hairs are abundant. x 8.

Fig. 6.—A lobed prothallus lithographed from a photomicrograph, and bearing two
embryos, em and em*. x 4.

Fic. 7.—A ycung sporophyte showing the first root. x 8.

Fic. 8.—A young sporophyte showing two roots; the cotyledon is still unex-
panded. x 4.

Fig. 9.—A young sporophyte with the primary root and the cotyledon. x 1,
Fig. 10.—A stouter sporophyte with three roots and the cotyledon. x %.
Fic. 11.—A lobed prothallus bearing two advanced sporophytes. x 1.

Fig. 12.—A prothallus bearing two further advanced sporophytes. x ¥.
Fic. 13.—A bifurcated sporophyte still attached to its prothallium. x 4.
Fic. 14.—An eight year sporophyte still attached to its prothallium. x %.

Fic. 15.—-A cross-section of a prothallus showing the antheridial ridge, x; the
fungiferous cells, 4; and the uninfected cells, a. At y are the archegonia, and 4, root-
hairs. x 16,

Fic. 16.—A long-section of the prothallus ; lettering the same as in the preceding
figure. af, apical region. x 16.

Fic. 17.—Apical meristem. a, apical cell of prothallus. x 250.

Fic. 18.—Showing the penetration of the fungus into the gametophyte. c, root-
hair; 6 and d, superficial cells, in which the cutinized sheaths have been produced ;
x, fungiferous cells ; y, uninfected cells ; a, conidia. x 250.

Fic. 19.—Fungiferous cells ; a, with purely filamentous mycelium ; 6 and c, mixture
of filamentous and vesicular mycelium. x 600.

Fic. 20.—Cell showing the formation of vesicles, /, as outgrowths from a
hypha, 2. x 1,000.

30 JEFFREY: GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM,

PLATE II.

Fic. 21.—Base of a broken root-hair ; s, cutinized sheath ; 6, hypha of penetrating
fungus. x 1,000.

Fic. 22.—a, formation of conidium; 0, ripe conidium; c, germinating conidium.
X 1,000.

Fic. 23.—Antheridial ridge showing three antheridia in different phases of develop-
ment, a1, a7, anda*. x 2650.

Fic. 24.—An older antheridum. x 250.
Fic. 25.—A still older phase in which the outer wall is undergoing division. x 250.
Fic. 26.—Antheridial ridge showing the formation of paraphyses, far. x go.

Fic. 27.—Development of antherozoids ; a, young sperm-cells. x 500. 4, definite
spermatic mother-cells ; c, a later phase of the same, the nucleus is beginning to
become crescentic ; d, young antherozoids within the mother-cells ; e, ripe antherozoid.
In e1, the protoplasmic vesicle is still retained ; in e?, it has disappeared. x 1,000.

Fic. 28.—Matured antheridia showing the doubled outer wail; within, the anthero-
zoids are swimming in a gelatinous matrix. Ina, they are escaping. x 250.

Fic. 29.—First stage in formation of the archegonium. x 250.

Fic. 30.—A later phase showing formation of the basal cell. x 250.
Fic. 31.—Anticlinal division of the cervical rudiment. x 250.

FiG. 32.—Periclinal divisions of the cervical portion of the archegonium. x 250.
Fic. 33-—Nuclear division of the axial cell. x 250.

FIG. 34.—The same completed. A paraphysis ata. x 250.

Fic. 35.—Nuclear division of the cervical canal-cel xX 250.

Fig. 36.—The same completed. x 250.

Fic. 37.—Ripe archegonium, showing the ventral canal-cell. x 250.
Fic. 38.—-Opened archegonium with penetrating’antherozoid. x 500.
Fic. 39.—Fertilized egg. x 250.

Fic. 40.—The same older and larger. x 250.

Fic. 41.—First division of the embryo. x 250.

Fic. 42.—Formation of the median wall of the embryo. x 250.

Fic. 43.—An older embryo in which anticlinal divisions are present in the upper
octants. xX 250.

Fic. 44.—Another embryo of the same age, with oblique walls. x 250.
Fic. 45.—The same age as the foregoing, showing irregular segmentation. x 250,

Fic. 46.—A more advanced phase showing periclinal activity in the upper cells of
the young embryo at a; 4 is the foot region. x 250.

JEFFREY : GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM, 3l

PEATE aL.

Fic. 47.—An older embryo ; y, the root; x, the shoot; f, foot; a, initial cell of
shoot ; 4, initial cell of root. x 250.

Fic. 48.—A younger, but larger embryo than the foregoing, with the same lettering.
xX 250.

Fic. 49.—An advanced embryo ; 7, root ; c, cotyledon ; s, shoot ; f, foot. x 160.

Fic. 50.—From a photomicrograph. Lettering as before; ca/, calyptra. This
embryo is considerably older than the foregoing. x 50.

Fic. 51.—Part of a prothallium containing tracheides ; a4, decayed spot where an
embryo has probably disappeared ; /, tracheides. x 250.

Fig. 52.—Apical region of the shoot in vertical section; a, the initial cell. x 250.
FIG. 53.—The same, in horizontal section ; a, the apical cell. x 250.

Fic. 54.—Longitudinal section of the apex of the cotyledon ; a, apical cell. x 250.
Fic. 55.—Apical region of the primary root ; a, apical cell. x 250.

Fic. 56.—Tranverse section of the same ; a, apical cell. x 250.

Fic. 57. Transverse section of two united embryos, aand 4. a? is second root of
a; a*, cotyledon of a; x, central cylinder of a; 67, second root of 4; 6°, cotyledon of
6; y, central cylinder of 6. x50. (From a photomicrograph).

Fic. 58.--The same, a section through a lower region. Lettering as in the pre-
vious figure. 1, first root of a; 61, first root of 6. x 50. (From a photomicrograph.)

FIG. 59.—Section through the foot-region of the same embryos. Lettering as
before. x 50. (From a photomicrograph.)

Fic. 60.—Transverse section of a diarchous primary root: a, endodermis; x
xylem ; 7, phloém ; 4, parenchyma. x 250.

Piceaciay We

Fic. 61.—Transverse section of the young stem, above the exit of the cotyledonary
trace: cot. cotyledonary trace; ¢.c., central cylinder; m, medulla. x 50. (From a
photomicrograph).

Fic. 62.—The same, more highly magnified. x 160. (From a photomicrograph).

Fic. 63.—Part of the central cylinder of the foregoing, more highly magnified ; ¢7,
endodermis; y, phloém; x, xylem; cam, cambium; s. 4, sieve-tube ; m, medulla.
x 220. (From a photomicograph).

Fic. 64.—Part of central cylinder of quite a young plant; ez, endodermis ; pA,
phloém ; cab, cambium ; x, xylem ; 7.7. medullary ray,

Fic. 65.—Longitudinal section of an attached sporophyte; 7, primary root; x,
remains of cotyledon ; /? and /*, developing leaves. x 20. (From a photomicrograph).

32 JEFFREY: GAMETOPHYTE OF BOTRYCHIUM VIRGINIANUM.

Fic. 66.—The base of the cotyledon from the preceding, more highly magnified,
showing the formation of abciss-periderm at 7. x 160. (From a photomicrograph).

Fic. 67.—Lamina of an attached sporophyte, eight years old, showing the fertile
segment, f s., and sterile segment, s. 5. x 8.

Fic. 68.—Longitudinal section of an attached young sporophyte ; /1, cotyledon ;
/ and /°, developing leaves ; 7, primary root ; x and y, endophytic fungus of the sporo-
phyte. x 20. (From a photomicrograph).

Fic. 69.—Cells of the primary root, containing the fungus of the sporophyte.
xX 420.

Fic. 70.—Transverse section of a prothallus : a7, antheridial ridge ; em, an embryo.
x 20. (From a photomicrograph).

—S ee

Trans. of Canadian Institute Vol.V. Pt 2. Plate I.

Fig 10.

F Huth, Lith? Edin®™

photo. et de

Jeffrey

@®

rr ee ee ee ae CU ee

Trans.of Canadian Institute. Vol V. P'2. Plate II.

6) fate

E.C Jeffrey del F Huth, Lith? Edin®

Trans. of Canadian Institute. Vol-V, Pt 2, Plate III.

EC Jeffrey photo et del

F Huth, Lith® Edin?

bs ' ave
¢2, Plate lV

Vol.V,P

Trans. of Canadian Institute

F Huth, Lith® Edin*

EC Jeffrey photo, et del

ou [ered th & hs Se?
i

Sores iis oo 1 Ne
Dey 8

BIOLOGICAL, SERIES

Ad ® im. ‘

ne et) 3 ; f

tir UNIVERSITY LIBRARY, 1902: PUBLISHED BY
HE LIBRARIAN | ae

we es

“<7 \ = me Bee hy

ah Pes.

Ny oa Rear ty the,

a . x \ 4
» TRE PINT AR ARLE it litters oe fem

oe va See asta “2: ae

UNIVERSITY OF TORONTO
STUDIES.

BIOLOGICAL SERIES

NO. 2. THE ANATOMY OF THE OSMUNDACE/E

BY J. H. FAULL

THE UNIVERSITY LIBRARY, tg0z2: PUBLISHED BY
THE LIBRARIAN

Adina ‘ Chirman: James Loupon, M. A, LL.D., President of +
: ve hee University.
ti fo ALEXANDER, M.A. Ph.D.
_ PROFESSOR PELHAM Epcar, B.A., Ph. eee
_ PRINCIPAL f: GALBRAITH, M.A. i

_ PROFESSOR R. RAMSAY Wricut, M.A., B. Se.
‘PROFESSOR Si M. Wrone, M. A. ;

Riniaraz:
age Hi me, Prd
Geos pertain ae

SED ter, AO) Siar “hogy i
i
ceili Pa he dia . . ? D
y ee gp ed t he
LAN The ee wt ‘ ew
ts Pty Sie ibecanmh
1 oy alae “ :
Yeah pee ee ’ ayia
er te :
MPR Le oo
PP TREND | tet
we REA
Se eS ee ,
A ee + -
apace ane
5 A
x

a sf

Pe y

4 4 pyo wee

“ ‘ Crew a

‘ a :

ie i -

* ‘
. rel ds > * f iar at »
,
a < Shi a dite

PREFATORY NOTE

The work embodied in this article was done by Mr.
J. H. Faull, as a graduate student in Botany, in the
Biological Department of the University of Toronto dur-
ing the winter of 1900-01.

E. C. JEFFREY,
Lecturer on Biology.

\

TPA
care
‘ane

THE ANATOMY OF THE OSMUNDACEAE

By fr Fauir, BA.

We

Ty ‘
iG et MWe

BUC a
/ i

THE ANATOMY OF THE OSMUNDACEAE.
(Reprinted from the Botanical Gazette, Vol. XXXII *)
(WITH PLATES XIV—XVII)
INTRODUCTORY.

THE cauline vascular system of the Osmundaceae has
attracted considerable attention on the part of morphologists,
since it is exceptional among the leptosporangiate ferns in
exhibiting a type of structure presented by the phanerogams.
Thus DeBary, the exponent of the ‘bundle system,” states that
“collateral bundles”’ are with rare exceptions characteristic of
the stems and leaves of the phanerogams, but are likewise found
in the Osmundaceae,’ and that in their arrangement in the stems
of the Osmundaceae they follow the ‘dicotyledon type.’’?
Later we find Van Tieghem, the first enunciator of the ‘‘stelar
theory” expressing himself as follows :3

La tige des Osmundes et des Todées différe de celle des autres Fougéres.
La stéle axile et sans moelle du jeune age, au lieu de se diviser en restant
gréle, demeure simple en s’élargissant progressivement 4 mesure que la tige
grossit; elle prend une moelle de plus en plus large, a la périphérie de
laquelle sont rangé en cercle un certain nombre de faisceaux libéroligneux
a bois séparés, mais a libers confluents, entourés d’un péricycle commun et
d’un endoderme général. En un mot la tige de ces plantes demeure monos-
télique a tout Age, comme celle de la plupart des Phanérogames.

Plainly enough, therefore, these eminent botanists, starting
from very different conceptions, have arrived at the same con-
clusion, namely, that the central cylinder of the Osmundaceae
resembles that of the phanerogams.

It is important to note, however, that heretofore all anatomical
researches in this family have been confined to the tropical genus
Todea and the cosmopolitan Osmunda regals,; and that hence
the conclusion just stated has been based on the phenomena

* The original pagination is as follows: page 3 of this separate corresponds to
page 381 of the Botanical Gazette, Vol. 32, December, 1901.

tDeBary: Vergleichende Anatomie der Vegetationsorgane der Phaneroga-
men und Farne 331.

2 DEBARY: of. cit. 246. 3 VAN TIEGHEM: Traité de Botanique 1373.
[35]

4 FAULL: THE ANATOMY OF THE OSMUNDACEAE

presented by these alone. When Van Tiegham proposed his
‘“‘stelar hypothesis” several cryptogams besides the Osmunda-
ceae were cited as exceptionally possessing medullated mono-
stelic central cylinders. Since then more extended researches
have been made which have yielded important results. Thus it
has been shown that the central cylinder of Ophioglossum and
of Botrychium instead of being medullated monostelic is in
reality “‘gamodesmic ;’’4 that the central cylinder in the entire
family Equisetaceae, some of whose species were included in
the exceptions, is of the same kind;5 and that the central
cylinder of the genus Helminthostachys is also of the ‘‘ gamo-
desmic”’ type.© It is true that Strasburger holds’ that the
internal endodermis and endodermal sheaths about individual
bundles are of intrastelar origin, and not of cortical as is the
external endodermis, and that therefore these exceptions still
stand; but this objection may be advantageously left for sub-
sequent consideration. Of the apparent exceptions, the family
Osmundaceae has remained untouched, and I have undertaken
the present research on this anomalous case, with the primary
object of furnishing data that will help determine the proper
morphological interpretation of its vascular system.

The family Osmundaceae is a very limited one in point of
numbers, consisting of but two genera, Osmunda with eight
species, and Todea with six, and therefore constitutes a very
small part of the fern flora of the earth. But this does not
seem to have always been the case,® for the Marattiaceae,
although overwhelmingly predominant in the Coal period, con-
stituted but 4 per cent. of the total filicineous flora in the Lower
Jurassic, the remainder being composed of Osmundaceae and
Cyatheaceae, with the related families Matonineae and Proto-
polypodiaceae. As to distribution, the first genus is confined to
the northern hemisphere, and the Todeas are with one exception
found only in Australasia. Five Osmundas belong exclusively
to restricted areas in east Asia and the adjoining islands; 0.

4POIRAULT: Ann. Sci. Nat. Bot. VII. 18:113. 1893.
SJEFFREY: Mem. Boston Soc. Nat. Hist. 5:155. 1899.
SFARMER: Ann. Bot. 13: 421. 1899.
7 STRASBURGER : Histologische Beitrage. 3:—. 1891.
8 ScotT: Studies in fossil botany 304. 1900.

[36]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 5

Claytoniana occurs in the Himalayas and North America; (0.
clnnamomea in eastern Asia, North and South America; and OQ.
regalis in every continent except Australasia. Of the Todeas, 7.
barbara is a native of Australia, New Zealand, and South Africa;
and the remaining species, the so-called “filmy” Todeas
(Leptopteris of some authors), belong to oceanic islands in the
eastern south-tropical region.

Of these species I have had the opportunity of studying
five, namely, O. regalis, O. cinnamomea, O. Claytoniana, T. barbara,
and 7. superba. Nevertheless, in the following pages most atten-
tion will be devoted to O. cimnamomea, not so much because
its anatomy has not previously been described, as because the
writer, for reasons which will become apparent, believes it retains
a more primitive type of skeletal axis than any of the family so
far investigated. The material of the species of Osmunda stud-
ied was collected from several different localities, and in large
quantities. Of O. cinnamomea specimens from fully a hundred
and fifty plants were preserved and examined, and of each of the
others perhaps one-third of that number. The more important
points were verified from specimens taken from three different
localities.

Observations have been mainly restricted to the mature root,
stem, and leaf trace. Some young plants of Osmunda were
studied, and the growing points of the older stems have been
sectioned. But the mature stem, especially the region at which
it branches, has proved to be of chief interest from the stand-
point of questions of comparative anatomy.

THE STEM.

GENERAL ANATOMY.—The mature stems are very stout rhi-
zomes, exceptionally so in 7. barbara, which grow in a direction
somewhat oblique to the horizontal. The leaves are in a closely
set tuft at the anterior end, for they are annual and the inter-
nodes are very short. The broadly winged, overlapping bases
with their sclerenchymatous sheaths resist decay long after. the
remaining portion of the leaf has perished, and these, together
with the roots, which are very numerous, greatly add to the
bulk of the stem. The stem usually bifurcates once into two

9 Diets: Engler and Prantl’s Natiirlichen Pflanzenfamilien 14: 377. 1900.

[37]

6 FAULL: THE ANATOMY OF THE OSMUNDACEAE

branches of equal size, which lie in a horizontal plane. A few
specimens of O. vegalis were found, however, in which one of the
forks was much larger than the other, but the larger almost
immediately divided again, so that there were three branches of
about the same size lying in the same plane. The forking bears
no relation to the number of leaves produced, counting from the
cotyledons, nor to the age of the plant. Occasionally there is
no branching at all, though maturity has long since been
attained, while in rare cases it has taken place comparatively
early in the life of the fern.

The rhizome exhibits a very characteristic appearance in
cross-section (jig. z). The outer portion, the thick external
cortex (ex. ¢.), consists of very resistant, dark-brown scleren-
chyma, in O. cinnamomea of a rich red-tinted brown, in O. regalis
and the Todeas of a black, and in O. Claytoniana of a dull brown
hue. The cortex is marked by leaf-traces (4), which form a
close spiral, and at the nodes by the escaping roots (7). In
O. cinnamomea sclerification of the cortical tissue is later in
taking place than in the other species. The internal cortex (2. c.)
is parenchymatous, comparatively narrow, roughly pentagonal,
and its cells are heavily loaded with starch grains. Passing
the pericycle and the bast region, which form a complete
sheath, the wood (2) of the stele is seen to be broken up
into bundles of various shapes arranged ina circle, and sepa-
rated from one another by the so-called medullary rays. These
medullary rays extend out from a large pith. The pith or
medulla in O. Claytoniana and T. superba is apparently homo-
geneous. In O. vegalis it is often discolored and may contain
one or more strands of brown sclerenchyma; in O. cimnamomea
it is very frequently characterized by some brown sclerenchyma-
tous tissue, and in 7. darbara there is a large axial strand of this
supporting tissue.

HisroLocy.— But we turn now to acquire a more intimate
acquaintance with the stem as revealed by a study of its histo-
logical features. For this purpose several sets of transverse and
longitudinal series were prepared, and a great many microtome
sections examined. The material cut included stems of various
ages. As development proceeds rather slowly, all the tissues are
mature only at a considerable distance from the apex of the plant.

[38]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 7

The cortical part of the stem has little of interest for us
other than in the respects already mentioned. The scleren-
chyma consists of elongated, thick walled cells, with a small
lumen containing starch grains. The walls are brownish, and
marked by simple pits, which are round or slit-like. According
to Strasburger, the endodermis is not the innermost cortical
layer, but I am unable to verify this. He has made the state-
ment that the innermost cortical layer at a certain stage divides
by tangential walls to form several layers of cells; of these,
the outermost becomes differentiated as the endodermis, and the
remaining layers lie between this and the phloem, filling the
place of a pericycle. The somewhat elongated cells of the
endodermis are marked in every case by the characteristic cuti-
cularization of the radial walls, which in transverse section shows
as) the) “radial |dot,{)\( joes 0, 2,.¢).)| Lhe radial dot’’ is, dis-
tinctively brought out by treatment with phloroglucin and
hydrochloric acid, and also with dilute sulfuric acid. In O}
Claytoniana the radial markings are generally not as distinct as
in the rest of the species studied, and the cells are reduced in
size in comparison with those of the layers in contact (fig. 8, e).
The contents in this species, too, are meager, consisting of
granular protoplasm, a nucleus which as a rule stains a deeper
red with saffranin than those of surrounding cells, and a few
starch granules as shown by treatment with iodin solutions.
Sometimes the endodermal cells of O. cinnamomea are likewise
apparent by the lack of contents, in contrast to the heavily-laden
cells, both ectad and centrad. Generally in this species, as in
the remaining ones, 7. superba excepted, the cells are filled with
tannin, so that the endodermis stands out very distinctly.

The pericycle is entirely parenchymatous and consists of
several layers—in O. Claytontana and Todea of two or three, in
O. cinnamomea of three or four, and in O. regalis of one to three.
The cells are elongated, cylindrical, provided with large nuclei,
and filled with finely granular contents, part of which is starch.
Haematoxylin imparts to this tissue a light blue color. Tangen-
tial sections show that the orientation of the cells is very irregu-
lar (figs. 3 and 9,/). Immediately opposite the point of origin
of a leaf trace, and for a short distance below, the long axes of

to STRASBURGER, Of. cit. 449.
[39]

8 FAULL: THE ANATOMY OF THE OSMUNDACEAE

the cells run parallel with the long axis of the stem, but for the
most part in the remaining regions of the stem there is consid-
erable disturbance, though only in tangential planes. This dis-
turbance is commonly so marked that the long axis of the cell is
at right angles to the stem axis, and between this and the paral-
lel position there is every gradation. Therefore in transverse
section these cells are either round or more or less tangentially
lengthened (fig. 8, ~). This variation in orientation is of
interest, as it is connected with a similar phenomenon in layers
lying nearer the cauline axis, namely in the phloem region.

XyYLEM.— Before dealing with the phloem, however, it will
be convenient to describe the xylem. The wood elements are
of two kinds, namely, small ringed and spiral elements consti-
tuting the protoxylem, and scalariform tracheids which are of
later development constituting the metaxylem. Occasionally a
parenchymatous cell is found among the tracheids. A transverse
section shows, as mentioned before, a ring of variously shaped
bundles ; and by tracing these up and down, or by boiling a piece
of stem in potash and then removing the softer tissues, there is
shown to be a network forming the wall of a hollow cylinder, the
strands being the “bundles” of DeBary, and the meshes the
spaces occupied by the ‘‘medullary rays.” Though there is a
great deal of regularity in the apparent construction of this net-
work, as proved by DeBary and Zenetti in O. vegalis, yet a study
of development shows that the ‘‘bundle theory” is inadequate
for giving the right conception of the vascular system. In the
young stem of the Osmundaceae the wood forms a completely
closed cylinder, and Van Tieghem, basing his conclusions on
Todea and O. regalis, has stated this to be the case for the
whole family. I am able to state that the phenomena in the
young stem of O. cinnamomea and O. Claytoniana are in accord-
ance with his general conclusions in this respect.

Now directly above the point at which a leaf trace leaves the
stele the wood is not developed for some distance. This gap is
filled by parenchyma chiefly, except at the outer part, which is
occupied by sieve tubes. There are exceptions in O. cimnamomea
to be described later. Thus a transverse section of the stele,
just above a node, shows a ring of wood broken at one place,
the break being occupied by the tissues just referred to; in other

[40]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 9

words, the stele here has one medullary ray. Mg. 23 shows a
transverse section of the stem of O. Claytoniana through this
region. Still further up the internode the ring is complete
again. There is the same sort of gap above the second node.
However, as the nodes become more frequent, that is, as the
internodes become shorter, a leaf gap extends through more than
one internode, and ina transverse section there is more than one
medullary ray, until in the full grown stem, where a leaf gap
extends through several internodes, a transverse section shows
several gaps cut across, or in other words shows several medul-
lary rays. It is therefore evident that the number of medullary
rays seen in any transverse section depends on the frequency of
the nodes and the length of the gaps. In well nourished stems
the number is greatest in O. Claytoniana ( fig. 7), there usually
being about twenty, and in 7. barbara (fig. 24) the fewest. In
this species the gaps are quite short, so that while the wall may
be thin in many places at any given level, there are not more
than two to six medullary rays seen in the cross section (jig. 24).

The persistent portions of the cylinder of wood, the
“bundles,” present various contours in cross section, the shape
of any particular portion lying between two adjacent gaps, that
is, of any strand, varying with the level at which it is cut. Just
below where the leaf trace is given off, the wall is hollowed out
on the side towards the pith, so that the transverse section of
the strand presents a horseshoe shape (fig. 77). The middle
of the inner surface of the strand at this level is occupied by
protoxylem, which consists of about a half dozen small ringed
and spiral vessels. Following the strand down, it is seen that
the arms of the horseshoe thicken on the sides facing one
another, especially towards the ends of the arms ( fig. #5 S).
Finally, the opening between the ends is fully closed and a
small group of parenchymatous cells lying exactly centrad of
the protoxylem is thereby enclosed (jig. 77,s). The parenchyma
is more and more encroached upon by the xylem, until lower
down it is seen no more. Not far below where the parenchyma
vanishes, the protoxylem in that strand likewise disappears.
Somewhat above the level at which the parenchyma is enclosed
the strand begins to thin out on the outer side, a sharp trough-
like indentation appearing, but not in the same radius as that in

[41]

Io FAULL: THE ANATOMY OF THE OSMUNDACEAE

which the protoxylem lies. This trough continues to deepen
until a few nodes down the strand is cut through, the point at
which the break occurs being, indeed, the apex of a leaf gap.
Thus neither the outer nor the inner surface of the cylinder of
xylem is smooth; the lower part of a leaf gap can be traced as
a hollow on the inner surface just below where the leaf is given
off, ending as a blind tube amongst the tracheids, while the
upper end of the gap may be traced as a furrow on the outer sur-
face of the cylinder, gradually becoming more and more shallow.

The protoxylem occurs in small groups of six to eight cells
each, and a transverse section of the central cylinder shows from
five to seven of these groups. Each group of protoxylem
elements passes out in its entirety into a leaf trace, and on fol-
lowing back from the leaf trace each vanishes as already
described. The protoxylem is therefore not continuous through-
out the stem, but is in small, discontinuous strands. This fact
has been recorded for O. vegalis by Zenetti.™

Lying externally to the wood are from four to six layers of
elongated parenchymatous cells, rich in protoplasmic contents
and in small starch grains. They are continuous with the
parenchymatous cells of the medullary rays and do not mate-
rially differ from them. Those occupying the middle of the
medullary rays have more meager contents, and towards the
stem axis they become larger. That there is a ‘‘xylem sheath”’
characterized by cells of greater size and richer contents such as
Zenetti describes for O. vegalis, 1 cannot affirm, and certainly
there is not such a sheath in O. Claytoniana.

PHLOEM.—The tissues that have just been described are
bordered by the phloem, which consists chiefly of sieve tubes
Parenchymatous cells are sometimes met with in isolated posi-
tions in the metaphloem, and between the metaphloem and the
protophloem they constitute a more or less broken layer, most
pronounced in O. Claytoniana, and least constant in O. regals and
T. barbara. The sieve tubes are strongly developed and are of
the ‘“‘type vigne” of Lecomte. They are large, have thin walls
of unmodified cellulose lined with a delicate layer of protoplasm,
and are devoid of nuclei. They are provided with oblique

™ ZENETTI: Das Leitungssystem im Stamm von O. rvegalis. Bot. Zeit. 53:63.
1895.
[42]

FAULL: THE ANATOMY OF THE OSMUNDACEAE I!

terminal walls and are furnished with sieve plates both simple
and compound. The sieve plates are covered with ‘globules
brillants,” and by treatment with proper reagents callus plugs
may be demonstrated in them. The sieve tubes of the proto-
phloem are smaller than those of the metaphloem and their
terminal walls are not as oblique.

As there has been considerable difference of opinion regarding
the disposition of the phloem in O. regalis, it will be well to
define a sieve tube. Both Russow and Janczewski have studied
sieve tubes very carefully, and Poirault has more recently rein-
vestigated the subject in the vascular cryptograms. The investi-
gator last named summed up his observations on sieve tubes in
the roots of vascular cryptogams in the following terms :”

Les tubes criblés peuvent se rapporter 4 deux types: le premier carac-
térisé par des cloisons transverses perpendiculaires aux faces principales et
ne portant qu’un seul crible (tye Courge, Lecomte) ; le second reconnaissable
4 ses cloisons transverses trés obliques portant d’autant plus de cribles que
leur obliquité est plus grande (tyfe Vigne, Lecomte). On trouve, en outre,
sur les faces longitudinales des ponctuations isolées ou réunies en trés petits
groupes, constituant rarement des cribles aussi développés que ceux des
faces transverses. Le contenu de ces tubes est un liquide hyalin tenant en
suspension de nombreuses sphérules réfringentes, rassemblés surtout au
niveau des cribles et des ponctuations isolées. Il n’ya pas denoyau. La
membrane est cellulosique.

He further adds that two substances occur as a rule — (1)
the ‘globules brillants’” already mentioned, and (2) “les
bouchons calleux qui font corps avec la membrane et peut-étre la
traversent entiérement.”’

In dealing with the sieve tubes of the stem and petiole he
does not point out any other peculiarities, but deals at length
with the callus plugs, and the perforation of the sieve plates.%
The observations of Russow, Janczewski, and Poirault agree for
the most part, except in reference to the callus plugs.

The following criteria would seem to be distinctive in deter-
mining the presence of sieve tubes in the Osmundaceae; the
existence of sieve plates, the absence of nuclei, and the presence
of “globules brillants.” Less distinctive and rather as a con-
firmatory test I have sought for callus. Russow made this test

12POIRAULT: Récherches sur les cryptogames vasculaires. Ann. Sci. Nat.
Bot. VII. 18: 138. 1893.

13 POIRAULT : of. ctf. IQI.
[43]

12 FAULL: THE ANATOMY OF THE OSMUNDACEAE

one of paramount importance, but it seems best in dealing with
the vascular cryptogams to give it a second place, and for the
following reasons: (1) the callus, so-called, in the seive tubes
of the vascular cryptogams may not be identical with that found
in phanerogams ; (2) it occurs in minute quantities only, and in
some plants (¢. g., the Ophioglossaceae) probably does not occur
at all; (3) its presence is determined by delicate microchemical
means, and then only by limited color reactions.

Janczewski*+ claimed to have found callus in Pteris aquilina
alone of all the vascular cryptogams he examined, and states
that it does not occur in O. vegalis. The reagents he used were
_ Schulze’s solution (or chlor-zinc-iodin) and rosolic acid. On
the other hand, Russow found callus in all of the sieve tubes he
examined. The reagent he used was a mixture in variable pro-
portions of chlor-zinc-iodin and potassium iodid-iodin. It
should be stated that in the vascular cryptogams callus occurs in
the wall of the sieve plate, appearing as if it were a part of the
wall. After staining with a suitable iodin solution, the callus
shows in face view as one or more round brown spots, and in
section as rods or granules occupying the entire thickness of the
lamella. Poirault has largely corroborated Russow’s observa-
tions. He disagreed with Russow’s generalization that it is a
constant feature of sieve tubes, for he states that he has been
unable to find any trace of callus in Angiopteris and Ophioglos-
sum."5 In view of these investigations, therefore, it becomes a
matter of interest to know if the sieve tubes of Osmundaceae
show the phenomena of callus as described by Janczewski for
Pteris, and by Russow and Poirault for many others of the vas-
cular cryptogams.

Accordingly tangential and transverse sections about five
microns thick, of the three Osmundas studied and of 7. barbara
were cut from mature pieces of stem embedded in celloidin. In
the present research the writer has tried several stains, such as
ruthenium oxid, Hofmann’s blue, rosolic acid, and Russow’s
mixture. These have been applied to sieve tubes of plants from
widely separated groups, such as Vitis (summer and winter sieve
tubes), Tilia, Pinus, Pteris, and the mixture of chlor-zinc-iodin

™ JANCZEWSKI: Tubes cribreux. Ann. Sci. Nat. Bot. VI. 14:50. 1882.
15 POIRAULT: of. cit. 192.
[44]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 13

and potassium iodid-iodin proved to be by far the most satis-
factory reagent for the demonstration of callus. The two con-
stituents of this reagent were prepared fresh, and then mixed in
different proportions until one giving the best results was
obtained. The proportions vary with the different kinds of
plants tested. In using this stain, though the presence of the
celloidin is not a serious objection, it is preferable to dissolve
out the celloidin, wash in alcohol, then in distilled water, and
examine in stain on the slide.

In face view it is difficult to make observations on account
of the ‘globules brillants,’’ hence the most reliable observa-
tions can be made on sectioned plates. Almost at once after
applying the stain, the callus plugs become evident, staining a
dark red-brown (fig. 7). They appear as more or less fine rods,
completely traversing the sieve plate, and their number in a sieve
plate depends on its size. The cellulose is slower in staining;
at first it is light blue or a violet, and later a deep blue. Hence
the callus plugs are to be seen most clearly in the early stages
of the staining process. The stain unfortunately is not perma-
nent. Callus was clearly demonstrated in the species under
investigation, but on account of the size of the cells and of the
sieve plates, 7. barbara proved the best subject for the purpose.
As one of the characters of the sieve tubes of the Osmundaceae,
we record, therefore, the regular occurrence of callus plugs in
the sieve plates.

The “globules brillants’” are exceedingly abundant in the
sieve tubes, and especially in the older ones (fig. 7). While
they adhere to the protoplasm of the cell and may be found in
any part of the cell they are by far most abundant about the
sieve plates, dotting their surface, filling the pits, and surround-
ing the entrance to the pits. They are evidently not homoge-
neous, but appear to consist of two substances, one of which is
more refractive than the other, for by slight focusing up and
down they change from a dark looking, opaque granule to a light
semi-translucent spherule. Jodin solutions stain them brown,
not appreciably different from the callus plugs. Occasionally
irregular fragments of matter are found in the cell, which also
stain a similar brown. The relation between these fragments,
the ‘‘globules brillants,” the callus, and the disappearance of the
nuclei calls for further investigation.

[45]

14 FAULL: THE ANATOMY OF THE OSMUNDACEAE

The sieve plates are very numerous, and vary in size and
form. The walls of the pits are abrupt, and the number of pits
varies with the size of the plate. The larger plates are irregu-
larly oblong and the smaller ones are round. In Todea they are
largest and most numerous. Having described the sieve tubes
at some length, we shall now examine their distribution.

The phloem forms a continuous sheath, the outer portion of
which is the protophloem. To this peripheral part of the phloem
I find that DeBary and Zenetti alone make specific reference.
The former states*® that this region of the central cylinder is
characterized by ‘‘ quergestreckte Zellen,” but he offers no opin-
ion as to the nature of the tissue in question. Zenetti divides
the zone into strands of typical protophloem and connecting
portions of ‘‘quergestreckte Zellen.”’ The typical protophloem
is in short strands one cell thick, lying ectad of strands of xylem
which are about to give off the leaf traces. The ‘‘quergestreckte
Zellen”’ he cannot recognize as sieve tissue, because of the ‘“ quer-
gestreckt”’ form and the position of the cells. This tissue in
Osmunda regalis, consequently, forms a cylinder interrupted by
the strands of protophloem only, but Zenetti found it to be gen-
erally two cells in thickness and opposite the medullary rays
often several cells deep.

I have ascertained that the ‘‘quergestreckte Zellen” are
devoid of nuclei; their walls are of cellulose, staining violet or
blue with iodin solutions. They are rounded, elongated ele-
ments, with more or less oblique terminal walls; and are char-
acterized by the possession of sieve plates which show the callus
reaction when treated with Russow’s reagent; the cells contain
an abundance of ‘globules brillants’” which are aggregated
especially about the plates; the protoplasm is reduced to a thin
parietal layer. The characters of the typical protophloem cells
are the same as those of the ‘‘quergestreckte Zellen”’ except in
regard to orientation (jig. 6).

Transverse sections show that the long axes of the so-called
“quergestreckte Zellen” are tangentially placed, and never radi-
ally to any degree. To determine the slant of the long axes,
therefore, with reference to the axis of the stem, tangential sec-
tions must be made. Ifsuch be examined it is seen that some are ~

16 DEBARY: of. cit. 360.
[46]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 15

exactly at right angles to the axis, others are almost or entirely
parallel, and between these extremes there is every gradation.
This at once explains the difference in ‘‘width” of the ‘quer-
gestreckte Zellen’’ in transverse section. It is further to be
noted from the tangential sections that the ends of typical pro-
tophloem cells never abut against the long sides of the ‘‘ quer-
gestreckte Zellen,” but there is a gradual change in the direction
of the latter so that their ends communicate with the proto-
phloem and it is quite impossible to say where the typical proto-
phloem ends and the ‘“‘quergestreckte Zellen”’ begin.

The root and leaf have been examined for ‘‘quergestreckte
Zellen,”’ for if these elements constitute a characteristic textural
feature of the Osmundaceae they would naturally occur elsewhere
thanin the stem. They are not present atall in the appendicular
organs. Further, in the young sporophyte, where the leaf gaps
are far apart, they are absent from considerable portions of the
stem. The real nature of the ‘“‘quergestreckte Zellen”’ will be
discussed after observations on their development and their rela-
tion to the leaf traces have been described.

The ‘‘quergestreckte Zellen” and the typical protophloem
cells form a continuous sheath in all the species studied. In O.
Claytomana the elements of this sheath are very much smaller,
and so it is easier to distinguish them from the metaphloem.
In 7. barbara their histological characters are best studied
because of their relatively large size. Frequently in O. cinna-
momea and O. regalis it is difficult to decide in the mature stem
whether or not certain cells belong to this sheath or to the
metaphloem. But evidently in all of the species the sheath is
rarely more than two cells in thickness, and often, especially in
O. Claytontana, there is but a single layer. Opposite outgoing
leaf traces the sheath is reduced to a single stratum.

The metaphloem forms a hollow cylinder consisting of large
sieve tubes such as have already been described. They are
thin-walled, and especially in O. vegaiis in the older parts of the
stem have often collapsed. The sheath is one or two cells thick
opposite the strands of xylem, and several cells in thickness
opposite the medullary rays (jig. 8, ph). Most of the tubes run
parallel with the long axis of the stem, but here and there
‘“‘quergestreckte”’ examples occur.

[47]

16 FAULL: THE ANATOMY OF THE OSMUNDACEAE

This cylinder of metaphloem has a smooth outer surface, but
the inner surface is rendered very uneven on account of the
wedge-like proliferations of the sieve tissue opposite the leaf
gaps. Since this isa phenomenon common to all the species
studied, we naturally seek an explanation of this peculiar dis-
position of the phloem. In his memoir on sieve tubes Janczew-
ski,*7 who could hardly have been prejudiced by any stelar
theories, noted that isolated sieve tubes occur occasionally here
and there in the medullary rays of O. vegas. The writer has
found undoubted cases of the same thing in O. cimnamomea.

Two such eminent botanists as DeBary and Strasburger have
disagreed as to the topographical distribution of the layer of
metaphloem sieve tubes in O. vegalis. The former states” that
the sheath is continuous, while the latter states’ that he puts
himself in opposition to DeBary on this point, for he considers
the phloem to be interrupted opposite the medullary rays.
Strasburger does not say for what reason he considers the cells
opposite the medullary rays not to be sieve tubes. My own
observations on O. regalis are precisely in accord with those of
DeBary and Janczewski. The cells opposite the medullary rays
differ in no way from the sieve tubes opposite the xylem strands.
I have found the same to be true of the other species studied,
with the additional observation that isolated sieve tubes occur
sometimes in the tissues filling the leaf gaps of O. cinnamomea.

To this last observation I have two others to add, namely
_ the occurrence of an internal phloem in which the sieve tubes
form a more or less continuous ring ( jigs. 2r, and 22), and in
rare cases the union of external and internal phloem through a
leaf gap. Ina certain rich, moist situation about a dozen well
nourished plants of O. cinnamomea grew, of which, on examina-
tion, five showed the phenomenon of a continuous layer of inter-
nal phloem. Search in an adjoining locality resulted in finding
specimens which showed the same feature. To extend the range
of observations, I visited a peat bog some twenty miles distant
from Toronto, where I knew the cinnamon fern grew, and
secured specimens characterized by the same peculiarity. yg.
21 shows a transverse section of a stem found in this last locality.

17 JANCZEWSKI: of. cit. 66.

18DEBARY: of. cit. 360. 19 STRASBURGER: Of. Cctt. 449.

[48]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 17

The sieve tubes of this internal phloem are as typical as those of
the external, and except for their position not be distinguished
from them. They do not always form a continuous ring as do
the sieve tubes of the external phloem, but are often in more or
less detached groups, embedded in small celled parenchyma.
The layer of sieve tubes is from one to three cells thick. It
should be added that internal phloem occurs only near where
the forking of the stem takes place.

O. cinnamomea shows likewise two other features which are
constant throughout every part of the stem, and at once distin-
guish it from other species: (1) an internal endodermis, and (2)
several layers of parenchyma between this and the xylem.

INTERNAL ENDODERMIS.—The internal endodermis possesses
the characteristic radial dot, though sometimes not as clearly
distinguishable as in the external endodermis (2. ¢., fig. 2). Its
cells are usually larger than those of the latter, but are filled
with similar contents, most frequently tannin (fig. 75). It is
further to be noted that it bends outwards opposite the leaf
gaps (jigs zo, etc.), and not infrequently connects through them
with the external endodermis. I have examined scores of stems
of the cinnamon fern, and in every specimen there was an internal
endodermis. On the contrary, it seems to be invariably absent
from the other species studied. As the central cylinder of the
family Osmundaceae has heretofore been classed as monostelic,
the existence of an internal endodermis in one of the species is
therefore a matter of considerable moment, especially if it be
regarded as a real phloeoterma.

Between the internal endodermis and the xylem there is a
cylinder of elongated parenchyma, rich in starch and protoplasm,
and from two to seven cells in thickness. This layer is continu-
ous with the medullary rays. In O. vegalis, O. Claytoniana, T.
barbara, and T. superba a similar but thinner layer it found as a
rule, and the cells are always smaller and richer in contents than
those of the medulla on which they border.

THE MEDULLA.—The medulla is very large in this family,
particularly so in O. Claytoniana, and consists of large-celled
parenchyma. Most of the cells are partly filled with large
starch granules, but frequently some of them contain tannin,
especially in Z. barbara. A brownish fluid may occur in inter-

[49]

18 FAULL: THE ANATOMY OF THE OSMUNDACEAE

cellular spaces, and in O. regalis within the cells themselves. In
these regards there is often a striking resemblance between the
parenchyma of the medulla and that of the internal cortex in the
same plant. But there yet remains to be described a still more
significant phenomenon, namely, the occurrence in the pith of
brown sclerenchyma of the same kind as is found in the external
cortex (figs. rg and 20). This is probably a primitive feature,
and in this, as in many other respects, O. cinnamomea proves to
be most interesting. Out of forty-four pieces of stem, chosen
at random, and representing a corresponding number of different
plants of this species, twenty-five of the examples showed brown
sclerenchyma in one or both ends. It occurs as a central strand,
varying in size from a few cells to almost the limits of the pith,
or as several small strands irregularly arranged. fag. 14 is a
photograph of the transverse section of a stem in which there is
a large axile strand, and fig. 75 of one in which the scleren-
chyma is entirely absent from the pith. Further it has the
peculiar habit of being present at one level, but perhaps not at
another; so it is likely to be found in nearly every plant if the
stem be sectioned from end to end.

This same habit is characteristic of its appearance in O.
vegalis (fig. 20), but more often it is not present at all. That
brown sclerenchyma occurred in the pith of O. regals did not
escape the observant DeBary,”° but elsewhere I find no reference
to this fact. Strangely enough, however, out of thirty-five or
forty plants harvested from one locality there was not a trace of
sclerenchyma to be found in the medulla of any of them, while
in one region not far distant 25 per cent. showed this phenome-
non, and in another a still higher per cent.

Parenchyma is the sole constituent of the medulla of O. Clay-
toniana (fig. 17). This is probably true of 7. superba too. lg.
25 is a cross section of J. barbara taken too near the growing
point to show sclerenchyma, but farther down the medulla was
occupied by a large strand of this tissue (jig. 24).

Thus medullary brown sclerenchyma is usually present in O.
cinnamomea, in O. regalis not uncommonly, and in O. Claytoniana
not at all. In J. darbara it also occurs, but apparently not in
T. superba. It is perhaps significant that such series can be

20 DEBARY: of, cit. p. 290.
(50]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 19

arranged, but of greater importance is the fact that the occur-
rence in the Osmundaceae of brown sclerenchymatous tissue,
apparently within the cauline central cylinder, has no parallel
among existing ferns.

THE FORK.—There yet remains to be described the anatomy
of one particular portion of the stem, the part in the region of
bifurcation. It has been stated that it is peculiar to the stem of
the Osmundaceae to fork once, and that in a horizontal plane.
We shall treat first of the phenomenon in O. cinmnamomea. Trac-
ing the main stem forwards, it is seen to become flattened and
then to become constricted in a median vertical plane. Imme-
diately anterior to the point of bifurcation of the vascular axis,
there is a wide ramular gap in the central cylinder of each
branch (fig. 70). Sections of the main axis immediately below
the fork show two bands of phloem, one on the upper and one
on the lower internal surface of the central cylinder (fg. 73).
Sections passing through just in front of the region of bifurcation
show similar bands of phloem along the inner wall of the
central cylinder of each branch (fig. zz). Cases have been
described above, in which there is a complete cylinder of inter-
nal phloem instead of the two isolated bands just referred to
(jig. 22). The internal and external phloem connect through
the ramular gaps (jig. zz). Likewise the internal and external
endodermis are in textural continuity through these gaps, so that
there is free communication between the cortex and the pith
(fig. To).

Sometimes the cortex lying between the two branches con-
tains brown sclerenchyma which is continuous through the ramu-
lar gaps with strands of the same tissue occurring in the medulla
of the branches. Frequently in less vigorous plants a transverse
section of the main axis posterior to the point of ramification
shows a diamond-shaped piece of cortex surrounded by endo-
dermis (jig. 72). Posteriorly this included piece of cortex
becomes continuous with the medulla of the main axis (fig. 73),
and anteriorly with the general cortex (fig. rr).

Twenty-five forks of O. ctnnamomea were selected at random
and sectioned. Twelve of them presented the phenomenon of
typical wide ramular gaps. Six of them were of the reduced
kind just described. In five cases there were gaps in the xylem

[st]

20 FAULL: THE ANATOMY OF THE OSMUNDACEAE

only, cortex and medulla never becoming. continuous; and in
two even the xylem did not open up (fg. 76). For reasons to
be outlined later, the writer believes the wide gaps to be the
most primitive.

O. regalis presents a much degenerated form of ramular gap,
for here only the xylem opens (fig. 79). In O. Claytoniana the
degeneration is carried still farther, for as a rule there are no
branch-gaps at all (fig. 78). In 7. barbara the xylem alone may
open up.

The phenomena of the fork may be thus summarized :

(1) Complete ramular gaps occur only in O. cénnamomea.

(2) Internal phloem occurs only in O. cinnamomea. It is
found in the branches just above, and in the parent axis just
below the point of bifurcation of the central cylinder.

(3) The internal phloem may form an entire cylinder.

(4) Where gaps are complete, the cortical and medullary
tissues connect through them.

(5) Thus sclerenchyma of the cortex is sometimes continu-
ous with sclerenchyma in the medulla of the main axis, and of
the branches.

(6) O. cinnamomea presents the following forms of ramular
gaps arranged in order of degeneration, (2) complete gaps, (0),
phloem and xylem only open, (c) the xylem alone opens, (@)
no gaps at all.

(7) O. regalis and T. barbara show gaps in the xylem only,
and in O. Claytoniana there are usually none at all. O. Clayton-
ana, therefore, presents the extreme case of degeneration.

THE LEAF TRACE.

The leaf traces pass very obliquely up through the external
cortex. A section of a leaf trace shortly before it passes into
the petiole presents some noteworthy characters. In the first
place there is. no pith, but a solid horseshoe-shaped mass of
xylem with the convex side turned outwards (jig. 5,7). The
xylem is made up of large scalariform tracheids with a protrud-
ing mass of a few small vessels constituting the protoxylem.
The protoxylem is situated on the inner face of the single strand
of xylem (fx), and is continuous with that of the stem. In 7,
barbara it frequently breaks into two or three groups.

[52]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 21

Surrounding the wood is a layer of parenchyma, which on
the concave side of the xylem quite fills the space between the
arms of the horseshoe. The phloem consists of a crescentic
band of sieve tubes, one to three cells thick on the external
side of the leaf trace (f2), and a smaller band on the opposite
side (fg). The protophloem consists of small elements which
form a ring, broken only on the concave side of the xylem.
Here the ring is completed, however, by the inner band of
metaphloem. In O. cimnamomea and T. barbara isolated proto-
phloem cells have been observed by the writer on the side of the
inner band of metaphloem towards the stem axis. On the convex
side the protophloem is separated from the metaphloem by paren-
chyma. There are no ‘‘quergestreckte Zellen.”” The pericycle
consists of two or three layers of cells, and is bounded by a well
developed endodermis continuous with that of the stem. With
reference to the attachment of the leaf trace to the cauline
vascular axis Zenetti has given a very careful and accurate
description.”

Strasburger has held” that the stele of the petiole of O.
vegalis is a collateral bundle. He has considered the inner band
of metaphloem to be a parenchymatous tissue. However, the
cells of this band prove to be characteristic sieve tubes, and are
continuous with sieve tubes in the stem opposite the medullary
rays. The leaf traces, therefore, are undoubtedly concentric.
Several botanists have arrived at the same conclusion for QO.
regalts,?3: 74

In summary, the most important features of the leaf trace
are: (1) the absence of a pith, (2) the endarch xylem strand.
(3) the concentric type of stele, (4) the absence of ‘“ querge-
streckte Zellen,” and (5) the cylinder of protophloem completed
on the inner face by a band of metaphloem.

THE ROOT.

The roots have a definite relation to the leaves, both in posi-
tion and in numbers. Two roots invariably originate from the
base of every leaf trace, or from the central cylinder immediately
below. They come off at the same level, one opposite each arm
of the horseshoe-shaped strand of xylem ( fig. 78) in every case

2 ZENETTI: of. cit. 69. 3%3SCOTT: of. cit. 319.

22 STRASBURGER : of. cit. 448. 24 ZENETTI: of. cit. 66.

[53]

22 FAULL: THE ANATOMY OF THE OSMUNDACEAE

where there are just two roots to a leaf. They grow almost
directly outwards, and so in a transverse section of the stem are
cut longitudinally. In such a section it is seen, likewise, that
the cortical tissues of the stem and root are entirely independent
of each other, and that, therefore, the root is of endogenous
origin. This fact is true of the secondary roots also.

The cortex is exceedingly thick, forming by far the main
bulk of the root, and consists of large celled sclerenchymatous
tissue. The cortical cells diminish in size towards the periphery,
and become thicker walled. In 7. barbara, however, there is a
discontinuous ring of exceedingly thick walled brown scleren-
chymatous cells immediately surrounding the vascular axis.
The endodermis, which is continuous with that of the stem and
leaf, is very pronounced in all of the species, and is at once
noted by the radial dot, and by the fact that its cells are filled
with tannin. JIn the second particular, exception must be gener-
ally made of O. Claytoniana.

The stele is comparatively small, and is typically protostelic,
since there is no pith. The wood has a narrow elliptical form,
consisting mainly of very large scalariform tracheids. At each
end of the ellipse there are a few small protoxylem elements,
which are especially evident in the young root, and which have
no connection with the protoxylem of the stem or leaf. The
root, therefore, is diarch. There are likewise two bundles of
phloem alternating radially with the bundles of xylem. In all
of the Osmundas, however, I have observed triarch steles in the
larger roots, which exception is of comparative frequence in O.
cinnamomea. The phloem consists of two flat bundles or bands.
These bands are made up chiefly of thin walled sieve tubes
which are of the same kind as occur in the stem. None of them
are ‘“quergestreckt.’’ The phloem is separated from the xylem
by three or four rows of parenchyma, and from the endodermis
by a two rowed parenchymatous pericycle.

DEVELOPMENT OF THE TISSUES FROM THE GROWING POINT.

In discussing this subject there are two points in particular
which will receive special consideration: (1) the statements of
Strasburger and Zenetti regarding the origin of the endodermis,
and (2) the real nature of the “quergestreckte Zellen.”

[54]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 23

The determination of the relation of the tissues to the apical
cells seems of little concern, and moreover in the study of the
apical region of the growing point there are serious difficulties.
Having described these for O. regalis, Professor Bower aptly
remarks :?5

The meristem being thus at times irregular, and the subdivisions of the
segments being variable, it is to be expected that the study of it (the apical
region of the growing point) in longitudinal section would present difficulties,
and I have not been able to trace any definite and characteristic mode of
segmentation. Longitudinal sections cut from a considerable number of
stems show that a conical apical cell is present. The relations of the sur-
rounding tissues, and their reference to regularly succeeding segments are
difficult to recognize.

To these observations on the extreme apical end of the growing
point we have nothing to add, but pass further down the stem.

A short distance from the apex of the stem, the various tis-
sues, though in embryonic form, become apparent. The cylinder
of wood, whose thin walled, unlignified cells are still provided
with protoplasm and nucleus, can be distinguished from the pith,
the parenchyma in the leaf gaps, and the immature phloem.
The pericycle is rich in protoplasm, and its cells are radially
arranged. At an earlier stage still, even before there is any evi-
dent differentiation in the vascular tissues, the leaf traces can be
seen coming off from the cauline vascular axis. ~

When the protoxylem can be first demonstrated by phloro-
glucin and hydrochloric acid, the endodermis (both internal and
external in the case of O. cinnamomea) is also demonstrable by
the same reagents, though not before. Zenetti has claimed”
that at the time the protoxylem is formed, the endodermis, peri-
cycle, “‘quergestreckte Zellen,” protophloem cells, and some
cortical cells are all in the same radial rows; and that, there-
fore, all have originated from the same mother layer. Stras-
burger has asserted’’ that the tissue lying in the stem between
the phloem and the endodermis and occupying the position of a
pericycle arises by tangential divisions with the endodermis out
of the innermost cortical layer. Therefore, not the entire

75 BOWER: The comparative examination of the meristems of ferns as a phylo-
genetic study. Ann. Bot. 3: 323. 1889.

26 ZENETTI: of. cit. 64, 27 STRASBURGER: of. cit. 449.
[55]

24 FAULL: THE ANATOMY OF THE OSMUNDACEAE

phloeoterma, he claims, but the outer division product is that
which gives origin to the endodermis.

Now, at the time the protoxylem elements appear, I did not
find, in the species examined, the cells of the endodermis cor-
radial with those lying centrad. It is true that in younger
stages the cells in this region are in radial rows; but nearer still
to the punctum vegetations this is approximately true of all the
cells of the stem. At this earliest stage one would hesitate to
say, because certain cells were corradial, that they were therefore
division products of the same mother cells; so Zenetti’s con-
clusion, based on this sole argument, scarcely seems conclusive,
even granting the correctness of his observation. If, too, such
a conclusion were correct there would be the curious anomaly of
certain phloem and cortical tissues having a common origin.

Evidently the study of transverse sections cannot settle the
matter. To attempt to follow these layers upwards is obviously
only possible in median longitudinal sections. But in the stems
of the Osmundaceae the leaf traces are exceedingly numerous,
and at the growing point are closely packed together, and
appear before the tissues of the cauline central cylinder become
at all differentiated. Hence, no matter what be the plane of
section, the endodermis cannot be traced continuously very far
anteriorly to the point at which it is differentiated, for a leaf
trace is certain to intervene; andI found it quite out of the
question to pick out an undifferentiated endodermis on the side
of the leaf trace turned towards the apex. Therefore, every
attempt failed to refer the endodermis and the rows of cells
‘occupying the place of the pericycle”’ to the same initial layer.

The'typical protophloem, and the ‘‘quergestreckte Zellen”’
begin to be differentiated simultaneously with the appearance of
the protoxylem. They are best examined in tangential sections.
Their walls at this time become pitted, and their contents much
less granular than those of the surrounding cells. Here, as in
the maturer parts of the stem, there appear to be no differences
between the typical protophloem and the ‘“‘quergestreckte Zel-
len.” Their relation to the leaf traces seems to explain their
irregularity in orientation. Immediately below the point of ori-
gin of a leaf trace they are arranged with their long axes parallel
to the long axis of the stem, and there is a gradual transition to

[56]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 25

the tangential position. More than this, the laterally placed
protophloem cells of the leaf traces can be directly traced into
the ‘ quergestreckete Zellen” of the stem. There seems little
doubt, therefore, as to their nature.

To summarize observations : (1) The ‘ quergestreckte Zel-
len”’ are sieve tubes, as has been demonstrated above; (2) they
become differentiated at the same time as the typical proto-
phloem, and (3) occupy the same relative position; (4) they
resemble the protophloem cells in form; (5) their orientation is
not uniform; (6) they pass imperceptibly into the longitudi-
nally orientated protophloem cells of the leaf traces. Hence
there seems no reason to regard them as anything else than pro-
tophloem.

CONCLUSIONS.

The question now remains, how to interpret the vascular sys-
tem of the Osmundaceae. To do this more intelligibly, it will
be well to recapitulate the main fibrovascular theories. We shall
begin with that of Sachs and DeBary.

These botanists regarded the bundle as the unit, and the vas-
cular system as a more or less simple complex of bundles
embedded in ground or fundamental tissue. Developmental
studies have shown that this theory is inadequate, for the unit is
wrong.

The hypothesis which at present obtains was supplied by Van
Tieghem and Strasburger. In this conception”® 93 the stele is
the unit. The primitive form of stele, the monostele, such as
occurs for example in most roots and in the stems of lycopods, is
a solid central strand of xylem, surrounded by a sheath of phloem,
and marked off from the cortex by the differentiated internal
cortical layer, the endodermis. Of this there are many modifi-
cations, of which mention is made of the most important. By
the repeated bifurcation of the monostele, the polystelic type is
presented, as in Primula and Pteris, each segment being in
every respect a stele. If these steles fuse laterally, thus forming
a ring with internal and external phloem, the gamostele is pro-
duced as illustrated by Marsilia. Again, when parenchyma

22VAN TIEGHEM: Traité de Botanique 673, 765.

79VAN TIEGHEM: Sur la polystélie. Ann. Sci. Nat. Bot. VII. 3: 275.

3° VAN TIEGHEM: Eléments de Botanique 1:84, 179.

[57]

26 FAULL: THE ANATOMY OF THE OSMUNDACEAE

segregates in the axis of the monostele, and the vascular ring is
broken into strands by ectad extensions of this pith (the medul-
lary rays), we have the medullated monostelic type, such as is
common in phanerogams. It is to be noted that the medullary
and cortical tissues are considered by both these botanists to be
of morphologically different value. Now by the bending in of
the endodermis of the medullated monostele between the
bundles, and the fusion of the ends of adjacent groups on the
centrad side of the bundle, so that each bundle has its endoder-
mal sheath, and medulla and cortex become continuous, the
schizostelic or astelic type results. Of this phenomenon Ranun-
culus and Equisetum afford examples. A modification of this
type, the gamodesmic-schizostelic, is produced by the lateral
fusion of these endodermal sheaths, so that there is a common
internal and a common external endodermis. If the internal
endodermis degenerates, as it does in #. arvense, then there is
evidently a simulation of the medullated monostele. It is fair
to add that Strasburger dissents 3* from the last two types des-
cribed, the astelic and the gamodesmic, for he regards the endo-
dermal sheaths about the bundles in the first of these, and the
internal endodermis in the second, as not morphologically
phloeotermal, but originating from specialized stelar cells.

The researches of Gwynne-Vaughan* and Jeffrey 33 have
shown that the phenomena said to lead up to polystely do not
occur in Primula and Pteris. If the polystelic conception falls,
obviously gamostely goes too. Further, astely has been shown,
where it occurs in Equisetum and Ranunculus, to be preceded by
the gamodesmic appearance. Later the internal and external
endodermis may fuse between the bundles, but in no case is there
an inward looping of the endodermis. Finally, the stelar origin of
the pith of the medullated monostele has been disputed, and the
question raised as to whether the medullary and cortical tissues
are in reality morphologically different. In other words, is the
medullated monostelic type primitive, as its simplicity might indi-
cate, or has it resulted by degeneration from more complex types ?

3t STRASBURGER : Of, cit. p. 442.

32 GWYNNE-VAUGHAN: Polystely and the genus Primula. Ann, Bot. 11: 307.
1897.
33 JEFFREY: Morphology of the central cylinder of angiosperms. Trans. Canad.
Inst. 6:—. (1-40) 1900.
[58]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 27

It is interesting to note that Potonié had discussed this last
question from the standpoint of fossil botany, and concludes*
that it seems evident in the case of certain groups, such as the
cycads, that the simple results from the complex (for example,
the cycads from the Medulloseae). Hence for these groups at
least he is inclined to reject this idea of segregation of parenchyma
in the center of the protostele to form the medullated monostele,
but holds that the medullated monostelic type has probably
arisen by degeneration from his ‘‘pericaulom.”’ Since this peri-
caulom was produced, according to his theory, by the lateral
fusion of leaf bases in the stem surrounding the originally solid
stele, the ‘‘urcaulom,” the medullated monostele has been derived
from a form of central cylinder such as Van Tieghem has
described as polystelic, preceded or accompanied by the dis-
appearance of the enclosed urcaulom. The paleontological
evidence, however, appears not to be conclusive, for in the very
group that Potonié cites, the cycads, so eminent a paleobotanist
as Dr. D. H. Scott takes a directly opposite view. He points
out35 that the vascular system of the Medulloseae was typically
polystelic, while in the recent cycads there is but one vascular
cylinder, and that hence ‘we should involve ourselves in unneces-
sary complications if we endeavored to derive the simple, primary
structure of the cycadean stem from the more elaborate organi-
zation of a Medullosa. It is far more natural to suppose that
the monostelic cycads arose from monostelic ancestors.”

In 1897, Dr. E. C. Jeffrey put forward another view of the
vascular system,3° based upon a study of the young sporophyte.
Here, too, the stele is the unit. According to this conception
there are two primitive types of vascular axes; the first the same
as Van Tieghem’s primitive type, and designated “ protostelic ;”
the second, one in which there is a hollow cylinder, or ‘siphono-
stele,” whose external wall abuts on the cortex, and whose
internal wall encloses the medulla, and which possesses internal
as well as external phloem. This is the “amphiphloic siphono-
stelic” type, called by Van Tieghem the ‘‘polystelic.” The

34PoTONIE: Die Metamorphose der Pflanzen im Lichte palzontologischer
Thatsachen 22.

35ScoTT: Studies in fossil botany 395. 1900.

3°JEFFREY: Trans. Brit. Assn. Toronto. 1897.

[59]

28 FAULL: THE ANATOMY OF THE OSMUNDACEAE

’

commonly called ‘‘astelic’’ modification results from the amphi-
phloic type by a degeneration of the internal phloem, and the
medullated monostelic type of Van Tieghem is derived from the
astelic by the loss of the internal phloeoterma or endodermis.
A study of development from the seedling is likely to show how
these and other modifications in the stellar structure have been
derived from the primitive types. Attention is also called to
certain portions of the wall of the siphonostele in which the
vascular tissues do not develop. These places lie above the
points of exit of branch traces, and of leaf traces, and are known
as ramular and foliar gaps respectively. Through these gaps the
tissues outside and inside connect. In transverse section, the
connecting tissues seen constitute the medullary rays, and the
segments of the woody cylinder with adjacent phloem and
parenchyma the bundles. A fact of great phylogenetic impor-
tance in dealing with ‘“‘gaps’’ was further pointed out, namely,
that in small leaved plants, as in the Lycopodiaceae, Equise-
taceae, etc., only ramular gaps occur. These plants are grouped
in the division Lycopsida, and their steles are said to be
cladosiphonic. In all other vascular plants there is a gap for
every leaf. These constitute the large leaved plants, the
Pteropsida, and their steles are said to be phyllosiphonic.

As a matter of theory, it is suggested that the siphonostele
arose from the protostele for mechanical causes, in the Lycopsida
to support the branches, and in the Pteropsida to support the
leaves. Potonié also explains the origin of his second primitive
type the ‘‘pericaulom,” the homologue of the siphonostele, on
mechanical grounds.

In the light of these theories we can now apply ourselves to
an interpretation of the anatomy of the vascular system of the
Osmundaceae, and likewise note if the facts already dealt with
throw any light on the theories.

First, we are in a better position now to decide whether the
internal endodermis of O. cimnamomea is phloeotermal or not.
It has been noted that in similar cases, that is, in gamodesmic
stems, Strasburger has denied the phloeotermal character of the
internal endodermis. With regard to the internal endodermis
the following facts have been observed :

1. There is present the characteristic cuticularized ‘radial dot.”

[60]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 29

2. The structure and contents of the cells are materially the
same as of the external endodermis.

3. The sheath is continued into the portions which in some
individuals present the phenomenon of internal phloem, just as
in any form called by Van Tieghem and Strasburger gamostelic.
In the gamostelic type the phloeotermal character of the internal
endodermis has been admitted.

4. It generally connects with the external endodermis through
ramular gaps, and by no means rarely through foliar gaps.
When this occurs, there is no point at which it could be said that
the one stops or the other begins.

Having verified these facts in a great many cases, I am there-
fore of the opinion that the internal and the external endodermis
are homologous tissues.

Second, are the medullary tissues morphologically equivalent
to the cortical? Again we recapitulate observations.

1. They do not differ in structure or in contents.

2. The medulla very often contains brown sclerenchyma, at
least in three species studied, a tissue which, in other ferns,
never constitutes a part of the stele.

3. Medulla and cortex connect more frequently than not
through the ramular gaps in O. cimnamomea, and occasionally
through foliar gaps; and neither is there a transition in the
nature of the connecting tissues, nor any line at which we can
say, the cortical tissues lie externally to this and the medullary
tissues internally.

4. The cortical and the medullary brown sclerenchyma some-
times fuse through ramular gaps in O. cimnamomea.

5. Portions of stem of O. cinnamomea have been found which
are of the ‘“‘gamostelic” type of Van Tieghem. The medulla
in gamosteles is granted to be morphologically a cortical tissue.

The conclusion is evident for O. cinnamomea at least, and if
it be granted that the medullary tissues of this species are
morphologically equivalent to the cortical tissues, then biological
principles alone would demand a like conclusion for the other
species.

Third, of what type is the vascular system of O. cinnamomea ?
Again the facts must form the basis for a decision:

1. The young stem of O. cimmamomea possesses an entirely

[61]

30 FAULL: THE ANATOMY OF THE OSMUNDACEAE

closed hollow vascular cylinder, sheathed with phloem and
broken only immediately above the exit of a leaf trace; and at
a level higher up the cylinder is entirely closed again. There is
a medulla and an internal endodermis.

2. In older plants the leaves are more frequent, and the gaps
extend through several internodes; but yet the cylinder is the
unit. The cylinder of phloem is quite rarely broken, except
where branching takes place.

3. There is an internal endodermis which is persistent
throughout the entire central cylinder of the stem.

4. As a rule the internal endodermis bends out opposite leaf
gaps.

5. There is an internal phloem in portions of some plants.

6. Not only does the cylinder of external phloem remain
practically unbroken, but opposite leaf gaps there is on the
inner side a proliferation of sieve tubes. In O. rvegals Janczewski
found isolated sieve tubes in the parenchyma filling the leaf
gap; and the same thing is true of O. cimnamomea.

According to Van Tieghem’s stelar theory, the last two
facts can be explained only by considering the central cylinder
of the Osmundaceae to be ‘“gamostelic.” The centrad exten-
sions of the phloem opposite the medullary rays could then be
explained by assuming that steles had united laterally, with the
disappearance of phloem on the medullary side, but with the
partial persistence of phloem on the radial planes. This would
also explain the occurrence of internal phloem, the union of
internal and external endodermis, and the homology of medul-
lary and cortical tissue. But from the study of development
there is not a shred of evidence to prove that there has been a
union of steles. In fact, such a study shows distinctly that
there is but one stele in the stem of O. cinmnamomea from the
very first. Van Tieghem’s observations on O. regalis have
already been quoted (see InrRopUCTION); so we cannot describe
the cauline vascular system as “gamostelic,” if this name
implies a union of steles.

There remains yet another interpretation, namely, that the
vascular system of the stem of O. cinnamomea is a siphonostele
in which some degeneration from the primitive type has taken
place. It has been pointed out in a description of the concep-

[62]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 31

tion of the vascular system held by Dr. E. C. Jeffrey that the
most primitive siphonostele is the amphiphloic siphonostele.
In this there is an internal phloem and phloeoterma, and in its
phyllosiphonic form there are wide leaf gaps and branch gaps
through which internal and external phloem, internal and
external phloeterma, and medulla and cortex connect with each
other. In O. cimmamomea the gaps in this primitive type have
closed somewhat, so that medulla and cortex rarely connect
except through ramular gaps. Also the phloem forms an
almost unbroken cylinder, and the centrad proliferations opposite
the medullary rays are the vestigial relics of connection between
external and internal phloem. The internal phloem has also
disappeared in greater part.

With such a conception of the cauline vascular system of O.
cinnamomea, the centrad accumulation of sieve tubes opposite
the medullary rays, the occasional presence of sieve tubes in the
medullary rays, the fact of the internal phloem, the connection
of medulla and cortex through ramular and foliar gaps, the
presence of sclerenchyma in the medulla, the bending out of the
internal endodermis into the leaf gaps, and the facts of develop-
ment, all become intelligible.

Fourth, which of the species studied possesses the most prim-
itive type of central cylinder ?

After a fairly comprehensive study there is one feature that
stands out prominently, the great similarity and uniformity of
vascular structure in the various species of Osmunda and Todea.
According to Solms-Laubach the stems of fossil remains of this
family, of which none earlier than the Tertiary have been found,
do not present any striking differences from the living represen-
tatives. Paleobotany, therefore, offers no solution to the prob-
lem. Inspite of the conservatism of the central cylinder, there
are, however, minor anatomical differences. On the basis of
these alone, without referring to the young sporophytes, I think
there is sufficient warrant for placing O. cinnamomea at one end
of the series, possessing as it does an internal endodermis, inter-
nal.phloem, and wide ramular gaps. It is difficult to say which
species is to be placed at the other end of the series. In view
of the fact that O. Claytontana never has sclerenchyma in the
medulla, that there are small or even no ramular gaps, no internal

[63]

32 FAULL: THE ANATOMY OF THE OSMUNDACEAE

sclerenchyma, and even a degenerated external endodermis,
we may not be far astray in putting it in the position farthest
from O. cinnamomea.. Now of these two, which retains a cen-
tral cylinder more nearly primitive ? If O. vegalis has a medul-
lated monostelic central cylinder, as has hitherto been claimed
for it, then O. Claytoniana has also, and therefore, according to
Van Tieghem, a more primitive form than that of O. cinnamomea.
Assuming the correctness of this for the moment, it will be in
order next to see if such phenomena as presented by O. cinna-
momea could be derived according to Van Tieghem’s hypothesis
from such a simple medullated monostelic form as that of O.
Claytoniana.

The phloem sheath must have broken into bundles, and the
endodermis must have looped in between the bundles, and con-
nected around them on the centrad side. With the formation
of this astelic type some of the cortex would have been included
in the medulla, in evidence of which the sclerenchyma in the
pith would stand as proof. Then next the bundles must have
fused laterally to produce the gamodesmic type in which there
is an external and an internal endodermis. Granting that the
central cylinder could be so plastic in a single species, there are
left yet to be explained the continuous sheath of phloem, the
proliferation of sieve tubes opposite the medullary rays, the
occurrence of isolated tubes in the medullary rays, the occur-
rence of internal phloem, and the phenomena of the ramular
gaps. Further, there are no facts in development that point to
such a series of changes.

Turning now to the other alternative, namely, the possibility
that O. cinmamomea has the more primitive form of central cyl-
inder, it will be granted that by the degeneration of internal
phloem, endodermis, and medullary sclerenchyma, and by the
closing of the ramular gaps the central cylinder such as we find
in O. Claytontana wouldresult. In proof that such degeneration
could have taken place, it is to be noted (1) that in O. cnmamomea
itself, it has been pointed out that the amphiphloic condition is
localized, that the internal endodermis has already begun to
degenerate, that medullary sclerenchyma is not a constant fea-
ture, and that closed steles above the point of branching are not
at all uncommon; and (2) in further proof, analogous cases of

[64]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 33

degeneration within the same genus are frequent. Thus within
the genus Equisetum two species suchas &. arvense and E. hiemale
may be chosen, the first long considered medullated monostelic
and more primitive, the second gamodesmic and considerably
modified. But a study of development and of nodal por-
tions of the stem has shown that &. arvense has a reduced cen-
tral cylinder, the product of degeneration from a gamodesmic
type, and that therefore &. Azemale is nearer the primitive. Sim-
ilar cases of degeneration have been pointed out by Van
Tieghem, Poirault, and Jeffrey, in the genera Ophioglossum,
Botrychium, Equisetum, Ranunculus, etc. Very lately Boodle, 3738
has called attention to an interesting series of central cylinders
in the family Schizaeaceae. Anemia Phylhtidis has a ring of
separate bundles, each with a band of xylem surrounded by a
phloem, pericycle, and endodermis of its own; A. Mexicana has
a complete ring of xylem in the internodes with external and
internal cylinders of phloem and endodermis ; Schizaea has a
ring of xylem surrounding a central pith, but no internal phloem
or endodermis. It is likely that here, too, the Schizaea type is
derived from the Aneimia type by degeneration. In the Hymen-
ophyllaceae likewise, every grade is found from the case in
which the phloem of the solid stele forms a complete ring to
that in which it is developed on one side only.

After examining a number of comparatively young specimens
of O. Claytoniana, 1 am somewhat doubtful if the study of
the development of this species will throw any further light on
the subject of morphology; but for O. vegalis I am more hope-
ful. Nevertheless, aside from further developmental proofs, I
incline to the view that O. cimnamomea possesses the most
primitive type of central cylinder. I again recapitulate the
reasons :

1. The opposite view demands a very plastic central cylinder
in one species alone, not differing very greatly in habit from the
others.

2. There would still remain phenomena that the opposite
view could not explain.

37 BOODLE: Stem structure in Schizaeaceae, etc. Brit. Assn. Dover, 1899.

3 BOODLE: On the anatomy of the Hymenophyllaceae. Ann. Bot. 14: 455.
1900. f

[65]

34 FAULL: THE ANATOMY OF THE OSMUNDACEAE

3. There are no facts of development even in analogous
cases to support the opposite opinion.

4. The view adopted here demands only slight changes, and
those are of degeneration, to explain all the phenomena.

5. There are precisely similar analogous cases of degenera-
tion.

6. Within the species O. cinnamomea itself, every phase of
degeneration except the entire disappearance of internal endo-
dermis is observable in suitable specimens.

When we attempt to orient the other species amongst them-
selves, the task is more difficult, and of little importance. As
already indicated, a closer study of development may afford
more precise proofs. In the mature stems we have seen that
O. vegalis occasionally has sclerenchyma in the medulla, that
there are ramular gaps, though usually small, and that the external
endodermis is well developed. In O. Claytoniana, on the other
hand, sclerenchyma is never found in the medulla, ramular gaps
are infrequent, and the external endodermis shows indications of
degeneration. In neither of these species is internal endodermis
or internal phloem present. The probability, therefore, is that
in the genus Osmunda there is a series, O. cimnamomea possessing
the most primitive type of central cylinder and O. Claytoniana
the most degenerate, O. regalis occupying a middle position, but
nearer to the latter. It is merely interesting to note in passing
that Professor Campbell concluded % from his study of the pro-
thallia of O. Claytontana and O. cinnamomea, that the gametophyte
of the former was more specialized in many particulars, in other
words, was less primitive in type than the latter.

Fifth, does a study of the vascular system help to determine
the phylogenetic position of the Osmundaceae ?

It was stated at the beginning of this paper that botanists
have regarded the Osmundaceae as possessing an anomalous
form of central cylinder among the Filicales, their reason being
that it seemed to present more of the features of a central
cylinder such as is typical for dicotyledons, that is,a medullated
monostele in Van Tieghem’s terminology. In determining the
position of the family, therefore, in any natural system of

39 CAMPBELL: On the prothallium and embryo of O. Claytoniana and O. cinna-
momea. Ann. Bot. 6:49. 1892.

[66]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 35

classification, it was hopeless to try to reconcile this single
dicotyledonous character with the remaining filicinean characters,
and so the vascular system in the family was regarded as
anomalous.

It is fair to note that Zenetti dissented ‘4° from the prevailing
view, and evidently for the reason that he attached some value
to the nature of the central cylinder from the phylogenetic
standpoint. Hence he sought to find the same type amongst
the vascular cryptogams. He rejected the ordinary fern type
because it is ‘‘polystelic,” and the lycopod type because there
is no pith, obviously overlooking Selaginella laevigata, Phylloglos-
sum, etc. So finding no living form with which comparison could
be established he turned to paleophytology. Among the Lepi-
dodendraceae he found the prototype sought for, especially in
such of these fossils as ZL. Harcourtiz, and the Sigillarians, because
in these the wood is broken into bundles between which there
are medullary rays. But he evidently did not grasp the signifi-
cance of bundles and medullary rays in relation to leaf traces
and branch traces. In O. vegas, too, the protoxylem is endarch,
while in those ancient lycopods it was exarch. The stele of the
Lepidodendraceae, as in all plants bearing palingenetically small
leaves, was cladosiphonic, while O. vegais is. phyllosiphonic, as
are all primitively megaphyllous plants. Hence any attempt to
establish a relation between the central cylinder of modern ferns
and of those ancient horsetails must fail. Indeed, of the early
fossil forms preserved, the one with a central cylinder most
closely resembling that of the Osmundaceae, as has been pointed
out by Dr. Scott,** seems to be the cycadofilicinean Lyginoden-
dron (fig. 26).

Further, we dissent just as strongly from the view that the
family is anomalous in the matter of its vascular system. The
typical fern stem possesses an amphiphloic siphonostele, as is
especially revealed by a study of development. But degenerated
forms of this are to be met with in almost every family, some
examples of which have been noted. The Osmundaceae, as has
been shown above, all exhibit some degree of degeneration from
this type. It is therefore evident that the cauline vascular sys-
tem of this family is neither primitive nor anomalous among the
Filicales.

4° ZENETTI: of. cit. 73. AX SCOTT: of. cz.

[67]

36 FAULL: THE ANATOMY OF THE OSMUNDACEAE

SUMMARY OF OBSERVATIONS.

1. An internal endodermis has been demonstrated in Osmunda
cinnamomea, but in none of the other species examined. This inter-
nal endodermis is in textural continuity with the external endo-
dermis through branch gaps, and sometimes through foliar gaps.

2. Internal phloem has been found in O. cimnamomea in the
region of branching. This is continuous with the external
phloem through ramular gaps.

3. The external phloem of the Osmundaceae forms a contin-
uous cylinder, a fact which De Bary has stated for O. regalis ;
and is not broken opposite the medullary rays as Strasburger
has affirmed of the same species. Isolated sieve-tubes have
been found in the medullary rays of O. cimnamomea.

4. The xylem forms a cylinder broken only by foliar and
ramular gaps. |

5. Brown sclerenchyma has been shown to be usually present
in the medulla of O. cinnamomea, not uncommonly in O. regals,
and not at all in O. Claytoniana. It occurs likewise in Todea
barbara, but has not been observed in 7. superba.

6. The medullary and cortical tissues of the Osmundaceae
are histologically equivalent. Brown sclerenchyma, which is
not an intrastelar tissue in other ferns, occurs in both medulla
and cortex; and in O. cinnamomea the brown sclerenchyma of
the medulla is in continuity with that of the cortex.

7. In O. cinnamomea the typical ramular gap is one through
which internal and external endodermis, internal and external
phloem, cortex, and medulla connect. Every stage of degen-
eration has been observed in O. cinnamomea, however, down to
the completely closed steles. O.vegals has a gap in the wood
only, and O. Claytoniana usually none.

8. The so-called ‘quergestreckte Zellen” pointed out by
DeBary in O. vegas, and more fully described by Zenetti, have
been found in all the species studied. They are sieve tubes,
possessing all the characteristic features of sieve tubes, even
that of callus plugs. Their irregularity of orientation is shared
by the other peripheral tissues of the central cylinder, and is
apparently due to disturbance caused by the exit of the large
leaf traces.

g. Callus plugs have been demonstrated in the sieve tubes.

[68]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 37

10, A study of the growing point has further shown that the
‘‘quergestreckte Zellen”’ and the typical protophloem are of the
same kind; but it has failed to verify Strasburger’s statement
that the pericycle and the endodermis arise from a common
maternal layer.

11. The phloem forms a continuous sheath in the leaf.

12. The root possesses a protostelic, diarch, occasionally
triarch, vascular axis.

SUMMARY OF CONCLUSIONS.

1. The internal endodermis in O. cinnamomea is to be
regarded as phloeotermal in nature, a fact denied by Strasburger
in homologous cases.

2. The medullary and cortical tissues seem to be morpho-
logically equivalent.

3. Observations on the anatomy of the Osmundaceae have
been confined heretofore to the cosmopolitan O. regalis, and the
subtropical Todeas. From these observations it was concluded
by Van Tieghem that this family possessed a type of central
cylinder anomalous among the vascular cryptogams, a type
(the medullated monostelic type) peculiar to the phanerogams.
The writer dissents from this view. It appears to be the case
that the central cylinder of O. cimmamomea is not medullated
monostelic, for the medulla is obviously extrastelar. Further,
it cannot be regarded as gamodesmic on account of the topo-
graphical distribution of the phloem. The most obvious inter-
pretation seems to be that it is a degenerate form of the amphi-
phloic siphonostelic type of central cylinder (polystelic type of
Van Tieghem). O. cinnamomea, O. regalis, O. Claytoniana form
a series arranged in order of degeneration of their central
cylinders, and the same is true of 7. barbara and T. superba.

The present research was carried on in the Biological Depart-
ment of Toronto University under the direction of Dr. E. C.
Jeffrey, to whom I wish here to express my obligations for his
advice throughout. My thanks are due to Professor R. Ramsay
Wright for the facilities afforded in the department. For some
of the material used I am indebted to Mr. Oakes Ames, Assistant
Director of the Botanical Gardens, Harvard University; Sir

[69]

38 FAULL: THE ANATOMY OF THE OSMUNDACEAE

William Thistleton Dyer, Director of the Royal Gardens, Kew;
Dr. Brodie, Toronto; and Mr. R. B. Thomson, B. A.

UNIVERSITY OF TORONTO.

EXPLANATION OF PLATES XIV-XVII.

Abbreviations used.

cp, callus plugs. PB, pericycle.

c, cortex. ph, phloem.

é. ¢., external cortex. Dp. ph., protophloem.

Zz. c., internal cortex. px, protoxylem.

e, endodermis. gu, ‘quergestreckte Zellen.”

é. é., external endodermis. ~¥7%, root.

z.é., internal endodermis. _ s. s. s., strands.

Zt, leaf trace. sc, sclerenchyma.
m, medulla. x, xylem.

m.r., medullary ray.

PLATE XIV.

Fic. 1. Transverse section of the stem of Osmunda cinnamomea.

FG. 2. Transverse section of part of central cylinder of O. cinnamomea.

Fic. 3. Tangential section of O. regadis.

Fic. 4. ‘Quergestreckte Zelle” of 7. darbara, showing sieve plates and
callus plugs.

Fig. 5. Transverse section of leaf trace of O. Claytoniana near the grow-
ing point.

Fic. 6. Transverse section of part of central cylinder of O. cinnamomea.

Fic. 7. Sieve tubes of 7. barbara, showing sieve plates, ‘ globules bril-
lants,”’ and callus plugs.

Fie, 8. Transverse section of part of central cylinder of O. Claytoniana.

PLATE XV.

Fic. 9. Tangential section of ZYodea barbara, showing “ quergestreckte
Zellen.”

Fic. to. Transverse section of O. cimnamomea, immediately above point
of ramification, showing open branch gaps.

Fic. 11. Transverse section of O. cinnamomea through nearly the same
region in another plant.

Fic. 12. Transverse section of same plant as in fg. zz, but lower down.

Fic. 13. Transverse section of same plant as in fg. 72, but lower down.

Fic. 14. Transverse section of the central cylinder of O. cimnamomea,
showing internal endodermis and brown sclerenchyma in the medulla.

PLATE XVI.
Fic. 15. Transverse section of central cylinder of O. cémamomea, show-
ing internal endodermis and an absence of brown sclerenchyma in the

medulla.
[70]

FAULL: THE ANATOMY OF THE OSMUNDACEAE 39

Fic. 16, Transverse section of the stem of O. cinnamomea in the region
of forking, showing absence of ramular gaps.

.Fig. 17. Transverse section of the stem of O. Claytoniana.

Fic. 18. Transverse section of the stem of O. C/ayfoniana in the region
of forking.

FIG. 19. Transverse section of the stem of O. vega/zs in the region of
forking.

Fic. 20. Transverse section of the central cylinder of O. regalis, show-
ing brown sclerenchyma in the medulla.

PLATE XVII,

FiG, 21. Transverse section of the central cylinder of O..cinnamomea,
showing internal phloem.

Fig. 22. A part of the central cylinder of O. ciéwmamomea shown in fig.
2z more highly magnified.

Fic. 23. A transverse section of the young sporophyte of O. Claytoniana,
showing one foliar gap, und the corresponding leaf trace opposite.

Fig. 24. Transverse section of the stem of 7. daréara, showing brown
sclerenchyma in the medulla.

Fig. 25. Transverse section of a part of the stem of 7. darbara nearer the
growing pcint.

Fig. 26. Transverse section of Lyginodendron Oldhamium, showing a
leaf gap, a leaf trace opposite, strands of sclerenchyma in the medulla, and
strands of primary xylem centrad of the cylinder of secondary xylem.

[71]

PLATE XIV

BOTANICAL GAZETTE, XXXII

a WwW,

mY.

es

Uh hy

iii

a -~ S
=

FAULL on OSMUNDACEAE.

Arvin

rt
Dae abies pie | :
i - ’
sd TP ig eal ee
"ae pe a Cee |
i fe ‘ iN 7 ;
+ i ' S f
ty A ® gpa ¥
ne z te i
: 4 7 be Fi
r= sal nn G z : '
vv ee ‘
:
ay & =

PLATE XV

BOTANICAL GAZETTE, XXXI11

= eo

=

a

wi

FAULL on OSMUNDACEAE.

BOTANICAL GAZETTE, XXXII PLATE XVI

19 20

FAULL on OSMUNDACEAE.

BOTANICAL GAZETTE, XXXI1 PLATE XVII

FAULL on OSMUNDACEAE.

UNIVERSITY OF TORONTO
STUDIES

BIOLOGICAL SERIES.

No.2:

ON THE IDENTIFICATION OF MECKELIAN AND
MYLOHYOID GROOVES IN THE JAWS OF
MESOZOIC AND RECENT MAMMALIA,

BY B. ARTHUR BENSLEY.

THE UNIVERSITY LIBRARY: PUBLISHED BY
THE LIBRARIAN, 1902.

COMMITTEE OF MANAGEMENT.

Chairman: JAMES Loupon, LL.D., President of the
University.
PROFESSOR W. J. ALEXANDER, Ph. D.
PROFESSOR PELHAM EDGar, Ph. D.
PRINCIPAL J. GALBRAITH, M.A.
ProFEssoR R. RaMSAY WRIGHT, M.A., B.Se.
PROFESSOR GEORGE M. Wronc, M.A.

General Editor: H. H. Lanetron, B.A. Librarian of the
University.

ON THE IDENTIFICATION OF MECKELIAN AND
MYLOHYOID GROOVES IN THE JAWS OF
MESOZOIG AND RECENT MAMMALIA

BY
B. ARTHUR BENSLEY, B.A.

LECTURER ON ZOOLOGY IN THE UNIVERSITY OF TORONTO

TRC ae aah, i
; ry re ee
> Te Cm, iy 1%, dainy TA if K t oT, +f j \
’ ral er Apes :
; wa 7 Me et ts i } i
id : i iA :
i \ “ : ’
‘ y } j tr}
iM ‘ r ’ i ‘ ' Wek r ‘
‘ i a { \
i Se cun shape
M » nity wt
1 Ca 4 Mud oe ,
ney ene \
: ) Re, ate i wit
+. Papo aes
l y amis © ys i n't
i { yu b
: *
ay
|
} A f
\ : fj '
| ad am
I \
: if
i iM {
‘ ‘ ry :
i
} J
. f | .
é )
A
i t * 'y i
A s
0
’
, ‘
vay
Ny {
wt
f ?
" “i i
chy
.
if )
; Wi
,
ik fi ‘ i
AY ms r h
\ i
aX) ! }
;
5 @ H
i} [ LL \ , i]

ON THE IDENTIFICATION OF MECKELIAN AND
MYLOHYOID GROOVES IN THE JAWS OF
MESOZOIC AND RECENT MAMMALIA

Owen, in his well-known Monograph of the Mesozoic Mammalia,
noted the common occurrence of a linear furrow on the inner surface
of the jaw in Jurassic mammals which he designated asa “mylohyoid
groove,” and Marsh (’87) has described a similar groove for several
species from the Jurassic deposits of America. The possible signifi-
cance of this structure was first commented upon in the discussion
which took place during the years 1838-39 as to the mammalian or
non-mammalian nature of the original specimens of Amphitherium
and Phascolotherium from the Stonesfield Slate, De Blainville (38, p.
733) having called attention to an inferior marginal groove, in Amphi-
therium, which he regarded as a suture and as indicative of the com-
posite structure of the jaw in question. De Blainville’s opinion was
criticized at the time by Duméril, but accepted by Grant (’39), the
latter considering the composite structure of the jaws to be obvious
“from the distinct deep fissure extending along their base between
the dental and opercular pieces.”” Owen (’38), on the other hand,
believed the grooves to be due to the pressure of a nerve or vessel,
although he appears to have been, at the time, a little uncertain of
his interpretation, partly, perhaps, on account of the presence of a
second upper groove in the specimen of Phascolotherium, and partly,
doubtless, on account of the criticisms of Ogilby (’38) who, while
having no explanation of his own to offer as to the significance of the
grooves, objected to their being regarded as of vascular origin. In
his ‘‘ British Fossil Mammals,” published some years later (’46), how-
ever, Owen replied to the opinions of Ogilby and Grant, and showed
that the groove in the specimens of Amphitherium possessed an
entire surface, and was therefore not referable to a suture. In his
subsequent monograph (’71) he repeated his opinion, referring to the
groove as ‘“‘mylohyoid.” Following Owen, Osborn (’88) has compared
the structure with the true mylohyoid groove in the human jaw, at
the same time considering it to be of little taxonomic value on account
of its variable presence in recent mammals. In the ninth edition of
the Encyclopedia Britannica we find the following statement by

[75]

4 BENSLEY : MECKELIAN AND MYLOHYOID GROOVES

Flower (’83, p. 376)—“the mylohyoid groove [is] persistent [in
Amphitherium|, as in some of the existing Marsupials and the Whale-
bone Whales. This groove, aremnant of that which originally lodges
Meckel’s cartilage, mistaken for a suture, was once considered evi-
dence of the reptilian nature of these jaws.” It thus appears that two
somewhat similar, but fundamentally distinct structures have been
confused under the designation ‘“‘ mylohyoid groove,” and it is chiefly
towards pointing out their distinction and distribution, and the
taxonomic importance of that occurring in the mesozoic Mammalia,
that the following remarks’ are directed.

The true mylohyoid groove appears to be typically developed
only in the primates and artiodactyl Ungulata, although it is by no
means always present in the former, as, for example, in Zorzs, Lago-
thrix, Tarstus, and Galago demidofi. It is fairly frequent in the
KEdentata (Zatusia, Bradypus, Cholepus). It is present in Lepus
among the Rodentia, and possibly also in other forms, but is difficult to
identify in this group on account of the presence of other mandibular
grooves. As to its general character it is usually a broad superficial
furrow (cf Plate, fig. 3, my.), beginning near and below the dental fora-
men and terminating abruptly a short distance forwards. It is fre-
quently double, and its extent of development varies somewhat in
different individuals. In the human jaw this groove is stated by
Quain (’82, p. 56) to lodge the mylohyoid nerve with its accompanying
artery and vein.

The relations of the groove which occurs in the mesozoic Mammalia
have been amply illustrated by Owen (’71), Marsh (’87), Osborn (’88),
and Goodrich (’94). I'wo of its modifications are here represented by
figs. 1 and 2 in the plate (Amdblotherium, Spalacothertum), copied
from Osborn’s memoir. Apart from its regular linear outline its more
important features are as follows: (a) its close relation with the
dental foramen ; in many forms (Amphitherium, Amblotherium, Am-
philestes, Spalacotherium, Tinodon) it appears to be simply an anterior
continuation of the latter ; (b) the fact that while sometimes confined
to the posterior part of the jaw (Spalacothertum, Phascolotherium,
Amphilestes) it frequently traverses the whole length from the dental
foramen to the symphysis (Amblothertum, Achyrodon, Dryolestes,
Docodon). As regards the distribution of this groove it is typically

1 Where not otherwise noted these are based on the osteological collection of the
British Museum (Natural History), London.

[76]

BENSLEY: MECKELIAN AND MYLOHYOID GROOVES 5

developed only in the presumably higher mesozoic forms, being absent
as far as known in the Multituberculata.

There can be no doubt that Flower’s opinion as to the relation
of this structure with Meckel’s cartilage is the correct one. An
exactly similar groove lodging Meckel’s cartilage may be seen in
embryos of existing mammals, and its somewhat close resemblance to
the true mylohyoid groove may be easily shown to be the result of
coincidence. In the plate (figs. roa-d) will be found illustrations of
four transverse sections through the lower jaw of a 6cm. pouch-fcetus
of Macropus'. Section a, taken immediately behind the symphysis,
passes through what is at this stage the anterior limit of Meckel’s
cartilage. ‘The symphysial portion of the latter has already been
‘reduced or the anterior part of the jaw has grown beyond it. This
section and section 4, which is taken further back,show Meckel’s
cartilage lodged in a groove on the inner surface of the bony man-
dible, and separated by a bony strand from the dental nerve and
artery in its interior. Section d, taken immediately behind the dental
foramen, shows the dental nerve and artery in close relation with
Meckel’s cartilage, while section c shows the condition at the foramen
where the nerve and artery become separated from the cartilage, the
two former passing into the body of the jaw while the latter enters
the groove on its inner surface. A short distance posterior to section
d in this series the nerve and artery are seen to give off mylohyoid
branches.

A comparison of these sections and of the dissection of a feetal
jaw represented in the plate (fig. 5) will suffice to show the identity
of the groove lodging Meckel’s cartilage in the embryo with that in
the mesozoic Mammalia. ‘The cause of its resemblance to the true
mylohyoid groove will also be apparent, since in the embryo we find
Meckel’s cartilage leaving the dental nerve and artery at the dental
foramen and passing into a groove on theinner surface of the jaw, in
much the same way that, at a later stage, the mylohyoid branches
leave the inferior dental trunks at the foramen and pass into the
mylohyoid groove. It is probable that the condition described above
for Macropus, namely, the lodging of Meckel’s cartilage in a groove,
represents the general one in the Mammalia. Parker (’85) has de-
scribed and figured it for several of the Edentata and Insectivora, and

‘For this specimen, with many others, the writer is indebted to Professor Bash-
ford Dean, of Columbia University, New York.

L77]

6 BENSLEY : MECKELIAN AND MYLOHYOID GROOVES

the writer has observed it in the case of several genera of marsupials,
including AZyrmecobius, Phascogale, Trichosurus, Phalanger, Dasyu-
rus, Perameles, and Thylacomys'.

It is obvious that in many cases the mylohyoid groove must,
during development, become superposed to the Meckelian groove.
Magitot and Robin (’62) have described what is apparently that con-
dition for man. But that such is not always the case may be seen
from the forms represented in figs. 87and 9 of the plate, in which
both grooves are present with similar relations to the dental foramen
but with different positions in the jaw.

Considering the nature of the groove represented in the mesozoic
Mammalia we can scarcely expect to find it fully developed in adults
of recent mammals. Owen (’38, ’71) described and figured a groove
in the jaw of Wyrmecobius, which he regarded as equivalent to that
in the mesozoic forms, but Osborn (’88) was unable to recognize this
structure in two specimens belonging to the Yale University collec-
tion, and he has further stated, on the authority of Mr. Thomas, that
it is absent in the British Museum specimens. Leche (’g91) also
failed to find it in three of his specimens, but has mentioned its pres-
ence in a fourth immature one. The fact of the matter is that a
short broad furrow does occur in MWyrmecobius exactly as Owen has
described and figured, but its great width almost precludes its being
spoken of as a groove, and it has obviously nothing to do either with
the mylohyoid or the Meckelian groove. Its presence is due simply
to the elevation of the internal alveolar edge. A much more definite
groove, due to the same cause, is frequently present in recent mam-
mals (cf Plate, fig. 4).

Owen also mentioned a similar groove for Phascolomys. ‘This
structure, which is amply illustrated in the British Museum speci-
mens, appears to represent a mylohyoid groove. In adult jaws it is
frequently found to be branched. Ina young animal of which the
writer dissected this region, the posterior portion of the groove was
alone developed, and it lodged the mylohyoid nerve. In the young
wombat the groove is placed just at the point where the anterior por-
tion of the inflected angle joins the body of the jaw. A simular
structure is frequently present in other marsupials. Its somewhat

1 The specimens representing these genera were kindly lent by the late Mr. Mar-
tin F. Woodward, of the Royal College of Science, London.

2 For the loan of this specimen—a foetal jaw of Propithecus—the writer is in-
debted to Dr. Forsyth Major, of the British Museum, London.

[78]

BENSLEY : MECKELIAN AND MYLOHYOID GROOVES 7

different appearance as compared with the mylohyoid groove of plac-
entals is due to the presence of the angular inflection.

Undoubted traces of the Meckelian groove are, however, to be
seen in adults of recent mammals, although in most cases only as
variations. Fig. 4 of the plate shows the appearance of it in an aged
specimen of Didelphys. ‘This may be compared with fig. 9 which
shows the normal condition in a young animal. Figs. 6 and 7 show
the condition in two other specimens representing Zatusza and
Chrysochloris ; the former is not fully adult. Similar conditions are
observable in some specimens of the following forms,—77zchosurus,
Phalanger, Perameles, and FPetauroides, among the marsupials,
Xenurus and Dasypus among the Edentata, MHemzcentetes and
Echinops, among the Insectivora. The groove is doubtless frequently
present in many other forms, but such a reducedand variable
structure almost defies recognition. As stated by Eschricht and
Reinhardt (’66) it is present in the adult of Balena mysticetus.

It is an interesting question why the Meckelian groove is not
present in the Multituberculata. Following Cope’s suggestion (’88)
as to the monotreme affinities of these animals, the writer examined
the condition in three specimens of Achzdua of 2, 9, and 16cm. head
and body lengths. The relations in this form, however, proved disap-
pointing. In the 2cm. egg-embryo there were only a few traces of
bone formation in the lower jaw, while in the two larger animals the
dentary element was well formed but was not in relation with Meckel’s
cartilage posteriorly. In both of the later stages the symphysial
portion of the cartilage was seen to be very much elongated, and for
a short distance behind the symphysis the cartilage was lodged in a
concavity of the dentary bone. Immediately posterior to this point,
however, the cartilage was found to leave the jaw, and to pass back-
wards independently of it. In the 9cm. embryo its position was in-
ternal and ventral with reference to the jaw, and in the older animal
its separation from the jaw and its internal position were still more
marked.

The condition in Achzdua is apparently the result of the great
reduction or degeneration of the jaw characteristic of this form. It
is possible that the conditions in Oruzthorhynchus might throw some
light on the question, but no embryos of this form were available.
It seems most unlikely that in the Multituberculata the Meckelian
cartilage could have had the same relations as in existing mammals,

[79]

8 BENSLEY: MECKELIAN AND MyYLOHYOID GROOVES

and was absent in the adult stage. There is a possibility that in the
mammalian prototype the cartilage was either completely enclosed in
the dentary bone or co-ossified with it. It is interesting to note that
Parker (of. cit.) has described a partial ossification of the cartilage in
the young of Centetes and Talpa, and a partial enclosure of it in the
latter form and in Er7zaceus. ‘The frequent exposure of Meckel’s
cartilage in the jaws of lower Vertebrata, however, warns against the
adoption of such an explanation before the acquisition of more definite
evidence.

BIBLIOGRAPHY

1888 Cope, E. D., The Multituberculata Monotremes. American Naturalist, March,
1888, p. 259-

'38 De Blainville, H., Nouveaux Doutes sur le prétendu Didelphe de Stonesfield.
Comptes Rendus, t. vii., pp. 727-736.

‘66 Eschricht, D. F., and Reinhardt, J., On the Greenland Right Whale (Balzena
mysticetus, Linn.) (Translation). Ray Soc., 1866, pp. 1-143,

'83 Flower, W. H., Article ‘‘Mammalia.’’? Encyclopedia Britannica, 9th Ed.,
vol. Xv.

’94 Goodrich, E. S., On the Fossil Mammalia from the Stonesfield Slate. Quart.
Jour. Micr. Sci., vol. xxxviii., N.S.. pp. 407-432.

'39 Grant, R. E., Thompson’s British Annual. Quoted in full by Owen (’46, p. 38).

’9t Leche, W., Beitrige zur Anatomie des Myrmecobius fasciatus. Verh. d. Biol.
Vereins in Stockhoim, Bd. iii., Nr. 8, S. 136-154.

‘62 Magitot, E., and Robin, Ch., Mémoire sur un organe transitoire de la vie foetale
désigné sous le nom de Cartilage de Meckel. Ann. Sci.Nat., iv me Série, Zoologie,
t. xvili., pp. 213-241.

‘87 Marsh, O. C.. American Jurassic Mammals. Am. Journ. Sci., vol. xxxiii., 3rd
series, pp. 327-348.

38 Ogilby, W., Observations on the Structure and Relations of the presumed Mar-
supial Remains from the Stonesfield Oolite. Proc. Geol. Soc. London, vol. iii.,
1842, pp. 21-23.

’88 Osborn, H. F.. The Structure and Classification of the Mesozoic Mammalia.
Jour. Acad. Nat. Sci. Philadelphia, vol. ix., No. 2, pp. 186-265.

’38 Owen, R.,Observations on the Fossils representing the Thylacotherium Prevost-
ii (Valenciennes), etc. Trans. Geol. Soc. London, vol. vi., 1841-42, pp. 47-65.

‘46 Owen, R., British Fossil Mammals. London, 1846.

7 Owen, R., Monograph of the Fossil Mammalia of the Mesozoic Formations.
Palzontographical Soc., London.

85 Parker, W. K., On the Structure and Development of the Skull in the Mamma-
lia. Part IJ., Edentata; Part III., Insectivora. Trans. Roy. Soc. London, vol.
176, PP. 1-275.

82 Quain’s Anatomy (Thomson, Schafer, and Thane), 9th ed., vol. 1.

[80]

(as

Io,

BENSLEY: MECKELIAN AND MYLOHYOID GROOVES 9

EXPLANATION OF PLATE

. Amblotherium soricinum. Right mandibular ramus, showing the groove for

Meckel’s cartilage. (After Osborn. )

. Spalacotherium tricuspidens. Weft ramus reversed, showing the relation of

the groove for Meckel’s cartilage with the dental foramen. (After Osborn.)

. Mycetes ursinus. Right ramus, showing the true mylohyoid groove.
. Didelphys marsupialis. Right ramus of an old individual, in which traces of

the groove for Meckel’s cartilage are present as a variation.

. Macropus sp. Moist preparation of the left ramus of a 7cm. pouch-fcetus,

showing Meckel’s cartilage lodged in its groove. The ear-bones are schemati-
cally represented.

. Tatusia novemcincta. Right ramus of an immature individual, showing the

groove for Meckel’s cartilage and the mylohyoid groove below it.

. Chrysochoris trevelyanus. eft ramus, showing traces of the groove for

Meckel’s cartilage in the adult.

. Propithecus sp. Right ramus of foetus ; the mylohyoid groove is here formed

below that lodging Meckel’s cartilage.

. Didelphys marsupialis. Right ramus of an young individual showing the nor-

mal appearance of the groove for Meckel’s cartilage in the later stages of its
reduction.

a-d. Macropus sp. ‘Transverse sections through the right ramus of a 6cm
pouch-fcetus, showing the relations of Meckel’s cartilage and the dental nerve
and artery to the jaw. For explanation see text.

Abbreviations
mg—groove for Meckel’s cartilage. ty—tympanic annulus.
my—mylohyoid groove. u—dental nerve.
mc—Meckel’s cartilage. a—dental artery,
ml—malleus. c—coronoid process of mandible.
i—incus. fm—masseteric foramen.

st—stapes.

[8x]

i j Mi a mii WAY a ein
mes “uae

, ty, i
tive: OLA
1s }

ee NP a Bary
Bak i Bi hi hate ”

4

STUDIE,

UNIVERSITY OF TORONTO
STUDIES

BIOLOGICAL SERIES

No. 4: THE MEGASPORE-MEMBRANE OF THE
GYMNOSPERMS, By R. B. THOMSON

THE UNIVERSITY LIBRARY, 1905: PUBLISHED BY
THE LIBRARIAN

COMMITTEE OF MANAGEMENT

\

Chairman : James Loupon, ay Ds ‘President of hee | ity
University. ;

-PRoressor W. J. ALEXANDER, Ph.D.
PRoFessorR PELHAM Epcar, Ph.D.

‘PRINCIPAL J. ‘GatpraitH, M.A.

PROFESSOR R. RAMSAY Wricut, M.A., B. Sc. |

PROFESSOR GrorcE M. Wronc, M.A.

ueaee

THE MEGASPORE-MEMBRANE OF THE
GYMNOSPERMS

BY
R. B. THOMSON, B.A.

INSTRUCTOR IN BOTANY IN THE UNIVERSITY OF TORONTO

anetnyt

. vid 7 ay
ia x ’
; 0 A eae Weil aE,
D ee | ae an Hy ,
4 : lor ara As | a y

i
-)

‘
4

Ae ey

THE MEGASPORE-MEMBRANE* OF THE
GYMNOSPERMS

Hofmeister, in his classic comparative researches of 1851,
indicated the fundamental connection between the ovule
of the gymnosperms and the sporangium of the higher crypto-
‘ gams. He showed the homology of the embryo-sac of the
former with the free spore of the latter in words which deserve
prominence in connection with the present work (Hofmeister,
p. 141): ‘‘Der Embryosack der Coniferen lasst sich betrachten
als eine Spore, welche von ihrem Sporangium umschlossen
bleibt.’’ Mettenius,? Solms-Laubach,*? Scott* and Worsdell 5
have directed attention to the cryptogamic nature of the
foliar and peduncular bundles of the gymnosperms, thus
carrying out in another direction the generalization which
Hofmeister reached in the case of the ovule. The occurrence
in palaeozoic strata of a group of forms,designated by Petonié ®
the ‘‘Cycadofilices,’’ forms which combine pteridophyte and
gymmnosperm features, and some of which have been recently
shown to be seed-bearing (the so-called Pteridospermae ’‘),
indicates the close connection that existed in past times
between these two phyla. The discovery within the last
few years of antherozoids in the cycads and Ginkgo by Hirasé, 8
Ikeno,? Webber,!° and Lang,!! affords more evidence of the
relationship of the gymnosperms to the vascular cryptogams.

The investigation with which the present account is con-
cerned deals with another of the cryptogamic features of the
gymnosperms. A megaspore-membrane has for a long time
been somewhat generally recognized as occurring in certain
forms of this group of primitive seed-plants. The extent of
its occurrence and the character of the coat in the whole group
have, however, not as yet received the attention they deserve,
in view of the comparison made by Hofmeister in 1851, as
indicative of the free-sporing ancestry of the gymnosperms,

* The term megaspore-membrane or megaspore-coat is applied ordin-
arily only to the coat of the uninucleate spore. It is here used to designate
as well the investing coat of the prothallium, undoubtedly the representative
of the former which has been delayed in development.

[85]

4 THOMSON : THE MEGASPORE-MEMBRANE

and from the phylogenetic importance which attaches to the
relative state of development of the membrane in the different
groups and subgroups of this division of the spermatophytes.
Indeed at the present time, information which will help to
indicate the phylogenetic positions, especially of the subgroups
of the conifers, is very desirable on account of the balanced
state of conflicting testimony from various sources, especially
of such as is based on the interpretation of the female ‘‘flower,”’
and also because of the lack of historical evidence, a lack
which Dr. Scott clearly states in his recent Fossil Botany (p.
483): ‘‘On the whole, it is impossible, in the present state of
knowledge, to say which tribeor family of the Coniferae is
the most ancient.’”’ It is then chiefly with the object of
affording evidence of the relative antiquity of the forms of
the Coniferales that this study of the megaspore-membrane
is concerned.

In addition to the investigation of the megaspore-mem-
brane, the tapetum has received considerable attention, since
a relationship between the two structures was observed
during the progress of the work.

Material for the purpose of the investigation was secured
during the last two years. Many of the forms required not
being native, I could not personally collect material of them
and am indebted to the kindness of the following persons
for rendering it available : Sir W. T. Thiselton-Dyer, Direc-
tor of the Royal Gardens, Kew; D. H. Scott, M.A., Ph.D.,
Honorary Keeper of the Jodrell Laboratory, Royal Gardens,
Kew; W. H. Lang, D.Sc., Lecturer in Botany, University of
Glasgow; Wm. Trelease, Director of the Missouri Botanical
Gardens, St. Louis; J. M. Coulter, Ph.D., Head Professor of
Botany, University of Chicago; C. J. Chamberlain, Ph.D.,
Instructor in Botany, University of Chicago; M. A. Chrysler,
Ph.D., Fellow in Botany, University of Chicago; A. A. Lawson,
Ph.D., Lecturer in Botany, Stanford University; B. L. Robin-
son, Ph.D., Asa Gray Professor of Systematic Botany and |
Curator of the Herbarium, Harvard University; E. W. Oliver,
A.M., Instructor in Botany, Harvard University; W. C. Coker,

[86]

OF THE GYMNOSPERMS 5

Ph.D., Professor of Botany, University of North Carolina; H.
J. Webber, Ph.D., U.S. Department of Agriculture; L. Cocayne,
Ph.D., Christ Church, New Zealand; Dr. M. Treub, Director
of the Botanic Garden, Buitenzorg, Java; J. P. Lotsy, Ph.D.,
State Botanist, Tjibodas, Java; Dr. Th. Valeton, Botanic
Garden, Buitenzorg, Java; and J. H. Faull, Ph.D., Lecturer
on Botany, University of Toronto. I gratefully acknowledge al-
so my indebtedness to Dr. Brodie, of Toronto, for much valu-
able information on the localities of the native species. Many
gaps in this present work remain to be filled, but whatever of
completeness it possesses is largely due to the kindness of
those who so liberally responded to the requests for material.
To Professor E. C. Jeffrey, of Harvard University, I am es-
pecially indebted for suggestions and assistance in the work.

In the following account the order in which the forms are
taken up is that indicated in Engler and Prantl’s ‘‘Die natur-
lichen Pflanzenfamilien.”’

CYCADALES.

I. Cycadeae: Ovules of Cycas revoluta at a comparatively
eatly stage of development have a well marked megaspore-
membrane. The coat is 4.5 » thick at the stage when the
prothallium consists of a parietal stratum of protoplasm with
a single row of numerous nuclei* embedded in it. It consists
of two layers (photo. 1, pl. 3),the endosporiumy and the exos-
porium, which are approximately equal in thickness at this
stage. The double nature of the coat is, however, usually
not apparent in sections from ordinarily prepared paraffin
material. The coat withstands the infiltration of paraffin
much more than the tissues of the ovule and in consequence
is often ‘‘dragged”’ in sectioning. In the ‘‘dragged’’ condition
it appears as a single layer with fine granules slightly projecting
from its surface. It is variable in thickness but ordinarily
about twice as thick as the norm. The true structure of the
coat is always apparent in sections from ovules which have

* About roo nuclei are present in a 5 # section through the megaspore,
and cell formation is beginning in the basal region.
+ The endosporium is always towards the left in the figures, next the
prothallium.
[87]

6 THOMSON : THE MEGASPORE-MEMBRANE

been subjected to prolonged infiltration of paraffin (m.p. 56°C.)
and cut 2 to 5» thick at alow temperature. In such prepara-
tions the endosporium and the exosporium are quite distinct
from each other, a clearly defined line of demarcation occurring
between them. Along this line in the sections a separation
of the layers sometimes occurs for a short distance (middle of
photo. 1, pl. 1). Moreover, the two layers of the coat are
different from each other in structure, and hence the double
nature of the coat is the more apparent (fig. 1, pl. 1). The
exosporium is granular towards its somewhat irregular external
border and a fine radial striation can be made out in this part,
while towards the inside it becomes gradually more homogen-
eous. The endosporium appears in section to be sub-divided
into two longitudinal strata, by a granular area (fig. 1, pl. 1).
The outer of these strata is very finely granular and yellow in
colour in the unstained condition, as is the exosporium, while
the inner is quite homogeneous, translucent and almost colour-
less. The zone separating the two sublayers is indistinctly
defined, the area in question presenting the appearance of
being radially and somewhat coarsely granular at its centre,
gradually merging into the strata on either side of it. In
very rare instances I have observed that, in the sections of the
coat, the two strata of the endosporium are torn from one
another at places along the granular area. The endosporium
besides being quite different from the exosporium in structure,
expands much more than the latter in certain fluids. A tension
is thus set up between the layers which, in sections, results
either in their separation, or in a ‘‘coiling’’ of the whole coat.
The former occurs in cases where the sections of the coat
remain attached to the slide (middle of photo. 1, pl. 1), the
latter where the sections are detached and float freely in the
fluid. In addition, it would seem that the endosporium is
more impenetrable to paraffin than the exosporium, since in
sectioning the former is often ‘‘dragged” and its structure
obscured, while the latter is clearly cut.

The double nature of the coat and the structural as well
as the chemical differences in its layers are clearly brought

[88]

OF THE GYMNOSPERMS 7

out by the action of stains and reagents. A number of stains
were experimented with, but the best results were obtained
from the use of a very dilute solution of Ehrlich’s acid haema-
toxylin, with a small amount of alum solution added, followed
by a dilute aqueous solution of safranin. The exosporium
stains cherry red, the endosporium a variety of colours passing
into one another gradually. From dark red along the outer
border the endosporium becomes reddish-violet, then dark
violet along the median granular area, and finally dark and
then lighter yellow towards the inner side. Sometimes a tint
of orange is seen in the yellow area. The action of the stains
on the layers of the coat gives an intimation at least of their
chemical composition. A more definite idea of this is obtained
by the use of certain standard reagents. Most reliance has
been placed on the action of chlorzine iodine. When thin
sections of the young membrane (photo. 1, pl. 1) are treated
with this fluid the exosporium becomes yellowish-brown with
a tint of red in it, while the endosporium exhibits a variety of
colours, one shading into another. The outer stratum of this
layer is somewhat darker than the exosporium but very similar
to it in colour, while the inner is slightly violet becoming less
so towards the inner side. It is usually more pink-violet than
the walls of the ordinary parenchyma cells of the ovule. Be-
tween the two differently coloured strata of the endosporium
is a narrow greenish-yellow area which shades gradually into
those on either side. The boundary line between the endos-
porium and the exosporium is dark and distinct after treat-
ment with chlorzine iodine, and the coat is somewhat swollen
by the fluid and appears less granular. After the action of
iodine and sulphuric acid upon sections of the membrane, the
outer layer becomes dark yellowish-brown, while the inner
has a yellowish-brown outer stratum which entad passes into
light yellow, greenish-yellow, and finally into blue. Within
the blue can often be seen a homogeneous inner greenish-yellow
margin into which the blue shades. From the well known
action of the stains and reagents used it is evident that we
have to do with a coat whose exosporium is suberized but whose

[89]

8 THOMSON : THE MEGASPORE-MEMBRANE

endosporium is of very complex chemical composition. The
inner part of this layer contains a substance which seems
closely related to pectin.* Towards the median area of the
‘endosporium this substance is largely replaced by cellulose.
The cellulose in turn, in the outer stratum of the endosporium,
gives place gradually to suberin, the outer border of the layer
being constituted chiefly of the latter.

In older ovules of Cycas revoluta the megaspore-membrane is
still distinctly double but it has increased considerably in
thickness. At the stage of prothallial development when cells
are forming (two layers thick in this case) the membrane is
of quite uniform distribution around the prothallium and
slightly more than 5 » thick. The increase in thickness does
not affect the two layers equally but is chiefly confined to
the endosporium which at this stage is about one and one-half
times as thick as the exosporium. The latter layer appears
more distinctly radially striated in cross-section than in the
case of the younger membrane. Sections of it cut parallel
to the free surface of the coat are very coarsely granular.
The granules are dark yellow in colour but differ from one
another both in size andin form. They appear to be surround-
ed by irregular spaces. The exosporium would thus seem to be
formed of little columns or fibrillae which are radially arranged
and so responsible for the appearance of radial striation in
cross-sections of the coat. This interpretation of the structure
of the exosporium is verified by the character of the cleavage
of the layer. When sections are broken across the cleavage
plane is always at right angles to the free surface of the coat
and along the line of the striation. Sometitnes two breaks
in the exosporium occur near one another and a small rec-
tangular mass is torn out of the layer. This cleavage feature
of the coat is characteristic of the younger membrane of Cycas
as well, though the latter is not so frequently found broken.
The inner layer of the megaspore-coat, as stated above, is
much thicker at this stage than at the previous one. It pre-

* This part, as was stated, does not colour a decidedly orange- yellow
in safranin (Strasburger !”, p. 135), but a rather dark yellow. At a later
stage, however, to be described next, the orange-yellow is more distinct.

[90]

OF THE GYMNOSPERMS 9

sents clearly the appearance of being subdivided into two
strata. These, however, vary considerably in relative thick-
ness at different places in the sections, but the outer portion
is always thicker than the inner, sometimes appearing as
thick as the exosporium. The boundary between the two
strata of the endosporium is, as in the young condition, a
somewhat radially granular part. Moreover, a slight differ-
ence in the chemical composition of the endosporium is to be
noted at this stage. A greater portion of the outer stratum
of the layer is completely suberized, while its inner part gives
a more decided indication (orange yellow in safrania) of the
presence of pectin, the cellulose being more definitely confined
to the median part of the layer at this than at the previous
stage. The coat “‘coils’’ more readily also in the older con-
dition, ‘‘coiling’’ being an especially marked feature at this
stage in sections which have been treated with potash solution.
The endosporium swells very much in this fluid and the coat
may attain a thickness of 8 pv.

In addition, there is present on the inner side of the coat
a thin layer which is in very intimate association with it, so
intimate in fact that it is readily mistaken for a part of the
coat. It is, however, formed of the outer walls of the super-
ficial prothallial cells. This apparent ‘“‘reinforcement”’ of
the coat may be very considerable in older ovules. Indeed
in ripe seeds of many forms a suberized band is present over
the free surface of the prothallial cells which is as thick or
even thicker than the coat itself and is difficult to distinguish
from it, since the two structures are alike in chemical compo-
sition and often closely in contact with one another.

The chief change in the megaspore-membrane of Cycas
revoluta at the stage when the archegonia are formed is in the
relative thickness of its two layers. The exosporium is fully
double as thick as the endosporium, though the thickness of
the whole coat is but slightly increased. The exosporium
is more plainly radially striated than at previous stages.

I obtained only a single ovule of Cycas Rumphii for exam-
ination. It was at a stage of development similar to the

[91]

10 THOMSON : THE MEGASPORE-MEMBRANE

earliest examined of C. revoluta, i.e., when the prothallium
is about to begin cell formation. At this stage the coat of
C. Rumphw is so similar in structure, in thickness, and in
chemical composition to that of C. revoluta as to be indis-
tinguishable from it. In fact had not the shape of the ovule
been different in the two species I should have been unable
to distinguish the sections of them.

In order to determine whether the structure and chemical
composition of the coat of the megaspore of Cycas has been
affected by the retention of the spore in the megasporangium
it seemed desirable to study the coat of the microspore for
comparison. Material of Cycas was not available, but a study
of the coat of the mature microspore of Pinus restnosa was
made instead. The coat in this form is about 5 » thick in
the region above the prothallial cells, its exine and intine
being about equally thick and separated by a dark line.
The outer layer of the coat is roughly radially striated and
the inner homogeneous. When thin sections (3 to 5 p» thick)
have been treated with chlorzine iodine the exosporium is
yellowish-brown and the endosporium shades from light
violet on the inside to darker violet about the middle, then
into greenish-yellow, and finally into light brownish-yellow
along the outside. In safranin the intine is somewhat orange
yellow on the inside and pink towards the middle and outer
parts. The exine is cherry red. In haematoxylin the inner
and outer borders of the intine stain less intensely than the
median part. It is thus seen that the strata of the intine of
the microspore of Pinus react similarly with stains and re-
agents to those of the endosporium of Cycas, though they are
not so distinct from one another as are those of the latter.
The enclosed and the free spore coats are thus in structure
and in chemical composition fundamentally alike.

For the purpose of comparison of the process of suberiza-
tion in spore-coats and in cell-walls an examination was made
of the suberized walls of the epidermal cells on the upper
surface of the megasporangium of Zama tntegrifolia, a form
whose megaspore-coat will be described later. When sections

[92]

OF THE GYMNOSPERMS II

are treated with chlorzine iodine, around the protoplasm of
these epidermal cells is a uniformly thick homogeneous layer
which is violet in colour. Over the free surface of the cells
is a dark brown band from which characteristic wedge-like
projections pass inwards between the outer parts of the violet
walls of contiguous cells. The dark brown of this band and
its projections never comes into intimate contact with the
violet of the inner layer but is always separated from it by a
greenish-yellow area just as in theinner layer of the spore-coats
described above. This colour is no doubt produced from a
blending of the colours of the adjoining strata, so that it would
appear that we have here to do with an area of composite
chemical character, a transitional area between strata of
different chemical composition. The suberized walls of epi-
dermal cells are undoubtedly formed from within,and since
the endosporium of the coat of Cycas consists of similar chemi-
cal strata and has a like transitional area between the cellulose
and the suberized layers, it seems probable that it, too, is
formed from the inside. Confirmation of this view is indicated
in the action of chlorzine iodine on the two suberized parts
of the coat. The outer stratum of the endosporium is darker
brown in colour than the exosporium and so according to the
well established colour reaction of this reagent is of more
recent formation.

The present work on the Cycadeae has demonstrated the
occurrence of a megaspore-coat in the subgroup and has
determined its structure, thickness, and chemical composition.
The coat is double, its outer layer suberized and composed
of radially arranged fibrillae, while its inner is fairly homogen-
eous and formed of strata which, though they differ funda-
mentally in chemical composition, are yet not distinct but
pass by gradual change of chemical components insensibly
into one another.

The tapetum of Cycas at the youngest stage of prothallial
development described above consists usually of a single
layer of large somewhat cubical cells, on the inside of which
there may often be seen here and there some more or less

[93]

12 THOMSON : THE MEGASPORE-MEMBRANE

collapsed ones (fig. 1, pl. 1). The distribution of the tapetum
is uniform around the prothallium at this stage just as is the
megaspore-coat. The large cells have one to several nuclei
in each, two being not uncommon in a section. In the cells
are irregular-shaped grains which in stained preparations and
in those treated with chlorzinc iodine, etc., appear as vacuoles
(fig. 1, pl. 1). These grains in iodine solution* colour at first
a brownish-pink or purple, some of them even a red brown.
On heating the colour fades, and on cooling and applying
fresh fluid the brown colour does not return but is replaced
by pink to bluish-purple. These grains are then not glycogen-
ous as at first seemed probable but consist largely of amy-
lodextrin (Strasburger!?, pp. 82 and 83). They are quite
different from the starch grains of the nucellar and integu-
mental tissues, the latter being much smaller and giving the
characteristic starch reaction. The tapetal cells at the later
stage of prothallial development when cell formation has begun
are devoid of amylodextrin and their protoplasm appears
to be fairly uniformly distributed. The most interesting
point, however, with regard to the tapetum in the present
connection is apparent when sections of it and of the nucellar
tissue are treated with chlorzinc iodine. All the megasporang-
ial cells, even the very thin-walled ones of the more or less
collapsed tissue immediately adjoining the tapetum, have
violet-coloured walls, while those of the contiguous large tape-
tal cells are dark yellowish-brown. The walls of the tapetal
cells are thus suberized and so afford a striking contrast be-
tween this tissue and that of the nucellus. On the other hand
the suberization gives indications of an association between
the tapetal cells and the megaspore itself, which, to say the
least, suggests very forcibly a common origin, and is thus
almost in the nature of a demonstration of the Spore
character of the tapetum.

II (a). Zamieae-Stangerieae. In some very young ovules
of Stangeria paradoxa, the only form included in the Stangerieae

* The solution used was made up as follows: 5 c.g. iodine and 20 c.g.
potassium iodide were dissolved in 15 c.c. of water.

[94]

OF THE GYMNOSPERMS 13

the megaspore-nucleus had already divided several times,
probably about eight free nuclei being present in the some-
what parietal protoplasm of the young prothallium. Around
the protoplasm at this early stage there is present an undiffer-
entiated, glassy-looking thickening which represents the young
megaspore-inembrane.

At a later stage, when a cellular prothallium has been form-
ed and the ovules are about ready for pollination* the megas-
pore-coat is a very conspicuous feature. It is about 4.5 » thick
and is very distinctly double (photo. 2, pl. 3), in fact it is very
similar to that of Cycas revoluia at the stage when the prothal-
lial cells are forming. The exosporium is fairly granular and
very deeply and distinctly striated. The endosporium con-
sists of two homogeneous bands, the outer of which is thicker
than the inner. These have a narrow granular area between
them just as in the case of Cycas. The coat, too, is “‘rein-
forced’’ from within by the outer walls of the superficial layer
of prothallial cells, so that it appears to consist of four layers, a
pair of inner narrow and a pair of outer relatively broad bands.
Haematoxylin stains both layers of the endosporium, the outer
less than the inner but still to a considerable extent. Safranin
stains the outer stratum also and colours as well the exospor-
ium. The outer part of the endosporium is intensely stained
as a result of the double staining, and a good contrast is ob-
tained between the layers as is shown 1n photograph 2, plate 3.
The thickness of the inner part of the endosporium appears
less in the photograph than its actual thickness, because of
the over-staining of the outer part of this layer and its con-
sequent encroachment on the inner. There is also but little
indication in the illustration that the inner lighter part is of
a composite character, since the ‘‘reinforcement’’ from the
cell-walls is very closely associated with the inner part of the
membrane. In sections treated with chlorzine iodine, how-
ever, this is evident, since the cell-walls are coloured yellow
while part of the endosporium is blue-violet. The transition

* For the material of Stangerta at this stage I here express my
especial indebtedness to Dr. Lang of Glasgow University, who was kind
enough to send me several embedded ovules.

[95]

14 THOMSON : THE MEGASPORE-MEMBRANE

area from the violet to the yellowish-brown (tinted red) of
the outer part of the endosporium is greenish-yellow as in
Cycas. This outer part of the endosporium is homogeneous
and coloured similarly to but darker than the finely granular
and radially striated exosporium. When sections of the coat
are treated with iodine and sulphuric acid the inner part of
the endosporium is greenish-yellow to blue in colour and shades
as in Cycas into the dark yellowish-brown of the outer parts of
the coat.

Dr. Lang has represented the coat of Stangerta, in his work
on the ovule of this form, as a single, thick, somewhat granular
band (Lang"!, fig. 16, pl. 17), at a stage of prothallial develop-
ment similar to that at which I have described it above. I
find that the coat of Stangerza, like that of Cycas, ‘‘drags’’
very frequently in sectioning, and that in this condition it
appears as an unstratified, fairly homogeneous layer, much
thicker than in normal section, fully as thick as Dr. Lang has

represented it. Dr. Lang was primarily interested in other
important features and possibly his sections, like many of

mine, while showing the structure of the tissues perfectly, gave
but a false impression of the coat.

The megaspore-membrane of Stangerza thus is similar in
every respect to that of Cycas, It is double, its endosporium
subdivided into two more or less homogeneous strata. The
exosporium and the outer stratum of the endosporium are
suberized, the latter containing a little cellulose, while the inner
part consists largely of this substance with which some pectin
is probably associated. Between the two strata of the endos-
porium is a transitional area, transitional both in structure
and in chemical composition, just as in the case of Cycas.
In thickness, however, the coat of Stangeria is somewhat less
than that of Cycas even when the latter is at a younger stage
of prothallial development.

Dr. Lang has described the tapetum of Stangerza in several
stages of development. His account of it when the prothal-
lium has become cellular is as follows (Lang!!, p. 287) :
“Around the megaspore a layer of cells was present which is

[96]

OF THE GYMNOSPERMS 15

clearly to be traced to the sporogenous group. The thick
zone of sporogenous tissue present in earlier stages has, how-
ever, become reduced to a single layer....The cells of this
persistent layer (fig. 16) are very large, and stand with the
longer axis at right angles to the surface of the megaspore.
Sometimes a single nucleus is present; often two are found in
each cell.’’ This layer is said ultimately to disappear in the
seed. In infertile ovules Dr. Lang has further observed that
the tapetum is well developed. My own observations on the
tapetum in Stangeria agree, so far as they go, with Dr. Lang’s
description. In the young condition when the nucleus of the
megaspore has undergone probably three successive divisions
the tapetal tissue is of considerable thickness. The walls
of its cells have not, however, as yet differentiated chemically,
colouring greenish-yellow in chlorzinc iodine. At the older
stage at which the megaspore-membrane has been described
the tapetum consists of a single, or at rare intervals of a double,
layer of cells. These have thin but yet distinctly suberized
walls, just as in the case of Cycas at a similar stage of develop-
ment. In sections of infertile ovules of the same age as the
fertile ones described above the tapetum is a very prominent
structure, being four to five layers of cells in thickness. Its
cells, too, are large and have very abundant protoplasmic
contents, while those of the surrounding tissue are in a more
or less collapsed condition. Further, the walls of the tapetal
cells are thick and suberized while those of the adjoining nucel-
lar tissue are thin and composed of cellulose. The greater
development of the suberized tapetum in ovules which develop
no prothallium would seem to be in keeping with Dr. Lang’s
view that the single-layered tapetum, which is characteristic
of the older stages of fertile ovules of Stangeria, results
from the gradual disintegration and absorption by the growing
prothallium of the inner cells of the originally many-layered
“sporogenous group’’ surrounding the uninucleate spore.
The sharp line of chemical distinction between the walls
of the large exterior tapetal cells and those of the adjacent
collapsed cells of the nucellar tissue affords evidence of a

[97]

16 THOMSON : THE MEGASPORE-MEMBRANE

similar character, while in addition, by laying emphasis upon
the definiteness of the boundary between these tissues, it
brings into more intimate association the tapetal cells and the
megaspore, affording thus clearer evidence of the sporogen-
ous nature of the tapetum.

II (b). Zamieae-Euzamieae. Dioon wmbricata at a stage
of prothallial development previous to the appearance of
archegonia has a well developed coat about 3.8 » thick and
very distinctly double. The exosporium is about one and one-
half times as thick as the endosporium, and the fibrillae of
which it is composed are quite regularly radially arranged
and coarse. The endosporium presents the appearance of
being subdivided into two strata, the outer thicker than the
inner, but both quite homogeneous. ‘The inner stratum of the
endosporium stains with haematoxylin, while the outer as well
as the exosporium are coloured by safranin. In chlorzine
iodine the exosporium and the outer part of the endosporium
are yellowish-brown, the former a little lighter than the latter.
The inner part of the endosporium is pinkish-violet and pass2s
by a darker coloured, somewhat granular narrow portion into
the yellowish-brown of the outer part of the layer. There
does not seem to be any shade of red in the suberized part of
the coat when it is treated with chlorzinc iodine, as has been
noted in the case cf Cycas and Stangeria, but this may be due
to the, unfortunately, poorly preserved material examined.

Warming in 1879" gave an illustrated description of the
mature coat of the same species of Dioon that Iexamined. He
states that the coat is layered like that of Ceratozamia robusta,
whose megaspore-membrane he described and figured two
years previously. There is, however, a great difference in
his figures of these forms, two layers more being represented in
Ceratozamia than in Dioon. (Compare Warming,’ fig. 23,
pb. 3, and: Warming, fig, 7, pl. 16.) Dhe statements wipe,
in the layering of the coat in the résumé and in the original
(Danish) of Dr. Warming’s article do not agree (Warming’’,
résumé p. 18, and the Danish p. 100). Yet it seems probable
that the coat of both forms is considered double since this is

[981

OF THE GYMNOSPERMS ey

the statement of its structure in the Danish and since, also,
Warming’s statement on the cycads in general (Handbuch der
Botanik, p. 168) would lead us to so regard it. Further, Dr.
Warming states that the outer part of the coat of Dioon appears
to be made up of a large number of small ‘“‘ prismatic bodies,’’
structures which are apparently considered the homologues of
ce-tain ‘‘fusiform crystal-like structures,” stated to be common
in other tissues of Dioon. It is in this way that Dr. Warming
accounts for the radial striation of the outer portion of the
coat. Undoubtedly the “prismatic bodies’’ correspond to
the little columns or fibrillae which I found present in the coat
of Dioon and other forms. The structures I have observed
are, however, too irregular to be termed prismatic though
they have something of a hexagonal! outline in cross-section,
and in this respect correspond somewhat to the representation
of them that Dr. Warming has given (Warming," fig. 6, pl.
6). The coats of Dioon and Ceratozamia are considered as
completely suberized, colouring dark yellow in chlorzinc iodine.
I find that the inner stratum of the endosporium contains no
suberin. In fact in chemical composition the coat of Dzoon is
very similar to that of Cycas and Stangeria, the inner stratum
of the endosporium giving a decided reaction for the presence
of cellulose, while its outer stratum and the exosporium are
suberized.

Dr. Warming has figured a younger stage of Dioon in which
the membrane is represented as a single homogeneous layer.
I have not examined any material of Dioon at a similar stage
but it seems probable that the coat does not differ from that
of other cycads that I did examine (See Cycas, photo. 1, pl. 1,
and Zama wntegrifolta, pp. 17-18).

Several stages in the development of the megaspore-coat
of Zama iniegrifolia were examined. When the prothal-
lium is in a comparatively early stage of free nuclear* division
a distinct coat about 3 » thick is present around the parietally
placed cytoplasm. The coat ‘‘drags’’ very readily in section-

* I counted as many as 16 nuclei in one section through the centre of
the prothallium. This section was not more than 5 p thick.

[99]

18 THOMSON : THE MEGASPORE-MEMBRANE

ing but appears distinctly double in well prepared sections.
Its inner half in unstained sections is transparent and homo-
geneous, while the outer is yellow, opaque and somewhat
granular especially towards the exterior. Moreover, there is
present a very dark and granular line about the middle of the
coat. Though the general appearance of layering is evident
in sections of the coat yet the boundaries of its strata and
substrata are less clearly defined than in the case of any other
cycad coat that I have examined. In sections stained in
haematoxylin and safranin the inner and somewhat irregular
border of the endosporium is light pink in colour and exterior
to this is a homogeneous violet band which externally becomes
darker, its outer boundary being formed by an intensely dark
granular line which separates the coat into two equal parts.
Exterior to this central line is a violet-pink band slightly
granular in appearance and stained more intensely than any
other part of the coat except the line above referred to. Appar-
ently continuous with this deeply stained part is one stained
a lighter pink. It is more granular than the latter and indi-
cations of a radial arrangement of the granules can be detected.
This part probably represents a poorly differentiated exospor-
ium about one-third the thickness of the whole coat. In
chlorzinc iodine the inner margin of the coat is light greenish-
yellow and after several days’ treatment separates from the
coat as a thin film. The greenish-yellow is succeeded extern-
ally by a yellowish-violet which in turn is replaced by dark
green when the median granular area is reached. Exterior
to the middle line of the coat is a dark yellowish-brown band,
neither border of which is clearly defined. It passes on the
outside into the lighter yellowish-brown of the exosporium.

At a stage when about two layers of cells are present in
the prothallium, the megaspore-coat of Zamza inlegrijolia has
increased somewhat in thickness and the layers are more
differentiated. The exosporium constitutes about one-half
of the coat and is more distinctly granular and radially striated
than it is in the younger condition. The endosporium is
composed of two strata and has an addition to its thickness

[100]

OF THE GYMNOSPERMS 19

from the walls of the outer prothallial cells. The so-called
‘‘reinforcement”’ is similar in colour when the sections of the
coat are treated with chlorzine iodine to the inner light yellow
border of the younger coat. Like the latter, too, after pro-
longed treatment it separates from the coat. The inner
stratum of the endosporium is violet in chlorzinc iodine
and the outer, which is about double as thick,is brownish-
yellow, as is the exosporium. The boundary line between
the exosporium and the outer part of the endosporium is more
clearly marked than in the younger coat.

At a later stage, when cell-formation is well advanced in
the prothallium the megaspore-coat of Zamza integrifolia is 3.8
to 4 » thick. The exosporium at this stage is thicker than
the endosporium even with its ‘‘reinforcement’’ which has in-
creased considerably. The outer layer of the coat, too, is
more distinctly radially granular.

In mature seeds of Zamza integrtfolia I found that the coat
is quite thick, 4.5 » at least, without the addition from the cell-
walls. The latter has differentiated into an inner “‘cellulose”’
and an outer ‘“‘suberous’’ layer so that when the coat is in
contact with the prothallial cells it is difficult to make out the
inner cellulose layer of the endosporium even when sections
are treated with chlorzine iodine. The whole endosporium
at this stage is very much reduced in thickness constituting
only between one-fourth and one-sixth of the whole coat. The
exosporium is plainly radially striated, the fibrillae of which
it is composed being quite fine and somewhat irregularly
radially arranged.

The tapetum in the mature seed is represented merely by
some suberized débris from collapsed cells. In the younger
stages it is very similar to that of Cycas and Stangeria, being
formed of a jacket of cells tolerably uniform in thickness,
investing the prothallium. The cells have suberized walls
which are more or less spongy-looking, as if they were irregular-
ly radially perforated. Usually they are uninucleate and con-
tain some small amylodextrous grains. The number of layers

of cells constituting the tapetum varies at different times.
[ror]

20 THOMSON : —THE MEGASPORE-MEMBRANE

The material of Zamia floridana, at a stage when ferti-
lization is about to take place, showed a megaspore-membrane
which in section is very similar to that of the other species
and about 4.2 » thick. The endosporium and the exosporium
were not so unequal in thickness as in the case of Z. integrtfolta.

At a stage when the archegonia have been fully formed,
Ceratozamia longifolia has a megaspore-membrane which is
4.5 » thick in the lateral and basal parts of the ovule. In the
archegonial region, however, it thins out fully one-third. The
exosporium is coarsely radially striated and fully three times
as thick as the endosporium. It has a very thin homogeneous
inner portion and the boundary line between it and the endos-
porium is distinct.

Warming,!’ as already stated, has described the coat of
Ceratozamia robusta. His figure indicates that it is in appear-
ance very like that of C. longifolia. The coat he describes is
mature, however, and is represented as much thicker (by my
computation from his figure) than the younger coat of the
latter species that I examined. The coat of C. robusta, too, is
said to be suberized, while I found the one of C. longifolia not
different in this respect from that of the other cycads.

The megaspore-membrane of the Cycadales in all the forms
examined shows the same fundamental characteristics of
structure and chemical composition. The coat is double,
its outer layer, especially in the later stages of development,
being composed of distinct, radially arranged fibrillae. This
layer is suberized in all cases. The endosporium presents the
appearance in cross-section of being composed of two more
or less equal and homogeneous longitudinal strata. These
differ in chemical composition, suberin, cellulose, and pectin
replacing one another gradually from the outside to the inside
of this layer. The character of the succession of the chemical
strata of the endosporium and the transitional nature of the
areas between thein have led to the conclusion that the coat
is formed from within—that the endosporium is in fact the
formative layer of the coat. This conclusion gains support
from the fact that as the exosporium increases in thickness,

[102]

OF THE GYMNOSPERMS 21

the endosporium decreases, the former apparently at the
expense of the latter. With regard to the increasing distinctness
of the striation of the exosporium which keeps pace with its in-
crease in thickness, it may be possible to consider it as an in-
dication of a progressive degeneration resulting from phys-
iological activity. The development of the striation is at
least associated with the growth of the prothallium and the
accompanying transfer of food-material from the surrounding
cells to it. No matter what view is taken of the mode of forma-
tion and differentiation of the coat, its uniformity in structure
and in chemical composition in the Cycadales as a group has
been demonstrated. Moreover, the coat is thickest and best
developed in the Cycadeae, and thinnest in the Euzamieae,
while in the Stangerieae it is of an intermediate character.

The tapetum is uniformly well developed in the Cycadales
judging from the representatives of its subgroups that I have
examined. It is evenly distributed around the prothallium
and in certain cases is as much as four to five layers of cells
in thickness. In all forms it probably thins out in the later
stages of ovular development, being practically eliminated
from the mature seed, though even here its pre-existence is
indicated by certain suberized remains. In all the forms that
I have examined the tapetal cells have suberized walls. The
suberization it would seem possible at first sight to regard
aS a cenogenetic feature developed in relationship to the
function of the tapetal cells, whose office it is to supply nutri-
tion to the growing megaspore. The study of various tapetal
layers, however, shows that the cells constituting such layers
may have cellulose walls (znfra pp.25-26). Againif such were the
case it might reasonably be expected that the suberization of
the tapetal cells would not be so marked in infertile ovules
where no megaspore is developed. As has been observed in
the case of Stangeria suberization is even more marked in
such cases. Thus suberization of the walls of tapetal cells
would appear to be a palingenetic feature and to afford
almost a demonstration of the sporogenous nature of this
tissue. The plurinucleate condition of some of the cells

[103]

22 THOMSON : THE MEGASPORE-MEMBRANE

constituting the tapetum in the various forms and the course
of development of this tissue as indicated by Dr. Lang’s work
on Siangeria affords further evidence of a similar nature.

GINKGOALES

As early as 1855 there is a record of the occurrence of a
megaspore-membrane in Ginkgo. Hooker and Binney state
that “an extremely delicate membrane (fig. 16) surrounds
the albumen of Salisburia.’’ It is represented in their figure
by a line. Williamson,!* in 1876, referred to the coat in the
same connection but with no more detail of its structure.
Recent investigators seem to have overlooked it entirely.
In fact in Seward and Gowan’s!’ résumé of the knowledge
of Ginkgo (up to 1900) there is an indirect negation of its pres-
ence. They say: ‘The upper part of the endosperm is
covered by a thin papery membrane which represents the
crushed remains of the nucellus.”’

On embedding and sectioning some mature seeds I found
that this ‘thin papery membrane”’ is of a composite character
and represents not only the ‘‘crushed remains of the nucellus’’
but tapetal débris and the megaspore-membrane as well. The
megaspore-coat in fact is in a fair state of preservation in the
mature seed (fig. 2,* pl.1). It is very distinctly double and
4.5 to 5 » thick, thinning out somewhat in the archegonial
region. The exosporium is from four to six times as thick
as the endosporium and composed of quite loose fibrillae which
are very irregularly radially arranged, so much so that in
sections the outer layer often presents the appearance of a
net-work, though the reticulation is more apparent in oblique
sections than in accurately transverse ones. Staining differ-
entiates the inner part of the endosporium from the other parts
of the coat. The former takes the haematoxylin stain while
the rest of the coat stains in safranin. Under the action of
chlorzine iodine this inner part becomes violet and the outer
part of the endosporium yellowish brown. There is present
in Ginkgo as in the cycads a greenish yellow but very narrow

* Because of technical difficulties I have not represented the stratifica-
tion of the endosporium in Ginkgo and many other forms.
[104]

OF THE GYMNOSPERMS 23

transitional area between the strata of the endosporium. The
exosporium is yellowish brown, the reticulation being very
distinct in this fluid. The outer walls of the superficial pro-
thallial cells are very thick and consist of a suberized outer
layer and a cellulose inner one, the former about one-half as
thick as the latter. Where the coat abuts directly on the
suberized part of the cell wall it is difficult to distinguish be-
tween the ‘‘reinforcement’’ and the coat itself.

When the megaspore of Ginkgo is in the parietal nucleated
condition, at about the same stage* as that of the younger
ovule of Cycas (photo. 1, pl. 1) described before, its coat is
quite thin, about 3 » thick. It is double, as in the mature
seed, but the exosporium is aot more than one and one-half
_ times as thick as the endosporium. The former is plainly
striated at its outer border but becomes less and less so towards
the endosporium, and finally blends almost imperceptibly
with the outer stratum of the latter. The endosporium is
formed of two chemically different strata, as the staining and
treatmeni with chlorzinc iodine indicates. Haematoxylin lays
hold of the immature coat more strongly than of the mature
one and more readily obscures the layering of the coat, espec-
ially of the endosporium. The action of chlorzinc iodine on
the membrane at this stage is similar also to that on the mature
coat. A thin film, of the same colour as the protoplasm in
chlorzine iodine, extends along the inner side of the coat some-
times in contact with it but usually separated from it in the
sections. It may be a part of the megaspore-coat or the be-
ginning of the outer walls of the first row of prothallial cells
which will shortly be formed. It is, however, characteristic-
ally present, and adds about .3 » to .4 » to the thickness of
the coat.

At a very early stage of nuclear division in the megaspore
of Gingko, when possibly sixteen free nuclei have been formed,
there is a slight thickening around the protoplasm in which

* In a section about 5 » thick through the centre of the megaspore I
counted over 80 nuclei, while not more than roo are present in Cycas at
the stage indicated in photo. 1, pl. 1.

[105]

2A THOMSON : THE MEGASPORE-MEMBRANE

the nuclei are embedded. In chlorzine iodine this young
membrane is coloured light yellow, possibly with a shade more
of green in it than the protoplasm to which it adheres.

The megaspore-coat of Gingko is thus very similar to that
of the Cycadales. It is quite uniformly distributed around the
prothallium in both groups, and in structure and chemical
composition presents essentially the same features, though
the endosporium of Ginkgo at the stages I examined does not
seem: to be so thick as that of the cycads.

The tapetumin Ginkgo at the youngest stage described above
is four to five layers of cells in thickness. The inner ceils
are square to oblong in longitudinal sections of the ovule and
have less dense contents than the outer ones, which are some-
what elongated tangentially. When the megaspore has many
nuclei in a parietal stratum of protoplasm(at the intermediate
stage described above) the tapetum consists of a jacket one
to several cells ia thickness. The cells have fairly dense con-
tents and the walls though not well developed are suberized.
Some of the cells, too, are plurinucleate. When the seed is
matured the tapetum is represented merely by some suber-
ized débris around the prothallium (fig. 2, pl. 1). The tapetum
of Ginkgo like its megaspore-membrane is not so well developed
as it is in the case of the cycads but undoubtedly it is an homo-
logous structure, originating from the sporogenous tissue.

CONIFERALES.

I (a). Pinoideae-Abietineae-Araucarunae: In the post-
humous account of Gnetum by William Griffth,'§ which
appeared in 1859, there is a reference to the megaspore-mem-
brane of Agathts, one of the two genera constituting the Arau-
cariinae. We are told simply that this structure, which he
termec the ‘‘amnios’’ and of which he probably did not recog-
nize the homology, forms a very distinct lining to the cavity im
the ‘‘nucleus’’ (nucellus) of Agathts. This is at a comparative-
ly early stage of prothallial development.

In the youngest stage of Agathis Austrairs* that I have
examined, the megaspore has about 16 nuclei in a 10 » longi-

“* Material of this form has become available re cently tl 1rough the kind-
ness of Professor Treub and Dr. Valeton.

[106]

OF THE GYMNOSPERMS 25

tudinal axial section. At this stage the megaspore-coat is very
variable in thickness, averaging possibly 4.5. It has a some-
what granular inner stratum and an outer one about equally
thick but hyaline in appearance. There is no sharp line of dis-
tinction between the strata of the coat though the inner stains
in haematoxylin and the outer in safranin. In chlorzince iodine
the inner stratum is violet, light yellowish towards the inside
and darker towards the outside. The latter portion passes
by a greenish-yellow area into the slightly brownish-yellow
outer part of the coat. Around the coat there are the remains
of a suberized tissue, in intimate contact with it. Possibly
its presence may indicate that a suberized tapetum was present
at an earlier stage. Numerous pollen-tubes (6-10), entering
the apical region of the nucellus, converge irregularly around the
megaspore, some passing even deeper than the latter which
occupies the middle half of the longitudinal axis of the sporang-
ium. The pollen-tubes and the megaspore resemble one
another very closely at this stage, so closely that in cross-
sections it is difficult to distinguish them unless a number of
consecutive sections are examined. They are alike in size,
similar in contour and irregularly arranged in the axial portion
of the sporangium. Their walls, too, are closely associated
with the surrounding tissue and in structure and in chemical
composition are very similar. The similarity amounts almost
to an identity in certain of the cross-sections, since numerous
nuclei and much protoplasm are present in the pollen-tubes
at this stage.*

Around the megaspore at the young stage described above
and filling up more or less the areas between the pollen-tubes
are groups of cells very full of starch grains and protoplasmic
contents. This tissue undoubtedly supplies nourishment to

* Six or seven nuclei are not uncommon in a tube at this stage, while
in one I counted as many as thirteen. This recalls the interesting con-
dition recently described by Juell9 in Cupressus Governiana, where, however,
not a nuclear but a cell-complex is present in the pollen-tube, derived in
this case from the body cell. A detailed account of the pollen-tube in
Agathis, and of certain other peculiar features to which my attention was
directed in the course of the present work will appear in a subsequent
paper.

[107]

26 THOMSON : THE MEGASPORE-MEMBRANE

the prothallium and must be considered as a tapetum. It
differs very markedly, however, from the tapetum in the
cyeads and Ginkgo. Its cells are uninucleate and their walls
composed of cellulose while the tissue itself has no sharp ex-
ternal boundary but passes by easy transition stages into the
ordinary nucellar tissue, being merely a differentiated inner
part of the latter.

At a later stage just after fertilization has been effected,
the coat does not present even so definite structural features
as it does in the younger condition. It is much thinner and
colours more yellow in chlorzine iodine. The prothallium is
relatively very large at this stage, much of the surrounding
nucellar tissues having been absorbed. A few of the inner
layers of the latter are starch-bearing but more collapsed than
in the younger condition. Remains of pollen-tubes are appar-
ent especially in the apical region of the prothallium. In the
ripe seed of Agathis the coat is still present. It is fairly thick
in some places but usually thin and much disorganized. The
prothallial cells of the mature seed have very thick cellulose
walls and over the superficial ones a suberized band extends,
which is, however, clearly distinct from the coat. In infertile
ovules collected at the same time as the ripe and fertile ones,
the prothallium has collapsed, though the integument and
nucellus are nearly as fully developed as in the fertile ones.
The shrunken prothallium is surrounded by the megaspore-coat
which closely-invests it. This structure is undifferentiated so
far as layering is concerned, but fairly uniform in character.
In chlorzine iodine the bulk of the coat is yellow to brownish-
yellow in colour, the inner border only being violet.

The coat of the megaspore of Agathzs is thus poorly differ-
entiated structurally at all stages. In the young and better
developed condition it resembles very closely the wall of the
pollen-tube, though it is slightly thicker than the latter. In-
deed so very close is the resemblance that the numerous
pollen-tubes surrounding it are readily mistaken for infertile
megaspores. The resemblance of the megaspore-coat of
Agathis to the intine of the microspore of Pinus, both in struc-

[108]

OF THE GYMNOSPERMS 27

ture and in chemical composition, is very intimate. This
becomes even greater when the microspore has shed its exine
and grown down into the tissues of the nucellus, for the wall
of the pollen-tube (intine) of Pinus is then of increased but
more variable thickness. The megaspore-coat of Agathis
never differentiates any farther, but in the later stages
becomes even less definite in structure and finally is much
disorganized in the mature seed. As it becomes older
the composition changes gradually, the cellulose being almost
completely replaced in the later stages by a substance resem-
bling suberin. The latter gives, however, not so dark a brownish-
yellow in chlorzinc iodine as ordinary suberized layers.

The functioning tapetum of Agathis differs very decidedly,
as has been noted, from that of the cycads and Ginkgo. It
is clearly of nucellar origin and in this respect may be dis-
tinguished as a secondary tapetum, that of the cycads and
Ginkgo, which is regarded as sporogenous, being considered
as of a primary nature. The latter is represented in Agathis
only by suberized remains.

The coat of Araucaria imbricata is very variable in thickness
at the stage indicated in photograph 3, plate 3, and is closely
applied to the surrounding tissue. In this young condition
it is in appearance very like that of Agathis at about the same
or possibly a little later stage of development. It is much
thinner, however, and in composition has a larger proportion
of cellulose in it. A careful examination of mature seeds of
Araucaria Braziliensis and of A. wmbricata, as well as of infertile
ovules of the former, failed to reveal the presence of even a
trace of the coat. The coat is then much less developed than
that of Agathis though undoubtedly similar in character
to it. In both the characteristic outer suberized layer which
is found in the cycads and Ginkgo is absent and the coat re-
sembles the wall of the pollen-tube both in structure and in
chemical composition. It is suggestive, in this connection,
that the wall of the megaspore of the Araucariinae grows in
intimate association with the surrounding tissues, just as
that of the pollen-tube does.

[1c9j

28 THOMSON : THE MEGASPORE-MEMBRANE

Around the megaspore of Araucaria in the younger stages
(see photo. 3, pl. 3) there is a mass of tissue of elliptical outline
which represents the tapetum. Its cells contain many starch
(amylodextrous) grains and their walls are suberized. The
tissue is chiefly aggregated around the base of the megaspore,
being very thin along the sides and fairly so at the apex. At
places its suberized cells shade gradually into the surrounding
cellulose-walled cells of the nucellus, while at other parts they
are distinctly marked out from them. The tapetum of Arau-
carta is very different from that of Agaihis. In the latter it is
of fairly uniform distribution while in the former it is massed
in the basal region. In Agathts, too, the only suberized part
consists of collapsed cells within the functional tapetum,
while nearly all the tapetum of Araucaria has suberized cell
walls. The tapetum, however, in each form looks as if it were
an integral part of the nucellus, merely a differentiated inner
portion of its tissue.

I (b). Pinoideae-Abietineae-Abretinae : An idea of the
distribution and prominence of the tmegaspore-membrane
in this subgroup may be gained by a glance at photo-
graphs 4 to 10 (pls. 3 and 4). The coat is thick in the
chalazal region and thins out gradually towards the micropylar
portion of the prothallium, being not more than one-third as
thick at the apex as at the base of the megaspore.

The coat of Pinus resinosa (fig. 3, pl. 1) at the stage indi-
cated in photograph 4, plate 3, just prior to fertilization, is about
4.2 ». thick at the lateral basal region of the prothallium from
which part the drawings have usually been made. It is dis-
tinctly double,the endosporium being about one-third as thick
as the exosporium (fig. 3, pl. 1). The former is homogeneous
and appears as in the cycads to be subdivided into two longi-
tudinal strata, the inner, in the case of Pinus, only about one-
half as thick as the outer. The former stains in haemotoxylin
while the rest of the coat takes the safranin stain. In chlor-
zinc iodine the inner stratum of the endosporium is violet
while the outer is yellowish-brown, considerably darker than
the exosporium. ‘The fibrillee of the latter are more or less

[r10]

OF THE GYMNOSPERMS 29

collapsed and irregularly radially arranged, as indicated in
the figure. The tapetum of P. resinosa at this stage is almost
confined to the basal part of the ovule and consists of uni- or
pluri-nucleate cells which are more or less broken down.
Their protoplasmic contents are sparse and their suberized
walls irregular and somewhat porous, like those of the cycads
but yet much thinner. It is rather a striking feature with
regard to this primary tapetum that it thins out with the
megaspore-membrane, thickest at the base and almost dis-
appearing around the apical region of the prothallium.
P. strobus differs from P. resinosa in having a more delicate
megaspore-coat and one which is more evenly distributed
around the prothallium. The coat at the stage just subse-
quent to fertilization is not more than 3.8 » thick. Its endos-
porium is nearly as thick as its exosporium. The latter is
so irregular and ‘‘ ragged’ along its outer border as to give
one the impression that the decrease in thickness may be due
to the destruction of the outer part of this layer. The tap-
etum like the megaspore-coat is fairly evenly distributed
around the prothallium, disappearing only above the arche-
gonia. It consists of a jacket of partially collapsed cells with
suberized walls which lie away from the megaspore-coat at
a considerable distance. In the intervening space some gran-
ular material is found. Outside the tapetum is a layer of
very loose open cells which are not compressed as in most other
cases. Pinus sylvestris and P. Austriaca have coats and tapeta
which at the stage when the archegonia have been fully
developed are similar to those of P. resinosa.

The coat of the larches differs from that of the pines in its
distribution. In both Larix Europaea (photo. 5, pl. 3) and
L. Americana there is scarcely a trace of the megaspore-coat
in the archegonial region. In the former the thinning out
process is somewhat abrupt while in L. Americana which has
a longer prothallium it is more gradual, the latter occupying
an intermediate position in this respect between the European
species and the pines. The coat of these two forms is some-
what thicker and coarser than that of Pinus. Its endosporium

[111]

30 THOMSON : THE MEGASPORE-MEMBRANE

does not constitute more than one-sixth to one-fifth of the
whole coat (figs. 4 and 5, pl. 1) and consists of two substrata
which are more nearly equal in thickness than in Pinus. The
exosporium is very regularly formed and the fibrillae composing
it are coarse, especially in the case of L. Americana (fig. 5, pl. 1).
The distribution of the tapetal celis agrees with that of the
megaspore-coat. In the chalazal region of the prothallium
there are a few cells with suberized walls while around the rest
of the gametophytic tissue there are only scattered traces of
the tapetum.

The megaspore-membrane of the spruces examined is quite
similar to that of the larches at the same stage of development.
It is slightly thicker, however, and is not so attenuate in the
micropylar region. Photographs 6 and 7, plate 3, are of Picea
migra and Picea excelsa respectively and indicate the distri-
bution of the coat around the prothallium. Figures 6 and 7,
plate 1, illustrate the structure of the membrane in the lateral
region of P. excelsa and in the basal region of P. nigra. The
thickness of the membrane in our black spruce is not quite so
great as that of the European species. Both are quite thick,
however, though the endosporium is extremely thin in each.
Picea alba has a coat which is possibly a little thinner than
that of P. mgra but otherwise like those of the other species.

The megaspore-membrane of Tsuga Canadensis looks
very thick, as photograph 8, plate 4, shows. It is, however,
not so thick as it appears. Some granular material is massed
against it and makes its outer layer appear thicker at a low
magnification than it really is. In figure 8, plate 1, this
material is represented as slightly removed from the coat,
which in thickness and general character resembles that of
Pinus resinosa very much (cf. figs. 3 and 8, pl. 1). The hem-
lock, however, has a thinner endosporium than P. resinosa and
its exosporium is somewhat more regular.

Abies balsamea (photo. 9, pl. 4) has a well formed megaspore-
membrane which is of about the same thickness as that of Pacea
nigra. Its endosporium is very thin and its exosporium thick
and coarse (fig. 9, pl. 2). All the membranes of the Abietinae so

[112]

OF THE GYMNOSPERMS ay

far described have been from ovules at approximately the
same stage of development, about the time of fertilization.
They are all quite similar to one another as is indicated in
figures 3 to 9 (pls. 1 and 2). The membranes are “‘ reinforced’”’
on the side next the prothallium by the adjacent cellulose walls
of its superficial cells. Two areas are recognized in the endo-
sporium of all forms, an inner one which stains with haemo-
toxylin and becomes violet in chlorzince iodine, and an outer
one which is stained by the safranin andcolours dark brownish-
yellow in chlorzinc iodine. The exosporium is finely or coarsely
fibrillar in all. It stains in the safranin and is coloured a lighter
yellowish-brown in chlorzinc iodine than the outer part of the
endosporium. The tapetum at this stage is distributed much
as is the megaspore-coat. It consists of at least a single layer
of loose cells at the base of the prothallium and thins out to-
wards the apical region. The cells have suberized walls and
some of them at the base are plurinucleate as well. In these
respects they resemble the tapetal cells of the cycads and
Ginkgo.
In the mature seed of Pinus ponderosa, which is very large,
the megaspore-membrane is 4.5 to 5 » thick. The exosporium
is five to six times as thick as the endosporium and consists
of fibrillae which are fine but which look at places as if they
were more or less disorganized. _ They are always much pressed
together and seem almost to have lost the radial arrangement
which characterized them at an earlier stage. In chlorzinc
iodine there is a violet inner border to the coat but the body of
it is yellowish brown. The mature coat of Pinus Banksiana,
whose seed is small, is quite thick, 4.2 ». Its endosporium
is thin and homogeneous and the exosporium is coarsely and
irregularly granular. In chlorzinc iodine the coat is yellowish-
brown with a violet inner border. Pinus insignis is very
similar to the former in the size of the seed and in the character
of the megaspore-coat. The exosporium is, however, some-
what finer and retains more of the radial arrangement of its
fibrillae. Ripe seeds of Cedrus Atlantica have a coat which
in structure, thickness, and in chemical composition is almost

[113]

a2 THOMSON : THE MEGASPORE-MEMBRANE

identical with that of Pinus ponderosa. The size of its seed
is, however, much less. The megaspore-coat in the ripe seed
of the European larch is slightly thinner than that of Cedrus,
but otherwise similar to it. From the comparative uniformity
of the thickness of the megaspore-membrane of the above
mentioned forms it would seem that there is no direct relation-
ship between the size of the ovule and the thickness of the coat,
though possibly the larger seed has slightly the thicker coat.
The tapetum in the mature seeds of all the forms examined
is entirely disorganized.

At a young stage in Pinus resinosa when the prothallial
cells are forming (three cells in depth) the megaspore-coat
is about 3 » thick and appears quite similar to those of P.
pumilio and P. sylvestris as Mile. Sokolowa”® has represented
them, incidentally, in her work on endosperm-formation, as a
means of indicating the orientation of the cells. I find that
this is true also of P. sylvestris at a slightly younger stage
(fig. 10, pl. 2). The endosporium is homogeneous and com-
posed of two substrata as in the more mature condition de-
scribed before. The exosporium is finely and indistinctly
radially granular, but only about one and one-half times as
thick as the endosporium. The exosporium and the outer
layer of the endosporium are suberized. The tapetum is better
formed and more evenly distributed at this than at the later
stage. Its cells are irregularly shaped and one to several
nucleated. Their walls are not differentiated but the wails
of the cells immediately adjoining them (to the right in fig.
10, pl. 2) are of cellulose. The coat, too, is of uniform distri-
bution around the prothallium at this stage (photo. 10, pl. 4).
In very young ovules of Pinus resinosa, when about six nuclei
are present, in the parietal protoplasm of a section 4 to
5 p» thick through the young prothallium, the only trace
of a megaspore-coat is a more homogeneous outer border
around the protoplasm. The tapetum is thick but its cells
have not acquired differentiated walls. The outer ones are
elongated tangentially and quite narrow radially while the
inner ones are nearly equiaxial. These ceils have a large

[114]

OF THE GYMNOSPERMS aa

granular nucleus. The whole layer is four to five cells in
thickness and the one form of cell gives place gradually to the
other. At a later stage when about twelve nuclei are present
in a section of the prothallium the tapetum is differentiated
into two layers of oblong cells with large open nuclei. The
long axes of these cells are directed radially. Similar stages
in the development of the megaspore-coat and of the tapetum
have been observed in Pinus sylvestris and in P. strobus and
Larix Europaea.

The Abietinae have a megaspore-membrane and a tapetum
which are very uniform in structure and in the character of
their distribution. With respect to distribution both the
tapetum and the megaspore-membrane are peculiar and strik-
ingly different from the cycads. The reduction of both in
the Abietinae towards the micropylar region about the time of
fertilization suggests that in these forms this part is the seat
of the activities which are adverse to the retention of these
structures. Probably the reproductive processes are con-
cerned in this matter since the coat and the tapetum develop
quite uniformly around the prothallium until shortly before
the time when reproductive activity is at its maximum. The
coat, too, in the Abietinae is not so thick even inthe basal
region as it is in the cycads but in structure and in chemical
composition the membranes in the two groups are similar.
The tapetum, though very much less developed and more quick-
ly disorganized than in the cycads, is of the same nature, and in
the course of its development passes through similar stages
to that of Stangeria (supra, pp.14-15). The cells,too, composing
it are sometimes several nucleate and their walls are suberized.
The tapetum of the Abietinae is thus a primary one, derived
as in the cycads and Ginkgo from the sporogenous tissue.

I (c). Pinordeae-A bietineae-Taxodinae: Photograph 4,
plate 4, is of Sctadopitys verticillata. The gametophytic
tissues have shrunk considerably and are widely separ-
ated from the tapetum. The latter consists of from four
to five layers of more or less collapsed cells which have as
yet thin walls and are very full of granular material. The

[115]

34 THOMSON : THE MEGASPORE-MEMBRANE

tapetum in turn has separated from the nucellar tissue.
The megaspore-membrane adheres to the prothallium and in
sections is a very prominent structure. It is double and 4.2 » to
4.5 # in thickness, the endosporium being from one-half to one-
third as thick as the exosporium (fig. 11, pl. 2). At places in
the sections these two layers are torn from one another. The
endosporium consists of two homogeneous bands with a narrow
dark and granular-looking area between them, the inner hyaline
and the outer light yellow and less homogeneous. ‘The exos-
porium is light yellow also. It is, however, indistinctly radially
striated and finely granular. When breaks occur in the sec-
tions of this layer they are always transverse just as in the
case of Cycas. In fact the coat is very similar to that of Cycas
revoluta at nearly the same stage of prothallial development.
The whole coat is somewhat thinner, however, while its exos-
porium is if anything thicker. In staining the coat shows a
tendency to take up haematoxylin readily and the structure of
the endosporium especially is quickly obscured. In chlor-
zinc iodine the layering comes out very distinctly. The inner
part of the endosporium is bluish-violet, lighter towards the
interior and dark along the narrow granular area between the
strata. The outer stratum is brownish-yellow as is the granular
exosporium. When ‘‘dragged’’ the coat in chlorzine iodine
appears as an irregular thick single layer consisting of a mass
of dark brown granules in a homogeneous yellow matrix.
I have not observed any “reinforcement’’* of the coat in Scza-
dopitys though this may be so closely associated with the coat
as to escape observation. In distribution the coat is uniform
around the prothallium, just as it is in the Abietinae at about
the same stage of development (cf. photos. 10 and 11,
pl. 4). But it is thicker than the latter at this stage by fully
one-half (cf. figs. 10 and 11, pl. 2),'and 1n> this respece
and in distribution as well, approaches Cycas.

The tapetum in Sczadopitys, though thick all around the
prothallium, is especially thick in the basal region (photo. 11,

* The inner border of the endosporium is certainly yellowish, but it was
not observed to separate from the coat as is the case in Cycas.

[116]

OF THE GYMNOSPERMS 35

pl. 4) and thus is suggestive of the much reduced tapetum of the
Abietinae at about the time of fertilization when the tapetum
in the latter is practically non-existent except in the basal
region of the prothallium. (Compare photo. 4, pl. 3, and photo.
11, pl. 4). The walls of the tapetal cells turn yellow in chlor-
zinc iodine.

A figure of a young ovule of Cunninghamia appears in Dr.
Arnoldi’s paper on the Sequoiaceae (Arnoldi,”! fig. 2, pl. 7).
The prothallium is represented as in the parietal multinucleate
condition and the tapetal tissue (‘‘archesporial tissue’’ of
Arnoldi) is four to five cells in thickness in the basal region,
almost as thick, but more disorganized than I found it in
the case of Sciadopitys even at a later stage. The megaspore-
coat is not represented in this nor in any of Arnoldi’s figures,
nor referred to even in such forms as Sciadopitys where I have
found that it is very thick. It seems probable then that, since
the tapetum, which is evidently of a primary nature, is so well
developed in Cunninghamia,a fairly thick coat is present, since
I have found that there is a correspondence in the state of
development of these two structures. Material of Cunning-
hamia, however, was not available for examination.

Considerable attention has been recently devoted to the
study of the Sequoias. In young ovules of Sequoia sempervirens
numerous megaspores, ten to twelve according to Lawson,??
are found. These acquire thick walls and begin germination.
Only two or three develop very far (beyond the first division),
and but one of these, growing at the expense of the others,
develops a cellular prothallium and later bears the archegonia.
This one, previous to the formation of cells, is shown in photo-
graph 12, plate 4, with some abortive megaspores around it.
The walls of the megaspores are quite prominent in sections and
their close association with the tissues of the nucellus apparent.
The large megaspore encroaches irregularly on the nucellar
tissue, there being little or no suberized remains present, to
indicate the presence of a primary tapetum. Shaw” states,
however, that a well developed tapetum is present at earlier
stages but if such is the case it must be disorganized quickly.

[117]

36 THOMSON : THE MEGASPORE-MEMBRANE

The coat at the stage indicated in the photograph ‘“‘drags’’
very readily in sectioning and its thickness and structure could
not be as definitely determined as is desirable. It is, however,
about 2.5 » thick and consists of two layers (fig. 12, pl. 2). In
chlorzine iodine the inner border of the coat is violet and ex-
terior to this is a yellowish-brown homogeneous stratum, the
outer part of the endosporium. The exosporium is yellowish.
brown also but finely and indistinctly radially granular. It is
about one and one-half times as thick as the endosporium.
The coats of the infertile megaspores seem to consist almost
completely of cellulose though they are apparently quite thick.

In Sequoia gigantea there is but a single megaspore present.
This has no coat in the material I examined, when the prothal-
lium is in an advanced stage of free nuclear division, probably
about 128 nuclei being present in the peripheral protoplasmic
layer. The outer border of this layer is more homogeneous
than the rest of it, and this is the only indication of the presence
of a megaspore-coat. In the mature seed of S. giganiea the
megaspore-coat is thin, about 1.5 to 2 »in thickness. It con-
sists of a homogeneous inner layer and a coarsely granular
outer one. The coat comes into intimate contact with the
suberized layer of the superficial prothallial cells and is about
equal to it in thickness. Within the latter there is a slightly
thicker cellulose layer, the ‘‘reinforcement’’ of the coat being
thus much thicker than the coat itself at this stage. When
‘‘dragged’’ the coat appears very thick. It consists of a
uniform ground-substance which stains considerably like the
cellulose walls of the endosperm cells and has large granules
more or less rectangular in outline embedded in it or partly
projecting from it. These granules stain like the suberized
layer of the cell walls but appear when the high power of the
microscope is focussed on the matrix as dark, somewhat rec-
tangular areas in the latter. The tapetum in the younger
stage consists of a more or less broken row of cells scattered
along the inner border of the nucellar tissue. The walls of
these cells are not differentiated chemically, but the tissue
probably represents a poorly developed primary tapetum.

[118]

OF THE GYMNOSPERMS an

With regard then to the difference between the Sequoias
which has been shown to exist, at least in the tame of develop-
ment of the megaspore coat, it may be said that a certain
amount of variability is to be expected in vestigial structures.
The difference in the present instance is, however, associated
with others, which seemed of enough importance to Arnoldi?’
to warrant a separation of the two species by the revival of
an old genus, Wellingtonia, for the reception of one of them,
though Karsten,?4 who has reviewed Arnoldi’s work, thinks
that there is not justification for the separation of the two
species. I have merely referred to the matter here because of
the additional point of diversity afforded by the study of
the megaspore-membrane.

Cryptomeria Japonica at a younger stage (photo. 13, pl. 4)
than that of the young ovules of ‘Sequota gigantea examined
has no trace of a megaspore-coat. A poorly developed tape-
tum is present, however, but the walls of its cells are not differ-
entiated at this stage. In the mature seed of Cryptomeria the
megaspore-coat is not so thick as it is in S. gigantea, but other-
wise is very similar to that of the latter.

In Taxodium distichum at a comparatively early stage of
free nuclear division a megaspore-coat is present, as the figures
in Coker’s recent work on this form indicates (Coker,” figs.
50, 51, and 54, pl. 4). Dr. Coker has also stated that at a
later stage, about the time of cell-formation in the prothallium,
“the wall of the spore’”’ is “furnished with pits’’ (Coker,?5
p. 20, and figs. 97, 98, and 99, pl. 7). From his figure of them
in surface view they are small perforations of the wall of rec-
tangular to roundish outline. The only material of Taxodium
that I examined was of mature seeds. The megaspore-coat in
these, though much collapsed, is about 2.5 » thick. It consists of
a homogeneous inner layer and an outer coarsely and irregularly
radially granular one, about double as thick as the inner. The
whole coat, except its inner border which is slightly violet in
chlorzine iodine, is suberized. When ‘‘dragged’’ the coat
appears as a somewhat thick band, very much like that of
Sequoia gigantea. With regard to Dr. Coker’s statement

[119]

38 THOMSON : THE MEGASPORE-MEMBRANE

(Coker,”° p. 20) that the coat is pitted, I find that this is very
true of the exosporium of the ripe seed, but not of the endos-
porium. The pits are really not pits but irregular spaces
around the fibrillae of this layer, as is the case in all the forms
that I have examined. The ‘‘dragged’’ coat does, however,
look very much as if it were pitted when examined at a low
focus.

Dr. Coker has very fully described the tapetum in Taxodium
(Coker,”> pp. 17-20). It consists of large starch-bearing cells
in the young condition (megaspore mother-cell stage) which
not until later become differentiated from the surrounding cells.
When the megaspore is in an early stage of free nuclear divis-
ion the tapetum is two to three cells thick in the apical and
lateral regions, and four to five in the basal part (Coker,”®
fig. 47, pl. 4). Ata later stage, but before the formation of
cells in the prothallium, the tapetum consists of a single definite
layer of well formed cells (Coker,?> p. 18, fig. 51). This layer
consists of collapsed cells when the prothallium has become
cellular (fig. 53) and is said to become ultimately disorganized,
when the prothallium is mature.

Dr. Arnoldi has described in Sequota gigantea, Cryptomeria
Japonica and in Taxodium distichum tapeta which are similar
to the singly layered tapetum that Dr. Coker has figured for
Taxodium. My own observations verify the presence of such
a tapetum in the first two species. I also found some suberized
material between the gametophytic and nucellar tissues in the
mature seed of Taxodium. It thus seems probable that a poor-
ly differentiated primary tapetum is characteristic of them all.
This is in keeping with the relatively poor state of development
of the megaspore-coat in these forms. Reference has already
been made to Arnoldi’s description of the tapetum in Cunning-
hamia. In Sciadopitys both tapetum and megaspore-mem-
brane are thick and very different from those of other members
of this subgroup which have been examined. Reference has
already been made to Arnoldi’s proposed separation of the two
species of Sequoia, and to the added feature of difference be-
tween them which the study of the megaspore-membrane has

[120]

OF THE GYMNOSPERMS 39

brought out. Dr. Arnoldi has also found that Sczadopitys is
so different from the other Taxodinae as in his opinion to be
better removed from them. The great difference in develop-
ment of the megaspore-membrane and the tapetum in Sczado-
pitys from that which was found to be characteristic of the
other Taxodinae examined lends support to Arnoldi’s view.

II (2). Pinotdeae-Cupressineae : In Biota (Thuja) orten-
talis, at a stage (photo. 14, pl. 5) when the young embryos are
forming, the megaspore-coat appears quite thick. It is not so
thick as it appears, however, since the outer walls of the super-
ficial endosperm cells add much (about one-half) to its apparent
thickness (fig. 13, pl. 2). At thickest it is not more than 2 » at
this stage. It consists of two layers, the inner homogeneous
and the outer granular (fig. 13, pl. 2), the latter perhaps slightly
thicker than the former. In haematoxylin and safranin the bulk
of the endosporium stains a very intense pink, its inner border
violet. The exosporium is coloured light pink. In chlorzinc iodine
the two layers of the coat are yellowish-brown, the inner darker
and with a slightly violet inner border. Scarcely a trace of
the tapetum is present. The coat is of quite uniform distribu-
tion around the prothallium at this stage as the photograph
shows. Thuja occidentalis (photo. 15, pl. 5) thins out normally ,*
perhaps a little more than Burota in the archegonial region.
The coat, too, is not so uniform in thickness at other parts as
is the case in the former, nor is its thickness ever so great
(usually about 1.5 »). Even the “‘reinforcement’’ of the coat
from the prothallial cells is less than in Biota. Very little
trace of a tapetum is found at this stage, and the megaspore-
coat comes directly into contact with the cellulose-walled cells
of the nucellus.

In the mature seed of Biota the coat is about 2 » thick, not
so thick as is the suberized layer of the superficial endosperm
cells. The cellulose part of the latter is thicker again than the
suberized zone, so that the ‘‘reinforcement’’ in the mature seed
is more than double the thickness of the coat. In Thuja the

* In one case the coat appeared thicker in the archegonial region than
at any other part. Fertilization had not been effected in this case.

[121]

40 THOMSON : THE MEGASPORE-MEMBRANE

coat is very thin, almost disorganized, when the seeds have
matured.

Two species of Cupressus,C. sempervirens and C. thujoides,
have a coat which in ripe seeds is very similar to that of Bzota
orientalis but intermediate in thickness between this form and
Thuja occidentalis.

Some young material of Chamaecyparts sp. (?), when about
16 nuclei are present in a 10 » axial section of the megaspore,
showed only a trace of a coat, very much like that of Sequoia
gigantea at the young stage described before. The tapetum,
too, is but poorly differentiated.

The prothallium of Juniperus sabina, about the time of
fertilization, has a moderately thick (3 » +) megaspore-coat
(photo. 16, pl. 5). It is fairly uniform in distribution up to
the archegonial region where it thins out about one-third. The
coat ‘‘drags’’ very readily in sectioning and in the photograph ap-
pears much thicker than it really is. When cleanly cut it is
seen to consist of two nearly equal layers, the outer granular
and the inner homogeneous (fig. 14, pl. 2). The exosporium
stains light pink in the safranin and the outer part of the en-
dosporium dark pink, while the inner border of the latter is
blue in haematoxylin. Inchlorzinc iodine the coat appears
dark yellowish-brown, the inner part darker than the outer and
with a distinct violet inner border. At a younger stage, when
the endosperm cells are forming, the coat of Juniperus is some-
what thinner (2.5-3 ») but uniformly distributed around the
prothallium, and otherwise much as it is in the older stage
described above. The tapetum at this stage consists of a
somewhat loose layer of cells. These are more or less equiax-
ial but somewhat irregular in outline. They contain from one to
several nuclei. In chlorzinc iodine their walls, which arethin, are
yellow, while the walls of the cells immediately adjoining them
externally are violet.

In the mature seed of Juniperus Virginiana the coat is not
more than 2+ thick. As in the younger stages it has a homo-
geneous inner and a granular outer layer. It is intimately
associated with the thick suberized layer of the prothallial

x22]

OF THE GYMNOSPERMS 41

cells. This together with the cellulose layer of their walls is
fully two and a half times as thick as the megaspore-coat.
When ‘‘dragged”’ the coat appears as a fairly thick band light
in colour and with yellow granules partly embedded in it,
very similar both structurally and chemically to that of Sequoza
gigantea, described above.

The megaspore-coat of the Cupressineae in contrast to that
of the Abietinae does not thin out gradually towards the mi-
cropylar region of the prothallium, but is of more or less uniform
thickness up to the immediate neighbourhood of the archegonia.
The difference in distribution in the two groups is associated
with a difference in the arrangement of the archegonia and
possibly to be accounted for on this basis, since, as was re-
ferred to in the discussion of the distribution of the coat in
the Abzetinae, the reproductive processes which centre around
the archegonia may have to do with the partial destruction of
this portion of the coat. The tapetum (in Junzperus at least)
approximates in distribution to the megaspore-coat. Both
structures are less fully developed thanin the Abietinae. The
coat in the young stages as well as in the mature seeds is much
thinner and not so well differentiated and the tapetum dis-
appears earlier in the Cupressineae. The tapetal cells also, so
far as I have observed, do not acquire such thick nor such
clearly suberized walls as they do in the Abietinae.

II (3). Taxotdeae - Podocarpeae: In Podocarpus corvacea*
at a stage when the suspensorial cells have elongated and
forced the embryo-cells somewhat into the prothallial tissues
I could find no trace of a megaspore-membrane, though the
material was in a good state of preservation. In chlorzinc
iodine the thick outer wall of the superficial prothallial cell is
deep blue. There is only a trace if any of a light yellow border
to the walls, nor is there present between the gametophyte
and the nucellus more than a vestige of suberized substance
and no indication whatever of a megaspore-coat. The ovules

* The species is the one from Darlington, S.C., whose gametophytes
and embryo Dr. Coker26 investigated, and with material of which he was
kind enough to supply me.

[123]

42 THOMSON : THE MEGASPORE-MEMBRANE

of the other species of Podocarpus examined, P. Makoy1, were
in the mature condition, some having viviparous embryos
(Lloyd?’) projecting from theseeds. In this species as in the
last I could find no trace of a megaspore-membrane. It is of
course possible that a megaspore-coat is present in the younger
stages. Since, however, no reference is made to such a struc-
ture in the literature on this genus and since I have found,
with but few exceptions, some trace of the coat in the later
stages of development and even in the ripe seed, where it
occurs in the developing ovule, I consider that the coat must
be either absent in Podocarpus or at least very poorly developed.
Again, Dr. Coker refers to the absence in Podocarpus cortacea
of the tapetum or ‘‘spongy tissue’’ which, he states, is charac-
teristic of so many conifers and which I have found to be cor-
related in its state of development with the megaspore-mem-
brane. This form then must be considered to have either
none or a very poorly developed megaspore-coat and tapetum.

In Dacrydium laxtjoivum, on the other hand, the megaspore-
membrane is well developed at the stage indicated in photo-
graph 17, plate 5, a similar stage to that of Podocarpus corzacea
at which I found no megaspore-membrane present. The coat
is 4.2 » thick and very distinctly double, the endosporium
being about one-third as thick as the exosporium. The former
is homogeneous while the latter is very irregularly and coarsely
striated (fig. 15, pl. 2). When stained there is a dark blue
inner border to the endosporium while the rest of the coat is
pink. Under the action of chlorzinc iodine this inner border
of the endosporium is violet to bluish while the rest of the
coat is yellowish-brown, the inner part darker than the outer.
The exosporium appears at places to consist of little globules
of suberin and to be disorganizing. Some of the appearance
is no doubt due to the fact that the material was in the her-
barium for four years before it was ‘‘revived”’ and fixed. The
‘‘reviving’’ process, however, seemed to be very successful
in this case.

I do not pretend to be able to explain fully the difference
between Podocarpus and Dacrydium with respect to the

[124]

OF THE GYMNOSPERMS 43

megaspore-coat, but wish to point out in the former certain
correlated specialized features andin the latter corresponding
primitive ones. In Podocarpus no “spongy’’ tissue is present
around the prothallium, while in Dacrydivum the remains of
such a tissue are evident in sections of prothallia with the
megasporangium intact (fig. 15, pl. 2). That this is a mere
coincidence cannot be granted, since in the forms which are
known to be primitive, e.g. the Cycadales, and Ginkgoales, there
is an association in the state of development of the tapetum
or ‘“‘spongy’’ tissue and the megaspore-membrane. In ad-
dition, in other forms which have no megaspore-membrane
the primary tapetum has also been found to be absent or poorly
developed, as will be seen in the case of forms to be described
next. In certain morphological features of the female ‘‘flower”’
these genera are very different from one another. In Podocar-
pus the ovule is anatropous and has two well differentiated and
fused integuments, an inner woody and an outer fleshy one.
The fertile scales, also, are united in some species to form a
“receptaculum’’ which becomes berry-like at maturity. In
Dacrydium on the contrary the fertile scales are not fused and
differ but little from the ordinary vegetative leaves even
when the fruits are mature. The ovules are never anatropous |
(orthotropous in D. laxifoltum) and the outer integument is
represented by an “‘arillus’’-like structure which may only
partly enclose the inner one but is never united to it, except
at the base. The absence of the megaspore-membrane and
of the tapetum in Podocarpus, the fusion of the parts of the
“flower’’ and the resultant complexity of its structure con-
trast very strikingly with the more primitive state of affairs in
Dacrydium where the megaspore-membrane and tapetum are
present and the “flower’’ maintains the distinct individuality
of its parts.

II (4). Taxotdeae-Taxeae: In Cephalotaxus Mannii at
nearly the same stage as that of the Taxus indicated in photo-
graph 18, plate 5, there does not seem to be even a vestige of a
megaspore-membrane. The outer prothallial cells in this
case have a superficial layer which is yellow in chlorzinc

[125]

44 THOMSON : —THE MEGASPORE-MEMBRANE

iodine. This is the only indication of suberized material be-
tween the latter and the cellulose-walled cells of the nucellus.
Another species of Cephalotaxus (an undetermined one) pre-
sents features similar to those of the one described above.
Mlle. Sokolowa has observed that the megaspore-membrane of
Cephalotaxus is single and thinner than in the case of any other
gymmnosperm except Ephedra, with which she associates it.
Her figure of the megaspore-coat of Cephalotaxus Fortunes
represents it as a single finely granular layer about 1 » thick,
but with indistinct borders. This is at a stage when cells are
forming in the prothallium, as we learn from her figures (So-
kolowa,” figs. 14 and 15, pl. 11).

The material of Torreya nuctfera®* that I examined is at an
early stage of embryo development. In sections it appears
on first sight that a well developed megaspore-membrane is
present. Marking out the tissues of the gametophyte from
those of the nucellus is a rather uniform band which on treat-
ment with chlorzince iodine is seen to be suberized. In thin
sections which were made to determine the structure of this
supposed megaspore-coat, it is evident that the layer consists
of an aggregation of fibres. The prothallium encroaches
irregularly on the surrounding tissues, and no doubt the band
referred to is formed of the walls of cells pressed closely together
by the growing prothallium which has absorbed their contents.
The cells of the inner layer surrounding the gametophyte are
here and there apparently empty though most of them are very
densely packed with granular substance and have large nuclei.
Two or three layers of cells similar to the latter form with the
inner layer a quite distinct but irregular jacket around the
prothallium. The outer of these cells contain usually a con-
siderable number of starch grains. This tissue undoubtedly
supplies nourishment to the prothallium and is of a tapetal
nature. Its cells are, however, uninucleate and their walls are
composed of cellulose. They are not fundamentally different
from the ordinary cells of the nucellus, though they have more

* Dr. Coker, who kindly supplied me with the material, says that it was
obtained in the Botanic Gardens of Pisa, Italy, and that the species seems
to be T. nuctfera.

[126]

OF THE GYMNOSPERMS 45

dense contents than the latter, into which at places they pass
over very gradually. ‘The tapetum here is thus of secondary or
nucellar origin as was found to be the case with the functional
tapetum of Agathis.

In Taxus Canadensis at the stage indicated in photograph
18, plate 5, it is possible that there is a trace of a megaspore-
membrane in the basal region of the prothallium. If so,
the structure is exceedingly thin. The superficial prothallial
cells have their free walls covered by a partly suberized layer
which is about .75 » thick. Beneath it and of about equal
thickness is the cellulose part of the wall. The pollen-tube,
which is much expanded when it has reached the prothallium,
cotitrasts (see photograph) very strikingly with the megaspore
in the development of its wall. The tube wall varies in thick-
ness at different parts from 3.3 to 6 w. In one quite young
ovule of Taxus examined two embryo-sacs were present, one of
these comparatively large with archegonia developed on it,
and the other small with its cytoplasm in a parietal layer and
having several nuclei (16 probably). Around the smaller one
which would probably remain infertile there is a slight but dis-
tinct thickening, while around the other no membrane could be
distinguished. This is in keeping with Dr. Scott’s recent
observation of the difference in thickness between the walls of
the fertile and infertile megaspores of Lepidocarpon (Scott, 78
p. 299). The later stages of Taxus up to maturity of the seed
gave no indication of the presence of a megaspore-coat, nor
was any trace of a tapetum observed at any stage.

The chief genera of the Taxeae thus present a striking
uniformity in the absence or poorly developed state of the
megaspore-coat and of the primary tapetum. In this respect
they contrast very strikingly with the Podocarpeae, in which
group there is much difference in the state of development of
these structures. The Taxeae from the present standpoint
are to be regarded as a specialized subgroup, while the Podo-
carpeae contain some specialized and some primitive forms.

GNETALES.

In ovules of Ephedra vulgaris Mile. Sokolowa represents
a megaspore-membrane, at a stage when the endosperm-
[127]

46 THOMSON : THE MEGASPORE-MEMBRANE

cells are beginning to form, as a single and finely granular
layer about 1 » in thickness (Sokolowa,? fig. 23, pl. 12).
In older ovules I found the membrane still present (photo. 20,
pl. 5). It is difficult to determine its structure, however,
since the coat is very closely associated with the suberized
walls of the surrounding collapsed and compressed cells of
the nucellus. Still in very thin sections it can be made out
as an attenuate double layer around the base of the prothal-
lium, while around the upper part it appears more as if it
were single, the exosporium not being distinct. The outer
part of the coat is suberized and its inner border contains
some cellulose, the whole coat in the apical region being
largely composed of this substance. No tapetum is present
at the stage I examined, but there is a band of the compressed
inner cells of the nucellus which is thick in the basal region
and thins out gradually towards the apex of the nucellar
cavity, being in distribution similar to the tapetum of Arau-
carva, and like the latter also having suberized walls.

Hooker?® in his paper on Welwitschia (1863) states that
in this form ‘“‘the embryo-sac is a delicate membrane’’ which,
when the nucellus has elongated, “‘is found to have disappeared
over the summit of the endosperm.’’ Sir Joseph Hooker com-
pares Welwitschia and Gnetum with regard to this characteristic
of the distribution of the coat. This investigator also states
that ‘‘the membranous remains of the embryo-sac may often
be found on the surface of the nearly mature endosperm.”’
In photograph 19, plate 5, the character of the distribution of
the coat is indicated, at a stage soon after fertilization has taken
place. The coat is very thin, about 1.3 », at the stage ex-
amined, but still distinctly double, the outer part granular and
the inner homogeneous (fig. 16, pl. 2). No trace of a tapetum
is present. The nucellar tissue in the basal region is different-
iated, however, into an inner more or less collapsed part (see
photo. 19, pl. 5), which if compressed would appear thinner
but probably very much like the band that has been described
in Ephedra.

[128]

OF THE GYMNOSPERMS 47

In Griffith’s'® paper on Gnetum (1859) there is a statement
that a megaspore-coat is present at a young stage of ovular
development but that it disappears in the later stages. Hook-
er’s”® (1863) statement that the coat of Gnetum is similar in
distribution to that of Welwitschia gives us further information
with regard to it. Reference is made in Karsten’s*4 paper
(1892) which deals with this genus to the presence of a mega-
spore-membrane around the young spore. Lotzy®? (1899)
has referred to the coats of the megaspores of Gnetum Gnemon,
“which make them appear much more cryptogamous than the
embryo-sacs of most higher plants.’’ The megaspore-mem-
branes are represented as single and finely granular layers.
Those of the prothallia which will bécome fertile are, I have
estimated from his figures, about 1.2 » thick when free-nuclear
division is well advanced (Lotzy,** fig. 27, pl. 4). Those of
the infertile prothallia are thicker, some of them at certain
places being represented as fully 2.3 » thick.

The coat in the three genera of the Gnetales is thus thin and
the tapetum poorly developed. In distribution the coat
is of an accentuated Abietinae-type, scarcely a trace of it being
present in the apical region of the prothallium. The coat is
double in the basal region in two of the forms about the time
of fertilization, and will probably be found to be similar at a
like stage in Gnetum. Material of this form was not examined,
however, and the point remains in doubt.

FOSSIL GYMNOSPERMOUS SEEDS.

Numerous seeds of a primitive gymnospermous character
occur in the palaeozoic rocks. Many of these with the structure
well preserved are found in the Carboniferous and Permian
strata. Hooker and Binney’ were the first to study their in-
ternal features. Later Brongniart®! and Williamson!® worked
out in detail their intimate structure. A glance over the
illustrations of these investigators gives one an idea of the
prominence of the megaspore-membrane in the early seed-
plants. Brongniart whose specimens were exceptionally well
preserved refers to the megaspore-coat as follows, (Brongniart,

[129]

48 THOMSON : THE MEGASPORE-MEMBRANE

31, p. 242) : ‘La membrane intérieure ou périspérmique est
trés différente de celle qui limite le nucelle ; elle est extréme-
ment mince et ne parait pas cellulaire, mais marquée d’aréoles
diies a l’application des cellules qu’elle enveloppait et dont
il ne reste généralement plus de trace.’’ Williamson’s speci-
mens were not in so good a state of preservation but the mega-
spore-membrane is represented in many of the figures in his
paper. In the text also he often refers to the coat, and com-
pares it with the similar structure described by Brongniart.
These Palaeozoic seeds belong to a great variety of forms, some
related to the cycads, some to the conifers, and some, probably
very many, to the Cordaitales, the dominant gymnosperm
group of this era. Others again are seed-bearing Cycadofilices,
the Pteriodospermae, a group recently established by Oliver
and Scott.’ Still others are lycopods with seed-like fructi-
fications which Dr. Scott has described under the generic name
of Lepidocarpon. ‘The latter have an especially prominent
megaspore-coat, which in some cases is plainly double as its
separation into two layers in the chalazal region indicates
(Scott,?8 figs. 27 and 28).

In the mesozoic rocks the abundant remains of gymnos-
permous plants belong chiefly to the Bennettitales,a group with
strong cycadean affinities. Carruthers? in 1870 described for
the first time the female strobilus in Bennettites, the type genus
of this group. Neither he, Solms-Laubach, nor Scott, who
have studied the European forms, have observed that a mega-
spore-coat is present. In fact, for the only structure which
might be interpreted as such Solms-Laubach distinctly claims
a nucellar origin (Solms-Laubach,®? pp. 441 and 442). The
American representatives of the Bennettitales are being worked
over by Dr. Wieland who has already made very important
additions to our knowledge of the reproductive organs of the
group. In correspondence in regard to the presence of a
megaspore-membrane Dr. Wieland states that he is unable to
affirm directly that such a structure is present in the material
he has examined. With regard to other mesozoic gymno-
sperms, the forms that are related to the modern ones, the

[130]

OF THE GYMNOSPERMS 49

ancestral Ginkgoales, Coniferales, etc., I have been unable to
get any information on the presence of a megaspore-membrane
that is of value.

It is thus evident that in these fossil seeds the more primi-
tive, palaeozoic forms have a much more prominent megaspore-
coat than the specialized and more recent mesozoic ones, in
which even the occurrence of a coat does not seem to have
been demonstrated with certainty.

GENERAL CONSIDERATIONS.

The present work has determined the extent of the occur-
rence of a megaspore-membrane in the gymnosperms, as well
as the structure of the coat and its chemical composition.
The megaspore-membrane is present in all the groups and sub-
groups of these seed-plants, except the Taxeae of the Conifer-
ales, from the ovule of whose forms it is entirely or almost
entirely eliminated. The coat is strikingly uniform in struc-
ture and in chemical composition throughout this division of
the spermatophytes with the exception of one subgroup, the
Araucariinae. It is double, its exosporium, in the later stages
of development, composed of radially arranged fibrillae, and
its endosporium presenting an appearance, in section, of being
subdivided into two more or less equal, homogeneous strata.
The exosporium is suberized while the endosporium is of com-
posite chemical character. The outer stratum of the latter
is suberized but contains cellulose towards its inner border,
while the inner stratum consists chiefly of cellulose with which
entad is associated, a substance resembling pectin. The
megaspore-coat in fact closely resembles that of a microspore
(e.g. that of Panus) both in its structure and in its chemical
composition, and thus affords additional evidence of the free-
sporing nature of the ancestral forms of the gymnosperms.

In the forms where the normal type of membrane occurs
there is present a more or less well-developed tapetum. This
tapetum is derived from the sporogenous tissue as is shown
by the course of its development, the plurinucleate condition
of its cells, and by the suberization of their walls. It is quite

[131]

50 THOMSON : THE MEGASPORE-MEMBRANE

distinct from that which is derived from the nucellar tissues
and has for convenience been designated a “primary’’ tape-
tum.

The abnormal type of megaspore-membrane present in the
Araucariinae is comparable to the wall of the pollen-tube both
in structure and in chemical composition, a typical suberized
exosporium not being present. This group in many other
respects occupies a somewhat isolated position among the
subgroups of the Coniferales. The tapetum is of a peculiar
character in both Agathis and Araucaria,—just as abnormal as
is the megaspore-membrane. The ‘female flower,’’ too, is
difficult to homologize with that of any of the other forms.
Some consider that the seminiferous scale is a very reduced
structure being composed of the almost completely fused fertile
and infertile bracts, while others regard the ovule-bearing
structure as a simple sporophyll, and on this ground consider
the Araucariinae as the most primitive of conifers. In support
of this view reference is often made to the evidence afforded
by early occurrence of fossil Araucaria-like wood. Dr. Scott**
(p. 483), however, finds the case in this respect ‘‘ emphatically
snot proven’ on existing evidence’. The character of the
coat and of the tapetum is in keeping with his finding and
indicates in addition that the Araucariinae are to be re-
garded as a specialized subgroup of the Coniferales.

Leaving out of consideration the Araucariinae whose mega-
spore-membrane and tapetum cannot at present be satis-
factorily associated with the other gymnosperms, certain
general features which are important from the phylogenetic
standpoint have been demonstrated. The coat is thick, well
developed, and of fairly uniform distribution around the pro-
thallium in the Cycadales, the group which is recognized as the
most primitive of the modern gymnosperms. In the Gink-
goales it is thinner than in the Cycadales but similar in
distribution to that of the latter. The group is a recently
established one for the reception of the single form Ginkgo
biloba, which was previously included in the Taxeae of the
Coniferales but which is now considered ‘‘as the one surviving

[132]

OF THE GYMNOSPERMS 51

member of an ancient stock, derived from the same cycle of
affinities as the palaeozoic Cordaiteae’’ (Scott,* p. 485), and
has been given a phylogenetic position below the Coniferales.
In the last mentioned group,which comprises the much diversi-
fied forms of the present-day gymnosperms, the coat is very
varied in thickness and peculiar in its distribution around the
prothallium. It is, however, though thinner than in Ginkgo,
on the whole much thicker and more fully developed than in
the Gnetales, the group of gymnosperms which is recognized
as having the greatest affinity to the angiosperms. Thus in
the living forms it is seen that there is a direct relationship
between the thickness and state of development of the mega-
spore-coat, and the primitive character of the group, a progress-
ive destruction of the coat having gone on as the forms become
more specialized. The development of the suberized primary
tapetum in the different groups parallels that of the megaspore-
coat and affords confirmation of the evidence derived from the
state of development of the latter. Again in the fossil forms
the primitive palaeozoic representatives have a much thicker
megaspore-coat than the higher and more specialized mesozoic
ones, since, as was seen, the coat in the former is described as
a well preserved structure, while in the latter it is so poorly
developed as to have escaped observation, or, at least, descrip-
tion. Thus in both the great modern and fossil groups of
primitive seed-plants we have evidence that the megaspore-
coat varies in thickness according to the primitive or specialized
nature of the forms. That this is true of the subgroups as well
as of the large divisions is indicated by the study of the mega-
spore-membrane of the Cyeadales, the Cycadeae having the
thickest and the Euzamieae the thinnest coat.

The interpretation of the inter-relationships of the sub-
groups of the Coniferales in the light of the above generalization
is the chief object of the present work, and in connection with
the statement of the results obtained reference will be made to
certain other features of general phylogenetic importance which
are in keeping with the evidence afforded by the state of devel-
opment of the megaspore-coat.

[133]

52 THOMSON : THE MEGASPORE-MEMBRANE

All the subgroups of the Pinoideae have a megaspore-mem-
brane and a tapetum. Those of the Araucariinae are of a
specialized nature and have been referred to separately. Of
the other subgroups of the Pinoideae, the Taxodinae, which
Eichler in his classification has placed between the Abietinae
and the Cupressineae, show affinity to each of these subgroups
in the character of the megaspore-coat and the tapetum.
Reference has already been made to Arnoldi’s proposed separ-
ation of the two species of the Sequoias (p. 37). In the same
paper he gives it as his opinion that all the Taxodinae (Sequoi-
aceae) except Sciadopitys would be better associated with
the Cupressineae. The evidence afforded by the study of the
megaspore-coat of these forms is in favour of Arnoldi’s view and
indicates that the Taxodinae is a composite group, all the
forms examined, except Scradopitys, being best associated with
the Cupressineae to which their megaspore-coat approximates
in its state of development. The associated forms of the
Taxodinae and the Cupressineae have a very thin membrane
and a poorly developed tapetum, being in this respect highly
specialized, the true Cupressineae on the whole somewhat more
so than the Taxodinae. The forms of the combined groups
resemble one another in certain other features, such as the
absence of male prothallial cells, the grouped arrangement
of the archegonia and the degenerate nature of the brachy-
blast, all which point to their specialized character. The
cyclic arrangement of the leaves and sporophylls in the Cupress-
ineae proper is evidence of a similar nature.

The megaspore-coat of Sczadopitys is in distribution and
in structure similar to that of the Abietinae at the same stage
of prothallial development (see p. 34). It is, however, fully
one-half thicker, being almost or quite as thick as the coat of
Cycas at a similar stage. The tapetum of Sczadopitys, too, is
thick, especially so in the basal region, and its distribution
thus suggests the state of affairs at a later stage in the
Abietinae, when the tapetum consists of only a few cells in the
basal region. The megaspore-coat and the tapetum of Scza-
dopitys would thus seem to be of a primitive abietineous type.

[134]

OF THE GYMNOSPERMS 53

Other features as well indicate its affinity to the Abietinae.
Arnoldi has referred to the distribution of the archegonia as
being similar to that in the Abietinae. Again, in the vegetative
parts the development of long shoots and short shoots is
characteristic of both. The so-called “‘leaves’’ of Scradopitys
are peculiar and lend confirmation to the brachyblastic theory
of the seminiferous scale, which upon anatomical grounds
is considered to hold good for all the Coniferales. This
interpretation of the character of the ovule-bearing structure
of the Coniferales gains support from teratological evidence,
and is ‘‘in favour of the view that the Abietinae, and
the Taxodineae as well, are somewhat primitive orders”
(Jeffrey**, p. 456) since in these subgroups alone do _ pro-
liferous cones occur. In this connection attention is directed
to the very common occurrence of proliferating female
cones in Sciadopitys, (see Sir W. T. Thiselton-Dyer’s refer-
ence to Master’s work, Ann. Bot., 1903, pp. 779-787), and
to the additional evidence which is thus afforded of the primi-
tive nature of this form. The presence of two prothallial
cells in the microspore, and the lack of differentiation in
the male cells themselves, only one of which functions in
fertilization, are features which point in the same direction and
corroborate the testimony afforded by the state of develop-
ment of the megaspore-coat and the tapetum.

The Taxoideae have one order, the Taxeae, which must
be regarded as very specialized from the standpoint of the de-
velopment of the megaspore-coat and the primary tapetum,
since no such structures, or only traces of them, are present
in the three of its genera examined. In keeping with this
condition of affairs the female ‘‘flower’’ is of a very specialized
type. The axillary buds (brachyblasts) are reduced in number
and degenerate in organization. Moreover, no prothallial cells
are present in the microspore, and the single functional male
cell (in Taxus at least) is relatively very large and specialized.
In addition Taxus is the only one of the Coniferales in which
no resin ducts are developed, a feature in which it resembles
the Gnetales. Again, in Torreya a secondary, nucellar, but no

[135]

54 THOMSON : THE MEGASPORE-MEMBRANE

primary tapetum is developed. In the other suborder of
the Taxoideae, the Podocarpeae, there is a great difference in
the stage of development of the megaspore-coat and the tape-
tum in its two chief genera. Reference (seep. 43) has already
been made to certain associated differences which are in keeping
with the former and from which it would appear that Dacry-
dium is a much more primitive genus than Podocarpus. Cer-
tain resemblances of the whole group to the Abietinae have
been recently pointed out by Dr. Coker, such as the occurrence
of winged pollen grains, the arrangement of the archegonia,
the presence of two prothallial cells in the microspore (Podo-
carpus and perhaps others) and certain other features which
have led him to conclude “‘that in the Podocarpeae are to be
found the nearest living relatives of the Abieteae’’ (Coker,”®
p. 103).

From the standpoint, then, of the relative state of develop-
ment of the megaspore-coat and the tapetum we are to regard
the Abietinae as the most ancient group of the Coniferales;
the Taxeae as the most recent; the Taxodinae andthe Podo-
carpeae as complex groups, with some forms as ancient as,
or even more ancient than the Abietinae, and other forms quite
recent,—while the Cupressineae are considered as occupying
a somewhat intermediate position in the phylogenetic series.

[136]

OF THE GYMNOSPERMS 55

LITERATURE.

1. Hofmeister, W., Vergleichende Untersuchungen. 1851.

2. Mettenius, G. H., Beitrage zur Anatomie der Cycadeen.
Abhandl. d. kénigl. Sachs. Gesell. d. Wiss. 7, pp. 565-
608, pls. 1-5. 1861.

3. Solms-Laubach, H., Fossil Botany. English Edition.
Oxford, 1891.

4. Scott, D. H., The Anatomical Characters presented by
the Peduncles of the Cycadaceae. Ann. Bot. 11, pp.
399-419, pls. 20-21. 1897.

5. Worsdell, W. C., On ‘“Transfusion-Tissue’’. Trans.
Linn. Soc., V,B, pp. 301-319, pls. 23-26. 1897.

6. Petonié, Lehrbuch der Pflanzen-Palaontologie. Berlin,
1899.

7 (a). Oliver, F. W., and Scott, D. H., (1) On Lagenostoma
Lomaxt, the seed of Lyginodendron. Ann. Bot. 17,
Dp625-020:.) 19034) 2) py Oovathe) Stricture of , the
Palaeozoic Seed, Lagenostoma Lomax, with a statement
of the evidence upon which it is referred to Lygino-
dendron. Before the Roy. Soc., Jan. 21st, 1904.

(b). Oliver, F. W., A New Pteridosperm. New Phyt.
4:32. 1904.

8. Hirasé, S., Etudes sur la Fécondation et 1’Embryogénie
du Ginkgo biloba. Jour. Coll. Sci. Imp. Univ. Tokyo 8,
PP. 307-322, pls. 31-32. 1895 ; 12, pp. 103-149, pls. 7-9.
1898.

9. Ikeno, S., Untersuchungen tuber die Entwickelung
der Geschlechtsorgane und der Vorgang der Befrucht-
ung bei Cycas revoluta. Jahrb. f. wiss. Bot. 32,
PP. 557-602, pls. 8-10. 1808.

10. Webber, H. J., Spermatogenesis and Fecundation
of Zamia. U.S. Dept. of Agriculture, Bureau of Plant
Industry, Bull. No. 2, 1gor.

11. Lang, W. H., Studies in the Development and Mor-
phology of Cycadean Sporangia, II. The Ovule of
Stangerta paradoxa. Ann. Bot. 14, pp. 281-306, pls.
E7-LS.) LOO.

[137]

56

12
13

14

15.

16.

7

18.

19.

20.

7410

Pid,

AR,

24.

THOMSON : THE MEGASPORE-MEMBRANE

. Strasburger, E., Das botanische Practicum. 1896.
. Warming, E., Recherches et Remarques sur les Cyca-

dées. Oversigt over d. K. D. Vidensk. Selsk. Forh. 1877.
. Warming, E., Contributions a |’ histoire naturelle des
Cycadées. Ibid. 1879.

Hooker, J. D., and Binney, Edward, On the Structure
of certain Limestone Nodules enclosed in seams of
Bituminous Coal, with a Description of some Trigono-
carpons contained in them. Phil. Trans. 145, pp. 149-
156, pls. 4-5. 1855.

Williamson, W. C., On the Organization of Fossil
Plants of the Coal-measures, Part viii. Ferns (con-
tinued) and Gymnospermous Stems and Seeds. Phil.
Trans. 167, pp. 213-270 pls. 5-16. 1876.

Seward, A. C., and Gowan, Miss J., The Maiden Hair
Tree (Ginkgo biloba L.). Ann. Bot. 14, pp. 109-154,
pls. 8-10. 1900.

Griffith, William, Remarks on Gnetum. Trans. Linn.
Soc., 22, pp. 299-312, pls. 55-56. 1859.

Juel, H. O., Ueber den Pollenschlauch von Cupressus.
Flora 93, pp. 56-62, pl. 3. 1904. Reviewed by C. J.
Chamberlain in the Bot. Gaz., March 1904.
Sokolowa, Mlle. C., Naissance de l’Endosperme dans
le Sac embryonnaire de quelques Gymnospermes. Mos-
cou, 1880.

Arnoldi, W., Beitrage zur Morphologie einiger Gymno-
spermen, V. Weitere Untersuchungen der Embryogenie
in der Familie der Sequoiaceen. Bull. de la Soc. de Nat.
de Moscou, 1901, pp. 449-477, pls. 7-8. Also review by
Karsten, Bot. Zeit. Nov. 16th, rgor.

Lawson, A. A., The Gametophytes, Archegonia, Ferti-
lization and Embryo of Sequota sempervirens. Ann.
Bot. 18, pp. 1-28, pls. 1-4. 1904.

Shaw, W. R., Contributions to the Life-history of
Sequoia. Bot. Gaz. 21, pp. 332-339, pl. 24. 1896.

Karsten, G., Beitrag zur Entwickelungsgeschichte
einiger Gnetum-Arten. Bot. Zeit. 50, pp. 237-246,

pls. 5-6. 1892.
[138]

25

26.

af.

28.

29.

30.

aE.

22.

33:

34.
35.

OF THE GYMNOSPERMS 57

Coker, W. C., On the Gametophyte and Embryo of
Taxodium. Bot. Gaz. 36, pp. 1-140, pls. 1-11. 1903.
Coker, W. C., Notes on the Gametophytes and Embryo
of Podocarpus. Bot. Gaz. 33, pp. 89-107, pls. 5-7. 1902.
Lloyd, F. E., Vivipary in Podocarpus. Torreya, Vol.
2, No: 8;pp.11321n7. 7 1902.

Scott, D. H., On the Structure and the Affinities of
Fossil Plants from Palaeozoic Rocks, IV. The Seed-
like Fructifications of Lepidocarpon, a genus of Lycopodi-
aceous cones from the Carboniferous Formation. Phil.
Trans., Vol. 194 B, pp. 281-333, pls. 38-43. 1901.
Hooker, J. D., On Welwitschia, a new genus of Gnet-
aceae. ‘Trans. Linn. Soc. 24, pp. 1-48, pls. 1-14. 1863.
Lotsy, J., Contributions to the Life-history of the
Genus Gnetum. Ann. Jard. Bot. Buitenzorg, II, 1,
pp. 46-114, pls. 2-11. 1899,

Brongniart, Ad., Etudes sur les Graines Fossiles trou-
vées & 1’Etat silicifié dans le Terrain Houiller de Saint-
Etainne. Ann. des Sci. Nat. Bot. 20, PP. 234-265, pls.
ZN 2A TOT A:

Carruthers, W., On Fossil Cycadean Stems from the
Secondary Rocks of Britain. Trans. Linn. Soc. London,
26, pp. 675-708, pls. 54-63. 1870.

Solms-Laubach, H., On the Fructification of Bennettites
Gibsonianus Carr. Ann. Bot. 5, pp. 419-454, pls. 25-26.
I18gI.

Scott, D. H., Studies in Fossil Botany. London, 1900.
Jeffrey, E. C., The Comparative Anatomy and Phylo-
geny of the Coniferales, Part I. The Genus Sequoia.
Mem. Bost. Soc. Nat. Hist. Vol. 5, No. 10, pp. 441-459,
pls. 68-71. 1903.

[139]

58 THomson : THE M&GASPORE-MEMBRANE

EXPLANATION OF PLATES.

Plates 1 and 2 consist of drawings which were outlined by
the aid of acamera lucida. The drawings were made from the
lateral basal region of the prothallium except where otherwise
stated,since the membrane in this region is more uniform and
assumes about its average thickness. The figures are oriented
so that the prothallial tissue is to the extreme left of each. An
attempt has been made both by the use of the camera lucida
and by actual measurement to reproduce the relative thickness
of the coats in the various forms. For accurate comparison
however, the measurements must be taken as the basis. Only
in a few cases has any attempt been made to indicate the
stratification of the endosporium.

Plates 3-5 are reproductions of photomicrographs, and
for details of structure should be examined with a lens.

Plate a:

Fig. 1. Cycas revoluta. (4.5 ») The endosporium and the

exosporium are about equal in thickness. The former
presents the appearance of being subdivided by a broad
and coarsely granular area ; the latter is finely striated.
The tapetal cells are large and contain many amylo-
dextrous starch grains. These appear in the unstained
condition as vacuoles.

Fig. 2. Ginkgo biloba. (4.5-5 #) Membrane of adult seed,
separated slightly from the endosperm ceils, the outer
walls of which are especially thick and have an outer
suberized layer. The exosporium is relatively very
thick and is finely and irregularly striated. The tape-
tal and nucellar tissues consist of collapsed strands.

Fig. 3. Pinus resinosa. (4.2 #) For the stage see photo. 4,
pl. 3. The endosporium is about one-third as thick as
the exosporium. There is but a slight ‘‘reinforcement’’
from the endosperm cell walls. To the right of the
membrane the tapetal tissue is represented and the
inner edge of the nucellus as well.

[140]

OF THE GYMNOSPERMS 59

Fig. 4. Larix Europaea. (4.4 ») The walls of the endosperm
cells are just forming. The tapetum consists of a single
layer of rather collapsed cells.

Fig. 5. Larix Americana. (4.7 » ) The drawing is from the
basal part of the megaspore. The membrane is like
that of L. Europaea, but of heavier structure and
slightly ‘‘reinforced’’ by the outer walls of the endo-
sperm-cells. The tapetal cells are large in this region.

Fig. 6. Picea excelsa. (4.7 ») The stage of development is
indicated in photo. 7, pl. 3. The exoporium is relatively
very thick. It is finely and very regularly striated in
structure. The reinforcement is slight.

Fig. 7. Picea nigra. (4.6 ») Basal region of megaspore-mem-
brane—very similar to that of P. excelsa. Collapsed
tapetal cells are present.

Fig. 8. Tsuga Canadensts. (4.5 ») Membrane as in last two
but thinner. To the right of the exosporium is a layer
of tapetal débris which is really closer to the membrane
than it has been represented and makes the membrane

_ appear quite thick (See photo. 8, pl. 4).

Plate 2.

Fig. 9. Abies balsamea. (4.6 +) The stage of the ovule from
which drawing was made is seen in photo. g. pl. 4.
The membrane is slightly “reinforced.’’ The endo-
sporium is homogeneous and about one-fourth as thick
as the exosporium which is very distinctly and coarsely
striated.

Fig. 10. Pinus sylvestris. (3 ») Membrane of an ovule ata
slightly earlier stage than that indicated in photo. ro, pl.
4, in the parietal nucleated condition. The endo-
porium is homogeneous while the exosporium is finely
granular and shows signs of an indistinct radial striation,
To the right of the megaspore the tapetal cells appear as
quite a distinct layer.

[141]

60 THOMSON : THE MEGASPORE-MEMBRANE

Fig. 11. Sciadopitys verticillata. (4.2 ») The stage of the
ovule is indicated in photo 11, pl. 4. Cell formation is
proceeding in the megaspore. The exosporium is very
finely granular and radially striated and about one and
one-half times as thick as the endosporium, whose two
substrata are fairly distinct.

Fig. 12. Sequoia sempervirens. (2-7 »). The drawing was
made from a part of an ovule (at about the stage indi-
cated in photo. 12, pl. 4) where the membrane was
free from the nucellar tissues. Usually it is more close-
ly pressed up against the latter than the figure would
indicate, no tapetum occuring between them.

Fig. 13. Biota ortentalis. (1.7 ») The drawing shows that
the apparently thick megaspore coat indicated in photo.
14, pl. 5 is not really a true megaspore-membrane but
that a great part of the thickness of the investing layer
is made up of the thickened outer walls of the peri-
pheral endosperm cells. The endosporium is homo-
geneous, the exosporium granular.

Fig. 14. Juniperus Sabina. (2.7 ») The stage of develop-
ment of the ovule is indicated in photo. 16, pl. 5. The
membrane seems to vary somewhat in thickness in
different parts of the section. The two layers of the
coat are about equally thick. The endosperm cells
‘“‘reinforce’’ the membrane considerably. To the right
the collapsed tapetal cells are evident and the inner
border of the nucellar tissues.

Fig. 15. Dacrydium laxifolium. (4.3 ») Membrane at the
stage indicated in photo. 17, pl. 5. It invests closely
the endosperm and is ‘‘reinforced’’ by the outer wall
of the peripheral endosperm cells, the thickness of which
is about equal to that of the inner layer of the membrane
proper. The exosporium is coarsely and irregularly
striated. The tapetal tissue has collapsed.

[142]

OF THE GYMNOSPERMS 61

Fig 16. Welwitschia mirabilis. (1.3 p») At the stage
indicated in photo. 19, pl. 5, the megaspore-coat is thin
even in the lateral basal part of the ovule. The inner
border of the nucellar tissue consists of very much
collapsed cells.

Plate 3.

Photo. 1. Cycas revoluta. (Meg-memb. 4.5 » thick.) The
ovule is at a stage just previous to cell-formation in the
prothallium. The coat is double, its two layers about
equally thick. (For details see fig. 1, pl. 1.) An arti-
ficial separation of the layers is indicated about the
middle of the photograph. The staining is such as to
bring out the layering of the coat but not the structure
of the prothallial and other tissues.

Photo. 2. Stangeria paradoxa. (Meg-memb. 4.2 p thick.)
The endosporium appears subdivided in this case. Over-
staining for contrast effect has obscured the structural
details of the layers. The large cells of the tapetum
are indicated to the right and the gametophyte to the
left.

Photo. 3. Araucaria wmbricata. (x 50) Archegonial ini-
tials have developed in the lower lateral parts of the
prothallium. The distribution of the tapteum is
fairly well indicated, though usually much more is
seen in the apical region.

Photo. 4. Pinus resinosa. (x 20) The megaspore-mem-
brane closely invests the prothallium, becoming much
thinner in the archegonial region. The archegonium
has not been fertilized. Remains of the tapetum are
apparent in the basal region.

Photo. 5. Larix Europaea. (x 25) Several archegonia are
i present, two with neck and ventral canal cells. The
megaspore-membrane is scarcely perceptible in the
archegonial region and very thin for some distance be-
low this, while in the chalazal region it is thick. The
basal portion of the tapetum and of the nucellus have
been torn away.
[143]

62

THOMSON : THE MEGASPORE-MEMBRANE

Photo. 6. Picea nigra. (x 50) Just prior to fertilization.
Megaspore-membrane somewhat broken but closely
investing the prothallium while the tapetal remains are
sparse but distinct in the basal region.

Photo. 7. Picea excelsa. (x 25) Some of the archegonia
have been fertilized and the first nuclear divisions have
taken place. Others have not been fertilized. The
megaspore-membrane thins out uniformly towards the
archegonial region. The tapetum is apparent in the
basal region.

Plate 4.

Photo. 8. Tsuga Canadensis. (x 40) The megaspore-coat
thins out somewhat in the archegonial region, but not
so much as in Lavex, ete,’ The (tapetum as, pretty,
well destroyed but can be detected in the basal region.
Granular material adds somewhat to the apparent
thickness of the coat.

Photo. 9. Abies balsamea. (x 20) ‘The ovule has been
fertilized and the megaspore-coat is very thick in the
basal region thinning out gradually towards the micropy-
lar part. Tapetal remains are evident in the basal
region.

Photo. 10. Pinus sylvestris. (x 80) Ovule in longitudinal
axial section. The coat is double and evenly distributed
around the prothallium, which is loosely invested by
tapetal cells.

Photo. 11. Scradopitys verticillata. (x 40) “Revived”
material. Longitudinal axial section of an ovule at
the stage when archegonial initials are appearing. The
megaspore-membrane is of uniform thickness all around
the prothallizm. The tapetum is thick, especially in
in the basal region.

Photo. 12. Sequoia sempervirens. (x 50). The megaspore-
membrane is attached more or less to the nucellar
tissue. The endosperm is becoming cellular and col-
tected towards the chalazal region. ‘The section passes
through several megaspores towards the micropylar end.

[144]

OF THE GYMNOSPERMS 63

Photo. 13. Cryptomeria Japonica. (x 100) No membrane
is to be seen at this stage. The tapetum consists of a
single rather loose layer of cells.

Plate Re

Photo. 14. Biota orientalis. (x 10) Apparently shows
quite a thick megaspore-membrane at the stage when
numerous rudimentary embryos are developing. The
explanation is given in fig. 13, pl. 2, where the “rein-
forcement’’ from the cell walls is seen to be very thick.

Photo. 15. Thuja occidentalis. (x 20) A young embryo
is seen in the upper portion of the prothallium. The
megaspore-membrane is thin even in the basal region,
and the tapetum not apparent.

Photo. 16. Juniperus Sabina. (x 20) The megaspore-coat
appears very thick, much thicker than it really is since
it is “‘dragged’’. The tapetum can be seen in the basal
region. In the archegonial region the contents of
several pollen-tubes are apparent.

Photo. 17. Dacrydium laxtfolium. (x 20) Sections of dry
herbarium material ‘‘revived’’ 4 years after collection.
The integument and nucellus have been removed. The
megaspore-membrane is thus the thick coat enclosing
the prothallium in whose axis embryonic and suspens-
orial cells are visible.

Photo. 18. Taxus Canadensis. (x 50) The nucellus and
contained structures only are present. Above the
embryonic and suspensorial ceils a pollen-tube without
any contents is apparent.

Photo. 19. Welwitschtia mirabilis. (x 25) The megaspore-
membrane lies free, midway between the endosperm and
the nucellus. It thins out very perceptibly towards
the micropyle. Suspensorial cells with a triangular
mass of embryonic tissue at the base are visible in the
longitudinal axis of the prothallium about one-third of
the distance from its apex.

[145]

64 THOMSON : THE MEGASPORE-MEMBRANE

Photo. 20. Ephedra distachya. (x 12) The megaspore-
membrane is fused with a felt-work from the adjoining
nucellar issue. Numerous free nuclei are present in the
egg-cells.

[146]

BIOLOGICAL SERIES NO. 4

~

R: By T. DEL.

HELIOTYPE CO.

12

Il

4

1ES NO.
O

i

>

BIOLOGICAL SEF

WREST

16

a

= a

Seg RI Fp aS

Sas ——, aise ee
CRM A Si yee” age

2258553 SIF

ee ee

eo a

a>

13

HELIOTYPE CO.

Re Be T. DEL.

v

VINIVE MOET VP ITVNVINIY OI VUE

"LA
» BIOLOGICAL SERIES NO. |

Be en ee
ee ae

-"ty

BS

ee
b fae se
UAT an “eal
Wine
ri Pi

a

—Se

See ee
a
Lag

ae

~

HELIOTYPE CO

PHOTOMICROGRAPHS

‘

TD.

LO

RAPHS

HOTOMICROG

dD
Pp

tas

rFLAIE

UNIVERSITY OF TORONTO SIUUICS

BIOLOGICAL SERIES NO, |

ee ae

»

ve)

HELIOTYPE CO.

-HOTOMICROGRAPHS

p

)

UNIVERSITY OF TORONTO
PLUDIES
BIOLOGICAL SERIES

No. 5: THE HOMOLOGIES OF THE STYLAR CUSPS
IN THE UPPER MOLARS OF THE DIDELPHYIDAE
BY B. ARTHUR BENSLEY

THE UNIVERSITY LIBRARY, 1906: PUBLISHED BY THE
LIBRARIAN

COMMITTEE OF MANAGEMENT

Chairman : James Loupon, LL.D., President ot the
University.

PROFESSOR W. J. ALEXANDER, Ph.D.
PROFESSOR PELHAM EDGAR, Ph.D.
PRINCIPAL J. GALBRAITH, M.A.

PROFESSOR R. Ramsay WRIGHT, M.A., B.Sc.
PROFESSOR GEORGE M. Wronc, M.A.

General Editor: H. H. Lancton, M.A., Librarian of the
University.

THE HOMOLOGIES OF THE STYLAR GUSPS
IN THE UPPER MOLARS OF THE
DIDELPHYIDAE

B. ARTHUR BENSLEY, Pu.D.

LECTURER IN ZOOLOGY IN THE UNIVERSITY OF

m=
c
‘

THE HOMOLOGIES OF THE STYLAR CUSPS OF THE
UPPER MOLARS OF THE DIDELPHYIDAE*

INTRODUCTION

In the course of my studies on the relationships of the
Australian Marsupialia I had occasion to examine the exten-
sive series of skulls of modern Didelphyidae preserved in the
British Museum, with the object of defining the ancestral
characters present in the dentition. My notes on this subject
included a thorough survey of the stylar elements in the upper
molars; but, since the main changes in secondary evolution
are played upon the main cusps of the molar crown, the modifi-
cations of these elements were not referred to in the general
results as published,t except where points of sequence seemed
to demand their consideration. In fact, the detailed study
which I devoted to these structures was undertaken largely
through interest in Winge’s theory of dental evolution, accord-
ing to which they are considered as normally three in number
and as forming the original elements of the molar crown.t
It has since appeared to me probable, however, that an account
of the modifications of the stylar cusps in such a primitive
group as the Didelphyidae would afford a basis for their com-
parison as molar elements in marsupials with similar structures
in placentals, and that possibly their characters might afford
a better means of discriminating the molar patterns of early
Tertiary or Pretertiary representatives of the two groups than
those of the main cusps of the molar crowns, in view of the
fact that both are of trituberculate origin. The identification
of stem forms depends on the distinction of the primitive
characters of marsupials and placentals from one another,
and from those which may be common to both; and no
adequate conception can be formed concerning them until the
habit of denoting the characters of early placentals as ‘‘ mar-

* Read at the meeting of the Society of Vertebrate Palacontologists, New
York, December 28th, 1905.

{ Bensley, B. A., On the Evolution of the Australian Marsuptalia, etc.
(Trans. Linn. Soc., London; Ser. 2, vol. 9, pp. $3-217.)

tWinge, H., Om Pattedyrenes Tandskifte, 1saer med Hensyn til Taendernes
Firmer. (Vid. Medd. f. d. Naturh. Foren., Copenhagen, 1882.)

[149]

4 BENSLEY: HOMOLOGIES OF THE STYLAR CUSPS

supial” rather than as ‘‘primitive,’’ or ‘“‘marsupioplacental ”’
has been finally abandoned.

GENERAL CHARACTER AND DISTRIBUTION OF STYLAR CUSPS
IN THE EXISTING MARSUPIALS

The stylar elements are accessory structures in the molar
crown. They are serially arranged and represent processes
of an external ridge or cingulum which passes along the outer
faces of the paracone and metacone. They are particularly
characteristic of the Didelphyidae, but are found among the
Australian marsupials in the Dasyuridae, Peramelidae, and
in the phascolarctine division of the Phalangeridae. It is
apparent from their associations that they are primitive and
more or less conservative elements belonging to the insect-
ivorous stage of evolution. Although reduced in number they
are retained in the carnivorous development of the Dasyuridae,
two of them being associated with the paracone and metacone
in the production of a double shearing edge. In the incipient
omnivorous development, as indicated in the Peramelidae,
they are retained as in more primitive forms. This is true
also, although to a less extent, of the herbivorous and seleno-
dont development as seen in the Phascolarctinae. In the
omnivorous and herbivorous developments of the Phalanger-
inae, in which there is a bunodont modification of the molars,
the stylar elements disappear and even in the primitive forms
the cingulum is barely indicated. Their function in the
insectivorous stage is apparently that of preventing the food
from slipping off the smooth concave outer faces of the para-
cone and metacone, although they are doubtless accessory
piercing agents as well. They are associated particularly
with the paracone and with the piercing tip of the metacone.
The trenchant spur of the metacone tends to be free of these
elements on its outer side apparently in order that the shearing
action may not be hampered.

THE RELATIONSHIPS OF THE EXISTING DIDELPHYIDAE
In considering the arrangement of the stylar elements in
the Didelphyidae it is advisable to bear in mind the probable
[150]

OF THE DIDELPHYIDAE 5

relationships of the various representatives of the family.
These relationships as deduced from a study of the dentition
and foot-structure and of other characters are indicated in the
appended plan, the details of which are more fully explained
in the paper already mentioned

(loc. cit. pp. 182-185). The two

Chironectes Didelphys subgenera Marmosa and Per-
ee ie - amys include the smallest and

most primitive forms of tlie
family. They show the closest

A GinC tenes correspondence in _ dentition,

Caluromys / ENA :
ont Peramys being if anything the
\ / more primitive, as seen in the
Peramys greater development of the pos-
Dromiciops WA terior premolar, a character
th we which belongs also to the Oligo-
Wea cene Peratherium, judging from

Marmosa

the examples which have come
to my notice. The subgenera
Chironectes and Didelphys, to-
gether with their prototype Metachirus, are to be considered
apart from Caluromys and Dromiciops. The former are larger
but, in dentition, conservative forms retaining the general
conditions of Marmosa and Peramys, while the latter show
special characters indicating the beginnings of omnivorous
specialization.

THE STYLAR ELEMENTS IN PERATHERIUM
In the estimation of primitive conditions in the stylar
formula the question naturally arises—what was the condition
of these structures in Perathertum ? Although through the
kindness of Dr. Smith Woodward I was able to examine in
detail the British Museum specimens, I was unable to decide
this question to my satisfaction. The majority of the speci-

mens represent mandibular rami.* Of the few fragments of
upper jaws only one shows the characteristics of the external

* Lydekker, R., British Museum Catalogue of Fossil Mammalia, pt. 5,
pp. 283-288. 1887.
[151]

6 BENSLEY: HOMOLOGIES OF THE STYLAR CUSPS

styles. In this specimen (No. 27807), however, they appear
in the molars of the right side in a beautifully preserved,
unworn and unbroken condition. They are moderately
developed, and present the same condi-
tion in mr and m2. They tend to be

beeen) reduced in m3,and in m4 they are
ee hits: Ne, absent, this tooth being reduced just as
b he c in modern Didelphyidae. In mr and

m2 the external cingulum bears in all
six elevations, the arrangement of
which is shown in the appended dia-
gram of the external profile of m2 of the
right side reversed. Three elements, a,b, andc, are conspicuous,
and three others, b1,c1, andc2,aresubsidiary. Styleais situated
at the tip of the anterior spur of the paracone, b opposite its
concave face; c is situated opposite the small anterior spur of
the metacone. Of the subsidiary styles b1 and ci are accom-
modated in the space between styles b and c and appear to be
related respectively to these elements. Style c2 is placed on
the outer edge of the enlarged metacone spur. This element
is so small as to be scarcely recognizable, and its recognition
is still more difficult in this specimen on account of the dark
colouration assumed in fossilization.

Apparently the stylar cusps are as well developed in
American specimens. Cope* remarks of Peratherium that
‘“‘the superior molars, excepting the last, present two median
V’s which would be termed external but for the fact that the
external basal cingulum is so developed as to constitute an
external crest.”

Fig. 1. Stylar Cusps
an Peratherium -

THe STYLAR ELEMENTS IN EXISTING DIDELPHYIDAE
Peramyst

P. dimidiata. Two specimens in the British Museum col-
lection show the stylar cusps in an unworn condition, and both

*Cope, E. D., Tertiary Vertebrata, pp. 789 et seq. 1884.
tIn this and the succeeding subgenera the descriptions refer only to young
specimens or those in which the external styles are quite unworn.

[152]

OF THE DIDELPHYIDAE 7

present the same pattern (Fig. 2A—97.1.1.4*). In m1 the
elements representing b and c of Peratherzum are conspicuous.
Style a is indicated as a slight projection, and in one of the
specimens there is a faint indication of cr. In m2 five pro-
jections are shown on the cingular ridge, and these are identical
in size and arrangement with those in Peratherium, the sole
difference being in the absence of c2. In m3 we find again the
same condition except that the whole ridge tends to be reduced.

P. scalops. In two speci-

A ae Ce ee KN ee mens mi and m2 show the

absence of the intermediate
Styles br) and) cr. In) me
rio Wee ln bp the element c2 is present.
In m3 style bi is present in

CLAP Ka, \u association with c, which is

reduced (fig. 2B).

D ee KA —_ P. iheringi. Two speci-

mens show in all three
molars the predominance of
vi LAP br WW b and c, and the presence of
both intermediate styles br

ae Ly LAp brn and ci. Style a tends to be

reduced, and style c2 is ab-

G Lae Wen Gai sent; otherwise the pattern

is much as in the specimen
of Peratherium (figs. 1, 2C—
- Ln bn) byw 61.12.2.9). In m3 the same
reduction of the posterior
£ WAY boy Wry styles is shown as in the
Fig. 2. Stylar Cusps in Peramys specimens of Peramys scal-
ops.

P. sorex. In one young specimen displaying the first two
molars the number of styles represents the minimum, only a,
b, and c being developed (fig. 2D).

P. americana. In one specimen four elements are present
in m1and mz. Style cr is seen in association with c, style b1

*The numbers indicated are those of the British Museum Catalogue.

[153]

8 BENSLEY: HOMOLOGIES OF THE STYLAR CUSPS

being absent (fig. 2E). In m3 is shown a reduction of all
elements, those present being a, b and c.

P. domestica. Two specimens show in all three molars
the minimum of three styles as described for the first two
molars of P. sorex (fig. 2F—52.2.22.10). <A third specimen
shows a conspicuous difference. In m2 styles b and c are
enlarged and approximated, as generally in the more special-
ized Chironectes, Didelphys, and Metachirus. Style c2 is present
also as in the latter. In m3 styles b and c are separated
by an element probably representing c1, but style c2 is still
present as a small element (fig. 2G).

P. brevicaudata. Two specimens show the reduced formula
in mi and m2, as in the specimens of P. sorex and P. domestica.
In m3 five elements are present and the posterior ones are
reduced, as in the specimens of P. theringi (fig. 2H—o.5.16.60).
A third specimen shows practically the same condition except
that styles b1 and ci are only faintly indicated in m3, and style
c2 is developed on the posterior face ofc. Ina fourth specimen
mi and m2 show the presence of style bi (fig. 2I—67.4.12.540).

Marmosa

M. simonsi. In four specimens the cingular ridge is
moderately developed and bears five projections giving a
pattern much like the specimen of Perathertum, except that
the intermediate elements are relatively larger in mr and m2,
and that in m3 style c1 tends to be divided (fig. 3 A—99.8.1.23).

M. elegans. In two specimens five elements are again
indicated in m2 and m3. In mz styles a and cr are more
moderately developed and bi is absent (fig. 3B—98.8.2.12).
In five other specimens the intermediate styles bi and cr are
absent, while in two others the same condition obtains except
that c1 is indicated in m3 (fig. 3C).

M. sinaloe. In two specimens only four styles are develop-
ed. Of these a, b, and c show their usual relations, while style
ci is in comparison enlarged (fig. 3D). This is a feature which
becomes prominent in the species M. murina and M. cinerea.

[154]

OF THE DIDELPHYIDAE 9

M. marica and M. dryus. In three specimens of each
species the same conditions are seen as in M. sinaloe except
that in M. dryus style c1 is still better developed and c reduced.

M. microtarsus. One specimen shows the condition of
M. sinaloe except that in m3 the intermediate styles br and c1
are both present as in M. simonsz.

M. murina. Four speci-

- (rn Pye rye bap mens show the presence in

mi and m2 of five styles, of

B . which a, b, and c are moder-
vw bay bp ately developed. Style cx is
more definitely enlarged and

rl LAY binge) LAY compares in size with ec.

Style b1 is barely indicated

D Bi) Agia Conan) paige. gE. B-—97.4..7.12,0.5-1,

59). In m3 a different

E \ pattern is shown in each
baw bru specimen. Inthree of them

a, b, and ¢ are indicated,

¥ ee Gary yes and in two cr is present

(fig. 3F). Tie fourth speci-

G ee Craw A men shows a more signifi-

cant arrangement,the cingu-
lum ridge being poorly
IT
foe) Lu) br developed and occupied by
a number of low crenula-

ys ee, Lay CORN, tions (eight) prophetic of

Fig. 3. Stylar Cuspsin Marmosa the condition in the next
genus Caluromys (figs. 3E,
4B). A fifth specimen shows the increased development of ci in
comparison withce, especially inmz. Inmg3therelation is primi-
tive, bi and cr being present together and of moderate size
(fig. 3G—97.6.7.52). In a sixth specimen a more primitive
and moderate development is seen in all three molars (fig. 3 H—
97.6.7.26).
M. cinerea. The general condition is much as in M.
mura. In mi and m2 of one specimen the main stylar
[155]

10 BENSLEY: HOMOLOGIES OF THE STYLAR CUSPS

elements are b, cr, and c, and these are developed to about
an equal extent. Style bi tends to be slightly cleft. The
same three elements are seen in m3, but they decrease from
before backwards from b (fig. 31). In four other specimens the
styles are less modified, five elements being typically present,
of which b and c show their normal predominance, while bi
and ci are present together, but of small size. There is a slight
reduction of style c in m2.

The relations in M. murina and M. cinerea are of interest as
showing among the variations tendencies on the one hand
towards the more primitive condition of the styles, as seen for
example in M. simonst, and on the other towards the reduction
of the cingulum, as seen in Caluromys. The apparent increase
of style c1 signifies not an increase in development of the cingu-
lum but a general levelling off of the stylar projections. In
the omnivorous and bunodont development of the molars, as
seen in the phalangerine division of the Australian Phalan-
geridae, the cingulum is obliterated, being only seen as a faint
ridge in primitive forms. Caluromys among the Didelphyidae
shows indications of embarking upon this modification, and
the two species of M.murina and M. cinerea which are
in many respects prototypal to Caluromys show in association
with the latter the very first stages in this reduction.

Caluromys

C. laniger. In one specimen the transitional characters
between this genus and M. murina and M. cimerea are well

exemplified. In mz four

sf Cus COO Loar elements are present pro-

jecting to almost the same

B eon) (aga) Cae extent. Particularly notice-

able is the increased size

( ) of c1 in comparison with c

c br Law and the absence of bi. In
m2 the condition is slightly

Be T A Ce changed by a peculiar reduc
Fig. 4. Stylar CuspsinCaluromys tion of style b. In m3 the

[156]

OF THE DIDELPHYIDAE II

cingulum is simply crenulated and the individual elements
are not apparent (fig. 4C—44.12.18.28). Ina second speci-
men showing the same characters in mi and m2 the individ-
ual elements are apparent, although greatly reduced, and
the ridge would still be described as crenulated (fig. 4D—
80.5.6.88). In the variety C. laniger pallida two specimens
show distinctly the reduction of the cingulum (fig. 4A, B—
v7.01. S2,:0.7.11.35).

C. trinitatis. Two specimens show minute crenulations
only.

Metachirus

The general conditions in this genus and in Chironectes and
Didelphys appear as modifications of those of the primitive
species of Marmosa and Peramys, and it is seen that the devel-
opment in Caluromys is a special one ending as far as the
modern Didelphyidae are concerned with that genus. The
general arrangement is seen in the approximation of styles
b and c, development of style c2, an element scarcely recog-
nizable in the smaller forms, through the shifting forwards of
style c, the variable indications of the intermediate styles
bi and cr.

M. opossum. In one

<4 Var Neue) lay specimen four elements

make up the series,

B isd namely, a, b, cand c2
i aad Lad (fig. 5A—46.2.2.4). This

Cc Specimen is abnormal as
nV Vac La br» seen in the development

of m4 like the remaining

me Ay? yay ali teeth. In m4 styles b

and c are present though
E us of small size. A young
ld ly specimen of M. opossum
var. melanurus shows the

dd
ay LY (~) : characteristic form of
Fig. 5. Stylar Cusps in Metachirus stylesb,c,andc2, but the

[157]

12 BENSLEY: HOMOLOGIES OF THE STYLAR CUSPS

original elements bi and ci are present together in m1, defi-
nitely showing the homologies (fig. 5B—97.11.7.60). Style
br is shown in m2 of this specimen and also in a second speci-
men (fig. 5C—o.7.11.79). In a third specimen it is seen in
all three molars. A fourth shows exactly the condition
described in the specimen of M. opossum in which the inter-
mediate elements are absent in all three molars.

M. crassicaudata. Two specimens show the conditions
represented in fig. 5H, F (85.11.26.11, 79.5.1.13), which are
easily referable to the general type in M. opossum. ‘The
cingulum is greatly reduced in m3 in this species.

Chironectes

C. minimus. One specimen shows the approximation
and extra development of styles b and c, and style c2 is evident
in m1 and mz, so that the type corresponds with that of
Metachirus. Intermediate styles are absent in this specimen
(fig. 6A—849.a), but in a second young specimen they are
indicated in m2 (fig. 6B—849_f).

Didelphys

D. marsupialis azarae. The predominance of styles b and

c is indicated in two specimens in all three molars, but especi-
ally inmrandmz2. In m2

and m3 is shown the pres-

ie nV, Vor Wee (~~ ence of twostyles posterior
to c, probably indicating

sis Wawa un av a division of style c2. In

m3 the intermediate bi is

C Se ee Cay indicated (fig. 6C—84.2.8.

, ; 25). This element was
Fig. 6. Stylar Cusps in Chironectes identified in several speci-
and Didelphys ; :

mens. It is sometimes pre-

sent in the deciduous premolar, which has a molariform

pattern.
GENERAL SUMMARY

In comparing the characters of the stylar cusps in an
extended series of specimens such as indicated above, three

features become apparent. In the first place, as compara-
[158]

OF THE DIDELPHYIDAE 13

tively small and subsidiary structures in the molar crown
are certain to exhibit signs of variation, they are surprisingly
constant in their relations. Secondly, they show throughout
the family indications of a general type, best seen in Peramys
and Marmosa. ‘Thirdly, even in such a poorly differentiated
radiation as that represented by the modern Didelphyidae,
they show signs of adaptive change. The variations which
appear seem to be for the most part significant. The general
type indicated is one in which there are three main elements
and three more subsidiary ones which I have designated
respectively as a, b, c, and bi, c1, c2. If the modern Didel-
phyidae reflect in the upper molar patterns the primary tri-
tuberculate type these main elements are the ones for which
homologies must be sought. The general type indicated seems
to conform with that described for Peratherr1uwm, but whether
this correspondence would be confirmed on examination of
other specimens of Peratherium it is not possible to say. That
the stylar elements are worthy of consideration in estimating
adaptive changes in the molars, or relative specialization, is
indicated by Sinclair’s* studies on the Santa Cruz marsupials,
and my own on the Australian forms. Finally, considering
the stylar elements of the modern Didelphyidae as structures
not wholly conservative but showing signs of adaptive change
and in comparing the family with other, supposedly primitive,
forms, the characters presented by Marmosa and Peramys
should be consulted rather than those of the larger specialized
forms such as Didelphys.

* Sinclair, W. J., The Marsupial Fauna of the Santa Cruz Beds. (Proce.
Amer. Phil. Soc., vol. 49, pp. 73-81, 1905.)

[159]

Peat A et ea
i, vi ;

«UNIVERSITY OF TORONTO
STUDIES

BIOLOGICAL SERIES

No. 6: ON POLYSTELY IN ROOTS OF
ORCHIDACEAE, BY J. H. WHITE

THE UNIVERSITY LIBRARY, 1907: PUBLISHED BY
THE LIBRARIAN

qs

‘¢ | Hf
\ Y if
RUAN he
COMMITTEE OF MANAGEMENT

4 ; ats "PROFESSOR W. H. ELLIS, M. i ee

Chairman: PRopesson Maurice Hutton, M. A,, LLD., fr
Urea - Acting-President of the University. tf ii

"PROFESSOR W. j. ALEXANDER, Ph. D.

~ PROFESSOR A. KIRSCHMANN, Ph. D.
_ PROFESSOR ae J. MACKENZIE, B. AL

a
‘ j
¥
ry
a
, }
i
4
1 @
oe ee ;
~ 4 %
j " *
i as
#)
h
} y
AR ‘ Pe
se
a
+ A
. ’
‘
‘
Tr.
+) 1 .
‘ + i

ON POLYSTELY IN ROOTS OF
ORCHIDACEAE

BY

PO WHE BS Bo

PREFATORY NOTE.

The work embodied in this article was done by Mr. J. H.
White, while pursuing graduate studies in Botany in the

University of Toronto, during the year 1906-07.

J... Pau

Lecturer in Botany.

ON POLYSTELY IN ROOTS OF ORCHIDACEAE.

In his essay on polystely Van Tieghem (’86) defined poly-
stely as follows: ‘Les faisceaux conducteurs peuvent étre
groupés en plusieurs cercles autour d’autant d’axes diversement

BIOLOGICAL SERIES No. 6
ERRATA

Page 7 [169] line 23 should read:

“ Split off from the large stele, itself underwent division. The”
Page 10 [172] line 9, for “monostele” read “ stele.”
Page 11 [173] line 2, after “root” insert “in H. hyperborea.”
Page 12 [174] line 14, after “H. orbiculata” insert “and H., blephariglottis.”
Page 14 [176] line 10, after “H. orbiculata” insert “and H. blephariglottis.”
Page 15 [177] line 3-4, omit the word “extra-stelar,” and line 21, after

“placed” insert “by these botanists.”

Page 18 [180], third reference from the bottom (Strasburger, Lehrbuch
der Botanik) for ’03 read ’04.

Plates I & II, heading, for “ Biological Series Neo. 7
Series No. 6.”

”

read “ Biological

aw awe CHIe ALLCIILIVIL OL SEVeFal Investi-

gators since Van Tieghem’s time, with the result that it has
been disproved. Leclerc du Sablon (790) has shown that the

numerous strands in the stem of Pteris do

not originate by

division of a monostele. Likewise, Gwynne-Vaughan (97)
has disproved Van Tieghem’s statement in regard to Primula.
Jeffrey (:00, :02), by an exhaustive study, has shown that
in none of the so-called polystelic Pteridophyta, nor in the few

[165]

: iy 7. sn LOR tps St OU ee:

cen RE Ry oR be RE bs hicte as

Ww: ie Y eS pete tist i eg om i
kn echt ye ‘ibe eieay fut eh a) |
i -" “a fee com, ale toyota ge oh at
ae Ae, sae) Bis Bs ele Aa ithe a th
: . “hemi eA WON) Ey Caio Re i) “pyaar a AAR ay wate ss
| * “al vabginauiict AM er ae inning AV? eis rand cs
ei Faria. Beye ays ns see BOW AE pee? ie ae TRS
; 2 ‘.) | et tax? ae dt oh fh paged Ohanety’® ef a ry
a nah teeta tr x hindu ea Say Me oui re ee: Dak (ts
Bot ae OR GHRED 1485
oS bandyad ou wy) guilds dl if
a bing of ene "

ON POLYSTELY IN ROOTS OF ORCHIDACEAE.

In his essay on polystely Van Tieghem (’86) defined poly-
stely as follows: “Les faisceaux conducteurs peuvent étre
groupés en plusieurs cercles autour d’autant d’axes diversement
disposés, de maniére 4 constituer tout autant de cylindres cen-
traux distincts, ayant chacun sa moelle, ses rayons médullaires,
son pericycle et son endoderme, tous reliés et enveloppés par une
écorce commune. . . . La disposition de l’appareil con-
ducteur est polystélique.”” This condition he conceived to be
derived by repeated division of the monostele. He based this
view on his conception of the anatomy of the stems of Primula,
Gunnera, etc., and extended his conclusions to the stems of most
vascular cryptogams. As regards polystely in roots, he
described but one example—some of the roots in certain
Lycopodinez.

Van Tieghem mentioned the “ tubercules”” of Orchis, Ophrys,
etc., which show numerous steles, only to state that they merely
simulate polystely. “Les tubercules des Orchis, Ophrys, etc.,
possédent; comme on sait, un plus ou moins grand nombre de
stéles distinctes dans une écorce commune, mais toutes ces
stéles s’attachent independamment, quoique en des points trés
voisins, sur le rameau qui porte le tubercule; elles ne dérivent
pas l’une de l’autre par voie de division. Ce tubercule est donc
constitué par un faisceau de racines concrescentes et non par une
racine polystélique. C’est un des exemples qui montrent le
mieux combien il est nécessaire de dégager la polystélique vraie
des illusions produites par la concrescence.”’

Polystely in stems has received the attention of several investi-
gators since Van Tieghem’s time, with the result that it has
been disproved. Leclerc du Sablon (’90) has shown that the
numerous strands in the stem of Pteris do not originate by
division of a monostele. Likewise, Gwynne-Vaughan (’97)
has disproved Van Tieghem’s statement in regard to Primula.
Jeffrey (:00, :02), by an exhaustive study, has shown that
in none of the so-called polystelic Pteridophyta, nor in the few

[165]

4 WHITE: ON POLYSTELY IN ROOTS

cases among the Angiosperms, is this condition due to bifurca-
tion of a single stele. Instead, it has been proved that the
central cylinder is always a monostele, and that the gaps caused
by the departure of the foliar and ramular traces from a medul-
lated monostele furnish the explanation of the so-called poly-
stelic type.

Polystely in roots has likewise received some attention, and
it is said to exist in the tuberous or abnormal roots of the
Leguminosz, Cycadacez, and Palmaceze. Of these, the roots
of the Palmacez have been most carefully examined, notably by
Cormack ('96) and Drabble (:04). Cormack discovered
what he considered to be polystely in this group, and his
observations have been corroborated by Drabble, and summar-
ized as follows: ‘Cormack grouped the arrangements met with
under several headings, commencing with the normal type
possessing a complete endodermis surrounding a central fibrous
or sclerenchymatous ring in which lie the xylem and phloem
elements. Passing through the condition in which the cylinder
is still complete, but longitudinally lobed, the endodermis dipping
into the depressions, but being quite continuous, he came to
those roots in which the central cylinder is composed of a series
of independently running strands of fibrous tissue presenting
the form of arcs of circles in transverse section; between these
the endodermis dips in and becomes discontinuous; and, finally,
he described the case of Areca, where one or more of these
independent strands presents a complete radially symmetrical
cylinder or ‘stele,’ as he terms it, round which the endodermis
is complete.” Drabble points out that the “ polystelic ” condi-
tion is restricted to the proximal portion of the roots; the distal
is always that of a normal monostelic root. Whatever inter-
pretation may be placed upon these observations, it will be seen
that the phenomena to be described in the roots of the Orchi-
daceze differ in some respects from those in the Palmacez.

Since Van Tieghem’s time no comparative investigations of
the multistelic roots of the Orchidaceze have been made to deter-
mine the homologies of their stelar system, and all later refer-
ences agree with his views.

Strasburger (:04), in his text-book of botany, says:
“ Einen eigenen morphologischen Aufbau zeigen die Knollen
der Orchideen. Sie werden der Hauptmasse nach aus fleischig

[166]

OF ORCHIDACEAE 5

angeschwollenen, nicht von einander gesonderten Wurzeln
gebildet, oben schliessen sie aber mit einer Stammknospen ab ”
(p. 40, also p. 435). Schoute (:02) says: “ Wenn man
den Namen Polystelie dennoch beibehalten will, so kann man
diesen anwenden auf diejenigen Falle wie die Orchideenwurzel-
knollen, wo durch Verwachsen mehrerer Wurzeln ein einheit-
licher Korper entsteht, der wirklich mehrere Stelen fihrt.
Bekanntlich bezeichnet aber gerade Van Tieghem diesen Fall
nicht als Polystelie, weil er hier durch Verwachsen entstanden
fet) (ps) 256):

These views, however, have not seemed conclusive, and the
present investigation was begun with a view to determine, first,
the structure of the roots of terrestrial Orchidaceze, and sec-
ondly, the interpretation of their structure in terms of modern
stelar hypotheses.

The roots of these Orchids are outgrowths of an adventitious
shoot—a fact due to the habit of vegetative reproduction char-
acteristic of these plants. Each year one or more stolons are
developed, each of which produces a new bud. The old plant
dies away, and the adventitious shoot perpetuates the species the
following season. In connection with this method of reproduc-
tion many of these Orchidacee show a differentiation of the
root system into a primary tuberous root—part of the so-called
“tuber” and having the characters of a storage root—and
_ somewhat slender lateral roots. This is especially true of the
Ophrydinz. It is in the roots of this class of Orchidacez that
there is an appearance of polystely, and it is with the anatomy
of these alone that we are here concerned.

THE LATERAL ROOTS

Among the lateral roots there are two types, the monostelic
and the polystelic. The monostelic type is characteristic of
Habenaria bracteata, R. Br. H. repens, Nutt, Orchis
spectabilis, L., and some of the. roots of H. orbiculata,
Torr. In these forms there is a single stele, consisting of
bast and wood elements arranged fairly radially around a cen-
tral pith, and surrounded by an endodermis. As a rule the bast
is not very well developed. The strands composing this stele
come from different parts of the vascular supply of the stem,

[167]

6 WHITE: ON POLYSTELY IN ROOTS

just as in the case of such monostelic genera as Goodyera,
Epipactis, etc. Plate I, figure 4, represents a transverse sec-
tion of the stele of H. bracteata. In the apical region there is
a well-formed root-cap, and just back of it the meristem from
which are differentiated the three embryonic regions, namely
dermatogen, periblem and plerome.

The polystelic (pseudo-polystelic, according to Van Tieghem)
type is not mentioned by Van Tieghem as occurring in the
lateral roots, but has been described by Holm (:04) in several
forms. Habenaria orbiculata furnishes examples of distelic
roots, and there are indications of distely in Orchis spectabilis.
In the latter species many cases were met with in which the
stele showed a distinct flattening, and while no case of two
steles was seen, it is very probable that this condition some-
times occurs. The lateral roots of H. orbiculata are always
monostelic at their base. Farther out, the stele in some of them
flattens, and still farther out, assumes a _ horse-shoe shape.
Plate I, figure 6, represents the stele of H. orbiculata when just
beginning to flatten, and Plate I, figure 7, the latter condition.
Farther out still, the stele is constricted into two equal parts
(Plate I, figure 8), or in other words, by an act of bifurcation,
the monostele gives rise to two steles, a phenomenon that Van
Tieghem apparently could not have observed, standing as it does
in direct opposition to his statement on this subject (Van
Tieghem and Douliot (’86). These strands continue side by
side throughout the greater part of the root. At the tip of the
root a typical root-cap is present, and just behind it a single
meristem giving rise to a single dermatogen, periblem, and
plerome. At some distance back the primary plerome strand
gives rise to two secondary plerome strands corresponding to
the two steles and to the ground tissue separating them. Each
of these steles is a typical root stele, possessing alternate bast
and wood elements, and each is enclosed by an endodermis. As
already stated, the roots of H. orbiculata are not always distelic,
those of the earlier generations and the smaller ones being
monostelic throughout.

Of the polystelic type none is of greater interest than the
roots of H. blephariglottis, Hook. In these the strands com-
posing the root stele bear the same relation to the vascular sup-
ply of the stem as in the case of any monostelic form, such as

[168]

OF ORCHIDACEAE Y

H. bracteata. Each of the lateral roots in most of the speci-
mens that were examined was found to be monostelic at its base
as in H. orbiculata. It is in many cases protostelic for some
distance out, and then becomes siphonostelic, owing to the appear-
ance of a pith and an internal endodermis (Plate I, figure 1),
a condition that is unique in the structure of root steles. Farther
out from the stem a gap appears in the stele, through which the
pith and cortex connect, and the internal and external endo-
dermes become continuous (Plate I, figure 2, a). As the
series is followed on in the same direction the gap enlarges,
giving the stele a horse-shoe shape. Successive acts of segmen-
tation then take place, steles being pinched off from the ends of
the shoe, more or less alternately (Plate I, figure 3), till finally
in place of the monostele there is a number of small steles of
approximately the same size regularly arranged in a circle.
Each of these small steles has the bast and wood elements
arranged in an irregularly radial manner (Plate II, figure 10).

The roots of this species that were examined showed only
minor points of difference among themselves, such as a smaller
or a larger number of steles, according to the size of the root,
or an at first scattered, irregular, instead of circular arrange-
ment of the individual steles. In some cases a stele, after being
roots of H. hyperborea, R. Br. (Plate II, figure 9). The
main fact held for most of them, namely, that the individual
steles were derived by a process of segmentation from a mono-
stele. Occasionally, however, in robust roots, the conditions at
the base approach those in H. hyperborea.

In more specialized forms the fibro-vascular supply comes off
from the stem in a small number of steles, as, for example, in the
roots of H. hyperborea, R. Br. (Plate II, figure 9). The
origin of the strands composing these steles is in no way differ-
ent, however, from what obtains in the three species already
described. The small number of steles seems to mean that the
division of the monostele is carried back to the stem connection.
Farther out these steles are to be found undergoing division,
resulting in a large number of small steles of approximately the
same size, just as in H. blephariglottis (Plate II, figure 10).

The tip in all these polystelic roots is clothed by a single root-
cap, just as in H. orbiculata, behind which is a small-celled
meristem from which all of the organs of the root originate.

[169]

8 WHITE: ON POLYSTELY IN ROOTS

It is worthy of note from a phylogenetic standpoint that in the
roots of plants of younger generations the number of steles is
very much smaller than in those of older generations. For
example, in the roots of plants of H. orbiculata belonging to very
early generations never more than one stele was present; in
H. hyperborea as few as two were found, and in H. blephari-
glottis the number was likewise small. The increase in the
number of steles in the roots of plants of each succeeding gen-
eration from the seed till the flowering plant is reached without
a proportionate increase in the number of steles entering these
roots, and without any indication of a disturbance of the condi-
tions described in the punctum vegetations of each, seems out
of accord with the theory of concrescence.

Examination of very young plants in a number of species.
where the roots were just beginning to grow out from the stem,
showed that but one meristematic mass was present, and in no
case was there any indication of a fusion of two or more roots.
In accordance with the view that the phylogeny of the race is
recapitulated in the ontogeny of the individual, one might have
expected to find, in some cases at least, discrete meristematic
areas corresponding to component roots, if the concrescence
theory were correct; but one for each root is the invariable rule.

The other polystelic species examined were H. clavellata,
R. Br., H. psycodes, Gray, H. virescens, Spreng, H. lacera,
R. Br., H. obtusata, Rich., H. leucophaea, Gray, and H. Hooker-
tana, A. Gray. The structure of the roots of these species
agrees very closely with that just described for H. hyperborea.

c

THE PRIMARY ROOT OF THE “ TUBER”

As already stated, vegetative reproduction is characteristic of
terrestrial Orchidaceze, and in many of them the adventitious
shoot is tuberous. A typical “tuber” consists of a stolon or
rhizome bearing a lateral bud, from which in turn there is
developed an axial tuberous root, occupying the position of a
primary root. Among the “tubers” there is much variation
in the length of the stolon, and the direction of the axis of the
root. Thus in H. bracteata and H. orbiculata, for example, the
stolon is very short, whereas in H. clavellata, H. blephariglottis
and Orchis spectabitlis it is comparatively long. In most of

[170]

OF ORCHIDACEAE 9

these Orchids the primary root is vertical, but there are all
gradations to the horizontal position assumed in the case of
H. dilatata, Gray. There may be even some variation in the
same species. Thus in H. hyperborea there is considerable
variation in the length of the stolon, and it is noticeable that an
increase in the length of the stolon is accompanied by a greater
departure of the root from its usual vertical position.

The stolon possesses a single stele of cauline structure. There
is a continuous ring of xylem around a central pith, with the
bast externally placed, and the whole is surrounded by an endo-
dermis, that is, there is an ectophloic siphonostele (Plate I,
figure 5). This seems to agree with the primitive monocotyle-
donous type of cauline stele as defined by Chrysler (:04), and
Plowman (:06). This siphonostele goes to supply the young
shoot and its root.

Characteristically the primary axial root is polystelic, the
number of steles varying from as few as two in Orchis specta-
bilis to twenty or more in H. blephariglottis. It remains now
to be determined whether these are steles of a fundamentally
unit root, as has been shown to be the case in the lateral roots,
or monosteles of concrescent roots. Among the lateral roots,
as has been seen, there are both monostelic and polystelic forms,
with instances of distely occurring in H. orbiculata as a connect-
ing link. All of the lateral roots, except the most specialized
forms, are monostelic at their base, the root tips of all show but
one small-celled meristem (Plate II, figure 15), and the young
roots originate each from a single meristematic area. A care-
ful examination should reveal the extent to which these condi-
tions prevail in the axial roots.

In the transitional region between stem and axial root, in a
form such as Orchis spectabilis, the stolon siphonostele gives off
the vascular axis of the adventitious bud, and a pair of steles
which turn down into the tuberous root. Almost immediately
in some plants a third root stele is given off. After supplying
the bud the siphonostele at once becomes much attenuated,
curves down, and ends in a minute projection of meristematic
tissue on the surface, on the way giving off a single trace to a
leaf which sheathes the new shoot. The meristematic projec-
tion is the tip of the stolon, and hence the adventitious bud is
distinctly lateral. The three root steles do not form a mono-

[171]

10 WHITE: ON POLYSTELY IN ROOTS

stele at their base, though they leave the stolon at practically the
same level. They recall the condition of affairs in the base of
the lateral roots of H. hyperborea (Diagram 1). Much the
same phenomena obtain in the case of H. orbiculata (Diagrams
2-13). Here a pair of root steles is given off just as in Orchis,
but the third stele is given off at a lower plane. Usually this
third stele soon bifurcates, so that of the four steles present in
the axial root two, at least, are obviously derived from a pre-
existing monostele by an act of bifurcation.

In those forms which have acquired a diageotropic habit, as,
for example, H. clavellata, the origin of the root steles is very
diffuse—a phenomenon on which Van Tieghem based his main
argument in support of the concrescence theory. The stolon has
the same mode of origin as in the preceding forms, and
possesses a similar siphonostele. Serial sections show that this
siphonostele soon begins to give off steles. These are very
small and are usually given off right and left alternately, while
the remainder of the siphonostele continues on in the same
relative position. They slant downward and run along close
to the ventral surface, the whole of them forming an arch when
viewed in transverse section. These small steles are usually all
derived from the main stele; only very rarely does one of them
bifurcate after being given off. A marked difference is observ-
able in the character of the cortex in the dorsal and ventral
regions, in that the cells of the dorsal region are smaller and
more compact than those of the ventral. From the dorsal por-
tion of the siphonostele the stele to the new shoot is given off.
As in the forms already described, the stele of the stolon then
becomes abruptly attenuated, gives off one or two more root
steles, curves up, and ends in the stolon tip on the dorsal surface,
on its way giving off the leaf trace. It is important to note that
in spite of their very diffuse origin all of the steles given off
along the stolon proceed to the tip of the tuberous root, and end
in one meristematic area in the growing point, just as in all the
other species examined (Diagram 14).

On comparing this form with Orchis and H. orbiculata it is
noteworthy that there is no difference other than that the root
steles have a more diffuse origin—a phenomenon that is pro-
bably to be accounted for by the unusual position of the root.
That this is the case is well shown by H. hyperborea. As

[172]

OF ORCHIDACEAE II

already stated there is much variation in the length of the stolon
and the position of the root, the two being associated more or
less. In those cases where the stolon is extremely short and the
axial root accordingly perpendicular, no steles are given off
throughout the length of the stolon proper, and the origin of the
root steles is no more diffuse than in Orchis or H. orbiculata.
But in those instances where the plant puts forth a long stolon
and the axial root takes on a slanting position, steles are given
off as in H. clavellata (Diagrams 15-26). Hence, it seems evi-
dent that the diffuse origin of the root steles in forms like
H., clavellata is to be ascribed to the diageotropism of the root.
It might be stated here that in H. blephariglottis, where the posi-
tion of the root comes nearest, of the species examined, to that
of H. clavellata, the degree of diffuseness of the root steles is
intermediate between the condition in H. hyperborea and that in
HZ. clavellata.

Plants of H. hyperborea of a very early generation possess
another feature, however, of even greater import. For several
sections below the region of the stolon stele, that is, at the
upper end of the axial root, all of the vascular elements are
enclosed within one endodermis. This single root stele then
simultaneously breaks up into the two or three steles which are
present throughout the remainder of the axial root to its tip
(Diagram 27), a fact that amounts almost to a demonstration
that the axial root is a unit, and not the result of concrescence.

Thus there is a fairly complete series, in the case of the axial
or tuberous root, from the monostelic to the most extreme poly-
stelic type, which parallels closely what has been already
described for the lateral roots.

It is further to be noted that the tip of the axial root has a
single root-cap, and behind it the meristem, which is differen-
tiated as in the lateral roots. Figures 11, 12, 13, of Plate II,
are from serial sections of the tip of the axial tuberous root of
a plant of H. hyperborea of an early generation, in which there
is a primary plerome strand that is continuous proximally with
secondary plerome strands. Occasionally the tip is forked, as,
for example, in H. bracteata. This forking, however, occurs
only in the older roots, and seems to be of the nature of a split- .
ting or dichotomy, and bears no relation to the number of
steles. Apparently the root tip is first flattened, and a meri-

[173]

12 WHITE: ON POLYSTELY IN ROOTS

stematic area in the centre of the tip ceases growing, while the
isolated areas on either side continue their growth, which results
in the forking just mentioned. Drabble (:04) has figured a
similar forking in Kentia, Sp.

The same fact has been observed with regard to the tuberous
roots as was seen in the lateral roots, namely, that in those of
the earlier generations the number of steles is much smaller than
in those of the later generations. Thus it is not uncommon in
H. bracteata to find three steles in the axial root of one genera-
tion, and six steles in the corresponding root of the next gen-
eration. Again, in the tuberous root of H. hyperborea of a
young generation (Diagram 27) two steles were found, while
the next generation showed very many more, and in early gen-
erations of H. orbiculata there is but one. It is altogether prob-
able that in the earliest generations of all of them there is but
a single stele.

THE “ TUBER ”

Van Tieghem looked upon the “tuber” as a branch carrying
a “‘tubercule” with a greater or smaller number of distinct
roots, the steles of which attach themselves independently,
though at points very close together, on the branch. According
to him, the ‘‘ tubercule” is a bundle of concrescent monostelic
roots, and not a single polystelic root. Others have gone further
and included more in the concrescence. Thus, Germain de St.
Pierre (’55) includes in it stem and leaf portions as well as
roots, and he has lately been followed by Holm (:04). But
according to the foregoing observations on the primary root of
the “tuber ” taken in conjunction with those on the lateral roots,
the “tubercule ” is simply a swollen root which is usually poly-
stelic. One other fact has been overlooked in previous accounts
of the “tuber,’’ namely, that the stolon continues beyond the
origin of the adventitious shoot, and that the leaf ensheathing
the upper part of the “tuber” derives its vascular supply from
the distal portion of the stolon.

CONCLUSIONS

The “concrescence”’ theory has been stated already. The
grounds on which this view is based are the diffuse origin of the
steles of a tuberous root, and the forking of the root tips in

[174]

OF ORCHIDACEAE 13

forms such as H. bracteata and Orchis latifolia. But in the
light of the foregoing observations these grounds are scarcely
tenable.

The forking of tuberous roots is apparently a matter of acci-
dent. There is no indication of it in young roots prior to the
differentiation of the steles. Moreover, it bears no relation to
the steles. Forking is not characteristic of the more slender
roots, and it is never present in lateral roots, nor even in the
markedly multistelic roots of H. clavellata, though not of infre-
quent occurrence in swollen roots of several other species. The
forking is probably caused by purely mechanical forces, and in
view of the facts recorded of the earlier stages of development
can be regarded in no way as indicating concrescence.

It is apparent from former accounts that the theory of con-
crescence has been based on an examination of the tuberous roots
only. To understand the structure of these, however, in regard
to the origin of the steles, the conditions obtaining in the lateral
roots are of material assistance. Thus it has been seen that the
lateral roots of H. bracteata exhibit typical monostely, and the
other species may be looked upon as presenting deviations from
this type. Of these, H. orbiculata shows the least deviation,
only some of the roots being distelic, and these for the central
portion only of their length. The roots of H. blephariglottis
are characterized by a still greater deviation, in that they possess
more than two steles, but although this is the case they are char-
acteristically monostelic at the base, and hence can only be looked
upon as single roots. H. hyperborea departs from the type
more widely still, in that the roots are never monostelic at the
base. However, the fact that they possess but a few steles in
this region, and that the larger number farther out is derived
from these by segmentation, shows the tendency towards the
condition in H. blephariglotiis, and may be taken to indicate
that the division of the monostele has been carried back into
the stem. To cite an analogous case, the carrying back of the
cleavage of a fibro-vascular strand to its origin is seen in the leaf
traces of certain ferns. Primitively the foliar fibro-vascular
supply in the Filicales comprises a single strand, but in many of
the more specialized ferns this strand is divided, and commonly
the segments are attached separately to the stelar axis of the
stem. Indeed, the facts relative to the origin of the steles

[175]

14 WHITE: ON POLYSTELY IN ROOTS

in multistelic lateral roots point to the conclusion that they are
the result of the segmentation of a single stele, and are not indica-
tive of concrescence of several monostelic roots.

The steles of the multistelic axial or tuberous roots are much
more frequently attached separately to the cauline stele than is
the case in the lateral roots, but this phenomenon is not as uni-
versal as Van Tieghem thought. In fact, segmentation of a
root stele is by no means infrequent, and sometimes monosteles
exist at the base of axial roots in earlier generations. Thus, the
tuberous roots of early generations of H. orbiculata are mono-
stelic throughout, and what may be considered of greater sig-
nificance, corresponding monosteles in H. hyperborea divide after
leaving the stem, plainly fulfilling the conditions expressed by
Van Tieghem in his definition of polystely. Moreover, the
most striking examples of a diffuse origin of the steles occur
in species like H. clavellata, in which the axial roots are charac-
terized by an extreme diageotropic habit, and, as has been shown,
the degree of diffuseness is measured by the extent of the defec-
tion of the root from the vertical position. Hence, a phylo-
genetic series can be constructed, beginning with the condition
noted in early generations of H. orbiculata and H. hyperborea,
and terminating with the condition in H. clavellata or H. dila-
tata. Such a series harmonizes the phenomena in axial roots
with those in the lateral, rendering it quite certain that poly-
stely is the true explanation in the one as it is in the other.

But there yet remain other facts that are incompatible with
any theory of concrescence. A single root-cap clothes the tips
of all roots, and behind this root-cap there is a single small-
celled meristem, consisting of dermatogen, periblem, and plerome
initials. Moreover, the first indication of a root is the appear-
ance of a single meristematic area, and from this alone, without
concrescence with any other, the root develops.

The interpretation of the phenomena in polystelic roots in
terms of the modern stelar hypotheses, especially of those that
attach importance to the differentiated meristematic areas at the
growing point is a matter of some interest. The view of Han-
stein that the three meristematic areas, dermatogen, periblem
and plerome, in the growing point of vascular plants corre-
sponded to the epidermis, cortex, and stele of the mature organ
is well known. But it is very evident in these roots that a

[176]

OF ORCHIDACEAE 15

primary plerome strand is not wholly young stelar tissue, for on
tracing it back in Habenaria orbiculata, hyperborea, etc., we have
found that it passes into secondary pltrome strands and extra-
stelar ground tissue of the same kind, continuous with, and
indistinguishable from the cortical tissues. The secondary
plerome strands are transformed into steles, each, it may be, with
its pith and endodermis, while the ground tissue undergoes no
further change. That the plerome is not necessarily the undif-
ferentiated stele has already been proved for certain forms by
Schoute ( :02), Campbell (:05), and more recently by Conard
(:06). But nowhere is this so strikingly the case as in the
examples we have described.

In this connection the homologies of the pith and the tissues
occupying the axis in polystelic Orchid roots are particularly
interesting. There are two theories in regard to the character
of pith. One of these regards the pith as always stelar, while,
according to the other, the pith in many cases is an extrastelar
tissue. The first position has been maintained by Van Tieghem
(786) (except in the case of “gamostely’’), Strasburger
(91), Boodle (:01), Tansley and Lulham (:05), and
others. | Considerable stress has been placed on the corre-
spondence of Hanstein’s meristematic areas in the growing point
with the areas in the mature portions of the plant. The other
view, first proposed by Jeffrey, that the pith is cortical in its
homologies in some cases, has for its support the similarity of the
cells of the cortex and medulla, their continuity at the gaps, the
“dipping in” of the endodermis, certain facts of degeneration,
and the incompatibility in certain cases of the Hanstein con-
ception, as noted in the last paragraph. Jeffrey ( :00; 02)
described the “intrusion ”’ of the cortex into the stele above the
point of departure of the leaf traces in seedlings of the Pterop-
sida, and above the branch traces in the Lycopsida. Faull (:o1)
described the continuity of the cortex and pith through the ram-
ular gaps in his account of the Osmundacez. Chandler (:05)
has added further corroborative proof drawn from a study of
an extensive series of sporelings.

Further corroboration of the view that the pith may be extra-
stelar in some cases has been found by the author in his study
of the lateral roots of H. blephariglottis—a proof amounting to
a demonstration. Inthe majority of Orchid root steles in which
a “ pith”’ occurs there is no connection between the pith and the

[177]

16 WHITE: ON POLYSTELY IN ROOTS

extrastelar tissues. But in H. blephariglottis (Plate I, figure
1) there is an internal endodermis surrounding a pith of quite a
different character. ‘ On tracing the pith forward from the base
towards the root tip (Plate II, figures 2, 3) it has been seen
that it is continuous with the extrastelar tissue separating the
steles into which the monostele breaks up—a tissue that is
unquestionably homologous with the tissue in a similar position
in lateral roots of H. hyperborea, psycodes, etc. (which are
polystelic at their base), and with the tissue separating the two
steles in the lateral roots of H. orbiculata. The extrastelic char-
acter of this tissue is especially evident in the last, for the mono-
stele flattens, bends into a horse-shoe, incompletely enclosing a
part of the cortex, and then breaks into two steles of equal size
by a constriction or “intrusion” of the cortex from opposite
sides (Plate I, figure 6; Plate II, figures 7, 8).

In conclusion, polystely adds another specialized feature to the
many possessed by various members of this remarkably special-
ized family. It is difficult to say how this habit may have
arisen, but there is evidently some connection between polystely
and the tuberous character of the roots in which it occurs, for
it is to be noted that with the increasing size of polystelic roots
in succeeding generations there is an increased number of steles
in each. The phenomenon, as it exists here, plainly answers to
Van Tieghem’s definition of polystely. Finally, the conditions
existing in these roots must be reckoned with in any stelar
hypothesis, for it has been established according to the observa-
tions recorded in this paper that the middle and certain more
or less radially placed cells belonging to the primary plerome or
to the plerome initials directly give rise to extrastelar elements.

SUMMARY

I. There are two types of roots among the terrestrial Orchid-
aceee—the monostelic and the polystelic, in Van Tieghem’s sense
of the terms. In reference to the latter, the term “ concres-
cence’’ is inapplicable.

II. Polystely has been found in both the lateral and the
tuberous roots of H. orbiculata, H. blephariglottis, H. hyper-
borea, H. clavellata, H. lacera, H. psycodes, H. virescens, H.

[173]

OF ORCHIDACEAE 17

obtusata, H. leucophaea, H. Hookeriana; and in the tuberous
roots of H. bracteata and Orchis spectabilis.

III. In the basal portion of the monostele which is ordinarily
present in the lateral roots of H. blephariglottis a pith and an
internal endodermis may occur. The pith in such cases is
plainly extrastelar in its homologies.

IV. The plerome initial cells in all polystelic roots are dif-
ferentiated into stelar and extrastelar tissues.

The above investigation was carried on in the Biological
Laboratory of the University of Toronto under the direction of
Dr. J. H. Faull, by whom the subject was suggested, and to
whom I wish to express my deep obligations for much patient
criticism and advice throughout, as also for some of the mate-
rial. My thanks are due to Professor R. Ramsay Wright for
the facilities afforded in the laboratory.

[179]

LITERATURE.

Boodle, L. A., :o1.
On the anatomy of the Schizaeaceae. (Annals of Bot., vol. 15, pp. 359-
421.)
Campbell, D. H., : 05.
The structure and development of Mosses and Ferns, p. 465.
Chandler, S. E., : 05.
On the arrangement of the vascular strands in the seedlings of certain
leptosporangiate Ferns. (Annals of Bot., vol. 19, pp. 365-410, pl. 18-20.)
Chrysler, M. A., :04.
Development of the central cylinder of Araceae and Liliaceae. (Bot. Gaz.,
vol. 38, pp. 161-184, pl. 12-15.)
Conard, H. S., : 06.
Morphology of the Fern stem as illustrated by Dennstaedia punctilobula.
(Johns Hopkins University Circular, May, 1906, pp. 95-103.)
Cormack, B. G., ’96.
On polystelic roots of certain Palms. (Trans. of Linn. Soc. of London, pp.
275-287, pl. 19, 20.)
Drabble, E., : 04.
On the anatomy of the roots of Palms. (Trans. of Linn. Soc., pp. 427-490,
pl. 48-51.)
Faull, J. H., : ox.
Anatomy of the Osmundaceae. (Bot. Gaz., vol. 32, pp. 381-420, pl. 14-17.)
Gwynne-Vaughan, D. T., ’97.
On polystely in the genus Primula. (Annals of Bot., vol. 11, pp. 307-325,
pl. 14.)
Holm, Th., : 04.
Root structure of North American Orchideae. (Am. Jour. of Sci., vol. 18,
Pp. 197-212.)
Jeffrey, E. C., : 00.
Morphology of the central cylinder in the Angiosperms. (Trans. Can.
Inst., vol. 6, pp. 1-40, pl. 7-11.)
Jeffrey, E. C., : 02.
The structure and development of the stem in the Pteridophyta and
Gymnosperms. (Phil. Trans. B. cxcv.)
Plowman, A. B., : 06.
Comparative anatomy and phylogeny of the Cyperaceae. (Annals of Bot
vol. 20, pp. I-33, pl. I, 2.)
Sablon, Leclerc du, ’90.
Recherches anatomiques sur la tige des fougéres. (Ann. Sci. Nat. Bot.,
vol. 7, pp. 1-16, pl. I, 2.)
Schoute, J. €:, : 02.
Die Stelar-Theorie. Groningen.
St. Pierre, Germain de, ’55.
Bull. Soc. de France, vol. 2, p. 659.
Strasburger, E., ’gI.
Histologische Beitrage, III.
Strasburger, E., ’03.
Lehrbuch der Botanik. Jena.
Tansley, A. G., and Lulham, R. B. J., : 05.
A study of the vascular system of Matonia pectinata. (Annals of Bot.,
vol. 19, pp. 475-519, pl. 31-33.)
Van Tieghem and Douliot, 86.
Sur la polystélie. (Ann. Sci. Nat. Bot., vol. 7, pp. 275-322, pl. 13-15.)

EXPLANATION OF PLATES.

PLATE I.

Fig. 1. A. blephariglottis. Transverse section through base of lateral root,
showing siphonostele with internal endodermis.

Fig. 2. Same stele farther out.
Fig. 3. Same stele still nearer the root tip.

Fig. 4. A. bracteata. Transverse section of lateral root, showing typical
monostele.

Fig. 5. A. clavellata. Transverse section of stolon, showing ectophloic
siphonostele.

Fig. 6. A. orbiculata. Transverse section of lateral root, showing monostele
beginning to flatten.

PLATE II.

Fig. 7. The same farther out, stele horse-shoe-shaped.
Fig. 8. The same still farther out, after bifurcation.

Fig. 9. H. hyperborea. Transverse section through base of lateral root,
showing the few large steles.

Fig. 10. A. hyperborea. Transverse section of lateral root, showing typical
polystelic condition.

Figs. 11, 12,13. A. hypferborea. Transverse serial sections of tip of axia
root of plant of an early generation, showing differentiation into
secondary plerome areas.

Fig. 14. A. dlephariglotits. Transverse section of axial root near tip.
Fig. 15. 4. hyperborea. Longitudinal section of tip of lateral root.

DIAGRAM I.

Orchts spectabilis. Diagrammatic representation of vascular system, show-
ing stolon stele supplying young shoot and its axial root ; it ends in the meri-
stematic stolon tip which is sheathed by a leaf containing a single stelar trace.
There are three root steles: sh. s., stele of adventitious shoot; s.s., stele of
stolon ; %, stele of axial root (%, on the opposite side and corresponding to, is
not figured) ; 7:, stele of axial root; s. 7. ., stele of sheathing leaf; s. #., tip of
stolon ; a. ~., axial root; s., stolon; s./., sheathing leaf; s%., adventitious
shoot.

DIAGRAMS 2-13.

H. orbiculata. Regional sections of the transition region between shoot
and root of a ‘‘tuber.” A reconstruction would give a figure like Diagram 1,
except that the stolon tip is relatively higher up, and there are four root steles
instead of three. S%. s., stele of adventitious shoot ; s. s., stele of stolon; »
2, Ys, %1, Steles of axial root; s. 7. #., stele of sheathing leaf; s. ¢., tip of stolon ;
a. r., axial root.

20 EXPLANATION OF PLATES.

DIAGRAM 14.

H1. clavellata. Diagrammatic representation of vascular system. Note the
diffuse origin of the root steles due to the diageotropism of the root. .S%. s.,
stele of adventitious shoot; s. s., stele of stolon ; %, 7, 7%, 7%) %, steles of axial
root; s. /. z., stele of sheathing leaf; s. 4, tip of stolon; a@. 7., axial root; s.,
stolon ; s. /., sheathing leaf; s%., adventitious shoot.

DIAGRAMS 15-26.

H. hyperborea. Regional sections of the transition region between shoot
and root of a “tuber.” Note that the stolon stele gives off four steles along its
length before the region of the root proper is reached. The diagram represents
a condition intermediate between Diagrams 2-13 and Diagram 14. Sz. s., stele
of adventitious shoot ; s. s., stele of stolon ; 2 7., lateral root; 7%, %, %2, %4y %y
rs, steles of axial root ; s. 7. ¢., stele of sheathing leaf; s. 2, tip of stolon; a. ~,
axial root corresponding to shoot.

DIAGRAM 27.

H. hyperborea—plant of an early generation. Diagrammatic representation
of vascular system, showing stolon stele supplying young shoot and its axial
root; it ends in the meristematic stolon tip which is sheathed by a leaf contain-
ing a single stelar trace. Note that the root stele is monostelic at its base, and
jater divides into two steles. Sh. s., stele of adventitious shoot ; s. s., stele of
stolon ; , % steles of axial root; s. /. ¢., stele of sheathing leaf; s. 4, tip of
stolon; a. 7., axial root; s., stolon; s. 7., sheathing leaf; s%., adventitious
shoot ; »., monostelic axial root, which lower down bifurcates into %, and 7.

UNIVERSITY OF TORONTO STUDIES
BIOLOGICAL SERIES NO. 7

v.09 O-a Nts ACL
ie a
ae Cos ah Sy =". <f
AS gs wy were was 39 Ne

OS

aes

EP

HELIOTYPE CO.

UNIVERSITY OF TORONTO STUDIES PLATE Il
BIOLOGICAL SERIES NO, 7

shee A> Cy

aR as
O
4.

a ) a ey Ne
— JON ‘ { h y ..—~ /
a a ee | hy | @
ie a) am ES ; }
1 y

LOT

x
apres oeal ti?

-
1a) fs
ators
Cy ve

Wd
asia

avsoe
a
“i
cae Tee
* fer
a

coos
eae
tH ee

PB aie
Seana J)
re AS (?

|

/ 1)

4)
kak

HELIOTYPE CO.

i

*

UNIVERSITY OF TORONTO STUDIES
BIOLOGICAL SERIES, NO. 6

Sh

SS

eS

Bette: DS

Wena =

SSA

SG

Gam

eee bY) Te

aes SS

yore

Yanga -
Pea ie Oe a eR IS Se Se

SS

SIENA ce

ue
F
S
Te
a

cy a
oh
\ Es
\ in
\ PY
NY a
TENE 2 : a
Sis SEUKS Ris PUTER Grn ag Ti gray een se
SERRE SSeS CS Saag Mi jos HSER
Biase
Seowe
,
, 7
y
‘ s
z Le
q. y ;
* 4, a
,
‘ A ,
a e
’

L4

DIAGRAMS I, 14, 27.

Ta

in

Pi
«
fi
|
L-
7 1
. i
iy
Ais
'
4 ate :
ree
vv
'
a1 “a
Pets So
| _
ei
PRE, F) os
- a ;

.
4
4 — -_ ;
=
: }
-
e -
is)
one se | us SAH OAH

\.

UNIVERSITY OF TORONTO STUDIES
BIOLOGICAL SERIES, NO. 6

DIAGRAMS 2-13.

as

UNIVERSITY OF TORONTO STUDIES
BIOLOGICAL SERIES, NO. 6

a

DIAGRAMS 15-26.

: _ a
fs i
Ne “—_ 4
= \ - i i '
a 4
é ; ; pay
& oT O a
7 ¥
: 6
if _ -— ~ —
Oy 7 i |
= a rz
r
f =
‘
oe
> —
a? ye t -
e)
Ho y 7
a i :
7 Z P ’
f
: J
:
—) . re
“iy ; 23 e
; =a
‘ - :
54
i r ” . /
'
i
;
i
if
F
=
Pe ‘ VA TELA fe
¥ _| ZUSSth AMAVALE
ra
ve ~
'

UNIVERSITY OF TORONTO
STUDIES

BIOLOGICAL SERIES

No. 7: AN EARLY ANADIDYMUS OF THE CHICK
spy R. RAMSAY WRIGHT

(REPRINTED FROM TRANSACTIONS OF THE ROYAL Society oF CANADA, 2ND SERIES, VOL. XJ.)

THE UNIVERSITY LIBRARY: PUBLISHED BY
THE LIBRARIAN, 1907

og

University of Toronto Studies —
COMMITTEE OF MANAGEMENT

Chairman: Maurice Hotton, M.A., LL.D.,
Acting President of the University, ‘

PROFESSOR W. J. ALEXANDER, Pu.D.
Proressor W. H. E us, M.A., M.B.
PROFESSOR A. KirscHMaNN, Pxu.D.
ProFessor J. J. MACKENZIE, B.A.
ProFessor R. Ramsay Wricut, M.A., B.Sc., LL.D.
PROFESSOR GEORGE M. Wrong, M.A. |

ee, Editor: H. H. Lancron, M.A.,
Librarian of the University

Sxcrion IV., 1906. [21] Trans. R. §. C.

Ill—An Early Anadidymus of the Chick.
By Proressor RAMsay WRIGHT,
Biological Department, University of Toronto.

Read May 28rd, 1906.

The embryo which is described in the following pages was prepared
and sectioned in June, 1995, for class purposes but its abnormality
did not attract attention until it was brought into the laboratory. I
am, therefore, unable to figure the surface view, and so far have not
had leisure to model out its most interesting features.

The series contains 200 sections of 15 microns in thickness, cor-
responding to a length of 3 mm. in the hardened conditiom The egg
had been in the incubator for 24 hours, but, 10 somites having been
observed, it was marked as practically equivalent in age to Duval’s
embryo of 29 hours (No. 1, Fig. 89 and Pl. XVI).

It was noted that the incubator was running at a temperature
somewhat higher than the normal, which may account not only for its
more rapid development, but also for its abnormality, as may be
inferred from Dareste (No. 2, page 121).

Hertwig (No. 3:—Vol. I, p. 993) and others have remarked on the
rarity of cases of Anadidymus in Sauropsida in comparison with the
Ichthyopsida. This case is of particular interest, because, unlike Hoff-
mann’s (No. 4, page 40) there appears to be no indication of a double
primitive streak, and, therefore, it is to be placed in the same category
with Dareste’s embryo (No. 2, Plate 16, Figs. 5 and 6), and possibly
that of Mitrophanow (whose paper I have not been able to consult)
cited by Kaestner (No. 5, page 88). The occurrence of such a case
does not, in my opinion, invalidate the argument of Kaestner that
all such cases are primitively double (No. 6, page 141), because it
depends entirely upon the degree, locality and method of the inter-
ference of the two components, whether an organ shall appear double
or single. My figure of section 131 (Fig. 13) would not be suspected
to come from an embryo otherwise than normal, while the inspection
of section 126 (Fig. 12)) at once shows that each half of it in reality
belongs to a different embryo. From this point, the interference
caudad has been more complete than cephalad, so that in the backward
growth of the primitive streak region (cf. Hertwig, No. 2, pp. 895
and 896) the embryo appears to be single.

22 ROYAL SOCIETY OF CANADA

Attention must be called to the contrast in the method of inter-
ference in the head-region of my embryo and that in Kaestner’s (No. 6,
Taf. VII) where the ventral surfaces have interfered more than the
dorsal, the result being a single heart and a double brain, instead of
a double heart and a single brain (cf. my figure 9). The plane of
interference becomes caudad more and more truly sagittal, so that, the
chorde, at first widely divergent (Fig. 10), eventually fuse, (Fig. 13).

I now proceed to the description of the various systems of organs.

Nervous SYSTEM.

As a starting-point, I select section 12 (Fig. 5) through the region
of the optic vesicles. It is easy to understand how the condition here
pictured is arrived at if we proceed from the normal state as seen in
Duval’s Figs. 253 and 254. The two embryos have been inclined with
their dorsal surfaces towards each other, and have interfered in such
away thatethe right and left lips of the neural groove of the one, have
fused with the right and left lips of that of the other. In this way,.
no room is left for the complete development of the “‘ median” optic
vesicles which, consequently, are very minute (ov’). The points of
fusion are still noticeable and it is obvious that that of the left and
right lips of the right and left components respectively (which now form
the floor of the composite neural canal), is less complete, in such a way
that some mesoderm cells have intruded into the neural canal at this
point. The double character of the neural canal is brought strongly
out by the two infundibula which diverge laterally towards the two
blind foregut ends (ph.) beneath which the slightly thickened patches
of ectoderm already indicate the hypophyses.

It is less easy to interpret the preceding sections (Figs. 1 to 4),
but if two components such as are represented in Duval’s Fig. 252
have interfered in such a way as materially to reduce in size the con-
tiguous halves, then it becomes apparent that the convex floor of the
composite neural canal in figure 4 is formed of the left and right
brain-halves of the right and left components which have fused in
the region of their dorsal neural sutures, while their ventral sutures
are still widely separated. Still further forward (Fig. 3) these brain-
halves are fused so that the most anterior end of the neural canal
(Figs 1 and 2) is formed of the lateral brain-halves only of the two
components. It is noticeable that the separation of the brain from the
ectoderm has apparently taken place sooner than is normal (No. 3,
Vol. 2, page 252).

In the diencephalic region (Fig. 6) the brain is much compressed
from side to side, but it soon widens out into the mid-brain (Fig. 7).

[wricHT] AN EARLY ANADIDYMUS OF THE CHICK 23

In the trigeminal region of the hind-brain the neural canal is open for
some thirteen sections, but before the auditory region is reached it is
again closed as far as section 84, near which point (Fig. 11) there
is again a failure to close for a few sections; thereafter, however, the
canal is closed as far as section 126, Fig. 12, behind which point the
groove is, at first narrowly, and then widely, open.

In section 160 (Fig. 16) the fusion of the ventral wall of the
neural groove and the notochord begins and is continued in the follow-
ing sections (Figs. 17-20), the complete fusion of the ectoderm, chorda,
mesoderm and entoderm being attained at the 175th section (Fig. 20).
Beyond this point we can hardly speak of a neural groove; the 181st
section (Fig. 21), indeed, shows an unsymmetrical fissure which is not
uncommon in the primitive groove of normal embryos, and by section
190 all traces of the primitive streak have disappeared and the germinal
area presents a normal appearance (Fig. 23). The comparison of my
Figures 15-22 with those of Hertwig (1 c., Figs. 536-545, page 891)
shows that there is little difference except in the less amount of closure
of the neural canal, and without an inspection of sections further
forward, it would be impossible to detect any symptom of “ duplicitas.”

NoOTOCHORD.

The conduct of the two notochords has already been sufficiently
referred to in the hinder region; it only remains to call attention
to their gradual increase in size from their first appearance in section
9 (immediately behind figure 5) till their fusion in section 131, also
to their gradual convergence to this point.

MESODERM,

As already remarked there are ten somites, and this is the case
with the “median” series of fused somites which lie exactly in the
same plane as the lateral ones. Of the “median” series, the seven
posterior are better demarcated than those further forward, and are
sometimes notched on their ventral surface. The rudiments of the
Wolffian body may be seen in the region represented in Figs. 12 and 13.

VASCULAR SYSTEM.

A convenient starting-point for the description of the vascular
system is the region depicted in Fig. 10 (section 67), where the vitel-
line veins are perfectly normal, and the only thing that arrests atten-
tion is the “median” descending aorta. Fig. 9 shows that the vitel-
jine veins have not become fused into a single heart as in a normal

24 ROYAL SOCIETY OF CANADA

embryo. Their endothelial tubes remain independent throughout, but
the splanchnic mesoderm? does not at first dip in very far dorsad so
as to furnish an independent wall for each heart. Further forward,
however, it does so (Fig. 8), and eventually the two bulbs of the
heart are widely separated and enclose between them a portion of the
common ceelome (Fig. 7). But the two heart-tubes as seen in Fig. 9
do not contract gradually into the condition seen in Fig. 8; on the
contrary, there is a marked constriction at the opening of each heart
into its bulbus, beyond which a ventricular cul-de-sac extends cephalad
for a few sections on each side.

The picture presented by Fig. 6 is best calculated to show the
anterior duplicity of the vascular system, because when each bulbus
approaches the stomatodeum it divides into two ventral aorte. Of
these the lateral aortz alone form arches up the sides of the pharynx,
for the median ones first anastomose below the pharynx, then subdivide
into four small vessels which bend round its anterior surface, and
finally open into the large vascular space represented in Fig. 5, situ-
ated between its anterior diverticula. Tracing this space backwards
dorsad of the composite pharynx, we first find four vessels similar to
those referred to above, which soon, however, fuse into the “ median ”
dorsal aorta. This retains its size until we reach the segmented region
of the embryo, in which it tends to be obliterated opposite the somites
and to expand again intersomitically. The “lateral” dorsal aorte con-
duct themselves as in a normal embryo, and the same may be said of
the veins as far as they are developed._

EnToprRrMiIc TRACY.

Proceeding cephalad from Fig. 11 in which the median ridge
formed of the median row of somites alone distinguishes this from
the entoderm of a normal embryo we find nothing remarkable until
about midway between Figs. 8 and 9, there the lateral pouches of the
pharynx reach a little nearer the ectoderm in the region of the first
ill-clefts, but a few sections further forward (Figs. 6 and 5) the two
stamatodsa at once arrest attention, as do the two anterior diverticula
ecrresponding to the pouches of Seesel of normal embryos.

I venture to enter a mild protest against Professor Kaestner’s note (No.
6, p. 128) on the usage of the words somatopleure and spanchnopleure.
Surely, if it is desirable to have mononyms for “somatic mesoblast,” and

“splanchnic mesoblast,”’ it would be easy enough to form them instead of
using terms which were invented and are constantly used to designate

something else. If the language of anatomists knows only one meaning for
tAevpa’ that of zoologists is not so restricted. A Pleuronectid does not swim on its
«se ”

pleura !

Trans. R. 8, C., Sec. IV., 1906.

fwricHT]

AN EARLY ANADIDYMUS OF THE CHICK

Ente

EOE

_
e " =
RR

(p. 24]

{ WRIGHT] AN EARLY ANADIDYMUS OF THE CHICK 28

In conclusion, in spite of the apparent posterior simplicity of this
embryo I am of the opinion that it can best be explained by assuming
a double gastrulation at points very close to each other on the surface
of the embryonic area.

LITERATURE CITED.

I have thought it unnecessary to cite all the papers consulted.
-Hertwig (No. 3) and Kaestner (No. 6) give a full list of papers to
some of which, unfortunately, I have not had access.

No. 1. Duval—Atlas d’Embryologie.

No. 2. Dareste—-Production des Monstruosités.

No. 3. Hertwig—Handbuch der Entwickelungslehre.
No. 4. Hoffmann—Arch. mikr. Anat. XLI.

No. 5. Kaestner, Arch. Anat. Phys., 798.

No. 6. Kaestner, Arch. Anat. Phys., 702.

26 ROYAL SOCIETY OF CANADA
EXPLANATIONS OF THE FIGURES ON PLATE.

The sections were projected and carefully outlined on the drawing paper
by means of the Zeiss Epidiascope and 20 mm. micro-planar, at such dis-
tances as to give an enlargement of 102 for figures 1 to 9, and 116 for figures
LOetowza:

Subsequently, the drawings, which were made by Mr. J. R. G. Murray,
student in biology, University of Toronto, were reduced rather more than
one-third, so that the magnification is respectively 63 and 72.

Figs. 1-4,— Nos. 4, 5, 6, and 8, of the series, through the fore-
brain.

Fig. 5— No. 12, through the anterior blind ends—ph.— of the
pharynx. Ov. and ov’ the right and left optic vesicles of the right
component.

Fig. 6— No. 19, Brain fad stomatodea of both components
and the diencephalic region; round the composite pharynx are grouped
eight arteries; two ventral, and two dorsal aorte on each side.

Fig. 7.—No. 33, through the mesencephalon. Ventrad of the
pharynx are the two aortic bulbs; dorsad, the median dorsal aorte have
united into a single vessel; re, ectodermic recess under the head.

Figs. 8, 9, and 10.---Nos. 47, 55, and 67, respectively, through the
fifth, seventh and eighth, and ninth nerves.

Fig. 11,— No. 80, through the second intersomite. The median
dorsal aorte have given place to a mass of mesoderm.

Fig. 12,— No. 126, behind the last somite. The chorde are gain-
ing in size, and the mesodermic mass diminishing. The rudiment of
the Wolffian body is seen in this and in Fig. 13.

Fig. 13,— No. 131, the chorde have fused.

Figs. 14 and 15,— Nos. 150 and 154, the chorda and the wall of
the neural groove gain in size.

Fig. 16,— No. 160, the beginning of the fusion between the floor
of the neural groove and chorda.

Figs. 17, 18, 19 and 20,— Nos. 164, 168, 171 and 175, respectively,
show the progressive fusion of the neural wall, chorda, mesoderm and
entoderm.

Figs. 21 and 22,— Nos. 181 and 186, are through the hinder end
of the primitive streak. The former shows traces of an oblique fissure.

Fig. 23,— No. 196, shows the nature of the mesoderm behind the
primitive streak.

——-- + -

gem

UNIVERSITY OF TORONTO
STUDIES

BIOLOGICAL SERIES

No. 8: THE HABITS AND LARVAL STATE OF PLETHO-
DON CINEREUS ERYTHRONOTUS sy W. H. Pirersoi

(REPRINTED FROM PROCEEDINGS OF THE CANADIAN INSTITUTE 1908-9)

THE UNIVERSITY LIBRARY: PUBLISHED BY
THE LIBRARIAN, rg10

University of Toronto Studies
COMMITTEE OF MANAGEMENT

Chatrman: ROBERT ALEXANDER FALCONER, M.A., Litt.D., LL.D., D.D
/ President of the University

PROFESSOR W. J. ALEXANDER, PH.D.
ProFEssoR W. H. Ettis, M.A., M.B.
ProFressor A. KirSCHMANN, PH.D.
ProFessor J. J. MACKENZIE, B.A.
ProFEssor R. Ramsay Wricut, M.A., B.Sc., LL.D.
PROFESSOR GEORGE M. Wronc, M.A.

General Editor: H. H. Laneton, M.A.
Librarian of the University

Tas
i ia
* a
'
,
8 im) , ‘
{
j . f
yi f
a
.
@ 4 i
\
i
'
'
‘

By W. H. PrersoL, B.A, MB.

vi -
pa
‘ a
eee. 2 eee
ae: = ee ee a

1908-9.] THE Hasits OF PLETHODON CINEREUS ERYTHRONOTUS. 469

_THE HABITS AND LARVAL STATE OF PLETHODON CINEREUS
ERY THRONOTUS.

By W. H. Prersor, B.A., M.B.

ALTHOUGH Plethodon cinereus is considered by Cope (89) ‘‘the most
abundant salamander in the northern and central United States” its
habits and development have received but little attention. Cope (’89)
gives a brief outline and Wilder (’94) a scant mention, Montgomery (’o1)
adds a few points and describes the larvae of one bunch ofeggs. Reed (’08)
discusses the coloration of adults. Kingsbury (’95) touches on the
questions of the transference of sperm and the season of egg-laying but
without coming to any definite conclusions. Sherwood (’95) givesa date
on which eggs werefound. As regards Plethodon oregonensis the
behavior in captivity of a female found with her eggs is described by van
Denbrugh (’98), and Hubbard (’03) deals with some protective devices.
There are several other papers in which Plethodon is mentioned but with-
out reference to habits or development. ‘he observations recorded
below have been made partly in the field, partly in the laboratory, and on
specimens from several localities, all however within a radius of fifteen
miles fror: Toronto. Unless recorded as single occurrences, all observa-
tions have been verified in at least one subsequent season. Cope (’89)
divides P. cinereus into three sub-species of which P. cin. cinereus and P.
cm. erythronotus only are common. For the sake of ensuring uniformity
the latter variety alone forms the subject of this paper. Many larve
of P. cin. cin. have also been found and comparisons made with similar
stages of P. cin. eryth. show that in the larva as in the adult the sole dis-
tinction between the two varieties is the coloration. The difference in
geographical distribution which Cope mentions is not invariable, one bit
of woodland may yield the two varieties in about equal numbers, another
quite similar but a few miles away may contain P. cin. eryth. in abundance
and but few P. cin. cin. No locality yielding either variety alone, or a
majority of P. cin. cin. has been met with.

The typical coloration of the two sub-species is closely adhered to, the
intermediates noted by Reed (’08) are very rare and even then approach
closely the types; and only one specimen with much more than the normal
amount of red wasfound. About 250 adults have been under examination,
a number not so great as that used by Reed but sufficient to show that in

470 TRANSACTIONS OF THE CANADIAN INSTITUTE. [Vor. VIII.

this region there is less variation in colour than in the one where his studies
Were made.

Plethodon cinereus may be found in almost any wood in which
the underbrush has not been entirely cleared away. ‘They seek shelter
under logs, fallen branches, flakes of bark separating from old stumps,
or even masses of decaying leaves; the cracks of rotting logs and
stumps are also favorite places. In the latter situation the variety ery-
thronotus receives some protection from the red in its coloration, the shade
of red is often that of the decaying wood and the dark edging is a fair
reproduction of the dark and narrow cracks that run through the log.
Occasional specimens may be found in the latter part of April; during
the next three weeks they become abundant and by the middle of May
can be found as plentifully as at any time in the year. Just where they
pass the winter has not been determined though search has been made
both early in the spring and late in the autumn; digging quite through
stumps and logs embedded in the soil will not expose them, though it will
probably result in finding the wood-frog, Rana sylvatica. A few trifling
things have suggested that the animals follow down the cracks in the
roots of old stumps to a considerable depth, but this has not been
verified. Farther south Montgomery (’o1) finds them in the same situa-
tions both winter and summer.

Plethodon is strictly nocturnal, so far at least as regards life above
ground; however, a specimen that is suddenly uncovered is by no means
dazzled by daylight; it may remain motionless for a minute or two or may
at once crawl beneath the nearest shelter. Nor does it stop when under
cover but keeping out of sight crawls rapidly to a considerable distance,
so that if not captured as soon as exposed it runs a very good chance of
escaping entirely. This sensitiveness to light has rendered disappointing the
results from keeping specimens in a terrarium, an amount of light sufficient
to reveal them at all either arrests all movement or sends them under
shelter.

An examination of the stomach contents shows the food to consist
of a variety of small insects; in one case the remains of a small spider
were found.

In handling living specimens it will frequently happen that the
end of the tail will apply itself closely to a finger in a semi-circular loop,
and hold thus for a few seconds. This power may perhaps be regarded
as the first step in the development of a prehensile tail such as is de-
scribed for Autodax (Ritter & Miller, ’99).

1908-9. | THE HABITS OF PLETHODON CINEREUS ERYTHRONOTUS. 471

Two habits of Plethodon deserve notice as being of rare occurrence
among Urodeles. When excited it occasionally aids its progress by
leaping. In such cases as have been observed under conditions that
admitted measuring the length of the leap it has been found about equal
to the length of the animal’s body. If the animal is running on a rather
even surface it lands on its feet and continues the run having gained by
itsleap; but it will just as readily leap into difficulties. Held in the open
hand it will frequently leap off, no matter what may be the height of the
hand above the ground. In jumping the back is slightly arched and the
front limbs with most of the trunk are raised in the air to about the height
of one centimetre; then with a snap the tail is slapped against the surface
over which the animal is moving and the body sharply straightens and
shoots forward. ‘The whole movement is so rapid that it cannot be
distinguished with certainty whether the limbs aid in the leap. Two
things suggest that they do; it is difficult to imagine a force to raise the
anterior part of the body to the height it attains if it does not chiefly
lie in a spring given by the anterior limbs. The posterior limbs are even
stouter and are in a good position to aid in the forward propulsion. ‘The
young as well as adults possess this power of leaping, indeed the only
specimen observed to give a succession of leaps, three in fact, was one of
24 mm. — The explanation of the greater development of the posterior
limbs in the later larval stages, noted by Montgomery (’o1), may lie in
this habit. In this connection Cope (’89), says: ‘‘It frequently climbs
to the summit of low vegetation, from which it springs by a sudden straight-
ening or curvature of the body, as the case may be, in the manner of a
caterpillar.” The curvature and straightening in leaping are evident;
the climbing of low vegetation to leap from it has not come under obser-
vation.

The second noteworthy habit is also connected with escaping
enemies. Plethodon will sometimes break off a portion of its tail. Two
things suggested the existence of this habit before it was actually observed.
In searching for Plethodons in decaying logs not infrequently as the cover-
ing is lifted the animal will be found crawling stealthily away from the
bit of tail that is making itself very conspicuous by its violent movements.
This will occur at times when so little force has been used in picking apart
the log that it is difficult to conceive how the piece could have been broken
or pinched off. Again about 10% of all specimens found show the end
of the tail in the process of regeneration, raising the suspicion tha* it is
a mutilation out of the ordinary. On one occasion when a young Pleth-
odon had been repeatedly touched by a small rod it suddenly gave a
jump breaking off at the same time the terminal third of its tail. Two

472 TRANSACTIONS OF THE CANADIAN INSTITUTE. {Voy. VIIT

adults while being held lightly by the tail cast off a portion of it, the
separation occurring in the part grasped and leaving in the one case a
stump I cm. and in the other 1.75 cm. long, measuring from anus to
end of stump. In no case observed has the length of tail retained been
less than .5 cm.; it was usually more. There does not seem to be any one
place where separation greatly tends to occur. Incomplete separation has
also been noted, the vertebrae and muscles separating but the skin failing
to break; this causes a slight grooving of the tail as though an invisible
thread were tied around, compressing it. Longitudinal sectioning was
used to examine both proximal and distal pieces in those cases that occurred
under observation, and the relation of parts in the wound was found
to be as follows, (Figure t.) Separation of the vertebral column occurs
between vertebrae; this is the condition also in P. oregonensis (Hubbard,
03). In the muscle the myomereseare not broken but separation occurs
in the myocomma opposite the middle of the last vertebra retained.
The skin however does not break here but at the myocomma opposite
the middle of the first vertebrae of the piece cast off. Two things are
accomplished by this; first the wound that now terminates the animal’s
tail is protected by the extra length of skin which collapses laterally upon
it and almost covers it in. Second, on the piece cast off considerable
raw surface is left exposed, the irritation from which is doubtless largely
responsible for the rapid contortions that occur. This piece must play
an important part in the protective device for by its violent movements
it would draw the attention and invite the first attack of an enemy.

A similar habit has been noted in P. oregonensis by Hubbard (’03)
but there are distinct points of difference. Briefly in P. oregonensis
on moderate stimulation the glands on the dorsum of the tail swell greatly
and pour out an abundant secretion. Only the most powerful stimu-
lation—the act of being swallowed by a snake, or being plunged into
a fixing fluid without previously being anesthetized—would rouse the
animal to the point of sacrificing its tail, Separation always occurs at a
constriction just behind the anus. No mention is made of the behaviour
of the piece separated and, the subject of the paper being correlated
protective devices, it is perhaps fair to assume that it presents no striking
peculiarity. In P. cin. eryth. the tail shows neither constriction nor
swelling nor does the great development of dorsal glands occur, the
thickness of the dorsal skin being one-twelfth to one-eighth of the diam-
eter of the tail, while it constitutes one-fourth of the swollen tail of P.
oregonensts.

Owing to the extreme aversion to light already mentioned it has been
impossible to ascertain in a terrarium the manner of fertilization of the

1903-9. ] THe HABITS OF PLETHODON CINEREUS HWRYTHRINOTUS. 473

eggs. That fertilization is internal, and that it precedes egg-laying
by a considerable time may be inferred from the following. ‘The female
is provided with spermathecee which are already filled with sperma-
tozoa at a time when the eggs are yet in the ovaries and have not attained
a greater diameter than two mm. Quite likely the spermathece are
filled long before this because at this time no spermatozoa are to be found
in the genital tract of the male. ‘The testes are filled with spermatocytes
which as Mongomery (’03) describes undergo their maturation during
the summer; spermatozoa may be found after the middle of August.
Further observations are necessary to determine whether transference
occurs in autumn or in spring.

Although there is a total lack of secondary sexual characters it is
usually possible to distinguish between the sexes owing to the thinness
and translucency of the muscle and skin. Because of the dark colour
of the testis and the light colour of the eggs the posterior part of the ab-
domen when viewed from beneath is dark or light according as the speci-
men is male or female. ‘This distinction is most evident in the spring,
the eggs being largest at that time but in most cases it can be made at
any season. In addition, the cloacal glands of the male cause a whiteness
and a swelling at the sides of the body just behind the posterior limbs.

There is considerable latitude for a Urodele in the time at which egg-
laying occurs. The earliest date on which eggs have been found was
June 16th—3 clusters, the eggs in early stages of segmentation; the
latest is July 3rd—1r cluster, in the same condition. The difference is
not due entirely to different seasons being early or late, nor to one bit
of woodland being better sheltered and warmer than another, for clusters
of eggs with the females on guard and other females with the season’s
eggs yet in the ovaries have been found in the same log on the same
day. ‘The date, Oct. 25th, given by Sherwood (’95) for the finding of
eggs of Plethodon is not accompanied by any description of them ; and
its exceptional lateness leaves the meaning of his observation doubtful.

The eggs of the one individual seem to be deposited all at the one time.
No female was found guarding a cluster of eggs having also any in the ovi-
duct or even in the ovary that were nearly full size. It is however not at
all uncommon to find an egg in one ovary about half-size; this should be
associated with the small egg of some bunches described below and con-
sidered as one that should have been laid this season but which, having
failed to develop its proper amount of yolk, is retained in the ovary.

As asituation for the eggs marked preference is shown for logs almost
entirely embedded in the humus, logs that have decayed to a point where
the substance is friable with large cracks running lengthwise. In one

474 TRANSACTIONS OF THE CANADIAN INSTITUTE. [Vor. VIII.

of these cracks the eggs are hung from the roof like a little bunch of grapes
(Figures 2 and 3). Rarely a similar cavity within a decaying stump is
selected. All the localities in which Plethodon has been found contain
conifers and almost without exception it is in coniferous wood that the
eggs have been found; though the adults may be found plentifully in
wood of all kinds. Usually the eggs are placed from three to five inches
beneath the surface of the log, at which depth its substance is con-
stantly moist, and it can well be imagined that the air is saturated
with water vapour. They are alwavs accompanied by the female
and if in exposing them she has not been alarmed she will be
found holding the dorsal surface of her head and neck against the under
side of the bunch. That it is the female that remains on guard was
determined by the dissection of over twenty specimens. In only one
case was a male found by eggs and then in company with a female; the
eggs were several days advanced in development and the presence of the
male was perhaps merely an accident. Cope (’89) and Mongomery
(or) speak of finding the animals and their eggs under stones but if the
above mentioned shelters are available they are always preferred.
Wilder (’94) notes that the adults are seldom to be found under stones.
Sherwood (’95) gives the habitat as beneath logs and stones, while the
eggs are to be found in damp moss and beneath the bark of decaying
stumps. i

Among Urodeles that do not lay their eggs in water contact between
the body of the female and the eggs seems to occur in all cases; Amphiuma
(Hay ’88) and Autodax (Ritter and Miller ’99, and Ritter ’03) coil round
them, Desmognathus (Wilder ’99) inserts herself among the eggs wearing
them like a necklace or belt. Plethodon oregonensts (van Denbrugh ’98)
is described as holding the bunch in a loop of her tail and moving them
from place to place. But as this was after removal from their natural
surroundings it is quite possible that the eggs had been torn from their
original support and that part of the mother’s uneasiness arose from their
unattached condition. The Cecilians (Gadow ’o1) also when not vivi-
parous coil round their eggs.

The number of eggs in a cluster varies from three to twelve.. The
number of ovarian eggs in advanced condition, in specimens taken just
before the egg-laying season is also within these limits and the length of
the incubation precludes a second brood in the one year. Consequently
the preservation of the species must depend upon the larval stages being
so perfectly adjusted to surrounding conditions that the mortality is
very small, rather than—as is the case with most Urodeles—upon the
production of large numbers of young. The slender and almost cylin-

‘1908-9. ] Tue Hapits OF PLETHODON CINEREUS ERYTHRONOTUS. 475

drical form of body (the proportion of length to greatest breadth being
fourteen to one) in an animal so small, along with the necessity for pro-
ducing eggs containing a large amount of yolk are doubtless factors
determining the small number of eggs; another will be mentioned later.

The necessity for the large amount of yolk in the egg arises from the
purely terrestrial development of the larva. Aquatic larve have at
command the minute and abundant fresh-water plankton as food supply
and are thus at an early age rendered independent of the nourishment
~ provided in the yolk. ‘The insect life that constitutes the early food of
the terrestrial Plethodon is of larger size than much of the plankton
and much less abundant. Consequently the animal on leaving the egg
must be able to wait for food through comparatively long intervals
and also to capture food of larger size than an aquatic larva need
do. Both of these things demand an advanced development that can
only occur when a considerable quantity of yolk is provided.

Each egg is surrounded by a series of mucous spheres as is customary
in Urodeles. In their natural condition the number of these is rather
difficult to determine but after soaking a few minutes in water they swell
somewhat and the following is plainly seen:—an innermost sphere very
close to the surface of the egg; a second enclosing this but separated
from it by a greater interval than that between the innermost sphere
and the egg; occasionally this sphere is represented by two, one of them
fitted very closely around the other. The outermost sphere—usually the
third—fuses with the outermost spheres of neighboring eggs at all points
of contact. On its surface are threads and bands of a milky white mucus
which seems tougher than the rest, which is transparent; these are es-
pecially numerous between eggs and at the upper part of the bunch where
several uniting form the stalk by which the cluster is suspended. This
mode of attachment is probably derived from one originally like that
of Desmognathus (Wilder ’99) in which each egg of the cluster is inde-
pendent of all the rest and has its own cord joining it to the common
stalk. The tension of the envelopes, especially the innermost, is needed
to preserve the spherical shape of the egg. These envelopes are very
tough; a weak hypochlorite solution will soften them so that they can
be removed but when their support is gone the egg, even in water, flattens
until the vertical diameter is little more than half the transverse. In
the younger stages it is therefore necessary to fix and harden the egg
first and remove the membranes later, they should be removed as soon
as possible for if left around the egg the latter will in time disintegrate.
The method devised by Morgan (’91) has proved the most satisfactory.

As development proceeds the amount of fluid between the egg and

476 TRANSACTIONS OF THE CANADIAN INSTITUTE. (Von. Visas,

the innermost sphere increases markedly, thus relieving the growing »
embryo frem pressure.

The eggs vary from 3.5 to 4.5 mm. in diameter. In many of those
bunches which contain several eggs there is one egg considerably smaller
than the others; with this exception there is great uniformity in the size
of the eggs of the one cluster. The considerable variation noted above
exists between the eggs of different clusters.

The colour of the egg is a very pale yellow, a tint due to the yolk
for no pigment is present. The upper part of the egg being the more
free from yolk is almost white. Variation exists in the tint of the eggs
of different clusters, the range being between very pale cream and light
orange-yellow.

The usual rotation of amphibian embryos in their earlier stages.
occurs also in Plethodon but is modified by the large amount of yolk
present. In the usual type of amphibian embryo, such as Ambly-
stoma or Rana, the amount of yolk is not so great as to render it impos-
sible for the embryo to turn over when the course along which the ciliary
moveinent urges it sodemands. Until the Plethodon embryo has reached
a length of 6 mm. it lies along a meridian of the yolk mass and a true
forward movement would require the yolk at times to be uppermost.
Whether because the ciliary movement lacks force to drive it into this
position or for some other reason less obvious the rotation that occurs.
is around the vertical axis. There is no uniformity of movement even:
in eggs of the same bunch; an occasional one will be motionless and the
others about evenly divided between motion to right and motion to left.
A little later when the embryo twists itself into the horizontal plane the
appearance in rotation is that of the ordinary amphibian embryo but
the same lack of uniformity of direction obtains, some few embryos will
not be moving at all, of the rest the head will be preceding in some cases.
but the tail in quite as many others. ‘This latter direction was not ex-
pected and caused an examination of Amblystoma eggs to be made sub-
sequently to determine if anything comparable was to be found there.
Eycleshymer (’95) notes an early irregularity of direction but this is not
the impression derived from a brief examiination of larve in later stages;
nor is it to be expected from the account of the ciliation of Rana and Triton
larve given by Assheton (’96). The stage of developement of Ambly-
stoma equivalent to the one of Plethodon in which the backward move-
ment is so marked is that in which the gills are beginning to appear and
the larva is curved laterally in a semicircle to accommodate itself to its
spherical envelope. At this stage nearly all embryos examined will be

1908-9. THE Hasits oF PLETHODON CINEREUS ERYTHRONOTUS. A477

found moving slowly forward very nearly in the horizontal plane of the
body; some few will be motionless and about 1% will be moving back-
ward. It is not difficult to mark these latter by thrusting small splinters
of wood through the jelly until their points almost touch the sphere
surrounding the larva; a few hours later these individuals will be found
moving forward and another set moving backward. Reversal of direction
has also been noted in some Plethodon larve. Evidently the condition
in Plethodon is nothing peculiar to the genus but an exaggeration of
one found elsewhere among Urodeles. How widely spread it may be
and whether it means a reversing of the usual movements of the cilia
are points left for future determination.

The rate of rotation varies greatly, extremes of one minute and three
and one half minutes having been noted under similar conditions.

The service of the rotation to the embryo probably is that it prevents
adhesion between the egg and the envelopes. Smith (’96) found this
occurring in eggs of Cryptobranchus lying in dishes in the laboratory;
interference with development resulted. He makes no mention of rotation
in the eggs but suggests that gentle rocking due to the current of the stream
in which the eggs are laid prevents adhesion under natural conditions.
Apparently in eggs laid in waters that have no current—as Amblystoma
or Rana—or in other equally quiet situation—as Plethodon—the rotation
suffices to prevent adhesions until the embryo beginning muscular move-
ments is past all danger of their occurrence. Besides rotation other
movements in embryos of 6 to 7 mm. are performed at intervals and
consist of elevating the head from the yolk and waving it from side to
side. As development proceeds these movements pass into occasional
wrigglings of the tightly coiled larva whereby the whole position within
the envelope often becomes altered.

Development and growth are very rapid for a few days, an embryo
of 5 mm. will increase in length to 9 mm. in six days; but at this point
- the processes become much slower and to increase from 9 mm. to II
mm. requires fourteen days. ‘This change in the rate of growth is shown
very plainly in sections by a sudden diminution in the number of mitotic
figures which from being very numerous in the smaller embryos become
rare after the length of 9 mm. is passed. In its later stages development
is retarded in the central eggs of a cluster where the embryos are smaller
and show but little reduction of the gills while the outer ones are larger
and have the gills almost entirely absorbed (compare Figures 13 and 14).
Measurements of a typical case gave central embryos 17.5 mm. with
gills 1.75 mm. long and outer embryos 20 mm. with gills reduced to

478 TRANSACTIONS OF THE CANADIAN INSTITUTE. || [Vov. VIII.

25mm. A similar condition is found in Amblystoma, Rana and other
amphibia where the eggs are laid in clusters; the cause being the difficulty
with which respiration is carried on in the central eggs.

Escape from the mucous envelope occurs when the larva has reached
the length of 20 to 25 mm.; this is about the first week of September.
The outer larve of a cluster escape first but remain with the mother
beneath their less developed brethren. Respiration is now more perfect
in these; in a few days their gills have dissappeared and they also escape.
However the interval does not suffice to bring them so far on in develop-
ment as were the outer ones when they escaped; they have absorbed
less yolk and are shorter. They also seem unfit to come in competition
with their more favored brethren of the outside of the bunch of eggs,
their movements being comparatively feeble and ineffective; many of
them probably die. Thus is supplied another factor tending toward
the production of small numbers of large eggs; for those young will be
best fitted for life that come from broods where all the nourishment
available as yolk is divided among but few eggs, so few that none are
crowded into the centre of the cluster.

The family does not scatter at once, an unusual thing among Urodeles
that have attained their adult form, but accompanied by the mother,
perhaps led by her, the brood leaves the interior of the log to live beneath
stones, fragments of wood or bark, lying on the surface of the ground,
or even among layers of mouldering leaves. If not disturbed when
uncovered the young will be found in contact with the body of the mother,
probably to obtain moisture; for the localities in which they now are
contain little moisture at this season of the year. The families do not
hold together for long. In the latter part of September a few solitary
young will be found and early in October all the broods seem to have
broken up. The rate of growth in Plethodon must vary enormously
in different individuals, for at this season it is easy to collect a series
beginning with young accompanying the mother and ending with full
grown specimens, the increase in size being so gradual that it is impossible
to draw with certainty a line between this year’s and last year’s broods.
Apparently the ability to withstand the winter is independent of size for
early in May an occasional specimen will be found that is little larger
than the young at the time they escape from the egg—that is less than
30 mm. in length.

The following description of the external appearance during develop-
ment is based upon the examination of over 170 embryos that had been
fixed in Zenker’s fluid and preserved in 80% alcohol, the members of each

1908-9.] THe HasBits OF PLETHODON CINEREUS ERYTHRONOTUS. 479

bunch being kept together and separate from other bunches. In the series
twenty different stages of development are represented, though only the
most important of them will be described. As the eggs vary somewhat
in size so do the individuals of the same stage of development; the
descriptions given are those of average specimens, extremes in any
direction (very few in number) being disregarded.

The embryo may be considered as well defined upon the yolk when
the medullary folds are complete and approximated through a little more
than the posterior half of their extent. (Figure 4). Together they give
this part of the embryo a width of .75 mm. except at the posterior end
where they separate slightly before becoming continuous giving here a
width of 1 mm. In their anterior part the folds are widely separated
giving to that part of the head an extreme breadth of 1.5 mm. In front
of this part an almost transverse piece connects them while behind it,
they converge, meeting at a point a little anterior to the middle of the
body. In living specimens placed in a strong light a delicate scalloping
of the inner margins of the folds in the cranial region indicates clearly
the neuromeres, but no method yet used has succeeded in preserving them
through fixation. In a corresponding stage in the eggs of some Amphibia
the areas of brain substance that will form the optic vesicles are indicated
by depressions, sometimes pigmented (Eycleshymer ’93), but in Plethodon
no trace of these can be found. The length of the figure defined by the
medullary folds is-three millimetres.

Larva5 mm. Figure 5.

The form of the body is still determined by the yolk and the central
nervous system. The medullary folds are in apposition throughout.
The head is raised off the yolk as far back as the midbrain; eyes and ears
are indicated externally. Behind the head the mesoblast makes a border
on each side of the spinal cord extending but a short distance laterally
over the yolk. The anterior three-fifths of it is divided into ten somites
opposite the third and fourth of which it is thickened forming the begin-
ning of the anterior limb. The posterior two-fifths of the mesoblast is
not yet divided into somites. The medullary folds flatten and a little
in front of the well marked blastopore disappear.

Larva 5.5mm. Figure 6.

The pharyngeal region is now distinct and shows four gill arches;
sections, however, show that the gill slits are not yet perforate. Pos-
teriorly a crescent shaped furrow, the horns pointing forward, defines
the termination of the body and marks its tendency to rise off the yolk.

480 TRANSACTIONS OF THE CANADIAN INSTITUTE. (Vou. VIII.

Though thus hidden from view by the beginning of the tail, sections
show the blastopore opening into the bottom of this furrow. At this
time the increase of fluid around the egg begins. The whole body is as-
suming a cylindrical form and standing out from the yolk which is slightly
flattened near it, and markedly flattened in the region against which
lies that part of the body from the anterior limbs forward.

Larva 6 mm. Figure of

During the growth from 5.5 mm. a twisting of the anterior part of
the body through 90 degrees has occurred so that one side of the head
is now turned toward the yolk; of eleven specimens this was the right
side in nine, the left side in two. That portion of the body attached to
the yolk, 2.e., all posterior to the pharyngeal region, shares in the twisting
and becomes curved laterally upon the yolk in such a fashion that the
anterior three-fourths of the dorsal surface is brought into the one plane.
Opposite and external to the last three somites now distinguishable is,
on each side of the body, a low mound of thickened tissue—the begin-
nings of the posterior limbs. The terminal millimetre of the tail is quite
cylindrical and free from the yolk. The first traces of pigmentation
now appear on the dorsal surface of the anterior part of the body.

Larvagmm. Figure 8.

The anterior part of the body has more than doubled in thickness and
has grown so far off the yolk that the anterior limbs are now free from
contact with it. This gradual freeing of the body from its attachment
to the yolk is brought about by growth of the central part of the body
crowding off the extremities, accompanied by but little pinching off of
the connection between body and yolk; for the length of body attached
to the yolk has remained, and will for some time yet remain, constant at
3.5 to 4 mm. As the body is thus projected forward those parts of it
that were spread out laterally on the yolk move ventrally and unite in
the middle line. This is well shown in the anterior limbs which from pro-
jecting dorsally as at first are compelled by this movement at last to
point outwardly. They are now .35 mm. long. Similar growth back-
ward has brought the posterior limbs to the verge of the attachment ~
of body and yolk posteriorly. These limbs are yet represented by
rounded thickenings only. All traces of the gill arches have disappeared
from: the surface while the external gills have appeared as three points
.25 mm. in length. The costal grooves are visible and pigment in a broad
band extends over the dorsal surface of the neck and anterior half of
the trunk.

1908-9. | THE HABITs OF PLETHODON CINEREUS ERYTHRONOTUS. ' 481
Larva 9.5mm. Figure 9.

The eye is now pigmented. ‘The mouth is well defined and the gular
fold projects a little over the base of the gills; these now appear as three
points on a couimon base. The anterior limbs have a length of .5 mm.
the posterior of .25 mm.; the latter are now attached on the line between
yolk and body, consequently the anus lies in the free part of the body
posterior to the yolk.

The accumulation of fluid between the egg and the innermost envelope
has increased the diameter of the egg and envelopes together from 4.5
or 5.5 to 6.5 or 7 mm. This is not due to a thickening of the walls of
the spheres but to such a distension of the inner one that yolk and larva
together occupy only the lower two-thirds of the cavity. No further
enlargement of the cavity occurs, consequently the larva becomes more
and more closely coiled as its length increases, for the diminution of yolk
that soon becoines noticeable does not compensate for the increased bulk
of the larva.

The rotation ascribed to cilia on the ectoderm has been noticed in
some larve of this stage of development, but in none later.

Although the blood is not yet red a hand lens will reveal the follow-
ing plan in the circulation over the yolk miass, the vessels appearing as
colourless lines against the white background of the yolk; the direction
of flow can be followed as the corpuscles are easily visible. Small arteries
metamerically arranged run laterally over the yolk; they branch and
anastomose in such fashion that a network two or three meshes in width
is formed between the margin of the body above and the vein in which
they terminate ventrally. (Figure 9). Usually there are two such
veins at first, a right and a left; at a variable time in later development
one disappears, as is usual in Urodeles. Sometimes the anterior part
of the network on one or both sides collects into a separate vein, making
for a time three or four separate trunks that unite just posterior to the
heart. Sections show these vessels all lying, as might be expected, in
the splanchnic mesoblast which has ere this completely surrounded the
yolk.

Larva 10.5 mm.

At about this length sections show for the first time perforation of
the gill pouches, two only—the first and second—ever become perforate.
‘The anterior limbs show the first indications of digits in that their extremi-
ties are flattened and show a terminal notch and a shallow groove leading
to it on each side. The posterior limb is not appreciably flattened.

482 TRANSACTIONS OF THE CANADIAN INSTITUTE. [Vor. VIIT..
Larva Il mm.

Each division of the anterior limb above noted is showing two tuber-
cles, the four digits being thus indicated. ‘The posterior limb has ex-
changed its rounded contour for an angular one but no notches or grooves.
appear on it.

Larva 12mm. Figures to and 11.

The digits of the posterior limbs appear as five tubercles, the second
and the third of which are the most prominent.

Larva 16mm. Figure 12.

The posterior limbs are growing the more rapidly; they are now as.
long as the anterior ones—2.5 mm.—and much stouter. Both pairs
are provided with well developed toes. The gills have developed greatly
each being two to three mm. long and consisting of a central stem with
six or eight side branches. Considerable variation in the length of the
gills is found in larve from different broods in what is otherwise the
same stage of development, the shorter gills having also the shorter side
branches. Some diminution of the yolk mass is for the first time notice-
able, its vertical depth having lessened, though the attachment to the
body is still along 3.5 mm. of the ventral surface. Sections show that
in larvae from 9 mm. onwards the body walls are actually complete,
but the more ventral part of the mesoblast is so much thinner than the
dorsal and the line along which the two parts meet is so sharp as to pro-
duce in the entire larva the appearance of having body walls developed
only over the upper part of the yolk mass. At this stage rather more
than the dorsal half of the yolk is covered by this thickened part of the
body wall. The pigmentation typical of the adult is complete.

Larva 17.5 mim.

The posterior limb has outgrown the anterior, the lengths being
3 mm. and 2.75 mm. respectively. The length of the gills is reduced
to 1.75 mm. or less. ‘The yolk mass has so diminished as ‘to be almost
covered in by the thicker parts of the body wall, only a portion 2.5
mm. long and .5 mm. broad projecting slightly in the mid-ventral line.

Larva 20 mm. to 25 mm. Figure 13.
(just on the point of escaping from the mucous envelopes.)

The anterior and posterior limbs are 3 mm. and 3.5 mm. long re-
spectively. The yolk mass is almost entirely covered in. The gills are

1908-9. | THe Hasits OF PLETHODON CINEREUS ERYTHRONOTUS. 483

reduced to a few small points not over .25 mm. long. Sections of several
larve of this general stage but differing slightly in development show
that both the gill slits become closed before escape from the envelopes
occurs, the second one being the first to close.

In their development the digits of the posterior limb but partly
bear out the expectations of Cope (’89). He would regard Hemidactylium
with its posterior foot possessing only the first four digits as having for
permanent form that which is larval in Plethodon. Dealing with P.
cinereus he finds the adult has the outer digit longer than the inner, but
in younger specimens it is shorter and in his youngest (18 mm. only
but having already lost its gills) it is but a minute tubercle, ‘‘and in a
little earlier stage cannot but be wanting though this I have not seen.”’
On this point fifty-one. larve were examined covering twelve stages, begin-
ning with larve of 11 mm. long and ending with the smallest found living
alone, its length being 23 mm. In larve of 12 mm. the whole five digits
appear at once and the fifth is no less prominent than the first. There
follows a brief period—vz. until the larva attains a length of 16 mm.—
in which the rate of growth of the fifth digit as compared with that of
the first, varies; of twenty-nine specimens within these limits the first
exceeded the fifth in sixteen, the fifth exceeded the first in three, and in
ten they were equal. (Seven broods are represented in this, in four of them
only were all the larve alike on this point). This period alone is in accord
with the argument of Cope. In all larve over 16 mm. (thirteen in number
representing four stages) the external toe was longer than the internal.

EFFECTS OF TERRESTRIAL DEVELOPMENT.

The influence of a purely terrestrial development is seen chiefly in
the following points.

The large amount of yolk in the egg in proportion to the size of the
animal, a point already dealt with.

The yolk mass retains its globular form until late in development
‘(arve of 13 to 15 mm.), when the absorption of its substance causes
it to become fusiform. In aquatic larve the mass early elongates to
produce a slender body capable of rapid darting movements, a necessity
not laid upon the yolk in the inactive larva of Plethodon.

The development of limbs occurs early. Traces of the anterior
limbs are distinct in larve of 5 mm. and of the posterior limbs in those
of 6 mm. In Amblystoma these traces appear in larve of 7.5 mm.
and 13 mm. respectively. From the large size of the posterior limbs

484 TRANSACTIONS OF THE CANADIAN INSTITUTE. (Vo. VIII.

in his larve Montgomery (’o1) was led to suggest the possibility of
their development before the anterior limbs, contrary to the usual con-
dition in Urodeles. ‘This proves not to be the case, though they do appear
earlier than usual and grow more rapidly.

The development of gills is retarded. No trace of them as separate
points is found until the larva reaches a length of 8 or 9 mm.; in Am-
blystoma the same occurs at a length of 6 mm. Looking at limbs and gills
together the contrasts are marked; when Plethodon first shows external
gills its posterior limbs have been for some time distinct; when Ambly-
stoma first shows posterior limbs its gills are 1.25 to 1.5 mm. long and
plentifully branched.

It is of general occurrence that the gills of such amphibian larve
as have no free aquatic life are proportionately longer than those of larve
that do. The condition in Salamandra atra as described by Chauvin
(77), shorter and stouter gills being assumed when aquatic life was forced
on the larva, is typical. This obtains also in Plethodon, the filaments
as well as the main trunks of the gills being much longer but much less
numerous than in aquatic larve of the same size, e.g. those of Amblystoma.
Such marked reduction in the number of filaments does not occur among
all Urodele larve of non-aquatic development. In aquatic larve the
gills are largely directed backward to afford as little resistance as possible
to passage through the Water; in Plethodon they spread out as widely as
possible, the direction being a matter of indifference. The point of im-
portance is that they apply themselves to as much of the mucous envelope
as possible and so place themselves where they can best obtain a supply
of oxygen.

Plethodon has neither the balancers nor the adhesive discs common
among other amphibian larve.

The body of a Plethodon larva is from the first that of a terrestrial
animal, cylindrical and without trace of median fins. Like the larva
of Autodax (Ritter and Miller ’99) it has lost the swimming instinct;
when placed in water it sinks to the bottom and falls on one side; often
indeed it twists its body and writhes violently but such movements never
result in any progression or even in temporarily regaining balance. They
seem to be only the wriggling that occurs periodically within the egg
envelopes; it can always be induced by any stimulation of the larva.

The complete darkness in which the development of Plethodon
takes place is partly responsible for peculiarities in colouring. Pigment
is entirely lacking in the egg nor does it appear in the larva until a length

o

1908-9. ] THe Hasits of PLETHODON CINEREUS ERYTHRONOTUS. 485

of about 6 mm. is reached. When pigmentation does begin it rapidly
assumes the pattern and colours of the adult; this is evidently related
to the fact that when the larva issues from the egg envelopes it at once
assumes the habits and habitat of the adult. ‘The lack of a free larval
life differing from that of the adult renders unnecessary the distinctive
larval colouration so common among Amphibia. At the same time the
surrounding darkness renders it safe for the larva to receive the colour-
ation of the adult. The reason why the larva of Autodax (Ritter and
Miller ’e9; Ritter ’03) should under somewhat similar circumstances show
a darker colour than the adult may lie partly at least in the fact that its
development occurs in dimly lighted cavities and not in absolute dark-
ness as does that of Plethodon.

Amblystoma (or Spelerpes), Desmognathus, Plethodon, and Autodax
form an interesting series the members of which, taking larval and adult
life together, show an increasing adaptation to terrestrial life. The first
is terrestrial only in adult life and returns to the water to lay its eggs.
Desmognathus (Wilder ’99) begins its development on land like Plethodon
but near the water; it leaves the egg while yet a larva completing its
development in the water and accordingly stands as an intermediate
between Amblystoma and Plethodon. The habitat of the adult Des-
mognathus, never far from a stream, also shows a less degree of adaptation
to terrestrial life than that attained by Plethodon whose habitat bears
no relation to bodies of water. On the other hand Autodax is less de-
pendent on moisture in its surroundings even than Plethodon. I have
had adult Plethodons die in confinement from conditions of no greater
dryness than those described as supported by Autodax (Ritter and Miller
’*99). Moreover the prehensile tail, the greater ability in leaping and
more intelligent use of the power would also indicate more perfect adap-
tation to terrestrial conditions than has been attained by Plethodon.
As larva, Autodax has entirely lost the fringed condition of the gills;
Plethodon still retains these fringes but in such varying degrees in differ-
ent individuals as to indicate their decadence as organs. The degree
of parental care of eggs and larve increases regularly in the series Am-
blystoma, Desmognathus, Plethodon and Autodax.

EXPERIMENTS.

To determine what part the mother may play in the incubation
various plans of development under artificial conditions were tried. If
removed from their natural surroundings and suspended in air the eggs
show signs of rapid loss of moisture and the larve soon die. To obtain
an atmosphere as moist as that in which they naturally hang, clusters

486 TRANSACTIONS OF THE CANADIAN INSTITUTE. [Vo.. VIII.

of different ages were suspended over water, each cluster in a wide-mouthed
bottle the cork of which had two shallow grooves cut down the sides to
admit small quantities of air. The bottles were kept in a cellar where
the temperature was practically the same as that of the natural situation
of the eggs. In all the clusters development proceeded without inter-
ruption, in some cases for as much as twenty-five days when an impending
absence from the laboratory rendered it necessary that the experiments
be terminated and the larve were fixed. About eight per cent. of the
larve always die under this artificial incubation, in some bunéhes no
deaths at all occurring, in others several. This is quite striking, for no
unfertilized eggs or dead larve were ever found under natural conditions.
Three things may account for this; injury in conveying to the laboratory,
the growth of mould during incubation, and lack of the normal increase
of fluid between the egg and the inner envelope. Mould on the eggs
although never encountered under natural conditions sooner or later
makes its appearance on all eggs reared as above described. It does
not however always have an early fatal effect, for all the larve in a bunch
have been found alive after having been quite obscured for two days by a
growth of mould. ‘The increase of fluid between egg and envelope can
hardly be said to occur under these artificial conditions and presumably some
pressure is exerted upon the larva. These things suggest that the female
may in some way prevent the growth of mould on the eggs and also supply
them with moisture and an endeavour was made to test this in the following
way. In a wide-mouthed jar pieces of the log in which the eggs were
found were arranged to form a little chamber in which the female was
placed and the piece suspending the eggs then added as a roof; more
fragments were placed upon this and the surface covered with a little
humus and moss, a few drops of water were occasionally sprinkled on the
surface. Jars so prepared were kept under the same conditions as the
bottles previously mentioned for three weeks, by which time mould
would certainly have appeared on eggs kept in bottles, but none was
found, and the amount of fluid surrounding the larva was as great as
natural. In a subsequent set of experiments it was found that even
thoroughly wetting the cluster two or three times a day, in addition to
keeping it over water as before, would not suffice to bring about the nor-
mal accumulation of fluid. This was only obtained by allowing the lower
end of the cluster to remain in contact with the surface of the water in
the bottle for about twelve hours out of each twenty-four.

No final statement should be based on such scanty experimental
evidence but such weight as it has is entirely in support of the supposition
that the mother in this case—and presumably in similar cases reported

1908-9.] THe Hapits OF PLETHODON CINEREUS ERYTHRONOTUS. 487

among Urodeles—does along the lines mentioned actively promote the
welfare of her brood.

In those last mentioned and in other experiments in which the clusters
of eggs were entirely immersed in running water it was noticeable that
the mucous envelopes would imbibe readily a small quantity of water
and swell slightly in consequence, but the limit to this process was usually
reached within the first half hour, nor did any second period of imbibition
occur. No softening of the envelopes was to be noted even after constant
immersion for three weeks. These qualities of the mucus mark it as
different from that of most amphibian spawn which continues to absorb
water and to soften until the larve escape. They probably are to be
considered adaptations to prevent rains of even most unusual duration
from softening the mucus and turning loose the larve prematurely.

The bottle method before described was used to test the possible
effect of light upon development. Several bunches were exposed to
daylight, which however never became direct sunlight, while others of
the same stage of development were kept in entire darkness as controls,
the temperature being in each case, within limits of variation of three
or four degrees, the same. At the end of twenty-five days no difference
in the degree of development of the two lots was to be detected. The
lack of pigment in the egg might be put forward but the whole expla-
nation cannot rest on this for early in the life of the larva pigment appears
in the skin and rapidly increases in amount; and in the experiment this
pigment was present for the last twenty days. In the experiments of
various investigators to determine the effect of light upon growth Am-
phibian larve in aquaria have frequently been used; in which case the
oxygen and part of the food have been derived from the water. Since
both these things would be more abundant in the better lighted aquaria,
factors must be allowed for, whose exact influence is unknown. In the
present case the food and oxygen factors are the same for both sets
of larve and the uncertainty of result due to several factors varying
simultaneously is lacking.

Points ARISING OUT OF THE LUNGLESS CONDITION.

In Amphibia generally the function of the skin as an organ of respira-
tion accessory to the lungs has long been recognized. Later Maurer
(98) drew attention to the advantagcous position for purposes of respira-
tion held by the capillaries of the bucco-pharynx, many of them being
situated in the epithelium itself. To these localities therefore the atten-
tion of investigators was naturally directed in seeking the means which

488 TRANSACTIONS OF THE CANADIAN INSTITUTE. [Vo. VIII.

would compensate for the lack of lungs. ‘The conclusions reached were
not uniform, some investigators accepting both skin and bucco-pharynx
with more or less of the cesophagus as sharing with something like equal
importance in the respiration, others concluding that the skin is no more
efficient in respiration in lungless salamanders than in those with lungs.
Some examinations of Plethodon along these lines was in progress but
ceased when the paper of Seelye (’06) on the Circulatory System of Des-
mognathus came to hand for the points already dealt with showed that the
conditions in Plethodon would be but a repetition of those in Desmogna-
thus: and would lead to the same conclusion, namely that as an organ of
respiration the skin is much more important in lungless than in lunged
salamanders. ‘The same paper also gives a sufficient review of the ques-
tion and its literature so all that will be attempted here is to bring for-
ward three additional pieces of evidence in support of the above
conclusion.

First, as noted in the paper itself the value of the cutaneous capillary
network for respiration will depend upon the permeability of the mem-
brane through which diffusion must take place. The fact that this mem-
brane is the epidermis and not the entire skin renders exact experimen-
tation impossible. Nevertheless the experiments performed by Seelye
indicate that the entire skin of lungless forms is much more permeable
than that of those with lungs and it would be strange to urge that the
difference in the cutis accounts for this for it is the lungless forms that
have the thicker cutis. This point of structure is, according to Seelye,
the only one of general distinction between the skins of the two types
in question; a conclusion that it is hard to understand unless it is due
to the presence of two European forms with lungs among those examined.
A more trustworthy comparison would be one between forms that live
in the same environment and in the case of Plethodon this is possible,
for small specimens of Amblystoma punctatum and of Dtemyctylus
viridescens in its terrestrial stage of life are occasionally found along
with Plethodon cinereus, The skins of specimens so found and of adult
Plethodons of about the same length were examined all being submitted
to the same procedure, viz. the entire animal was fixed in Zenker’s fluid
with the usual after treatment, then from similar regions of head, trunk,
and tail, pieces from dorsal, ventral, and lateral aspects were sectioned
perpendicularly to the surface. The only considerable and constant
difference in the epidermis is one of thickness. To estimate this correctly
several measurements in micra were made from each piece of skin and
these were averaged. Finally the figures thus obtained for each of the
areas investigated were averaged to obtain a figure that would fairly

1908-9.] THE Hasits oF PLETHODON CINEREUS ERYTHRONOTUS. 489

represent the average thickness of the epidermis over the whole body.
These averages were 22.4 micra for Plethodon, 46 micra for Amblystoma,
and 44 micra for Diemyctylus. Consequently as regards this factor
unless we assume a most improbable thing, namely that the epidermis
of Plethodon is of a material less permeable than that of Amblystoma
and Diemyctylus, we must conclude that of the three the skin of the
lungless form Plethodon is fitted to be by far the best respiratory
organ.

Second. The bucco-pharyngeal respiration is indeed established
very soon after escape from the egg yet even in the adult it may be sus-
pended for a considerable time without serious inconvenience. When
confined in a glass vessel the animal will occasionally rest the ventral
surface of head, neck, and more or less of the trunk against the glass;
the adhesion between the glass and the moist and sticky skin is sufficient
to prevent the respiratory movements. I have frequently seen this
continue for two or three hours and have found the position of the body
apparently unchanged after even much longer intervals, but in these
cases observation not being continuous it is possible that the animal
may have moved for a time and then resumed its exact original position,
but such a thing is unlikely.

Third. There is a period of a few days in the life of a Plethodon
during which such respiration as occurs at all must take place through
the skin. ‘This is the period just prior to the escape from the egg envelopes.
The gills attain their maximum of development—a length of 3 mm.—
in larve of about 15 mm.; after this they decrease in size and for some
time before the escape of the larva are reduced to small points not over
.25 mm. in length. No movements of the ventral pharyngeal wall are
to be observed at this time so it cannot be that there exists a mode of
respiration similar to the aquatic pharyngeal respiration noted in Diemy-
ctylus, both viredescens (Gage ’o1) and torosus (Ritter ’97), the fluid
within the egg envelope playing for Plethodon the part of the surrounding
water for Diemyctylus. Consequently whatever oxygen is used by
the larva must be absorbed through the skin. The amount of oxygen
required by the larva may well be less than that required by the free-
living animal yet it is by no means an inconsiderable fraction of it. The
muscular activity in the beating of the heart and the wriggling of the
larva within the envelopes that occurs not infrequently is probably little
less than that of the free-living animal, which though capable of active
movement rarely indulges in it, forming in this respect a marked contrast
to Diemyctylus. If the larva can carry on its respiration through its
skin alone, hampered as it is by the surrounding fluid and envelopes,

490 TRANSACTIONS OF THE CANADIAN INSTITUTE. [Vox. VIII.

it is difficult to escape the conclusion that in the free-living stage as well
the skin must be an important factor in respiration.

The development of gills in the first place is governed by the same
necessity as exists in other Amphibia; sections show that the beginning
of their degeneration coincides with the skin becoming sufficiently de-
veloped for the circulation in it to become somewhat extensive; and
that their diminution keeps pace with the elaboration of the cutaneous
circulation, just as in most Amphibia it accompanies the increasing
activity of internal gills or lungs.

EarRLy DEVELOPMENT AND DEVELOPMENT OF INTERNAL ORGANS.

Studies in these fields have revealed several points of interest which
it is proposed to consider in a future paper.

1908-9. ] Tue HasBits OF PLETHODON CINEREUS ERYTHRONOTUS, 491

LITERATURE.

ASSHETON, R., ’96. Notes on the Ciliation of the Amphibian Embryo. Quart. Journ,
Micros. Se., Vol. xxxvili., p. 465.

Barrows, ANNE I., ’00. The Respiration of Desmognathus. Anat. Anz., Bd. xviii.,
p- 461.

BRAUER, A. ’97. Beitrage zur Kenntniss der Ertwicklungsgeschichte und der
Anatomie der Gymnophionen. Part I., Zool, Jahrb., Anat., Bd x., p. 389.

wee oos batt 11., Zool. Jahrb. Anat., Bd xii.. p. 477:

CHAUVIN, M. v., ’77. Ueber das Anpassungsvermégen von der Larven Salamandra
atra. Zeitschr f. wiss Zool., Bd. xxix., p. 324.

CLARKE, S. F, ’80. The development of Amblystoma punctatum, Pt. I, External.
Studies from the Biol. Lab. of Johns Hopkins Uriv., No. 2, p. 105

Corr, E. D., ’89 ‘The Batrachia of North America. Bull. U.S. Nat. Mus. No. 34.

DENBRUGH, J. van, ’98. Herpetological Notes. Proc. Acad. Nat. Sci. Phila., vol.
XXXVIi., p. 139

EYCLESHYMER, A. C., ’93. The Development of the Optic Vesicles in Amphibia
Journ. Morph., vol. viii., p. 189.

—’95. The Early Development of Amblystoma. Journ. Morph., vol, x., p 241

’06. The Habits of Necturus maculosus. Amer. Nat., vol. xl., p 123.

Gapow, H., ’or1. Amphibia and Reptiles. Camb. Nat Hist, vol. viii.

Gace, S. H., ’o1. The Life History of the Vermilion-spotted Newt. Amer. Nat.,
vol. xxxv., p 1084.

Gé.pig, E. A., ’99. Ueber die Entwicklung von Siphonops annulatus. Zool. Jahrb.,
Syst , Bd. xii, p 170

Hay, O. P., ’88. Observations on Amphiuma and its Young. Amer Nat., vol. xxii,,
P- 315.

HusBarp, M. E, ’03. Correlated Protective Devices in some Californian Salamanders,
Pub. Univ. Calif., Zool. I, p. 175.

Jorpan, E. O., ’91. The Spermatophores of Diemyctylus. Journ. Morph., vol. v.,
p. 263.

—’93. The Habits and Development of the Newt—Diemyctylus viridescens. Journ.
Morph., vol. viii., p. 269.

Kincspury, B.F., ’95. The Spermatheca and Methods of Fertilization in some

American Newts and Salamanders. Proc. Amer. Microscop. Soc., vol. xvil., p.261.

Maurer, F., ’98. Blutgefasse in Epithel. Morph. Jahrb., Bd. xxv., p. 190.

492 TRANSACTIONS OF THE CANADIAN INSTITUTE. [Von VIII.

MontTcomMery, T. H., Jr., ’o1. Peculiarities of the Terrestrial Larva of Plethodon
cinereus. Proc. Acad. Nat. Sci. Phila., vol. liii., p. 503.

——’o3. The Heterotypic Maturation Mitosis in Amphibia and its General Signifi-
cance. Biol. Bull., vol. iv., p. 259.

MoreGan, T. H., ’91. Some Notes on the Breeding Habits and Embryology of Frogs.
Amer. Nat., vol. xxv., p. 753-

’

——’g97. The Development of the Frog’s Egg.

REED, H. D., ’08. A Note on the Coloration of Plethodon cinereus. Amer. Nat.,
vol. xlii., p. 460.

RITTER, W. E., ’97._ Diemyetylus torosus. Proc. Calif, Acad. Sci., Zool. I, 3rd. Series.

03. Further Notes on the Habits of Autodax lugubris. Amer. Nat., vol.
XXXVil., p. 883.

RITTER, W. E. AND MILLER, L., ’99. A Contribution to the Life History of Autodax
lugubris. Amer. Nat., vol. xxxiii.. p. 691.

SEELYE, ANNE Barrows, ’06. The Circulatory and Respiratory Systems of Des-
mognathus fusca. Proc. Soc. Nat. Hist. Boston, Mass., vol. xxxii., p. 335.

SHERWOOD, W.L., 95. The Salamanders found in the vicinity of New York City

with notes upon extra-limital or allied species. Proc. Linn. Soc. of New York.
vol. vii., p. 21.

SMITH, B. G., ’06. Preliminary Report on the Embryology of Cryptobranchus al-
legheniensis. Biol. Bull., vol. xi., p. 146.

’o7. The Life History and Habits of Cryptobranchus allegheniensis. Biol.
Bull., vol. xili., p. 5.

’o7. The Breeding Habits of Amblystoma punctatum. Amer. Nat., vol. xli.,
DSO

WuipPLe, I. L., ’06. The Naso-labial Groove of Lungless Salamanders. Biol. Bull.,
WO; stony JOH Ue

WIEDERSHEIM, R., ’oo. Brutpflege bei: niederen Wirbeltieren. Biol. Centralbl.,
Bd) <x. upp. 304 and: 32m.

Wivper, H. H, ’94. Lungenlosen Salamandriden. Anat. Anz, Bd ix., p, 216.

——’g9. Desmognathus fusca and Spelerpes bilineatus. Amer. Nat,, vol. xxxiii.,
puesk

‘or. The Pharyngo-Cisophageal Lung of Desmognathus fusca. Amer. Nat.,
VOM ex NEV ee Loss

Wricut, A. H., ’08. Notes on the Breeding Habits of Amblystoma punctatum. Biol. _
Bull., vol. xiv., p. 284.

1908-9. ] THE Hasirs OF PLETHODON CINEREUS ERYTHRONOTUS. 493

ILLUSTRATIONS:

Figures 1 and 9 are camera lucida drawings, 2 and 3 from photographs, 4-8 and
10-14 from projections by Zeiss epidiascope. Figures 4-8 are enlarged to the same
extent, which is twice as great as that of figures 10-14. Figures 4 and 5 from eggs of
maximum size, 6 and 7 from eggs of minimum size. The outlines of the feet, Figs. 10,
12 and 13, are camera lucida drawings; they are right feet viewed from the dorsal sur-
face under low magnification.

Figure 1. Horizontal section of tail through the proximal wound of an autotomy.
v—centrum of vertebra.
a—perivertebral adipose tissue.
m—myotome.
c. t.—connective tissue.
s—skin.

Figure 2. Cluster of eggs suspended in a large crack in a decaying log.
Figure 3. Cluster of eggs.

Figure 4 Stage 3mm. Egg viewed from side, the dotted line being the equator
of the egg when the latter is in its natural position.

Figure 5. Stage5 mm _ Egg viewed from above.

Figure 6. Stage 5.5mm. Egg viewed from the side. The outline of the mucous
sphere was added from a living specimen.

Figure 7 Stage6mm Egg viewed from above.

Figure 8. Stage 9 mm. Viewed from above, the head and tip of the tail are
much bent towards the ventral surface.

Figure 9. Vitelline circulation in larva 10 mm. long.
a—anterior

Figures ro and 11. Stage 12 mm.
Figur 12 Stage 16 mm. Taken from a small larva, actual size, 14.5 mm.
Figure 13. Stage 20-25 mm Actual size of larva 20 mm.

Figure 14 Specimen from the centre of the same bunch of eggs as that of Figure
13, which is one of the outer larve

MY

ey
mite

IN A ND LARVA OF AMBYSTOMA JEFFER-

‘

Y AWS HL ‘Prerso

(Repanvren FROM THE | Anenican NATURALIST, VOL. XLIV)

ary

% :
“
" \
4
-
ae
‘ ¥
SA 2
a
*)
t
— Ng
5
Be e
vs
Sree

| 1: PUBLISHED BY ; eS a
HE LIBRARIAN, 1911 ¢ , . = eae

Chairman: ROBERT ALEXANDER FALconer, hea Litt.D., L
Mathes tang tc an i President of the Univers

+f

eee PROFESSOR w. J. “ALEXANDER, Pu.D. ee
a PROFESSOR w. H. Etus, M.A., M. B. |

PROFESSOR A. -Kirscumann, Pu. D,;

PROFESSOR J. qs MACKENZIE, B. A.

LLD. —

General Editor: H. H. pee M. a Renta dy di
. Librarian of the ‘University

;
a

[ Reprinted from TH& AMERICAN NATURALIST, Vol. XLIV., December, 1910. ]

SPAWN AND LARVA OF AMBYSTOMA JEFFER-
SONTANUM

PROFESSOR W. H. PIERSOL

UNIVERSITY or ToRONTO

SPAWN

Amone the various accounts of the habits and spawn
of Ambystoma punctatum occasional mention may be
found of Ambystoma jeffersonianum but always in such
connections as to suggest that A. jeffersonianum is by
far the less common species in the locality. This along
with the considerable similarity existing between the
spawn of the two species may explain why no account
of the spawn of A. jeffersonianum has as yet appeared.
Descriptions of the spawn of A. tigrinum sufficient for
distinguishing it from that of the other two species is
given by B. G. Smith (1907).

In most localities near Toronto A. punctatum is a
much more common species than A. jeffersonianum,
however in one piece of woodland that is quite isolated
from all the others examined, the former species is
rarely to be found, while the latter is very abundant.
This woodland contains four pools that last throughout
the year, although they become heavily choked by vege-
tation during the late summer and autumn. The
value of these pools as a collecting ground for spawn,
Branchippus, ete., was discovered some years ago by
my colleague, Dr. Huntsman and his observations on
the Ambystoma spawn suggested to him the possibility
of distinguishing in it two kinds. Later the writer also
became familiar with this woodland in connection with
observations on Plethodon and with the consent of Dr.
Huntsman undertook also the investigation of the Amby-
stoma spawn of the pools.

The writer first visited these pools in spawning time

732

(33 THE AMERICAN NATURALIST [Vou. XLIV

three years ago and found a small amount of spawn of
a type already familiar to him for some years from its
abundance in pools in other localities. But the greater
amount was of a type that differed from this in the
points detailed below. These two types have proved to
be the punctatum type and the jeffersomanum type,
respectively. The predominance of the latter subse-
quently found its explanation in the fact that 31 of the
33 individuals captured in the woodland since then have
been of the latter species. It is impossible to determine
accurately the proportions in which the two types of
spawn occur, but estimating roughly, the jeffersonanum
type is at least ten times as abundant as the other.

As will appear below, a small percentage of the eggs
of A. punctatum will approach in size, or color, or mode
of deposition—but rarely in more than one of these
points at a time—the eggs of A. jeffersonianum. Con-
sequently, the separation of the latter as a type when
found in a pool where the punctatum spawn greatly
predominates, is not an obvious thing. But when the
proportions are reversed, as in the special pools men-
tioned, the distinction is most easily made. Observa-
tions in the field have agreed in all four seasons and
have been supplemented by the capture of females just
previous to egg-laying and comparison of mature ovar-
ian eggs and eggs laid by them in the laboratory, with
those obtained in the pools; and finally by the rearing
in the laboratory of larve from the two types of spawn.

The points of difference in order of constancy are as
follows:

1. Size—The eggs of A. jeffersonianum are distinctly
the smaller, the usual diameter being 2—2.25 mm.

2. Color—The eggs of A. jeffersonianum are much
the darker, the pigment being but little removed from
a true black and covering a much larger proportion of
the surface of the egg than in A. punctatwm; even the
lower surface is usually as dark as the upper surface of
many of the eggs of the latter species.

3. Time of Laying—The deposition of most of the

No. 528] SPAWN AND LARVA OF AMBYSTOMA 134

spawn by A. jeffersonianum precedes that by A. puncta-
tum by a few days. It has been impossible to visit daily
the pools where the spawn of A. jeffersonianum is most
abundant, owing to their distance from the university;
one pool much nearer has yielded a small amount of it
and has provided more accurate although more scanty
data. In general the deposition of the bulk of the jef-
fersonianum spawn coincides with that of the first punc-
tatum spawn. Variations from this oceur—for instance,
this year the spawn in the single pool just mentioned
followed the above rule, while in the group of four pools
nearly all the jeffersonianum spawn had been deposited
three days before any punctatum spawn appeared; and
to complete the irregularity the last spawn of all to be
deposited was that of A. jeffersonianum. It was in
small quantity and probably all from one female.
(These eggs and the larve from them were unusually
small, the larve seemed vigorous, but could not be kept
alive many days after their own supply of yolk was ex-
hausted.) Another check on the time is furnished by
the spawning of Rana sylvatica. This year—an un-
usually early season—the writer observed the first
deposition of spawn in these pools by the wood-frog. It
began at 10.30 a.m., March 31. Spawn of A. jefferson-
ianum had appeared seven days previously.

4. Spawn-masses.—The typical spawn mass of A.
jeffersonianum is a small one, the number of eggs being
usually about twenty; the extremes encountered have
been small masses of jelly without any eggs and a mass
containing forty-one. A. punctatwm does indeed deposit
masses of spawn containing as few eggs as this, but the
number is usually much larger. The complement of
ripe ovarian eggs carried by two females of average
size was 128 and 161. These are probably representa-
tive numbers and indicate a rather smaller complement
than that possessed by A. punctatum—130 to 225—
(Wright and Allen, 1908) which in turn is much smaller
than that of A. tigrinwn—1,000 or more (Powers, 1907).

5. Hardly less characteristic than the small masses is

735 THE AMERICAN NATURALIST [Vou. XLIV

the manner in which they are frequently to be found at-
tached in succession to long slender twigs, each mass
being usually in contact with its neighbors. A sentence
in one paper on A. punctatum (Wright, 1908), ‘‘one
stem—had within a length of one and a half feet 14
bunches of eggs, 15-20 eggs to the bunch,’’ reads very
much like a description of spawn of A. jeffersonianum.
Many stems so laden have been found each year in the
special pools mentioned. The largest piece in Fig. 1 is
a portion of one of them. The twigs selected by A. jef-
fersonianum are, as a rule, very slender. A. punctatum
will make use of both stout and slender twigs indiffer-
ently, and no small quantity has been found attached to
the margins of leaves and to grass, even in the presence
of such twigs as are generally preferred. Eggs of A.
jeffersonianum have not been found except attached to
twigs or stems of water plants.

The low vitality of much of the spawn of A. jefferson-
ianum is a feature that has been noticed in each year.
No accurate estimate of the proportion that dies has
been made, but judged roughly by the conditions found
in the pools it is probably not overstating the loss to say
that three fourths of the eggs do not live to begin gastru-
lation. The same proportion of loss has occurred in
spawn reared in the laboratory, while spawn of A. pune-
tatum brought from the same pools a little later and
kept under the same conditions has suffered practically
no loss. The egg does not die, as a whole, but cells here
and there precede, the others going on dividing as usual
one or more times, only to die at last. The surface view
of such an egg when death is complete shows an irregular

mingling of minute cells with many others two or three

times as great, and at intervals others even up to eight
or ten times as great, in diameter. These dead eggs
imbibe considerable water, and become very much larger
than the living ones and under natural conditions are
soon infected by fungi; but in the laboratory they have
been kept for weeks and have remained free from it;
showing that death has not been caused by a fungus that

No. 528] SPAWN AND LARVA OF AMBYSTOMA 736

only later becomes visible. All the eggs of a mass
either die or develop properly; one or two of the eggs
may prove exceptions to this, but whatever the defect
may be it involves practically all the eggs of a bunch.
Whether it may extend to all the eggs of a female it has
not been possible to determine. This loss has also been
observed in spawn of A. jeffersonianum from a second
locality and is not likely to be due to any quality of the
water, for in the pools of each locality spawn of A. pwinc-
tatum has been found developing with very little loss,
and that apparently due to infection by fungus. Neither
can it be ascribed to low temperatures from early de-
position, for the earliest is no more liable to ,die than
that which comes later along with or after the spawn of
A. punctatum.

Larva

Spawn of A. jeffersonianum brought to the laboratory
has been allowed to develop and the larve fed until the
larger specimens had attained a length of 30-40 mm.
In these it has been possible to detect a peculiarity of
marking not present in similar larve of A. punctatum.
This peculiarity consists of a massing of dark chromato-
phores into three or four spots placed in a row along
each side of the mid-dorsal line, giving the animal, when
viewed from above, the appearance of being banded
(Fig. 2). Viewed from the side the same can be de-
tected, but is less conspicuous (Fig. 3). Incipient band-
ing 1s often indicated as soon as the chromatospheres
are well differentiated (Wig. 4).

In looking over a large number of larve all gradations
will be found between individuals in which the above
shows distinctly and those in which it is impossible to
detect it. For example, in 115 laboratory-reared larve
examined at one time, 80 (69 per cent.) showed the dis-
tinctive marking. Of the balance, some individuals
under different conditions showed it also (either ex-
treme expansion or contraction of the chromatophores
obscures the pattern), but some never did. Exact num-
bers for this division of the 31 per cent. are not available.

Fic. 1. Spawn of A. jeffersonianum. Natural size. The eggs of some of
the masses are dead, others are in various stages from blastopore to medullary

groove formation.
rT

eo

Fig. 2. Larve of A. jeffersonianum. Enlarged three diameters. Chromato-
phores considerably but not extremely contracted.

Fic. 3. Larva of A. jeffersonianum. Enlarged three diameters. Chromato-
phores considerably but not extremely expanded. The viscera and right side of
the trunk have been dissected away and the photograph taken by both direct
and transmitted light.

lic. 4. Larva of A. jeffersonianum. Enlarged four diameters.

ee

No. 528] SPAWN AND LARVA OF AMBYSTOMA 138

This year an attempt was made to remove all the
spawn of A. punctatum from the special pools. In the
middle of June larve 30-35 mm. long were collected
from them and examined as to this pattern; it was found
in but 35 per cent. Two causes may have contributed
to this—the abundance of brush in the pools may have
caused some spawn of A. punctatum to be overlooked,
and the great expansion of the chromatophores—much
greater than ever attained in the laboratory, the pools
being very dark, probably disguised it in some cases.
It was found impossible to put these larve under ob-
servation in the laboratory to test this point, for owing
to the long journey or to the change of water they in-
variably died within a few hours.

Little importance would have been attached to a point
of coloration so variable as this had it not been found to
be uniformly lacking in similar larve of A. punctatum,
whether raised in the laboratory or taken from the pools
known to contain little, if any, spawn of A. jefferson-
tanum. In view of the range of coloration for A. tigri-
num as larve (indicated by Powers), the degree of con-
stancy noted is perhaps the most that could be expected.

PAPERS CITED

Powers, J. H., 1907. Morphological Variation and its causes in Amblystoma
tigrinum. Univ. of Nebraska Studies, Vol. 7, No. 3.

Smith, Bertram G., 1907. The Breeding Habits of Amblystoma punctatum.
Amer. Nat., 41, pp. 381-390.

Wright, Albert H., 1908. Notes on the Breeding Habits of Amblystoma
punctatum. Biol. Bull., Vol. 14, pp. 284-289.

Wright, Albert H., and Allen, Arthur H., 1909. The Early Breeding Habits
of Amblystoma punctatum. Amer. Nat., Vol. 43, pp. 687-692.

if

vie

hh

|
. University of Toronto. Biological

T68 series
no.1-9
| Biological] | (|
_ & Medical iM |
Serials One
Qt

PLEASE DO NOT REMOVE
CARDS OR SLIPS FROM THIS POCKET

UNIVERSITY OF TORONTO LIBRARY