I

•yr-iX

r ;

•£P \ 0 ^929 y

TRANSACTIONS

ROYAL SOCIETY OF EDINBURGH.

VOLUME LII, PART III.— SESSION 1919-20.

CONTENTS.

ye?

XXIII. Amphicheiral Knots. By Mary Gertrude Haseman, Ph.D.
■ " C. G. Knott, General Secretary.' (With One Plate), -
(issued December 22, 1919.)

Communicated by Dr

XXV. Theoretical Determination of the Longitudinal Seiches of Lake Geneva. By A. T. Doodson,
R. M. Carey, and R. Baldwin, Tidal Institute, University of Liverpool. Communicated
by Dr E. M. Wedderburn, ........

(Issued April 2, 1920.)

PAUB ^

XIX. Contributions toivards a Knowledge of the Anatomy of the Lower Dicotyledons. II. The
Anatomy of the Stem of the Berberidacese. By R. J. Harvey-Gibson, C.B.E., D.L.,

M.A., Professor of Botany, University of Liverpool ; and Elsie Horsman, M.Sc. (With
One Plate), .......... 501

(Issued May 21, 1919.)

XX. Contributions towards a Knowledge of the Anatomy of the Lower Dicotyledons. III. The
Anatomy of the Stem of the Calycanthacex. By Christine E. Quinlan, M.Sc.,
University College, Cork. Communicated by Professor R. J. Harvey-Gibson, C.B.E.,

D. L., M.A. (With One Plate), . ....... 517

(Issued May 21, 1919.)

XXI. The Comparative Myology of the Shoulder Girdle and Pectoral Fin of Fishes. By Captain

E. W. Shann, B.Sc., Oundle School. Communicated by Professor W. C. MTntosh,

F. R.S. (With Four Plates and One Figure in the Text), . . . .531

(Issued September 26, 1919.)

XXII. The Morphology of the Stele of Platyzoma microphyllum, R. By. By John M'Lean

Thompson, M.A., D.Sc., F.L.S., Lecturer on Plant Morphology, Glasgow University.

(With Three Plates and Three Figures in the Text), . . . . .571

(Issued September 29, 1919.)

597

XXIV. On Old Red Sandstone Plants showing Structure, from the Rhynie Chert Bed , Aberdeenshire.

Part 11. Additional Notes on Rhynia Gwynne-Vaughani, Kidston and Lang ; ivith
Descriptions of Rhynia major, n.sp., and Hornea Lignieri, n.g., n.sp. By R. Kidston,

LL.D., F.R.S. , and W. H. Lang, D.Sc., F.R.S., Barker Professor of Cryptogamic Botany
in the University of Manchester. (With Ten Plates), ..... 603

(Issued January 20, 1920.)

629

XXVI. On Old Red Sandstone Plants showing Structure, from the Rhynie Chert Bed, Aberdeenshire.

Part 111. Asteroxylon Mackiei, Kidston and Lang. By R. Kidston, LL.D., F.R.S.,
and W. H. Lang, D.Sc., F.R.S., Barker Professor of Cryptogamic Botany in the
University of Manchester. (With Seventeen Plates), ..... 643

(Issued May 11, 1920.)

EDINBURGH:

PUBLISHED BY ROBERT GRANT & SON, 107 PRINCES STREET,

AND WILLIAMS & NORGATE, 14 HENRIETTA STREET, COVENT GARDEN, LONDON, W.C. 2.

MDCCCCXX.

Price Forty-three Shillings.

( 501 )

XIX. — Contributions towards a Knowledge of the Anatomy of the Lower
Dicotyledons. II. The Anatomy of the Stem of the Berberidacese. By
R. J. Harvey-Gibson, C.B.E., D.L., M.A., Professor of Botany, University
of Liverpool; and Elsie Horsman, M.Sc. (With One Plate.)

(MS. received November 22, 1918. Read January 20, 1919. Issued separately May 21, 1919.)

Introductory.

The present paper is the second of a series of investigations into the anatomy
of the lower Dicotyledons (l), aiming at the accumulation of anatomical data
which, it is hoped, may prove of service in elucidating the phylogeny of the
complex of orders usually regarded as situated somewhere near the base of the
Angiosperm phylum. The continuance of the War has again materially retarded
these investigations, and has also rendered it almost impossible to obtain material
from abroad that would, doubtless, have been available in peace time. We
are greatly indebted for substantial assistance in this relation more especially
to Sir David Prain, F.R.S., Director of the Royal Botanic Gardens, Kew, and to
Sir Frederick Moore, F.R.S., Director of the Glasnevin Gardens, Dublin, and to
them we tender our grateful thanks.

According to Prantl (2), the order Berberidacese includes eight genera —
Podophyllum, Jeffersonia, Diphylleia, Achlys, Nandina, Bpimedium, Leontice, and
Berberis (including Mahonia). To these Lotsy (3) adds Hydrastis and Glaucidium,
which Prantl places among the Ranunculacese. The Lardizabalaceae are included
in the Berberidacese by Solereder (4), as also by Bentham and Hooker (5),
while Berberidopsis, sometimes placed in Berberidacese, is included in Flacourtiacese
by other authors. In the following pages seven of the eight genera of Berberi-
dacese, as enumerated by Prantl, and representatives of four genera of the
Lardizabalacese, viz. Lardizabala, Stauntonia, Akebia, and Decaisnea are discussed.

For detailed treatment the common Barberry, Berberis vulgaris , L., has been
selected, and its anatomy is compared with that of several other species of the
genus. This has provided a basis on which to 'found a discussion of the other
genera referred to, and has suggested certain general remarks on the relationships
of the various types dealt with, both to each other and to members of other orders.

The- Anatomy of the Stem of Berberis.

Berberis vulgaris is a shrub with small simple leaves, with solitary spines
formed by prolongation of the mid-ribs. Occasionally two additional spines

TRANS. ROY SOC. ED1N., VOL. LII, PART III (NO. 19). 77

are

502

R. J. HARVEY-GIBSON AND ELSIE HORSMAN ON

developed as marginal outgrowths towards the apex of each leaf. The flowers
are yellow and trimerous, suggesting an affinity with Monocotyledons.

The most prominent feature shown in a transverse section of the stem is the
deeply-seated ring of cork cells with remarkably large lumina. The tissues are
arranged in the following sequence (figs, f , 2) : — ■

1. An epidermis of regular- rectangular cells with cuticle and numerous
stomata.

2. One or two layers of thin-walled cells with intercellular spaces. In older
stems these cells persist in that condition, but in other cases they are crushed by
the growth of tissues more centrally placed.

3. These cells are bounded internally by a ring of sclerotic tissue three or
four layers thick. In very young stems this ring is not quite continuous, but is
interrupted by thin-walled cells, usually in the interfascicular regions, agreeing in
this respect with the similar ring found in the Papaveracese (l). In older stems
the sclerotic ring increases considerably in thickness and complete continuity is
established, though later it is again segmented into patches as a result of the
tension set up by the growth of the internal tissues and as a preliminary to being
cast off by the cork.

The individual fibres are long, and their endings are either truncate, oblique,
or tapering. They all show simple pits. Occasionally this tissue, by hypertrophy,
produces ridges, although other species show this to a much greater extent than
B. vulgaris. Solereder considers this ring as belonging to the outer pericycle.

4. Immediately within this sclerotic band is the deeply-seated cork, which at
first consists of one layer of regular, large, thin-walled cells with remarkably wide
lumina. These cells are two, three, or more layers deep in older stems.

5. Next follows a region of cortical cells, the outer being small and compact,
the inner larger and having intercellular spaces. All are thin-walled and have
protoplasmic contents. In the medullary ray region, “berberin” is present in
the form of yellowish-green granules, and often the whole cell contents are im-
pregnated with it ; some cells in this region contain also klinorhombic crystals of
calcium oxalate, which almost completely fill the cell lumina (fig. 3). Berberin is
an isoquinoline alkaloid, insoluble in water, and hence can be best seen in sections
of fresh material mounted in water. Spirit in which Berberis material has been
preserved becomes yellowish-brown owing to the dissolved “ berberin,” which
remains as a resinous deposit on evaporation. Further research is at present
being carried out on the chemical composibion and reactions of this substance.

According to Solereder, the fibrous ring, the cork, and the cortical region
just mentioned form a composite pericycle.

No endodermis is differentiated, for ordinary cells of the cortex abut directly
on the vascular bundles.

In the young stem the vascular bundles are distinct, the interfascicular regions

THE ANATOMY OF THE STEM OF THE BERBERIDACEiE.

503

being composed of sclerotic tissue ; but later the bundles, together with the broad
primary medullary rays, form a complete ring of lignified tissue (fig. l).

The phloem consists of very small elements arranged in radial rows. In older
stems it shows sieve tubes with bevelled ends occupied by sieve plates and callus,
with slime strings across the plugs from segment to segment ; but no grouping
of sieve plates into areas is observable. This may be due to the very small size
of the sieve tubes, which also prevents the definite location of plates on the lateral
walls, though they appear to be present. Some of the sieve tubes are accompanied
by companion cells, although this is by no means an invariable rule. Phloem
parenchyma is abundant, and the cells show intercellular communications. Ligni-
fied fibres with deep simple pits are also present. They are short and spindle-
shaped, and in cross section show strongly thickened walls and ■ small lumina.
thus differing from the pericycle fibres, which are much wider and have walls less
strongly lignified. These bast fibres are formed in tangential rows in the inner
secondary phloem and border the lamellae into which the phloem is divided as
growth proceeds, thus recalling the alternating bands of Vitis, though they are
developed to a much less extent. A few scattered fibres are also -found in the
outer primary phloem, but no stone cells are present.

Crystals and “berberin” granules are found in the medullary ray region of
the bast, as above mentioned. The bast fibres are not continued across these
regions.

In the. early stages the xylem elements are regular and similar in size and shape,
with wTide lumina, some showing remains of protoplasmic contents. Later the
lignification becomes more pronounced, the lumina decrease, and the arrangement
in radial rows becomes more obvious. Comparatively large polygonal vessels are
distinguishable in the protoxylem and in groups towards the centre of each
vascular bundle. In the protoxylem the first formed vessels are spirally thickened,
the innermost spirals being right-handed and later ones left-handed, as pointed out
by De Bary (6). The other vessels possess bordered pits with slit-like apertures
(fig. 5), the slits lying at right angles to the axis of the vessel ; many have
faint spiral thickenings in addition to pits, similar to those recorded for Romney a
(Papaveraceae) (l) and for Dendromecon by Leger (7). Most of the xylem consists
of prosenchymatous fibres with simple pits on all their walls and often containing
starch, particularly in the winter condition. These fibres always occupy the
regions abutting on the medullary rays. Sanio (8) says that septate fibrous cells
are found only in the first annual ring. This does not seem to be invariably true,
although the septa are not so frequent in the fibres of the second and subsequent
years. Comparatively few tracheids are present, and these have bordered pits
accompanied by spiral thickenings, as in Taxus. There seems to be little difference
between the pitted fibres and the pitted tracheids. Spiral tracheids resembling-
fibres are also present.

504

' R. J. HARVEY-GIBSON AND ELSIE HORSMAN ON

The predominance of stereome and the reduction in the number of conductive
elements is to be correlated with the xerophytic character of the plant. The
small amount of leaf surface, and the consequent reduction in transpiration, the
necessity for protecting the cell contents from the intense insolation to which
Berberis is subjected in its native habitats (Asia, South and Central America),
may also be associated with the predominance of sclerosis and the large develop-
ment of wood fibres.

Broad primary , medullary rays persist, and consist of markedly sclerotic ceils
which, in transverse section, closely resemble the xylem elements^ but are often
larger than the adjacent xylem fibres. The rays are three or four cells wide, and
the cells have rounded ends. They are elongated radially, the radial diameter
being in many, cases almost twice the tangential one. All the walls show numerous
simple pits, and most of the cells are well filled with starch grains during the
winter. No secondary rays are produced. The contents of the medullary ray
regions of the cortex have been previously described.

The medulla is heterogeneous ; the cells near the protoxylems are smaller than
the rest and -have well-thickened walls. They show nuclei when young, and starch
grains in the winter condition at all ages. Numerous simple pits occur on all their
walls. The central cells are larger and become polygonal by mutual pressure ; they '
are thin-walled and devoid of contents.

The petioles are slender and wiry. In the main portion of the petiole there are
four vascular bundles surrounded by sclerenchyma, but these unite as the lamina
is approached. In the basal region there are only three vascular bundles, which
later on divide into four. As the petiole expands into the lamina the vascular
bundles again rearrange themselves into three bundles.

Other Species of Berberis.

Most of the species of Berberis exhibit features similar to those described for
Berberis vulgaris , among which the following structural variations may be noted.
Spines occur on the stems of some species. They are outgrowths from single
epidermal cells, and are present in B. Yunnanensis, Franch., B. stenophylla x hort.,
B. angulosa, Wall., B. Wallichiana, DC., B. empetrifolia , Lam., and B. Darwinii,
Hook. (fig. 6). In the last-mentioned species they are much longer, are curved, and
have transverse partitions.

The sub-epidermal parenchyma in B. aristata, DC., B. stenophylla x hort.,
B. empetrifolia , Lam., and B. aquifolium, Pursh., shows foldings in the cell walls
as described for Corydalis raeemosa (l), but in the present instances the folding
appears to be caused by the pressure of internal tissues acting in opposition to the
resistance offered by the epidermis. In the older stems the compression increases
until these cells are cast off along with the other, tissues external to the deeply-
seated cork.

THE ANATOMY OF THE STEM OF THE BERBERIDACEH3.

505

The sclerotic pericycle in some cases is increased so much as to form ridges in
the stem, e.g. in B. angulosa, Wall. (fig. 8), and B. Yunnanensis, Franch.

The spines so characteristic of the Berberidaceae consist mainly of sclerenchy-
matous fibres and are obviously modified leaves, as shown by their position at the
nodes and by their structure. The sheath of sclerenchyma almost entirely encloses
two cavities filled with loosely arranged thin-walled cells and one or more vascular
bundles (fig. 9). Stomata are found only on the incurved surface, where they are
abundant. The sclerotic fibres are continuous throughout the spine to the tip, which
is formed solely of fibres covered by epidermis.

The inner layers of the heterogeneous pericycle, which consist of thin-walled
parenchyma, are often so loosely arranged as to produce conspicuous cavities crossed
by cellular bridges (fig. 10). In some species these layers are increased locally,
contributing to the production of the stem ridges.

Resin ducts are present in the phloem in some cases. In transverse section they
appear to consist of single large cells with lignified walls showing oval pits (fig. 11).

In the young stem of B. Jamesoni, Lindl., vessels and tracheids with com-
paratively large lumina predominate, but as the stem increases in age and girth
lignified fibres -are developed bordering the medullary rays, as in B. vulgaris.

According to Haberlandt (11), “medullary rays always abut directly against
xylem parenchyma on their flanks as well as on their upper and lower borders.”
This suggests that the prosenchymatous fibres which always border the medullary
rays may be parenchyma cells modified to give additional support to the stem
(fig. 12). They are well supplied with simple pits, while some show septa and many
contain starch grains.

Several species show oval bordered pits longer than those of B. vulgaris, e.g.
B. Sieboldii, Miq. Others show bordered pits in conjunction with faint reticulate
markings, e.g. B. Wallichiana, DC., and B. Yunnanensis, Franch. The former
also shows loose reticulate markings in vessels without pits.

A few species show a slight development of normal xylem parenchyma, e.g.
B. stenojphylla x hort., B. Jamesoni, Lindl., B. Sieboldii, Miq. The parenchyma
when present always surrounds the larger vessels in the protoxylem regions, and is
usually lignified and similar to the xylem fibres in transverse section.

In B. angulosa, Wall., the greater portion of the wood consists of tracheids with
round bordered pits resembling those of Pinus, but smaller in size.

Solereder states that scalariform markings occur occasionally in Berberis ; no
such thickenings were found in any of the species of Berberis examined, though
they are to be found in Podophyllum and Diphylleia.

There is very little variation in the medullary rays ; most of them closely
resemble those of B. vulgaris. The ray cells are well lignified, thus rendering them
hardly distinguishable from the surrounding xylem. They vary in width from two
to eight cells, the widest being found in B. Wallichiana, DC., where they have

506

R. J. HARYEY-GIBSON AND ELSIE HORSMAN ON

very abrupt endings. In the very few cases where secondary rays have been found,
they are by no means uniform in their occurrence, a transverse section showing
perhaps only three or four secondary rays to a large number of primary ones. The
presence of secondary rays is always associated with wide bundles and narrow
primary rays.

The pith in the majority of cases is heterogeneous, as in B. vulgaris , L. Many
species show solitary klinorhombic -crystals of calcium oxalate in addition to starch
in the peripheral cells, e.g. B. aquifolium, Pursh., B. stenophylla x hort., B. Jame-
soni, Lindl., B. empetrifolia, Lam., B. Darwinii , Hook.

Some species show a pith sclerotic throughout, but here also the central cells
are often larger than the peripheral ones, e.g. B. Darwinii , B. angulosa.

The general plan of structure of the stem in Berberis ( Mahonia ) aquifolium,
Pursh., is the same as that of other species ; a sclerotic pericycle and cork cells with
wide lumina are present ; the xylem shows the same predominance of lignified
fibres and the same broad primary medullary rays. The bordered pits have round
apertures, as in B. angulosa , in contrast to the slit-like pits most frequent in the'
genus. No lignified fibres occur in the bast. The pith is sclerotic throughout.

Summary of Anatomical Characters of Berberis.

1. There is a continuous’ sclerotic ring of pericycle fibres, similar to that present
in the Flacourtiacese.

2. Cork cells with remarkably wide lumina occur, as in some Menispermacese.

3. Bast fibres occur in tangential rows in the secondary phloem, agreeing in form
and arrangement with those of Lauracese and some Magnoliacese.

4. The xylem is composed chiefly of strongly lignified fibres with occasional thin
transverse septa, as in Lauracese and Flacourtiacese.

5. The largest vessels and some of the tracheids exhibit bordered pits with slit-
like apertures, while most species show a combination of bordered pits and spiral
thickenings.

6. Very little parenchyma occurs in the xylem, in which particular it again
agrees with that of the Menispermacese.

7. Broad lignified primary medullary rays are characteristic of the genus, with
only very occasional secondary rays. In the Menispermacese the rays are similar
in all respects, save that some cells in the broad rays may not be lignified.

8. The pith is usually heterogeneous, though in some species the cells are
sclerotic throughout. Heterogeneous pith also occurs in Magnoliacese, some
Lauracese, and Menispermacese.

9. The phloem consists of sieve tubes, companion cells, phloem parenchyma, and
bast fibres ; the two former occur principally in the primary phloem, while the
secondary phloem consists mainly of parenchyma and bast fibres.

THE ANATOMY OF THE STEM OF THE BERBERIDACEYE.

507

10. Spines occur both, as epidermal outgrowths from the stem and as modified
leaves.

The Epimedese include Epiniedium, Yancouveria, and Nandina, and doubtfully
also Jeffersonia and Achlys. With the exception of Nandina, none of the genera
investigated are woody.

Nandina domestica, Thunb., is a Japanese shrub. In the young stem there are
distinct ridges formed by the development of sclerenchyma, especially in the
neighbourhood of the vascular bundles. The vascular bundles are well separated
by small cells which afterwards form medullary rays, 5-8 cells in width, in the
older stages. The ridges gradually disappear owing to the development of scler-
enchyma levelling up the depressions. The four or five large vascular bundles have
rather smaller ones alternating with them, but all traces of this distinction disappear
in older stems.

When in the young state the phloem areas are distinctly rounded in section and
are destitute of lignified fibres. The xylem consists mainly of vessels with wide
lumina, together with a few parenchymatous cells, while simple pitted fibres border
the medullary rays, as in Berberis. The vessels show a variety of markings : spiral
in the protoxylem, in other parts reticulate ; slit-like pits merging into reticulate
and bordered pits with slit-like apertures. Occasionally tracheids with bordered
pits are also present.

When young most of the pith consists of thin-walled cells, but later it becomes
entirely sclerotic. The pith is relatively large, comprising about 80 per cent, of
the area in the transverse section of an old stem. Some of the cells are elongated
and show simple pits ; they appear to act as storage organs. Some near the
periphery show peculiar forked endings.

Vancouveria hexandra, C. Morr. and Dec. — The species possesses an underground
stem or rhizome in addition to the normal aerial stem.

The aerial region in the young state has a slightly cutinised epidermis followed
by two or three rows of cells which are elongated longitudinally and show no inter-
cellular spaces. Next follow three or four rows of small sclerotic fibrous cells forming
a distinct ring, but merging into larger thin-walled cortical parenchyma, as a rule
devoid of contents. Through this tissue run widely separated vascular bundles. In
the young condition more than half of the bundle is composed of phloem without
fibres. The xylem consists entirely of spiral and pitted vessels with comparatively
wide lumina, while the protoxylem is flanked by small sclerotic cells in older stems.

The xylem of the vascular bundles in the rhizomic region is so embedded in
sclerotic tissue that together they form a complete ring of stereome, except in some
old specimens where the continuity is interrupted and the internal and external
parenchymatous areas are united by broad bands of non-sclerotic tissue. The
phloem occurs in patches outside the stereome ring, and consists mainly of sieve
tubes and companion cells. The sieve plates are in some cases transverse, in others

508

R. J. HARVEY-GIBSON AND ELSIE HORSMAN ON

oblique, while some show sieve plates on their external walls. No lignified fibres
occur. The vessels occupying the centre of the bundles show various forms of
lignification — spiral, reticulate, or bordered slits. On the lateral borders of these
groups of vessels are thickened prosenchymatous fibres which occupy the whole of
the interfascicular regions, thus completing the ring of mechanical tissue.

Outside the phloem areas are occasional sclerotic cells among the ordinary cortical
cells, and beyond these lies a ring of slightly thickened cells corresponding to the
sclerotic ring in the aerial stem. The rhizome thus contains two concentric rings of
mechanical tissue — an inner one consisting of the xylem of the vascular bundles and
sclerotic fibres, and an outer of sclerotic tissue only. The epidermal cells are thin-
walled, and the sub-epidermal tissue consists of fairly large and irregular cells, many
of which contain a brown deposit.

All the material examined wras herbaceous and showed no signs of any develop-
ment of cork.

Epimedium alpinum, L., shows a structure very similar to that of Vancouveria,
but some of the vascular bundles are more peripheral, so that their phloems are
embedded in the peripheral sclerotic ring. The medulla is markedly fistular.
Many of the cortical cells are elongated with large oval pits in their walls, and
hence are probably conductive.

In the rhizome the ring of mechanical tissue is incomplete, the bundles being
separated by broad plates of parenchyma. In the aerial stem there are two rings
of vascular bundles.

Epimedium sagittatum agrees with Vancouveria so far as the structure of the
aerial stem is concerned.

Jeffersonia dubia, Benth. and Hook., and Jeffersonia binata, Bart., agree with the
three previous genera in having their vascular bundles arranged in a ring, some of
them partially embedded in the ring of sclerenchyma ; in the structure of the xylem
— having vessels with bordered slit-like pits, in having a lacunar cortex, and in having
a phloem very rounded in section. In Jejfersonia dubia the sieve tubes are well
developed, with well-marked callus plugs and faint canals across the callus, and the
nodes on the median lamella. Solereder says the arrangement of bundles in
Jeffersonia is Monocotyledonous ; but, apart from the fact that some bundles are
smaller and rather nearer the periphery than others, the material we have examined
presents no evidence in support of this statement.

The chief characters of stem anatomy of Bpimedese may be summarised as
follows : —

1. There is a continuous ring of sclerenchyma, pericyclic in origin.

2. The phloem bundles are very rounded in outline and have no fibres. The
sieve tubes are larger than in Berberis, and in many cases show sieve plates on the
lateral walls.

3. The xylem consists chiefly of vessels with fairly wide lumina. Those of the

THE ANATOMY OF THE STEM OF THE BERBERIDACEH3.

509

protoxylem are spirally thickened, the others have reticulate or bordered slit-like
apertures.

4. Prosenehymatous fibres occur only in Nandina and in the rhizome of Van-
couveria ; in the other genera xylem parenchyma replaces fibres.

5. Except in Nandina and in the rhizome of Vancouveria the bundles remain
isolated, and the interfascicular tissue is unlignified.

6. The central parenchyma in Nandina becomes sclerotic, but in Vancouveria
and Epimedium only the cells in close proximity to the protoxylem show sclerosis.
In all save Nandina the cortex tends to become lacunar.

7. The vascular bundles are arranged in a ring, although developmentally con-
structed from two rings interlocking.

Achlys trijphylla, DC. — In the aerial stem the bundles are irregularly arranged
(fig. 13), and only a few are in contact with the ring of sclerotic tissue, which is
similar in position to that in Vancouveria, Epimedium, and Jeffersonia. The smallest
bundles are toward the periphery and the largest ones toward the centre, but no
definite arrangement in rings is observable. The general appearance recalls the
arrangement in Podophyllum rather than that in the Epimedese.

The xylem consists chiefly of spiral vessels without fibrous elements. The
phloem is rounded in section and consists of sieve tubes with occasional companion
cells. The sieve tubes show sieve areas on their lateral walls.

The bundles in the rhizome are arranged in two rings, as noted by Tischler (9).
The main bundles form an inner ring, while the others, which consist chiefly of
sclerenchyma, alternate with them and are nearer the periphery. The cork cells
resemble those of Berberis ; they have wide lumina, are deep-seated and of pericyclic
origin. The xylem contains pitted vessels and tracheids. Pericyclic fibres do not
occur in the rhizome.

The Podophyllese include Podophyllum and Diphylleia, and also, according to
Lotsy, Jeffersonia, Achlys, Hydrastis, and Glaucidium, although the two latter genera
are placed by most' systematists in the Ranunculaceae. Podophyllum Emodi, Wall.,
very closely resemble P. peltatum in anatomical structure, as described by Holm (10).

Stems of different ages show irregularly arranged bundles, widely separated
from each other, the largest being central (fig. 14). With the exception of the
cambium present in each bundle, the stem in section has the appearance of a
normal Monocotyledon.

In the bud stage no sclerosis is present in the cells surrounding the bundles, but
later a crescentic mass of sclerenchyma arises on the phloem side. Towards the
periphery there is a ring of sclerenchyma in which the phloem of the outer bundles
is embedded, as in Epimedium and Jeffersonia.

At first only a few of the xylem elements are lignified and show spiral markings,
later formed units show annular and reticulate thickenings. Xylem parenchyma
occurs but no fibres* The phloem appears to consist solely of sieve tubes with

TRANS. ROY. SOC. EDIN., YOL. LII, PART III (NO, 19).

78

510

R. J. HARVEY-GIBSON AND ELSIE HORSMAN ON

lateral plates and companion cells ; as in P. peltatum, no fibres and no parenchyma
are present.

The material examined was herbaceous and no cork was noticeable, although it is
given by Solereder as sub-epidermal in origin.

The rhizome shows a distinct ring of large bundles, with a few smaller ones
outside in the cortex, but the arrangement is not as definite as "in Achlys. There is
no peripheral ring of pericyclic fibres such as is shown by the rhizome. In the main
bundles the xylem consists of a few scalariform tracheids, wide vessels showing a
variety of markings — spiral, loosely reticulate, long oval pits, and bordered slits, —
and many parenchyma cells showing no lignification. The phloem is very little
differentiated and consists of radially arranged elements, a few of which are sieve
tubes, while most seem to be parenchymatous in their nature. The tissue bordering
on both phloem and protoxylem consists of smaller cells which are slightly sclerotic,
with simple pits in their walls. The central parenchyma cells, which are slightly
elongated, contain resin.

The petiole resembles the aerial stem in having a ring of bundles towards the
periphery and irregularly scattered bundles in the centre, formed from the bundles
nearest the upper surface of the leaf.

Diphylleia cymosa, Michx., is very similar to Podophyllum in anatomical
structure. In the aerial region there is the same ring of sclerenchyma towards
the periphery and a ring of bundles with their phloems partly embedded in the
sclerenchyma. The other vascular bundles are scattered irregularly, as in Podo-
phyllum, and have sclerenchyma on the outer margins of their phloems. The
phloem and xylem elements, as also the anatomy of the rhizome and petiole, agree
with the description given for Podophyllum.

The chief characteristics of the stem of the Podophylleee : —

1. The' bundles are scattered, the larger being towards the centre, as in Mono-
cotyledonous stems, while the outer and smaller bundles have their phloems embedded
in a sclerotic ring.

2. In the aerial axes the main bundles are in the form of a ring with smaller ones
more peripherally placed.

3. The xylem consists chiefly of vessels with annular, reticulate, and scalariform
thickenings and long, oval bordered pits.

4. No prosenchymatous fibres were found, but unlignified parenchyma cells are
present among the vessels.

5. The central parenchyma is practically homogeneous, although a few cells near
the protoxylems are slightly sclerotic.

In comparing the stem anatomy of the Berberidese, Epimedese, and Podophyllese
it may be noted —

1. Thai all three subdivisions are characterised by possessing a ring of pericyclic
sclerotic fibres in their aerial stems.

THE ANATOMY OF THE STEM OF THE BERBERIDACE2E.

511

2. That the remarkably wide-lumined cork cells which arise on the inner border
of the composite pericyle are present in Berberidese and in the aerial stems of some
of the Podophyllese, but are absent in the Epimedese and in the aerial stems of
Podophyllese, where the cork is normal in form and sub-epidermal in origin.

3. That in the aerial axes of Podophyllese the bundles show a “ scattered ” arrange-
ment, as in Monocotyledons, but that in the subterranean regions the chief bundles
form a definite ring, while others are scattered in the cortex. In Epimedese they are
arranged in a ring which appears to be formed by the interlocking of two rings of
bundles. In Berberidese the bundles exhibit the normal Dicotyledonous arrangement.

4. That bast fibres are present in tangential rows in the secondary phloem of
Berberidese, but are absent from the Epimedese and Podophyllese.

5. That the rigid texture of the Berberidese is produced by the relatively large
amount of prosenchymatous fibres, not present in herbaceous Epimedese (though in
the rhizomes of Nandina and Yancouveria), nor in the Podophyllese ; while unlignified
parenchyma is present among the vessels of Epimedese and Podophyllese, but not of
Berberidese.

6. That the vessels in all three groups possess bordered pits with slit-like
apertures, their greatest diameter being always at right angles to the axis of the
vessel. In the Berberidese these are often accompanied by spiral or reticulate
thickenings, but this combination does not occur in the Podophyllese and Epimedese.

7. That broad primary medullary rays consisting of lignified cells link up the
xylem of the vascular bundles into a complete stereome ring in some, whilst in
others (except Nandina and Vancouveria rhizomes) the bundles are widely separated
by undifferentiated parenchymatous cells.

8. That the central parenchyma is similar in all three groups, being composed
usually of central thin-walled cells, with thick-walled cells towards the protoxylems.
In both Berberidese and Epimedese some species show uniformly sclerotic parenchyma.
The Berberidese alone show crystals in the central parenchyma.

9. That the sieve tubes in all genera have sieve plates on their bevelled ends, and
occasionally on the lateral walls as well.

10. That the endodermis is not well marked, and that stone cells are conspicu-
ously absent.

In stem anatomy Hydrastis canadensis does not resemble Podophyllum, for the
vascular bundles are not scattered and the ring of sclerotic pericycle, which is
invariably present in all the Berberidaeese examined, is entirely absent. The
anatomical characters suggest that this species is not closely allied to the
Berberidaeese.

The Lardizabalaceas are regarded as a tribe of Berberidaeese by Bentham and
Hooker. They are included in that order also by Solereder, but separated from it
by Engler and Prantl and by Lotsy. This order comprises Lardizabala, Stauntonia,
Akebia, Decaisnea, Holboellia, Parvatia, and Boquila.

512

R. J. HARVEY-GIBSON AND ELSIE HORSMAN ON

Stems of the following have been examined, viz : — Lardizabala biternata, Ruiz
and Pav., Holbcellia latifolia , Wall., Akebia quinata * Dec., and Decaisnea Fargesii,
Franch. They all show marked differences from the Berberidacese, but many of
these differences may be accounted for by their climbing habits. For instance,
vessels of the xylem are much more numerous and have considerably larger lumina
than any Berberidacese, while prosenchymatous fibres are very few, being replaced
by xylem parenchyma which is often lignified and shows bordered pits with slit-like
apertures. Thus the supporting tissue which predominates in woody Berberidacese
is replaced in the Lardizabalacese by conductive elements. In common with many
other climbers these plants show large sieve tubes. In Decaisnea the sieve plates
occur on the bevelled ends, the polygonal sieve areas being separated by bands of
cellulose, while sieve fields also occur on the lateral walls. Phloem fibres are absent,
but on the external margins of the phloem areas are semi-lunar bands of sclerenchyma
with very thick walls and packed with calcium oxalate crystals ; they appear to be
of pericyclic origin and correspond to the ring in Berberidacese. The ring in this
case, however, is much more irregular since the sclerenchyma is formed in depres-
sions in the interfascicular regions, producing the ridged appearance so characteristic
of lianes.

The cork arises sub-epidermally and is normal in character.

Akebia shows a well-marked endodermis, the cells being thickened on three sides
and thin-walled on the inner, which abuts on the parenchyma.

In many respects the Lardizabalacese appear to form a link between the
Berberidaceae and the Menispermacese.

The Lardizabalacese and Menispermacese consist chiefly of climbing shrubs, whilst
the Berberidacese are not climbers.

1. The cork in Lardizabalacese is sub-epidermal in origin, as also in some Berberi-
dacese and Menispermacese.

2. The pericyclic sclerenchyma occurs in arcs outside the phloem bundles, and
since the bundles are close together, the arcs of sclerenchyma unite more or less into
a ring. In this respect the Lardizabalacese differ from the Berberidacese, where the
ring is continuous, and resemble the Menispermacese, which do not possess a con-
tinuous ring but have sclerotic arcs round the vascular bundles connected by groups
of stone cells.

3. Bast fibres are absent from the phloem of Lardizabalacese and Menispermacese,
but present in woody Berberidese.

4. The xylem of Menispermacese consists chiefly of vessels with large lumina, and
this is true also of Lardizabalacese, whilst in Berberidacese prosenchymatous fibres
predominate.

5. All three orders agree in having broad primary medullary rays. In all the
non-herbaceous species of Berberidacese the medullary ray cells are well lignified.
In some species of Lardizabalacese all the cells are strongly lignified, but in others

THE ANATOMY OF THE STEM OF THE BERBE RIDACE2E.

513

the cells in. the centre of the rays are thin- walled, flanked by sclerotic cells.
The Menispermacese agree with the Lardizabalacese in often having unlignified
medullary rays.

Berberidopsis. — Hallier describes Berberidopsis as a syncarpous Berberis with
points in common with the Flacourtiacese, while Engler and Prantl and Bentham
and Hooker place it definitely in the Flacourtiacese.

An examination of the stem structure of Berberidopsis corallina, Hook., presents
the following characters (fig. 15) : —

1. A thick cuticle is present on the epidermal cells.

2. The external cortex often contains stellate crystals of calcium oxalate.

3. The ring of sclerenchyma is very open and is interrupted in older stems. The
cells are short and rectangular and have numerous simple pits in their thick walls
(figs. 16, 17). These sclereides correspond to the stone cells in the cortex of young
twigs of Acer and iEsculus, but in these two genera the sclereides accompany
sclerenchymatous fibres, the latter being much more numerous, whereas in Berberi-
dopsis the sclereides predominate and very few fibres occur. Haberlandt (ll)
suggests that these sclereides are cortical parenchyma cells which have penetrated
gaps produced in the mechanical cylinder of bast fibres as a result of tensions set up
by the growth in thickness of the twigs. The cortical parenchyma cells afterwards
become thick-walled and so restore the unity of the now composite mechanical
cylinder. The varied shape and arrangement of the sclereides seem to support
this view.

4. The cork arises sub-epidermally and is normal in character, the cells being
rectangular.

5. The xylem contrasts markedly with that of Berberis, the elements are larger,
and the main portion consists of large vessels which are slightly thickened and have
very loose reticulations forming large irregular pits (fig. 18). Tracheids are present
with oval bordered pits.

6. Most of the cortical cells are thin-walled, but a few near the protoxylems are
sclerotic, and all of them* show simple pits. The more peripheral cells often contain
starch grains or crystals of calcium oxalate.

7. No bast fibres are present in the phloem. The information available as to
the anatomy of the Flacourtiacese is at present very meagre, so that a detailed com-
parison with the Berberidacese is not possible. The following points of resemblance
may, however, be noted : —

1. In the Flacourtiacese the sclerotic ring may be continuous, or in patches
joined by bridges or broken up completely into islands.

2. In both the xylem shows wide reticulate thickenings.

3. The medullary rays are numerous and vary from one to four cells in width in
both families, but in Berberidacese the rays are wider and less numerous.

514

R. J. HARYEY- GIBSON AND ELSIE HORSMAN ON

Summary of the Anatomical Characters of the Berberidacea;.

1. Scattered vascular bundles occur in the aerial steins of the Podophyllse, the
largest bundles being centrally placed. Save that cambium is present, the vascular
anatomy closely resembles that seen in Monocotyledonous stems.

2. Bordered pits occur very frequently in the secondary wood, both in vessels
and tracheids ; the apertures vary in shape, some being circular, others more or less
oval with slit-like apertures. Where circular bordered pits occur in tracheids they
closely resemble those of Pinus, although both tracheids and pits are smaller than
in the Coniferse.

3. The combination of a double spiral thickening and bordered pits is of common
occurrence in the woody species of the order; while bordered pits combined with
reticulate thickenings occur in some (conf. Taxus).

4. The xylem parenchyma is as a rule unlignified. Some species show a few
lignified cells among the vessels of the protoxylem, but in none are there continuous
plates or areas of lignified parenchyma.

5. Lignified fibres are common in the phloem of the woody species.

6. The cork cells have remarkably wide lumina and are very deeply seated,
arising from the inner layers of the pericycle. This is true only of woody species
and of the rhizomes of some herbaceous forms.

7. There is a continuous ring of sclerotic fibres in the pericycle.

8. Sclereides are absent from the Berberidacese, unless Berberidopsis be included
in the order.

9. In some woody species spines are present, either epidermocortical or foliar
in origin.

10. No endodermis is distinguishable save in Akebia among the Lardizabalacese.

BIBLIOGRAPHY.

(1) Harvey-Gibson and M. Bradley, “The Anatomy of the Stem of the Papaveracese,” Trans. Roy. Soc.

Edin., vol. li, 1916.

(2) Engler and Prantl, Die naturlichen Pfianzenfamilien, 1891.

(3) Lotsy, Vortrage fiber botanische Stammesgeschichte, vol. iii.

(4) Solereder, Systematic Anatomy of Dicotyledons, Oxford, 1908.

(5) Bentham and Hooker, Genera Plant arum.

(6) De Bary, Comparative Anatomy of Phanerogams and Ferns, Oxford, 1884.

(7) Leger, “Recherches aux I’appareil vegetatif des Papaveracees,” Bull. Soc. Linn. Norm., 1891.

(8) Sanio, Botanische Zeitung, 1864.

(9) Tischler, “ Berberidaceen nnd Podophyllaceen,” Bot. Jahrb., 1902.

(10) Holm, “ Podophyllum peltatum,” Botanical Gazette, 1899.

(11) Haberlandt, Physiological Plant Anatomy, 1914.

(12) Hill, “Histology of Sieve Tubes of Angiosperms,” Annals of Botany, xxii, 1908-

THE ANATOMY OF THE STEM OF THE BERBERIDACEiE.

515

EXPLANATION OF PLATE.

Fig. 1. Transverse section, stem of Berberis vulgaris (diagrammatic).

Fig. 2. Transverse section, young stem of B. vulgaris, showing deep-seated ring of large cork cells,
x 450.

Fig. 3. Transverse section, stem of B. elegans, showing calcium oxalate crystals in medullary ray
cells. x 450.

Fig. 4. Longitudinal section, stem of B. vulgaris. Sieve tubes showing sieve plates and callus
plugs, x 450.

Fig. 5. Longitudinal section, stem of B. vulgaris. Xylem and medullary ray. x 450.

Fig. 6. Epidermal cells and hairs of B. Darwinii. x 40.

Fig. 7. Transverse section, stem of B. aristata, showing folding of the walls of sub-epidermal
parenchyma, x 450.

Fig 8. Transverse section, stem of B. angulosa (diagrammatic).

Fig. 9. Transverse section, spine of B. setnensis (diagrammatic).

Fig. 10. Longitudinal section, stem of B. empetrifolia. External cortex, showing bridging of inter-
cellular spaces. x 450.

Fig. 11. Transverse section, stem of B. Bosclianii (diagrammatic).

Fig. 12. Longitudinal section, stem of B. Wallichiana, showing short blunt medullary rays. x 450.
Fig. 13. Transverse section, rhizome of Achlys triphylla (diagrammatic).

Fig. 14. Transverse section, stem of Podophyllum Emodi (diagrammatic).

Fig. 15. Transverse section, stem of Berberidopsis corallina. x 450. '

Figs. 16 and 17. Transverse section and longitudinal section, stem of B. corallina. Stone cells of
pericycle. x 450.

Fig 18. Longitudinal section, stem of B. corallina. Xylem elements with loose reticulate markings.
. x 450.

Fig. 19. Transverse section, rhizome of Vancouveria hexandra (diagrammatic).

Fig. 20. Transverse section, stem of Hydrastis canadensis (diagrammatic).

Fig. 21. Transverse section, stem of Jeffersonia dubia (diagrammatic).

Vol. LII.

Trans. Roy. Soc. Edin.

R. J. Harvey-Gibson and Elsie Horsman on “ The Anatomy of the Stem of the Berberidace;«.”

E. H. del.

( 517 )

XX. — Contributions towards a Knowledge of the Anatomy of the Lower
Dicotyledons. III. The Anatomy of the Stem of the Calycanthacese.
By Christine E. Quinlan, M.Sc., University College, Cork. Communicated
by Professor R. J. Harvey-Gibson, C.B.E., D.L., M.A. (With One Plate.)

(MS. received November 22, 1918. Read January 20, 1919. Issued separately May 21, 1919.)

The present paper is the third of a series of researches on the anatomy of the
lower Dicotyledons undertaken with the object of seeing what, if any, aid anatomy
might give in determining the point of origin of the so-called Monocotyledons. It
deals more especially with the anatomical and histological peculiarities of the stem
of the Calycanthacese.

Historical Summary.

Mirbel (l), in 1828, drew attention to the occurrence of four cortical vascular
bundles in the stem of Calycanthus jloridus, L., each consisting of phloem and
xylem, inversely orientated, and with an arc of sclerenchyma on the outer margin.
In 1833, Gaudichaud (2) found that, of the three petiolar bundles, the large median
one enters the normal vascular ring, while the lateral bundles join the nearest
cortical bundles; and his work was confirmed by Treviranus (4). In 1836,
Lindley (3), and, later, Henfrey (5), discovered this was common to all the
Calycanthacese. In 1860, Woronin (6) described the course of the cortical vascular
bundles and their relation to the foliar bundles, confirming Gaudichaud as regards
their mode of fusion, and pointing out that, at the node, a series of anastomoses
occurs : (l) between the central vascular ring and each of the cortical bundles,
(2) between the two opposite pairs of cortical bundles, (3) between the median and
the lateral foliars at the base of the petiole, and (4) between the median foliar and
the cortical bundles of the axis. In 1884, Lignier (7) showed that the median
petiolar bundle gives off lateral branches which go to form the cortical bundles of
the internode immediately superior, and pass out at a higher node.

In 1885, Her ail (8) stated that the peripheral bundles of the stem of the
Calycanthaceee were pericyclar in origin and eventually become pushed out into
the cortex, and that the peripheral bundles of each internode supply the lateral
bundles of the leaves situated at a lower level, the median one coming directly
from the central cylinder and receiving a branch from the lateral bundles. In
1887, Lignier denied the pericyclar origin of these bundles and stated that the
bundles originate, and are always situated, in the cortex. In 1904, Van Tieghem (9)
held that in Chimonanthus the peripheral bundles were pericyclic, but in all species
of Calycanthus purely cortical.

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 20).

79

518

CHRISTINE E. QUINLAN ON

While investigators have paid attention chiefly to the peculiar anatomy of the
Calycanthaceae, the histological features have been generally overlooked ; the present
paper deals with the structure of the stem more especially from the histological
standpoint, emphasising features not usually met with in Dicotyledons.

According to Prantl (10), the order Calycanthaceae includes two genera:
Eucalycanthus and Chimonanthus. The former genus is primarily a native of
North America, and comprises three species, viz. Calycanthus jioridus, L. , C. fertilis,
Walt., and C. occidentalism Hook and Arn. Chimonanthus is a native of Japan, and
is monotypic, i.e. Chimonanthus fragrans , Lindl., a few varieties of which are in
cultivation in Europe, var. grandiflora being the best known.

Calycanthus fertilis has been chosen for detailed examination ; the other species
will be referred to only in so far as they differ from it.

A transverse section through an internode of a very young stem (fig. l) of
Calycanthus fertilis shows the following arrangement of tissues : centrally a green
heterogeneous pith, the cells becoming smaller towards the periphery,; the vascular
bundles arranged in an oval, in groups of four or five, and separated by well-marked
medullary rays especially" between the bundle groups. The xylem is not so well
developed on the flanks of the oval. Outside the phloem is a small-celled tissue,
representing the pericycle and endodermis, though neither of these layers is in-
dividually distinguishable. The cortex is differentiated into an inner layer of thin-
walled green parenchyma with prominent intercellular spaces, and an outer layer
of collenchyma. Some of the parenchyma cells of the cortex are differentiated into
oil-secreting cells. In the cortex lie four vascular bundles with inverse orientation
of the phloem and xylem and with a meristematic layer between. Each bundle is
bordered externally by an arc of slightly thickened cells, which, later, become so
sclerotic that their cavities are almost completely obliterated. External to the
collenchyma is a unilamellar epidermis, bearing numerous elongated, unicellular hairs,
whose walls are impregnated with silica. Solereder (ll) states that the cell bases
are also silicified, but this I am unable to confirm. The hairs are often bent at
right angles at their bases.

The larger central cells of the pith are faintly pitted and exhibit distinct fold-
ings on their radial and longitudinal walls. The protoxylem consists of parenchyma,
in which are scattered spiral, annular, and reticulate vessels and large spiral
tracheides. The spiral thickenings are single. The cambiform cells are barely
distinguishable, but, here and there, may be seen to consist of about two layers
of cells not so elongated as the xylem parenchyma. The phloem consists chiefly of
phloem parenchyma. Sieve-tubes and companion cells are present, but the sieve-
plates are very inconspicuous. The companion cells are prominent and contain
nuclei and granular contents. Usually three layers of parenchyma cells contain-
ing chlorophyll are intercalated between the phloem of the central -cylinder and

THE ANATOMY OF THE STEM OF THE CALY C A NTHACEiE.

519

the phloem of the peripheral bundle.. The sieve-tubes in the latter are very
numerous, but the phloem parenchyma is less abundant than in the central stele.
Sieve-plates usually occur on the end walls, but one or two occur on the lateral
walls also. The xylem of the peripheral bundle consists of spiral, annular, and
reticulate vessels, and spiral tracheides, but there is no xylem parenchyma. On
the outer side of the xylem lies a mass of elongated prosenchyma whose cells
undergo sclerosis as age increases. Externally are several layers of cortical pitted
parenchyma, while scattered in the parenchyma are large spherical oil-secreting
cells (fig. 2). Outwardly the cortical parenchyma merges gradually into four or five
layers of subepidermal collenchyma' containing a few chloroplasts. Silicified bent
unicellular hairs, their points directed towards the apex of the stem, lie deeply
embedded among the epidermal cells. Stomata appear to be entirely absent.

As the shoot elongates, the outer cells of the pericycle become uniformly
thickened and sclerotic in small isolated groups (fig. 5). These cells must not be
confused with the sclerotic cells with U-shaped thickenings which one sees clearly
in the layer continuous with the sclerenchymatous arc of the peripheral bundle.
The fibres on the outer margin of the peripheral bundles also become strongly
lignified, and the pits become . extremely narrow canals. The sieve-tubes increase
greatly in number, and tracheides with bordered pits and spiral thickenings appear.
In the normal cylinder, the phloem parenchyma increases in volume and constitutes
the greater part of the bast ; pitted wood parenchyma is also present. The cortex
and pith remain unaltered.

In the older stems all the vessels and tracheides of the xylem, as also the xylem
parenchyma, become greatly lignified. Two kinds of tracheides are met with, viz.
long, narrow tracheides with pointed ends, whose walls bear uniseriate bordered pits,
and larger but shorter tracheides with two or three longitudinal rows of bordered
pits. In some, the pits become so elongated transversely that they appear scalari-
form. The bordered pits are either in longitudinal rows or scattered irregularly.
Here and there, among these elements, are elongated, strongly lignified fibres
without pits.

The medullary rays, seen in transverse section, are very numerous and uniseriate,
and from three to five cells deep. The cell walls have numerous simple pits, or
bordered pits on the side abutting on a tracheide ; in a few cases pits are absent.
Tyloses are frequent in the large pitted vessels. Most of the tracheides are inter-
mediate between the elements with ordinary bordered pits and the scalariform
tracheides so characteristic of Pteridophyta.

The secondary phloem shows a large development of phloem parenchyma and
comparatively few sieve-tubes, while the sieve-tubes in the peripheral bundles are
numerous and are accompanied by relatively little phloem parenchyma. The
sclerotic cells of the peripheral bundles have greatly thickened walls, as have also
the pericyclic fibres. The secondary cells are conspicuous, and show slight traces

520

CHRIST] NE E. QUINLAN ON

of lignification. The cortical parenchyma,, seen in transverse section, is arranged
in definite layers, which readily separate on sectionising. The individual cells are
connected by distinct pits in the walls. These pits are seen everywhere in the
ground tissue, and become more noticeable as the tissue increases in age. The
collenchyma cells are much thickened, and towards the periphery become flattened
and merge into the subepidermal phellogen. The cork cells become elongated
radially and show marked foldings on the radial walls, probably due to the pressure
exerted by the growth of the internal tissues. Deposits of tannin are found in the
dead epidermis, and particularly around the bases of the silicified hairs. Lenticels
of the ordinary type are common.

The floral axis exhibits the same structure anatomically as the ordinary stem.
On approaching the thalamus, the normal vascular ring divides into about fourteen
distinct bundles. The pedicel bears two small opposite leaf-like bracts. The first
pair of sepals are also opposite and decussate with the bracts. The leaf-trace bundles
of the latter are similar in structure to those described below for the ordinary leaves.
On nearing the thalamus the cortical bundles bifurcate, and one branch divides
again. At this stage two of the vascular bundles of the central cylinder jut out at
two opposite points into the peripheral ring, becoming the median vascular cords
of the first sepals. Anastomosis occurs between the contiguous members of the
last division of the peripheral bundles on the flanks. The occurrence of these
anastomoses in the peripheral ring is very irregular ; several may be seen in one
axis, and none in another. When the first sepal is given off, it receives a median
bundle from the central vascular ring, and two laterals from the peripheral, ring ;
but, before entering the sepal, anastomosis takes place between the median and the
lateral bundles, and between the lateral bundles themselves. Similar anastomoses
occur at the points of origin of all sepals, petals, and stamens. The vascular cords
supplying the carpels arise from the central stele only. They receive no strands
from the peripheral ring. Every time a branch of a peripheral bundle passes out
into a floral segment, the xylem and phloem, which were inversely orientated in the
pedicel, become normally arranged by a revolution through an angle of 180°, while
the sclerenchyma disappears from the peripheral bundles as they approach the
thalamus.

The mode of entry of the leaf-ttace bundles in C. fertilis is complicated. In
the petiole, there may be five or seven bundles, one large median and one or
two smaller laterals on each side. The laterals gradually decrease in number by
joining the median or the extreme laterals, till there are, eventually, only three left.
This is the condition found at the base of the petiole near its insertion on the axis.
The vascular ring of the stem opens, and, as the leaves are opposite, two gaps
occur at each node. The foliar gaps occur between the peripheral bundles. The
median foliar bundle enters the central vascular system through the gap, and is
accompanied by the vascular cords from the two axillary buds, which arrange them-

THE ANATOMY OF THE STEM OF THE CALYCANTHACEiE.

521

selves crescentwise, each joining the adjacent margin of the median bundle, thus
forming a more or less continuous arc. This gradually approaches, and finally
becomes inserted in, the central cylinder. Before doing so, however, the median
foliar bundle gives off two lateral strands, one on either side. Each lateral strand
divides into two segments, one of which joins the lateral foliar bundle, the other
unites with the peripheral bundle of the axis. The lateral foliar bundle joins the
peripheral bundle of the axis a little lower down, and has no connection with the
central stele. Immediately after the median foliar bundle has joined the central
cylinder and the laterals have united with the peripheral bundles, anastomosis takes
place between the flanking- peripheral bundles. The vascular strands of the foliar
bundles are normally orientated, but the lateral segments, on approaching the
peripheral bundles of the stem, become inverted. The arc formed by the median
on entering the stem becomes less and less conspicuous as the node merges into
the internode below, where the central cylinder becomes cylindrical once more.

The stem of Calycanthus jioridus shows very little difference in external appear-
ance from that of C. fertilis.

In transverse section it also resembles in all essential respects C. fertilis. The
normal vascular bundles fuse to form a complete ring by an early development of
an interfascicular cambium, leaving only narrow medullary rays between. These
rays are only one cell wide, even in -the young stem, and contain chloroplasts.
The cambium ring is very conspicuous and four or five layers in thickness. Both
in the normal vascular ring and in the peripheral bundles, crystals of calcium
oxalate are deposited in radiating masses, either attached to the cell membrane
or lying free in the cells. Double spirals are often met with both in the protoxylem
of the central stele and of the peripheral bundles. Tracheides with bordered pits
occasionally occur in the primary xylem. The sieve-tubes are very indistinct in
the central bundles, but are clearly differentiated in the phloem of the peripheral
bundles. Companion cells with dense contents are very prominent.

The secondary xylem consists mainly of reticulate tracheides with uni- or multi-
seriate bordered pits, or with scalariform markings. The secondary xylem also
includes lignified pitted parenchyma cells. Sanio (12) states that the walls of
these cells become broken down by the disintegration of the membrane to form a
mucilage or gelatinous layer which turns violet on treatment with iodine and bright
crimson with phloroglucin and hydrochloric acid. As in C. fertilis , there are also
lignified prosenchyma fibres without markings or pits, interspersed among the other
elements, probably mechanical in function.

The peripheral bundles in older stems are surrounded by a layer of slightly
sclerotised cells. The collenchyma cells have thick walls with canal-like pits.
The walls of the cortical cells are also thickened, but not to so great an extent.
The pericycle consists of isolated groups of sclerotic fibres between the phloem of
the central stele and the parenchymatous cortex. Between the groups some of

522

CHRISTINE E. QUINLAN ON

the cells are thickened on the inner and lateral walls, and thus form a hetero-
geneous pericycle. The sclerotic groups themselves, however, never fuse to complete
the pericycle, even in an old stem. In the old stem, traces of lignin are found in
the walls of the oleiferous cortical cells. Secondary thickening of the cortex has
its origin in the subepidermal region or the “ exodermis ” of Van Tieghem, and
the radial walls of the cork cells show well-marked foldings.

The pericyclic origin of the peripheral vascular bundles is seen more clearly in
C. fioridus than in C. fertilis. It has been shown that the cells with U-shaped
thickenings occur between, and in some places in a direct line with, the uniformly
sclerotic cells, and so form part of the pericycle. The sclerotic cells with U-shaped
thickenings are also found in the layer of cells surrounding the peripheral bundles
on the outside. Since the peripheral bundles are seen to be inserted between the
outer layer of the pericycle and the phloem of the central stele, it may be concluded
that they are pericyclic in origin.

The leaf-trace insertion on the whole resembles that of C. fertilis. The number
of vascular bundles in the petiole varies, but is usually four or, more commonly,
five. In no case did I find three bundles, as recorded by Gaudichatjd. As a rule,
in each petiole, there is one large median bundle and two smaller laterals on either
side. The median bundle enters the vascular ring through the gap, but before
doing so it gives off laterally two strands which join the lateral bundles, before
the latter unite with the peripheral bundles of the axis. The lateral foliar bundles
immediately' flanking the median bundle unite sooner or later with the extreme
lateral ones. The latter often unite with the peripheral bundle of the stem, before
receiving the strand from the median foliar ; and in some cases there may be a
simultaneous fusion of the strands from the median foliar bundle, the lateral foliars,
and the peripheral bundle of the axis. At a slightly lower level anastomosis takes
place between the peripheral bundles on the flanks of the node. The axillary bud
receives a very minute strand from the peripheral bundle of the stem. It passes
inside the lateral foliar bundles and gives off a smaller strand to the younger bud
on the outside.

The thalamus on examination shows a more regular arrangement of the bundles
than in C. fertilis. Each of the peripheral bundles divides into two, and the number
remains constant till the perianth segments begin to come off. After this many
divisions take place which cannot be followed. Each sepal, petal, and stamen receives
a median strand from the central stele and two laterals from the peripheral ring.
There is always anastomosis between the median and the lateral bundles, but there
is no transverse commissure between the laterals. The method of distribution is
the same for all the floral segments except the carpels, which receive vascular
cords from the central cylinder only.

The young stem of C. occidentalis in transverse section is practically identical
with C. foridus, save that the endodermis is well defined and is easily recognised

THE ANATOMY OF THE STEM OF THE CALYCANTHACEiE.

523

by the abundance of starch grains in its cells. This endodermal layer encircles
the so-called cortical bundles on the outside, showing clearly that they are pericyclic
in origin. The pericycle consists of isolated groups of fibres. The cortex divides
as usual into two layers, the inner containing numerous oleiferous cells. The
epidermis bears silicified hairs, though not so numerous or so well developed as in
C. fertilis or C. floridus.

In longitudinal section the primary xylem consists of annular, spiral, and
reticulate vessels, and spiral tracheides, while tracheides with uni- or multiseriate
bordered pits are very frequent in the secondary wood. Scalariform tracheides are
not common. In the older wood the different varieties of tracheides give place
almost completely to the peculiar tracheides described in C. floridus and compared
by Lignier to those found in Taxus. Tyloses are frequent in these tracheides. The
xylem parenchyma is small in amount and the medullary rays are uniseriate, con-
sisting of cells containing abundant starch. They may be from twelve to sixteen
cells deep. In the xylem their walls are pitted ; in the phloem the walls are slightly
thickened but bear no pits, and merge gradually into the pericyclar parenchyma.

In a one-year-old stem the sieve-tubes are scarcely distinguishable, either in the
phloem of the central cylinder or in the peripheral bundles, though they are readily
seen in both tissues of a second-year stem. In all cases, the fusiform companion cells
are prominent. De Bary (13) says, in describing the phloem of the peripheral bundles :
“ It only consists of soft bast and in the main, at least, of parenchymatous elements ;
sieve-tubes still remain to be sought for.” Herail (8) makes no mention of sieve-
tubes in the peripheral bundles, though Lloyd W. Williams (14) states that, in
comparing the bast of the peripheral bundle with the bast of the normal vascular
ring, both contain sieve-tubes and companion cells, but the sieve-tubes in the former
are far more numerous than in the latter. I found sieve-tubes in both tissues, and
I agree with Lloyd Williams as to the great difference in the number of sieve-tubes,
and I would further draw attention to the difference in the size of the sieve-plates
themselves.

Oleiferous cells occur frequently in the parenchymatous cortex but not elsewhere,
though Lignier states that they occur in the pith and bast, and even in the cork
underneath the old epidermis. He also states that the walls of the oleiferous cells
are both lignified and suberised. Only in a few cases could I find traces of lignin.
The periderm is formed in the secondary stem by the activity of an exodermal
cambium. The walls of the cork cells outside the phellogen show marked foldings.
Tannin deposits are found in the dried epidermis, and Lignier also finds resin. The
cambial layer in the central cylinder is not so clearly marked as in C. floridus , and
the same is true of the cambium of the peripheral bundles. The pith is hetero-
geneous, and its small lignified peripheral cells contain large starch grains. The
vascular tissue, both of the central cylinder and the peripheral bundles, contains
abundant radiating needles of calcium oxalate. Lignier finds them also in the

524

CHRISTINE E. QUINLAN ON

pith, where, he says, at least one needle may be seen in every cell. I found no
trace of crystalline masses in the pith.

Lignier regards the irregularly thickened pericyclic fibres as an inner lignified
layer of cortical parenchyma, and the Uniformly sclerotic cells as the bast fibres.
This can scarcely be correct. Both layers undoubtedly belong to the pericycle,
as is evidenced by the fact that a well-defined endodermis is seen immediately
outside. Both kinds of fibres form, in cases where they are developed, a layer
continuous with the arc of sclerenchyma of the peripheral bundle. Herail states
that the peripheral bundles, after differentiation from the pericycle, pass out into the
cortex and remain there. This is not the case in any species of Calycanthus, even
where the endodermis is not visible, since the sclerotic crest of the bundle is seen
to be in direct continuity with the layer of pericyclic fibres ; and where the endodermis
is visible, it is situated on the outer, and not on the inner side of each bundle.

In the older stem the peripheral bundle, seen in transverse section, resembles in
shape a wide sector of a circle, with the angle directed outwards. The rounded
angle is filled with sclerenchymatous tissue ; this crest is connected at each side
with one of the isolated groups of pericyclic cells. It is in this region that the
irregularly thickened cells most frequently occur ; they are rarely seen interspersed
among the ordinary pericyclic fibres, as is the case in C. jloridus. The xylem
occupies a semicircle immediately within the sclerenchyma, and consists of elements
like those seen in the younger tissue of the central stele. There are no reticulate
tracheides with bordered pits, nor wood parenchyma. The phloem forms two wings
on the inner side, and is separated from the xylem by a Y-shaped layer of cambium
which gives rise to secondary tissue. The phloem is arranged almost in two strands ;
in fact, Van Tieghem regards them as separate strands united by a common mass of
xylem and sclerenchyma. Groups of sclerotic cells occur between the peripheral
bundles and the central stele, among pericyclar parenchyma. The latter might
easily be mistaken for cortical parenchyma, were it not that the endodermis stands
out prominently after treatment with iodine, and completely separates the two layers.

A series of sections through the node shows the usual ramification of vascular
strands. In addition, the peripheral bundle of the stem gives off a definite strand
which passes inside the lateral foliar bundles towards the axillary buds. Here the
inner bud receives the main strand, and gives off one very minute branch to the
outer bud. These, later, become the peripheral bundles of the axillary branches,
and always remain connected with the mother axis at the node.

The stem of Chimonanthus fragrans differs very little from that of Calycanthus.

A transverse section of a one-year-old stem shows the usual arrangement of
vascular tissue. The epidermis bears numerous silicified hairs, which are short and
stout, and inserted by broad bases between the ordinary epidermal cells. Very few
show the bend so characteristic of Calycanthus. The cortex is divided into the usual
two layers, one collenchymatous, the other parenchymatous. Here and there,

THE ANATOMY OF THE STEM OF THE CALYCANTHACEHd].

525

glandular cells occur, but they are not nearly so numerous as in Calycanthus. The
cells of the innermost layer of the parenchyma contain abundant starch grains, and
represent an endodermis, though in other respects they do not differ from the outer-
lying cells. The endodermis surrounds the outer border of the peripheral bundles,
leaving no doubt as to their pericyclic origin, though Lignier regards them as
cortical. The peripheral bundles are elliptical in outline, and consist of an internal
phloem followed externally by xylem, and a mass of sclerenchyma. The phloem is
well developed, but does not form wings as in Calycanthus. The xylem is poorly
developed. It consists entirely of annular, spiral, and reticulate vessels and a few
spiral tracheides, there being no xylem parenchyma. Double spirals occur both in
the vessels and the tracheides. The pericyclic layer consists of isolated groups of
uniformly sclerotic fibres, situated immediately inside the endodermis. Cells with
irregular thickenings are rarely seen in a first-year stem. Occasionally glandular
cells are found in between the cortical stele and peripheral bundles. The phloem
consists almost entirely of parenchyma, sieve-tubes and companion cells being few
in number. Frequently, on the inner sides of the parenchyma cells bordering the
inner margin of the peripheral bundles, radial foldings occur, usually extending about
half way across the cell. Glandular cells are frequent in the bast, and the cambial
layer is barely distinguishable. In the older parts tracheides with multiseriate
bordered pits are found, but the pits are only faintly developed. Abutting on the
protoxylem is a small-celled chlorophylliferous pith, merging into a central large-
celled colourless pith. Oleiferous cells are found towards the periphery, but are
rare, though Van Tieghem states that they do not occur in the pith. Medullary rays
are uniseriate and resemble closely those found in Calycanthus.

The pericyclic ring in stems two or more years old is completed by irregularly
thickened sclerotic cells which fill up the gaps between uniformly sclerotic fibres,
while the secondary xylem consists mainly of lignified fibres in which the tracheides
are embedded. The commonest type of tracheide is that with bordered pits and
spiral or reticulate markings, while tyloses are often found partly blocking their
lumina. Xylem parenchyma is of rare occurrence in the old wood, though it is
fairly abundant in the younger tissues.

The leaf-trace insertion differs slightly from that occurring in Calycanthus.
There are three foliar vascular bundles, one large median, and two smaller lateral.
Two gaps are formed on opposite sides of the vascular ring of the axis (a) imme-
diately above the point of insertion of the petiole. Two buds are seen in the
axil ; the inner is large, the outer very small. At a lower level the inner bud
receives a strand from the peripheral bundle of the axis ( b ), and the vascular tissue
of the axillary buds arranges itself into two semicircles on either side. These fuse
by their adjacent ends. A gap occurs first on the inner side, then on the outer side,
and each half fuses with the vascular tissue of the central axis. Meanwhile, the
peripheral bundles become elongated tangentially, and receive the lateral foliars as

TRANS. ROY. SOC. EDIN., VOL. fill, PART III (NO. 20). 80

526

CHRISTINE E. QUINLAN ON

they approach the axis (c). On the flanks, the peripheral bundles become elongated
tangentially (d), and the median foliar bundle approaches and gives off two lateral
strands which diverge and fuse with the peripheral bundles. The latter separate
and gradually contract till they reach their normal size. The median foliar finally
becomes inserted in the central vascular ring ( e , f), and is eventually lost in the
lower internode. The lateral foliar bundles, before joining the peripheral bundles,
become inversely orientated. The same inversion takes place in the strands which
leave the median foliar bundle to fuse with the peripheral bundles. The latter
differ from those of Calycanthus in that they retain their sclerenchyma throughout
the node.

Summary.

The most noteworthy anatomical peculiarities in the structure of the stem of the
Calycanthacese are as follows : —

1. The occurrence of two basipetal axillary buds.

Two axillary buds occur in all the species of both genera. In Calycanthus,
they expand into leafy shoots almost simultaneously, the outer being later in
development. In Chimonanthus, the opening of the outer bud is delayed till the
stem is in its fifth or sixth year.

2. The occurrence of four vascular bundles in the pericycle.

These bundles were found in all the species examined. Herail states that they
are situated in the cortex of Calycanthus, and that in Chimonanthus they originate
in the pericycle and move out later into the cortex. Lignier holds that they are
both cortical, while Van Tieghem regards the bundles in Calycanthus as cortical,
and those of Chimonanthus as pericylic. So far as Chimonanthus is concerned he
is correct, but in Calycanthus an endodermis encircles the peripheral bundles along
their outer margins, and hence they must be pericyclic in origin. In C. fertilis and
C. jioridus the layer of pericylic sclerotic cells is continuous with the arc of scleren-
chyma bordering the bundles exteriorly, again demonstrating the pericyclic origin of
these bundles. As already stated, the bundles are inversely orientated.

3. The presence of a heterogeneous pericycle.

The pericycle consists of two kinds of fibres, viz. uniformly sclerotic fibres, and
sclerotic cells wdth U-shaped thickenings. The latter are few in number in Caly-
canthus, but are very numerous in Chimonanthus. In Calycanthus, the groups of
uniformly sclerotic fibres never fuse to form a continuous pericycle. In that genus,
the irregularly thickened cells occur either in isolated groups or joined to the groups
of uniformly thickened fibres. In Chimonanthus, the gaps between the groups of
uniformly thickened cells are filled with the second type of sclerotic cells in the
second and third years.

THE ANATOMY OF THE STEM OF THE CALYCANTHACEHi.

527

4. The pith is heterogeneous.

The periphery of the pith consists of a small-celled tissue containing chloroplasts.
The central cells are larger, and after the first year lose their contents.

5. Foldings in cell walls.

The radial walls of the first-formed cork cells are elongated, and, almost immedi-
ately after formation, show marked foldings (fif. 15), especially in the region outside
the peripheral bundles. In Chimonanthus, the phloem of the central stele is crushed
beneath the peripheral bundles, and the inner tangential walls show foldings.
Foldings were also observed in the pith of C. fertilis.

6. The occurrence of oleiferous cells.

Glandular cells occur in the cortical parenchyma of all species of Calycanthus.
In Chimonanthus they are found in the bast, the pericyclic parenchyma, and in the
pith, as well as in the cortical parenchyma. In old stems the walls are slightly
lignified.

7. The occurrence of pits on the walls of the cortical cells.

Pits invariably occur on the walls of the cortical cells both in the parenchyma
and collenchyma.

8. The presence of unicellular silicified hairs.

These hairs occur on the epidermis. They are long and bent near the base in
both Calycanthus and Chimonanthus.

9. The occurrence of scalariform tracheides.

Scalariform tracheides similar to those occurring in Pteridophyta are found in
C. fertilis, C. jloridus, C. occidentalism and Chimonanthus fragrans var. grandifiora.

10. The occurrence of spiral or reticulate tracheides with bordered pits in the
secondary xylem.

These are found in every species examined, except C. fertilis. In C. occidentalis
they are very abundant, to the exclusion of all other elements except the libriform
fibres. The bordered pits may be uniseriate or multiseriate. The lumina of these
tracheides are often blocked by tyloses.

1 1 . Multiseriate bordered pits.

Tracheides with multiseriate bordered pits are frequent in C. fertilis, C. foridus,
and Ch. fragrans var. grandifiora. Here tyloses are also frequent.

12. Double spirals in the primary tissues.

Double spiral thickening is seen on the walls of the vessels of C. jloridus and
Ch. fragrans ; they also occur on the tracheides of the latter.

13. The occurrence of lateral sieve-plates.

Lateral sieve-plates were found in C. fertilis, C. occidentalis, and Ch. fragrans
var. grandifiora.

528

CHRISTINE E. QUINLAN ON

14. Anastomosis at the nodes.

At the node the vascular system is complicated. In all species there are trans-
verse strands uniting the peripheral bundles on the flanks. The median foliar bundle
at the base of the petiole gives off lateral strands which unite with the lateral foliar
bundles. The peripheral bundles give off a minute branch which enters the axillary
buds, in all species except C. fertilis, in which it appears at a later stage of
development.

15. Peculiarity of leaf- trace entry.

The vascular ring of the axis opens to admit the median foliar bundle, while the
lateral foliars unite with the nearest peripheral bundles. Each of the axillary buds
possesses a vascular ring, which divides into two crescentic masses. The crescents
belonging to the outer bud fuse laterally with the respective crescents of the inner
bud. In Calycanthus, each set unites with the arc of the median foliar bundle, before
its insertion into the ring of the central stele ; in Chimonanthus, they unite laterally
with the central vascular tissue on each side of the gap, separating on the outer side
to admit the median foliar bundle.

16. Inversion of the lateral foliar bundles at the node.

On entering the mother axis at the node, the lateral foliar bundles, when fusing
with the peripheral bundles, become inversely orientated. This change in orientation
takes place also in the lateral strands given off by the median foliar bundle.

17. Anastomosis in the floral axis.

In C. fertilis and C. floridus, each floral segment, except the carpels, receives
from the central vascular ring a median strand, and from the peripheral ring, two
smaller lateral cords. The median cord gives off two branches which join the lateral
cords ; in C. fertilis, there is also anastomosis between the lateral cords. The carpels
receive branches from the central cylinder and not from the peripheral ring.

I should like to express my thanks to Professor P. J. Harvey-Gibson, C.B.E., of
Liverpool University, at whose suggestion the work was undertaken, for his constant
help and advice ; to Professors B. E. Duke and Hartog of University College, Cork,
and also to Professor H. H. Dixon, F.R.S., and Sir Frederick Moore, for supplies
of material.

BIBLIOGRAPHY.

(1) Mirbel, Ann. des Sciences Nat., t. xiv, ser. 1, 1828.

(2) Gatjdichaud, Arch, de Bot., t. ii, 1833.

(3) Lindley, Nat. Syst. of Bot., 1836 ; and Vegetable Kingdom.

(4) Treviranus, Bot. Zeit., 1847.

(5) Henfrey, Ann. of Nat. Hist , vol. i, ser. 2.

(6) Woronin, Bot. Zeit., 1860.

THE ANATOMY OF THE STEM OF THE CALYCANTHACEiE.

529

(7) Lignier, Bull. Soc. Bot. de France , t. xxxi, 1884.

(8) Herail, Ann. Sci. Nat., ser. 7, t. ii.

(9) Van Tieghem, Ann. Sci. Nat., ser. 8, t. xix, 1904.

(10) Prantl, in Naturl. Pflanzenfam., iii Teil, Abfc. 2, 1888.

(11) Solereder, Systematic Anatomy of the Dicotyledons.

(12) Sanio, Bot. Zeit., 1863, p. 105.

(13) De Bary, Anat., Eng. ed., p. 584.

(14) Lloyd Williams, “Sieve Tubes of G. occidentalism Ann. Bot., vol. viii.

(15) Harvey-Gibson and Bradley, “Anatomy of Papaveracese,” Trans. Roy. Soc. Edin., 1916.

(16) Harvey-Gibson and Horsman, “Anatomy of Berberidaceae,” Trans. Roy. Soc. Edin., 1919.

(17) Knight, “Anatomy of Magnoliacese,” Ann. Bot., 1915.

EXPLANATION OF PLATE.

Fig. 1 . Transverse section through internode of a very young stem of C. fertilis (L).

Fig. 2. Longitudinal section through parenchymatous cortex of G. fertilis (H).

Fig. 3. Transverse section through region below node of C. fertilis after elongation (L).

Fig. 4. ps-tZ, -series of sections through thalamus of G. fertilis (L).

Fig. 5. a-d, series of sections through node of C. fertilis (L).

Fig. 6. Longitudinal section of secondary stem of C. floridus, showing U-shaped cells and fibres of
peripheral bundle (H).

Fig. 7.. Successive sections of node of Oh. fragrans, showing leaf-trace entry (L).

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 20).

81

Vol. LII.

Trans. Roy. Soc. Edin.

Miss Christine E. Quinlan on “ The Anatomy of the Stem of the Calycanthacese.”

^ as* <> ^

■*/*// "■- -77'

rrm’n iTiTmTrr\'VTfv - -

• >-£-
>. -y-’r --

if ?

: I m ‘

if€i'

i ■ /■ 5

»1

iv

! : S.

^7f ~"'h ' r« V;.f

"1 ft ' i'h

4 $ h

.2.

:V.(

w '

( f

I A *t> .'5

i '■ g '! 'S

» ■'<./-■ f

, ■# W

3} 4%

A. ^

^ & -t v &

/ /

/

\ ' * ■
V! S' y

3?

\

v ft -ri ;

If# Jf

; :

/ # m
> nf- sc? (to

/ &

ni m

\ 9 ''

<z\ -

i /

r:

€#■

;>

■s \

) .5%P '

# e, , %

; #1

t f

f £

® ’ K

%

.

%r, / \%

.a

®> _ #

•\ -A

v

C. E. Q. del.

( 531 )

XXL — The Comparative Myology of the Shoulder Girdle and Pectoral Fin of
Fishes. By. Captain E. W. Shann, B.Sc., Oundle School. Communicated by
Professor W. C. MTntosh, F.R.S, (With Four Plates and One Figure in the Text.)

(MS. received October 16, 1918. Read February 3, 1919. Issued separately September 26, 1919.)

CONTENTS.

PAGE

Introduction 531

The Study of Natural Groups of Fishes . . 532

Generalised Comparative Survey (Table) . . 565

PAGE

List of Literature 566

Index to Figures 569

Description of Plates 570

Introduction.

Some months ago it was suggested that I should publish my thesis in its unfinished
state. I refrained from doing so at that time, as 1 was unwilling to allow my work
to appear until it was ready for publication. When it became apparent that the war
might continue indefinitely. I decided to publish the thesis in such a form as the
exigencies of the case would allow.

For some years I had been fascinated by that elusive problem, the evolution of
the pentadactyle limb of the higher Vertebrates. An extensive list of literature
bearing directly or indirectly on this subject was gathered, and notes were made from
upwards of a hundred monographs. It soon became clear that a necessary preliminary
to the solution of the problem was a thorough knowledge of the limbs of fishes and
their manifold modifications. The hind limb had been most ably handled by Davidoff
as early as 1873, and the skeleton of the fore limb had been extensively treated by
Braus, G-egenbaur, Owen, Parker, and others ; what still remained in a very unsatis-
factory state was the musculature of the fore limb. Numerous authors had dealt
with this latter subject, notably Braus (ll), Furbringer (31, 32, 33), Hamburger (42),
Hartmann (44), Humphry (50, 51, 52), Jaquet (55), M‘Murrich (60), Marion (62),
Owen (73), Tiesing (87), and Vetter (88), but few of these had treated it from the
comparative standpoint ; moreover, the nomenclature was found to be so much at
variance as to obscure probable homologies on the one hand, and, on the other, to
suggest homologies where none in all probability exist. My object, then, became to
elucidate the musculature of the fore limb of fishes.

I am indebted to Professor Graham Kerr for specimens of Lepidosiren and
Polypterus, and to Dr W. M. Tattersall for Periophthalmus ; the remainder of the
types, with the exception of a few from my own collection, were generously placed
at my disposal by Professor W. C. MTntosh, from the Zoological Department and
Museum at St Andrews University. I also wish again to record my indebtedness to
the Carnegie Trust, which provided me with the low-power Zeiss binocular microscope
under which many of my dissections were carried out.

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 21).

82

532

CAPTAIN E. W. SHANN ON

THE STUDY OF NATURAL GROUPS OF- FISHES:
List of Types.

ELASMOBRANCHII . -

OSTEICHTHYES

Selachii

Holocephali

Dipnoi

Chondrostei .
Polypterini .

Scyllium canicula.

Galeus canis.

• Acanthias vulgaris.

Rhina squatina.

Raia batis.
j Chimsera monstrosa.

\ Callorhynchus antarcticus.
Lepidosiren paradoxa.
Accipenser rubicundus.
Polypterus bichir.
j Ammodytes lanceolatus.
Motella mustella.

Zeus faber.

Periophthalmus koelreuteri.
Cottus scorpius.

Trigla gurnardus.

Blennius pholis.

Pholis gunnellus.

Zoarces viviparus.

Lophius piscatorius.

Teleostomi

Teleostei .

Note. — The above types were all actually dissected, and the list does not include
a still wider range of forms which I originally intended to investigate.

Each of the above groups will be taken under the following heads : —

1. Shoulder Girdle.

(1) A Description of the Lateral Muscle.

(2) The Posterior Muscles :

(a) Latero-dorsal.

(b) Latero-ventral.

(c) Mesio-ventral.

(3) The Anterior Muscles :

(a) Connections with the Skull (or with Lateralis Muscle).

(b) Connections with the Visceral Arches.

2. Pectoral Fin.

(1) Adductor (dorsal or inner).

(2) Abductor (ventral or outer).

'HE SHOULDER GIRDLE AND PECTORAL FIN OF FISHES.

533

Key to the Musculature of the Pectoral Arch and Fin in Elasmobranohii.

Gegenbaur.

Humphry.

Marion.

Owen.

WlEDERS-

HEIM.

Vetter.

Remarks.

(a) Retr. lat.-dors. pect.

Not always present.

(6) R. lat.-vent. pect. .

Lat.-scapul.

Serratus.

Neuro-lat.

Serratus of Braus.

Posterior

M. du ventre of

fixation

(c) R. mes.-vent. pect.

muscles.

(i) Superior

Latiss. dors.

Jaquet.

(ii) Medius .

Axili. por-

tion.

(iii) Inferior

Pectoralis.

' (a) Protr. lat.-dors. pect.

Not always present.

(6) P. lat.-vent. pect. *.

Protr. scap.

Wanting in Holo-

Trapezius

cephali.
Trapezius of

(c) Mesio-vent. derivs. .

(pars).

Jaquet.

(i) Levator pect.

Trapez.

Levator scap.

Trapez.

Anterior

(pars).

(pars).

fixation

(ii) Coraco-arc. comm.

Cervic. prof.

Coraco-arc.

muscles.

(pars).

comm.

Coraco-

(iii) Coraco-brancli.

Deepest part

Coraco-

Coraco-

of c.p.

brancb.

branch.

branch.

(iv) Coraco-hyoideus .

Cervic. prof.

Coraco-

Coraco-

Coraco-

(pars).

hyoid.

hyoid.

hyoid.

(v) Coraco-hyomandib.

Coraco-

hyomandib.

Only in Raia.

(vi) Coraco-mandib. .

Coraco-

Depressor-

Coraco-

Coraco-

mandib.

mandib.

mandib.

mandib.

• (a) Adductor.

(i) Superficialis.

Fin

(ii) Profundus.

muscles. "

(6) Abductor.

(i) Superficialis.

(ii) Profundus.

I. Selachii.

(Scyllium canicula, Galeus canis, Acanthias vulgaris , Rhina squatina,

Raia clavata.)

1. Shoulder Girdle.

(l) The Lateral Muscle.

Throughout its length the lateral muscle of Selachians is divided into a dorsal and
a ventral moiety by a horizontal septum running from the lateral line to the trans-
verse processes, and, further back, to their equivalents, the haemal arches. In the
dorsal moiety of the tail region (text-fig. l) the myocommata may be traced super-
ficially from the mid-dorsal line, running first forwards, then sharply backwards, then
gently forwards again ; following the same myocommata through the ventral moiety,
they are seen to run first backwards, then gently forwards, lastly sharply backwards.
On examining a transverse section of the tail, it is seen that the myomeres enclosed

534

CAPTAIN E. W. SHANN ON

by the above-noted myocommata form a series of concentric cones. In all, five such
series of cones occur (C 1-5), of which the dorsalmost have their apices directed
forwards, the next have their apices directed backwards, and so on in alternate order.
This is clearly seen in the well-known diagram of Scyllium canicula in Marshall and
Hurst’s Practical Zoology ; but the completeness of the division at the horizontal
septum is not adequately brought out, nor is the position of the lateral-line nerve
indicated. This occupies the furrow between the third and fourth series of cones.
The directions taken externally by the myocommata are manifestations of the internal
conformation. • There is an apparent incongruity in that, though three series of

In the lateral view the deep portion of the l.d. muscle, which, owing to the obliquity of the so-called “horizontal septum, ”
underlies the l.v. muscle, is shown by shading in the two posterior myomeres.

C 1-5. Muscle cones : 1, concave ; 2, convex,
and so forth alternately.
l.d. Latero-dorsal muscle.
l.l. Lateral line.

l.v. Latero- ventral muscle.
m.d. Mesio-dorsal muscle.
m.v. Mesio-ventral muscle.

complete cones are seen in section in the dorsal moiety, only two and a half cones
appear in corresponding areas on the surface. This incongruity is explained by
the fact that the ventral moiety overlaps the dorsal to a certain extent ; if the
ventral moiety be dissected away from the dorsal, the myocommata of the latter
are found to take a sharp bend backwards in the hidden area, thus completing
the superficial manifestation of the third, anteriorly directed, series of cones (see
text-fig. l).

Let us now trace these muscles forwards to their insertions. The series of cones
are indissolubly connected up to their insertion on the skull, though the upper two
were designated mesio-dorsal (M.D.) and the lower one latero-dorsal (L.D.) portions
respectively of the lateral muscle by Humphry (51), who distinguished them by the
directions of their superficial fibres. While agreeing with Maurer (Morjph. Jahrb.,

THE SHOULDER GIRDL E AND PECTORAL PIN OF FISHES.

535

1892) that this division of the dorsal moiety into mesio-dorsal and latero-dorsal portions
has no morphological value, I have found it sufficiently useful for descriptive purposes,
and it is retained under this proviso.

The ventral moiety in the anterior region is much modified owing to the presence
of the abdominal cavity, whose walls it forms. The conical structure is almost lost,
though, as we shall see later, evidence of it is retained in the lateral walls. The two
portions described in the tail region are still recognisable externally by their myo-
commata. Internally there is, however, a much closer relationship between the lower
half of the C 4 section and the C 5 section than exists between the two halves of the
C 4 section in the anterior region. This led Maurer to follow Humphry in his
designation of the latero-ventral (L.V.) and mesio-ventral (M.V.) portions of the
ventral moiety, the division being indicated in text-fig. 1. It is seen, then, that the
latero-ventral portion comprises the upper half of the C 4 section, while the mesio-
ventral portion comprises the lower half of the C 4 section together with the entire
C 5 section.

In the latero-ventral portion the fibres throughout their depth run approximately
parallel to the long axis of the body, the ventral limit of the muscle being defined
externally by the backward slope of the myocommata, and internally by the distal
extremities of the ribs. In the mesio-ventral portion the fibres run obliquely from
downwards anteriorly to upwards posteriorly, but their obliquity gradually disappears
as the ventral surface is peached, till they come to lie parallel to the long axis of the
body, in which condition they reach the ventral septum which separates them from
their fellows of the opposite side. These two superficial areas of the mesio-ventral
portion were declared by Maurer (ibid.) to be the homologues of the obliquus
internus and rectus respectively of the Amphibia. Moreover, on dissecting away the
superficial oblique fibres, Maurer observed a second layer of oblique fibres, which
took the opposite direction (namely, from upwards anteriorly to downwards posteriorly),
and this layer he took to represent the transversus of the Amphibia. Humphry also
noted this stratification of the mesio-ventral portion. There is, however, no septum
between these two layers ; indeed, they pass by gradual transition one into the
other. I have given in detail elsewhere * my reasons for regarding this pseudo-
stratification of the lateral muscle in the anterior region as a result of the conical
conformation that occurs posteriorly.

(2) The Posterior Muscles.

The lateral muscle when it reaches the pectoral girdle becomes interrupted in its
course by that structure, either throughout its depth or in the superficial region only.
The portions of the lateral muscle which are thus attached to the girdle constitute
its posterior system of muscles, and, from the nature of their position, act as retractors.

Proc. Zool. Soc., 1914, pp. 319 et sqq.

536

CAPTAIN E. W. SHANN ON

They will be described in terms of the division of the lateral muscle which were
advanced in the preceding section, as follows : — -

(а) Retractor latero-dorsalis pectoralis.

(б) R. latero-ventralis pectoralis.

(c) R. mesio-ventralis pectoralis :

(i) Superior.

(ii) Medius.

(iii) Inferior.

(a) Retractor latero-dorsalis pectoralis ( r.l.d.p .). — I can find no mention of this
muscle in the works of previous authors. Humphry, whose observations were based
on Mustelus lasvis, was of opinion that no fibres of the latero-dorsal portion are
inserted on the tip of the scapula. The latter, he says (51, p. 272), “projects in
between the fibres of the dorsal muscle, where it terminates in a ligament which is
lost in one of the transverse septa.”

After the examination of a considerable number of specimens of Scy Ilium, it still
remains doubtful whether this muscle may be regarded as present in that form. It
is usually absent, the tip of the scapula being embedded in a pocket of connective
tissue ; but in certain cases a few fibres of the latero-dorsal portion undoubtedly took
origin on the inner aspect of the scapular horn. In the one example of Galeus
dissected no such fibres were observed.

It is in the compressed Selachians that this muscle attains considerable develop-
ment. ■ In Rhina it is well marked in the superficial area, while in the Rays the latero-
dorsal portion originates in a tendon from the inner aspect of the postero-dorsal angle
of the scapula.

In the Rays the outer section of the mesio-dorsal portion (C 2 section in text-
fig. l) is firmly attached to the postero-ventral aspect of the cartilage which joins
the scapula to the vertebral column ; while the dorsalmost section (C l) of the mesio-
dorsal portion runs freely beneath the said cartilage to its attachment on the
chondrocranium. In these fishes, moreover, there is very little trace of myocommata
in the anterior region of the lateral muscle, so that the fibres run uninterruptedly in
longitudinal series. Owen speaks collectively of the whole dorsal moiety of the
lateral muscle in Batoids (Torpedo), which has been resolved above into its three
constituent parts, as the “ neuro-medial mass.”

( b ) Retractor latero-ventralis pectoralis ( r.l.v.p .). Neuro-lateral muscle, Owen ;
Serratus, Humphry ; Latero-scapularis, Gegenbaur. — Humphry observed in Mustelus
that the scapula lies upon the latero-ventral portion, its lower edge being connected
with one of the transverse septa of this muscle ; moreover, that some of the fibres of
the muscle pass into its under surface, constituting a “ serratus.”

In Scyllium, Galeus, and Acanthias, the superficial fibres of the latero-ventral
portion are attached to the posterior border of the scapula, while a few of the deeper

THE SHOULDER GIRDLE AND PECTORAL FIN OF FISHES.

537

fibres are attached to its inner aspect ; these constitute the retractor latero-ventralis
pectoralis, and are distinguished by their peculiar dark colour. In actuality this
muscle is more or less covered by the musfcle next to be considered — the r.m.v.p.
superior.

In the above types the muscle under discussion is divided by a fascia from the
next of the series of retractors. In Rhina, on the contrary, there is no such distinction,
the r.l.v.p. passing conformably into the r.m.v.p.

In Raia this muscle is present as a thin ribbon attached to the postero-ventral
aspect of the scapula external to the lateral nerve. It differs from the above types
in that it is entirely interrupted by the scapula, no part of it running forward beneath
that cartilage.

(c) Retractor mesio-ventralis pectoralis (r.m.v.p.). — The mesio-ventral portion
encounters in its upper reaches the fin, and in its lower reaches the coracoid portion
of the pectoral girdle, and is interrupted, throughout its depth by these structures.
It is more or less readily divisible into three regions of insertion : (i) ‘ superior —
from the scapula and upper surface of the fin ; (ii) medius — from the axil of the fin ;
and (iii) inferior — from the coracoid and under surface of the fin.

(i) R.m.v.p. superior. Latissimus dorsi, Humphry ; Serratus, Braus (14).

Humphry described this muscle in Mustelus as follows : It is inserted into the

scapular part of the girdle, and expands upon the dorsal surface of the fin, reaching
to its anterior edge. Upon the fin it lies upon, and to some extent blends with, the
proper muscle of the fin.”

I find that this condition obtains in Scyllium, Galeus, and Acanthias. The fibres
run obliquely throughout the depth of the muscle upwards from their insertion ; the
muscle never attains a great thickness. If we take Humphry’s description of the
insertion, the origin is found to be, dorsally, from the surface of the r.l.v.p., while,
ventrally, the fibres are continuous with those of the upper section of the mesio-
ventral portion of the lateral muscle. Arching upwards, the end-line of the fibres
extends as far up the side as the lateral line a short distance (about 10 mm. in
Scyllium) behind the girdle. Posteriorly, the r.m.v.p.sup. rapidly diminishes.

In Raia this muscle does not cover the surface of the fin, but finds its insertion
along the upper posterior rim of the metapterygium, as well as upon the scapula.

(ii) R.m.v.p. medius. Axillary portion, Humphry.

That area of the mesio-ventral portion of the lateral muscle which meets the
articular region of the pectoral fin is here implied. The fibres run obliquely in the
same direction as those of the r.m.v.p.sup. The muscle is sometimes (Scyllium,
G-aleus) separated at its insertion from the last-named by a patch of connective tissue,
in which case the muscle appears to concentrate upon a limited area of the coracoid
just in front of the glenoid border.. In Galeus and Mustelus (Humphry) the insertion
is more extensive and spreads upon the axillary aspect of the fin muscles, both

538

CAPTAIN E. W. SHANN ON

abductor and adductor, as well as upon the scapula (Galeus), in addition to the
typical insertion.

The Rays differ from the shark-like types in that in them this muscle is inserted
along the whole length of the metapterygium, beneath the r.m.v.p.sup., as well as
to the scapular and coracoid portions of the girdle.

(iii) R.m.y.p. inferior. Pectoralis, Humphry ; Rectus (?).

This muscle is confined to'the ventral aspect, and is very constant in form through-
out the Selachii. In appearance it recalls the rectus of higher Vertebrates. It arises
from the posterior border of the coracoid, and usually spreads to a slight extent over
the abductor muscles of the fin. For a short distance behind its origin this muscle
is separated by a low ridge from the axillary portion of the lateral muscle which has
been described above.

(3) The Anterior Muscles.

The anterior muscles which act upon the shoulder girdle are derived from the
more or less modified forward prolongations of the lateral muscle.

The deeper fibres of the dorsal moiety and of the latero-ventral portion of the
ventral moiety are continued forwards beneath the scapular region of the girdle,
except in the Skates and Rays, and are still recognisable up to their points of origin
on the chondrocranium ; the superficial fibres in these areas were observed to be
interrupted by the girdle. As these interrupted fibres behind the girdle serve as
retractors of that structure, so their anterior homologues serve as protractors. The
anterior fibres of the mesio-ventral portion are modified to such an extent by the
development of the visceral arches that they require special consideration. The
anterior muscles of the pectoral girdle will be dealt with in the following order : —

(a) Protractor latero-dorsalis pectoralis.

(b) Protractor latero-ventralis pectoralis.

(c) Muscles derived from the mesio-ventral portion of the lateral muscles :

(i) Levator pectoralis.

(ii) Coraco-arcualis communis.

(iii) Coraco-branchiales.

(iv) Coraco-hyoideus.

(v) Coraco-hyomandibularis.

(vi) Coraco-mandibularis.

(a) Protractor latero-dorsalis pectoralis ( p.l.d.p .). — This muscle has, so far as I
am aware, received no specific designation from previous writers. When present, it is
derived from the interrupted superficial fibres of the latero-dorsal portion of the lateral
muscle, which, arising from the anterior border of the scapula, are inserted immediately
into the main mass of that muscle. It is necessary to make the proviso “ when present,”
for this muscle is developed in precisely the same degree as its equivalent the r.l.d.p.

THE SHOULDER GIRDLE AND PECTORAL FIN OF FISHES.

539

Now, it has been seen that the latter is absent in Mustelus (Humphry) and, at least,
in some specimens of Scyllium, and in these cases the p.l.d.p. is also wanting.

In Rhina, where the r.l.d.p. is well marked, the p.l.d.p. is also well developed,
taking origin by a tendinous sheet from the knob-like extremity of the scapula.

In Raia the anterior border of the cartilage which joins the scapula to the
vertebral column is wider than the posterior border. For this reason all the fibres
of the dorsal moiety of the lateral muscle which are attached anteriorly to the
pectoral girdle take origin. on the said cartilage.

(6) Protractor latero-ventralis pectoralis (p.l.v.p.). Protractor scapulae, Owen.—
Actually the p.l.v.p. is overlain near its insertion by the levator pectoralis and,
further forward, by the superficial branchial muscles ; so that only a very small
portion of it is visible without dissection (at its insertion). In Scyllium, Galeus,
Acanthias, and Rhina this muscle shows marked constancy, and the following
description is applicable to any of these types.

Immediately below the lateral line fibres are inserted into the outer and inner
aspects of the anterior border of the scapula. Tracing these fibres forwards, after
reflecting the overlying muscles noted above, they are found to take origin from the
deeper fibres of the latero-ventral portion of the lateral muscle. Thenceforward the
entire muscle is continued as a dorso-ventrally compressed mass which terminates
in a tendinous insertion on the base of the chondrocranium. It is familiar to all
who have .had occasion to dissect the Dogfish as that muscle which separates the
lateral nerve from the anterior cardinal sinus.

In Raia, though very much reduced, this muscle is present as a triangular strip,
whose apex is inserted on the ventral aspect of the anterior upper portion, of the
scapula, while its base takes origin along the outer rim of the great lateral process
of the first vertebra. The muscle was described by Marion (62) as “ portion C of
the trapezius.”

Doubtless the great development of the vertebral process in Raia accounts for
the fact that no part of this muscle reaches the cranium, which is the case in the
shark-like forms.

(c) Muscles derived from the mesio-ventral portion of the lateral muscle.

(i) Levator pectoralis (l.p.). Levator scapulae, Humphry ; Trapezius (pars),
Vetter, Gegenbaur, Marion.

The levator pectoralis is a compact triangular muscle whose base runs along the
lateral septum from just in front of the scapula to the level of the third branchial
arch. Its apex is curved slightly forwards, so that the posterior side of the triangle
is slightly convex and the anterior slightly concave. The posterior convex border
is inserted into the outer face of the scapula, which is usually grooved (markedly in
Acanthias) to receive it. The body of the l.p. forms the outer wall of the anterior
cardinal sinus.

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 21).

83

540

CAPTAIN E. W. SHANN ON

The above description holds good for all the shark-like types. In Raia the l.p.
is also present, though considerably reduced in extent.

The origin of the levator pectoralis, whether in the Sharks or Rays, is continuous
anteriorly with that of a small muscle which runs to the last branchial arch ; these
two muscles constitute the trapezius of earlier writers, which from its superficial
position and double function has received abundant notice.

(ii) Coraco - arcualis communis (c.a.c.). Cervicalis profundus pars,
Humphry ; Coraco-arcualis communis, Marion.

On the ventral aspect of Elasmobranch fishes, anteriorly to the pectoral girdle, the
lateral muscle is broken up into a complex series of separate muscles coming from
the branchial, hyoid, and mandibular arches to be inserted on the coracoid portion of
the girdle. These muscles, however, anastomose to a greater or less extent before
reaching the girdle; so that a pair of muscle masses, derived from the separate
portions of each side, occur immediately in front of that structure on either side
of the. mid -ventral line ; each of these is named coraco-arcualis communis. The
anastomosis spoken of above is not gradual, but takes place at a well-defined junction,
namely, at a strong fascia — the coraco-arcual septum. The coraco-arcuales communes
form internally the floor of the pericardium.

In Scyllium the c.a.c. is intercepted by four transverse myocommata, which
enhances its resemblance to the portion of the lateral muscle immediately behind the
coracoid — the retractor mesio-ventralis pectoralis inferior. Humphry regarded this
muscle, which he designated, together with the coraco-hyoideus (to be described
below), cervicalis profundus, as the anterior continuation of the “pectoralis” ( = r.m.v.p.
inf.); moreover, he adds that the cervicalis profundus “presents transverse septa
corresponding in number and position with the branchial cartilages.” The former
surmise is highly probable, since Humphry observes that this muscle “ extends with
the septum for some little distance upon the ventral aspect of the fin,” and a similar
condition has been noted above in treating of the insertion of the r.m.v.p.inf. ; in
fact, the lateral range of these two muscles on either side of the median line at the
level of their insertions on the girdle is almost identical. With regard to the trans-
verse septa recorded by Humphry for Mustelus, it is to be supposed that the most
anterior of these is the coraco-arcual septum, while the remaining four are equivalent
to the four myocommata which have been noted in Scyllium ; they thus correspond in
number to the branchial arches, but, in the latter type, they cannot be described as
corresponding in position. The coraco-arcual septum in Scyllium is wedge-shaped,
the thin edge of the wedge extending in a horizontal plane between the coraco-
mandibularis and coraco-hyoideus on the ventral aspect, and the first coraco-branchialis
on the dorsal aspect.

In Galeus the coraco-arcuales are exceedingly short, and are not traversed by
distinct myocommata. Acanthias differs from the foregoing types in the considerable

THE SHOULDER GIRDLE AND PECTORAL FIN OF FISHES.

541

length of these muscles, which extend fully one-third of the distance between
the girdle and mandible ; here again the muscles are divided by four myocommata,
which observation is corroborated by Marion, though Vetter noticed only three.
Marion’s remark that the muscle in question is attached (in part) to a fascia
dorsal to the origin of the cera'tohyal” is unintelligible, unless “ceratohyal” is
a misprint for “ coracohyal” (i.e. the coraco-hyoideus muscle), in which case the
remark is accurate enough ; the coraco-arcualis is very far removed from the
hyoid arch.

A marked deviation is found in the case of Rhina. The coraco-arcuales are of
enormous strength ; they extend half the distance between the girdle and the mandible,
and, in the single specimen examined, are traversed by nine myocommata.

In Raia the coraco-arcuales are very thin straps of muscle which only touch one
another at their anterior extremities ; they extend not quite half the distance between
the girdle and the mandible. Their insertions on the coracoid are widely separated
by the coraco-mandibularis, which in this form is inserted directly on the coracoid
in the middle line. In no specimen have I been able to identify myocommata. In
Tiesing’s work the c.a.c. are figured for Torpedo, Rhinobatis, and Raia, though no
description is given in the text ; no trace of myocommata occurs in these figures.

(iii) Coraco-branchiales (c.br.). Deepest portion of cervicalis profundus,
Humphry ; Coraco-branchialis, Wiedersheim, Marion.

The coraco-branchiales occur with marked constancy in the Selachians as a series
of muscles which, inserted by separate tendons on the branchial arches, traverse the
lateral walls of the pericardium to take origin on the -coraco-arcual septum and on
the coracoid cartilage. In the region of their insertion these muscles form columnar
masses, but near their origin the distinct muscles become more or less fused with one
another and become thinner and flatter in appearance. In the shark-like forms the
C.br. bear a constant relationship to the afferent branchial arteries ; the fused trunk
of the first and second of these arteries appears immediately in front of the first
coraco-branchialis muscle (i.e. in the space between it and the coraco-hyoideus), the
third artery appears between the second and third c.br., the fourth between the third
and fourth c.br., and the fifth between the fourth and fifth c.br.

Although Scyllium is perhaps the commonest fish to be dissected in the laboratories
of this country, its coraco-branchial.musculature, so far as I am aware, has not hitherto
received any adequate attention. It differs markedly from that of the Piked Dogfish
(Acanthias), which has been taken in hand by Vetter, and Marion, and evinces
certain characters which, to my mind, indicate a more primitive grade of evolution.
I refer to the separation throughout their length of the muscle bundles from each
branchial arch, and to the fact that all the bundles, with the sole exception of the
first, are inserted directly on the coracoid bar.

In Scyllium (fig. 2), then, c.br. is a stout columnar muscle which, arising from the

542

CAPTAIN E. W. SH'ANN ON

dorsal surface of the coraco-arcual septum, from just behind the fused trunk of the
first and second afferent arteries to the level of the third afferent artery, is inserted
on the inner border of the first hypobranchial and on the dorsal aspect of the basihyal.
The c.br. of either side are in contact in the middle line from a very short distance
behind their insertions to their origins on the sdptum. I agree with Vetter, who
observed a similar condition in Heptanchus, that the part-insertion of this muscle on
the basihyal is a secondary relation, formed in consequence of the diminution of the
first hypobranchial, and does not prevent this muscle from being regarded as that
of the first gill-arch. C.br. 2, inserted mainly on the second hypobranchial, but by
a few fibres also on the second ceratobranchial, runs as a column of muscle to the
anterior rim of the coracoid. It is in contact with its fellow of the opposite side,
with which it blends to a certain extent, in the region of its origin ; dorsally the two
divide, allowing passage for the ventral aorta. C.br. 3 and 4 resemble c.br. 2 in
origin and insertion, and are found in sequence behind it. They, however, are
separated throughout their length from their fellows of the opposite side by the
presence in this region of the pericardium. C.br. 5 presents a greater surface to the
exterior than any of the foregoing muscles, but is also considerably thinner. It is
fairly distinctly divided into three portions, all of which, taking origin on the broad
anterior face of the coracoid, opposite the abductor muscles of the fin, are inserted on
the inner aspect of the fifth ceratobranchial.

In Acanthias, the stout columnar muscle which arises on the most anterior portion
of the coraco-arcual septum, and from it runs forwards and upwards, must be regarded
as c.br., although its insertion is found to be entirely on the dorsal aspect of the
basihyal cartilage. As Vetter maintained, the connection of this muscle with the
hyoid arch is almost certainly a secondary condition, consequent upon the extreme
reduction of the copula of the first branchial arch; the muscle corresponds in its
general topography with the muscle which we have described as c.br., in Scy Ilium
and Heptanchus. Marion also adopts this nomenclature, and therefore, presumably,
this view of the homology of these muscles. O.br. is separated by a strong aponeurosis
in the middle line from its fellow of the opposite side.

The remaining coraco-branchiales (c.br. 2-5) form the lateral wall of the peri-
cardium. With the exception of the most posterior one (c.br. 5), they are attached
ventrally not to the coracoid, as in Scyllium, but to a strong fascia which covers the
dorsal face of the c.a.c. and extends laterally to the superficial branchial muscles. In
this region they are difficult to distinguish one from another ; Marion regards them
as a single sheet of muscle. Certainly I have never found them so distinct as is
indicated by Vetter in his drawing of Acanthias (op. cit., Taf. xv, fig. 8).

C.br. 2 is inserted into the second hypobranchial ; c.br. 3 and 4 are inserted
mainly on the third and fourth hypobranchials, but to a slight extent also on the
corresponding ceratobranchials. C.br. 5 is inserted on the inner aspect of the fifth
ceratobranchial and (wherein it differs from the corresponding muscle in Scyllium)

THE SHOULDER GIRDLE AND PECTORAL FIN OF FISHES.

543

on the outer ventral aspect of the basibranchial plate. It shows no sign of division
into segments such as we have observed in Scyllium.

Turning now to Rhina, we again approach new ground, but, with a few trivial
exceptions, the characters of the coraco-branchial musculature are readily conformable
to type. In accordance with the flattened shape of their possessor, these muscles
are more flattened than in the shark-like forms. The arteries pass outwards in the
normal manner.

C.br. 1, inserted mainly on the first hypobranchial and in part also on the dorsal
aspect of the basihyal, runs backwards almost in the same horizontal plane to its
origin on the anterior part of the dorsal aspect of the coraco-arcual septum.

The remaining coraco-branchiales, inserted in turn into the succeeding branchial
arches, take origin from the coracoid itself beneath the coraco-arcuales communes.
C.br. 2-4 become inextricably fused with one another for a short distance in front
of their origins ; they then separate to go to their insertions on the appropriate arches.
These insertions are, as usual, mainly on the hypobranchials, but there is an increasing
tendency to send fibres to the ceratobranchials also as we proceed from before back-
wards. The hindermost muscle of this series, c.br. 5, is composed-of a very broad
sheet which is distinctly divided into two portions.

In Raia we meet with a very specialised condition of the coraco-branchial muscular
system. From its attachment to the angle of pectoral girdle, immediately lateral to
the origin of the c.a.c. and contiguous dorsally with the l.p., the muscle runs obliquely
inwards, keeping its horizontal plane and showing no signs of dividing into separate
bundles. It passes external to the trunk which divides into the third, fourth, and
fifth afferent branchial arteries, to its main area of insertion on the anterior process
of the basihyal and the membranous floor of the mouth. In the dorso-lateral region
it also gives off a few fibres which are attached to the hypo- and ceratobranchials ;
these are the only vestiges of the c.br. 2-4. Posteriorly there is a strong connection
with the fifth ceratobranchial ; this represents c.br. 5. I have been unable to detect
a trace of c.br. 1.

(iv) Coraco-hyoideus (c.hy.). Cervicalis profundus pars, Humphry ; Coraco-
hyoideus, Vetter, Wiedersheim, MIrion.

The coraco-hyoideus is usually a massive muscle which extends from the coraco-
arcual septum to the basihyal cartilage ; its length varies inversely with that of the
coraco-arcualis communis. The c.hy. is not as a rule marked by myocommata, though
Vetter figures one such division in Acanthias and Heptanchus.

In Scyllium, Galeus, and Acanthias the c.hy. arises from the ventral face of the
coraco-arcual septum. Externally, only its lateral surface can be seen, the median
portion being covered by the coraco-mandibularis, to be described later. It runs
forwards, in contact with its fellow of the opposite side, and finds insertion on the
posterior and postero-dorsal border of the basihyal. Its relative length is greatest in

544

CAPTATN E. W. SHANN ON

Galeus, where it extends almost to the coracoid bar, and least in Acanthias,
where it extends less than three-fifths of the distance between the basihyal and
the coracoid.

In Rhina this muscle is exceedingly short, but of great strength, and is inserted
on the ventral aspect of the ceratohyal as well as upon the basihyal. On the latter
it is inserted through the medium of a strong tendon ; moreover, the insertion is
diagnostic in showing an unpaired median portion, formed by the union in that area
with its fellow of the opposite side.

In Raia this muscle is very much reduced, probably in correlation with the
development of a muscle, the coraco-hyomandibularis, to be described later. I have
been able to verify Marion’s observation that, arising from the fascia covering the
c.h.m., it is inserted upon the minute hypohyal cartilage.

(v) Coraco-hyomandibularis (c.h.m.). Coraco-hyomandibularis, Marion.

Of the types examined, this muscle occurs only in Raia. In that form it is
present as a large strip of muscle — the strongest, in fact, of all the ventral muscles —
which, arising from the fascia limiting the coraco-arcuales communes as far forward
as the fork of the ventral aorta, runs outwards and forwards (passing dorsal to
the first hypobranchial) to its insertion on the antero-ventral aspect of the hyo-
mandibular cartilage.

(vi) Coraco-mandibularis ' (c.m). Depressor mandibuli, Owen ; Coraco-

mandibularis, Vetter, Wiedersheim, Tiesing, Marion, etc.

This muscle arises from the coracoid bar in the ventral middle line, or, more
frequently, from the coraco-arcual septum ; it is the most superficial of the series of
muscles under consideration, and, passing forwards in a straight line, 'becomes inserted
on the posterior border of the mandible. - The muscle may be either azygous or
divisible into two portions ; it is not divided by myocommata.

In Scyllium and G-aleus the c.m. is tolerably distinct from its fellow of the
opposite side, owing to the presence of a longitudinal septum in the ventral middle
line ; this is particularly noticeable in the region of insertion on the mandible.
In Acanthias the c.m. is an azygos muscle. The length of this muscle varies
inversely with that of the coraco-arcualis communis ; thus, in Galeus the muscle
is of great relative length, almost touching the coracoid bar, while in Acanthias
it is relatively short.

In Rhina the c.m. is entirely wanting.

In Raia the c.m. is a thin strip of muscle, separated by a fascia from its fellow of
the opposite side, extending from the mandible direct to the anterior border of the
coracoid itself. Right up to its origin on the coracoid it is distinct from the other
muscles of the coraco-arcual system.

THE SHOULDER GIRDLE AND PECTORAL FIN OF FISHES.

545

2. Pectoral Fin.

(l) Adductor {dorsal or inner) Musculature.

The adductor muscle of the fin is a fan-shaped structure, composed of a number of
muscle bundles which radiate from the axil to the blade. With the exception of a
few of the outermost (the first three in Scyllium), each bundle is divided horizontally
into two portions, so that the whole muscle appears to be divided into two super-
imposed layers. Most authors recognise but one layer in this muscle; Davidoff (20),
who studied very thoroughly the musculature of the pelvic fin of Elasmobranchs,
distinctly states that in that region there are two layers in Selachians (Acanthias).
The evidence which I have been able to collect from the large amount of embryo-
logical work published on this subject militates against the view that the division
of the muscle into superimposed layers is of real morphological value. Balfour’s
figure of the development of the pectoral fin of Scyllium (81, pi. xviii, fig. 7)
certainly appears to show two layers in the adductor muscles. This may, however,
only be an apparent layering ; the “radial muscles” ( = adductor bundles) are admit-
tedly oblique in older embryos (such as that figured), and this involves a certain
amount of overlapping of adjacent radial muscles ; thus, the segments appearing in
the section may represent parts of different muscles, not different parts of the same
muscle. That two layers exist in the adult, at all events in the proximal region,
there can be no doubt ; the outer layer is attached to the girdle, the inner to the
proximal cartilages of the fin, and between the two a well-marked cavity is found.
Proceeding outwards the layers approximate to one another, and finally coalesce so
closely that it becomes difficult to separate them. The muscle will be described
here in terms of two layers, namely, adductor superficialis (add.s.) and adductor
profundus (add.p.).

(i) Adductor superficialis (add.s.).

Arises from the scapula and is inserted into the adductor profundus. In Scyllium
the three preaxial bundles are not divided ; that is to say, they may be regarded as
belonging entirely to the superficial layer, since they arise on the scapula. These
three bundles are inserted on the propterygium and propterygial radial. The
remainder of the add.s. is divided into two regions, an anterior and a posterior.
Near their origins on the scapula the anterior (preaxial) portion passes beneath the
posterior, and its innermost fibres become attached to the inner border of the scapula.
This arrangement of the superficial muscles allows a certain degree of rotation, or
curling, of the fin.

The general arrangement of the superficial layer, with regard, to origin and inser-
tion, holds good in all the types under examination ; but the division into two
overlapping portions is only found in the shark-like forms, while in Rhina and Raia
the muscle bundles radiate in perfectly regular sequence.

546

CAPTATN E. W. SHANN ON

(ii) Adductor profundus (add.p.).

Arises from the proximal cartilages of the fin and is inserted into the distal
radials and the horny fin-rays (ceratotriehia): Bratjs (9) maintains that the

adductor is inserted only into the horny rays, never into the radials ; but my
observations do not confirm this statement. The same author has described in great
detail the musculature and innervation of the pectoral fin of Acanthias (op. cit.,
Taf. xii, fig. l), and he shows very clearly the remarkable fact that the muscle
bundles correspond neither in number nor' position with the distal radial cartilages.

(2) Abductor (ventral or outer) Musculature.

The abductor muscle is in most points the counterpart of its antagonist the
addmctor, though the number of component muscle bundles is not necessarily the
same. It may with equaA justice- be divided into two layers, abductor superficialis
(abd.s.) and abductor profundus (abd.p.).

(i) Abductor superficialis (abd.s.).

Arises from the coracoid and is inserted into the abductor profundus. The first
three preaxial bundles, which are inserted on the ventral aspect of the propterygium,
show no division into layers, and may be regarded as belonging entirely to the
superficial layer ; they are best seen from the lateral aspect. The fourth superficial
bundle, also lateral in position, has a double insertion, its deepest fibres being
inserted into the proximal border of the mesopterygium, while its superficial fibres
are inserted into the proximal portion of the propterygial radial and into the fascia
covering the first bundle of the abd.p. A similar arrangement will be noted in two
of the superficial bundles in Chimsera. The first four bundles serve as depressors of
the propterygium and its radial. The remaining twenty bundles are visible from
the ventral aspect and are inserted in the typical manner.

(ii) Abductor profundus (abd.p.).

Arises from the metapterygium and is inserted into the distal radials and the
ceratotriehia. There is a small muscle which, arising from the outer portion of the
mesopterygium, is inserted into the most preaxial horny rays ; this must be regarded
as the first (preaxial) deep bundle, and corresponds with the fourth bundle of the
superficial series. The second bundle of the abductor profundus is inserted into all
the remaining ceratotriehia articulating with the propterygial radial.

The above description of the abductor musculature applies to Scy Ilium, and
lack of time has ^rendered impossible anything but a superficial examination of
the other types. So far as my observations go, they indicate that the condition
of the musculature in Scyllium is typical of the general arrangement in Selachian
fishes.

THE SHOULDER GIRDLE AND PECTORAL FIN OF FISHES.

547

II. Holocephali.

(Chimera monstrosa, Callorhynchus antarcticus.)

1. Shoulder Girdle.

(l) The Lateral Muscle.

The lateral muscle conforms, both in structure and external appearance, to the
typical condition in Elasmobranch fishes. This condition has already received
sufficient notice in the section on Selachii.

(2) The Posterior Muscles.

The parts of the lateral muscle which go to form the retractors of the pectoral
girdle in the Holocephali have attained a much higher grade of specialisation than
has been encountered among the Selachians. The retractor muscles under considera-
tion are not merely interrupted portions of the segmented lateral muscle, but
separate sheets, enclosed in their own fasciae, and devoid of all trace of myocommata
(except in the two lower sections of the mesio-ventral portion). They will be taken
in the following order : —

(a) Retractor dorsalis pectoralis.

(b) Retractor later o-ventralis pectoralis :

(i) Externus.

(ii) Internus.

(c) Retractor mesio-ventralis pectoralis :

(i) Superior.

(ii) Medius.

(iii) Inferior.

(a) Retractor dorsalis pectoralis ( r.d.p .). — This is a broad thin sheet of muscle
which, arising from the portion of the lateral muscle immediately above the lateral
line, runs in a direct, antero-posterior direction to be inserted on the inner aspect of
the posterior rim of the scapula and of the backwardly projecting distal horn of that
cartilage. In Chimeera the muscle is of considerable length, extending to the third
myocomma behind the scapular horn, from which myocomma the majority of its
fibres take origin. In Callorhynchus, on the other hand, the muscle is relatively
short, its fibres taking origin from the myocomma which touches the tip of the
scapular horn. In both genera the scapular horn extends above the level of the
latero-dorsal portion of the lateral muscle ; the retractor muscle, therefore, comprises
part of the mesio-dorsal portion in addition to the entire latero-dorsal portion ; it
thus corresponds to the r.m.d.p. and r.l.d.p. of Selachians. As, however, it is not
divided longitudinally, it is designated here simply as the r.d.p., or the retractor
muscle which is derived from the dorsal moiety of the lateral muscle.

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 21).

84

548

CAPTAIN E. W. SHANN ON

This and the muscle next to be described are figured by Vetter (88, Taf. xii,
fig. l), but erroneously, for they are represented as parts of one and the same muscle
sheet, whereas they are in reality divided very distinctly by the lateral line ; more-
over, their fibres take different directions. '

( b ) Retractor latero-ventralis pectoralis ( r.l.v.p .). — Two straps of muscle (fig. 12)
are given off from the latero-ventral portion of the lateral muscle, one of which
(externus) is inserted on the outer aspect of the scapula, while the other (internus)
is inserted on its inner aspect.

(i) R.l.v.p. externus.

In Chim sera this muscle is covered to such an extent by the elevator muscle of
the propterygial radial and the r.m.v.p.sup. that it is scarcely visible until the latter
are reflected ; in Callorhynchus, on the other hand, where the first of these muscles
is wanting and the second is much more limited in extent, a large tract of the
r.l.v.p.ext. in the region of its insertion is visible without dissection. Inserted by a
broad tendon in a slight groove on the outer aspect of the scapula, and extending
a short distance above the level of the lateral line, the muscle fibres run obliquely,
especially in Callorhynchus, downwards.

(ii) R.l.v.p. internus.

This muscle lies slightly ventral as well as internal to the foregoing in Chimeera,
but in Callorhynchus is directly beneath it. Taking origin from the latero-ventral
mass, its fibres pass, gently downwards in Chimsera, parallel to the major body-axis
in Callorhynchus, to their insertion on the inner aspect of the scapula.

(c) Retractor mesio-ventralis pectoralis ( r.m.v.p .). — As in the Selachians, the
mesio-ventral portion of the lateral muscle is attached by three tolerably distinct
areas of insertion to the girdle ; but, except in Callorhynchus, it shows no tendency
to spread upon the fin muscles. The entire mesio-ventral portion is intercepted at
the pectoral girdle.

(i) R.m.v.p. superior.

In Chimsera this is an extensive sheet of muscle, whose fibres, arising from the
fascia covering the lateral muscle, immediately below the lateral line, run obliquely
downwards and forwards to their insertion on the posterior rim of the scapula. The
dorsal fibres take origin from the fascia covering the surface of the latero-ventral
portion, and for 20 mm. extend as far as the lateral line, this extent beginning
.19 mm. behind the scapula; thenceforward the end-line bends downwards, arc-wise,
and the ventral fibres gradually fuse with the latero-ventral portion and the middle
section of the mesio-ventral portion. It is no longer recognised as a girdle muscle,
but as the upper section of the mesio-ventral portion of the lateral muscle. Ventrally,
the muscle is separated by a distinct fascia from the r.m.v.p.med.

Callorhynchus differs from Chimsera in that the r.m.v.p.sup. is continuous ven-

THE SHOULDER GIRDLE AND PECTORAL FIN OF FISHES.

549

trally with the r.m.v.p.med., in which region also it show traces of myocommata.
The latter disappear entirely in the dorsal area of this muscle. The uppermost end-
line of the fibres, though arc-shaped, is less convex, and at its nearest point only
extends within 19 mm. of the lateral line. In the region of insertion the superficial
fibres of this muscle pass clear of the girdle, external to the scapula, to blend with
those of the superficial constrictor muscles. This condition recalls that which is
described by Humphry for Mustelus, and which obtains in numerous other shark-like
Selachians ; in Chimsera this junction does not take place.

(ii) R.m.v.p. medius.

As in Selachians, this muscle is formed by the entire middle section of the mesio-
ventral portion of the lateral muscle where it comes in contact with the girdle. The
insertion is upon the inner aspect of the posterior rims of the scapula and coracoid
from the ventralmost insertion of the r.m.v.p. sup. (and, in Chimsera, for some distance
above and internal to that insertion) to the ventro-lateral angle of the body-wall.
The muscle shows no tendency to spread upon the fin-muscles. The fibres run
obliquely upwards from their insertions, and are intercepted a short distance behind
the girdle by myocommata.

(iii) R.m.v.p. inferior.

This muscle only appears on the ventral aspect, and its fibres, inserted on the
posterior face of the coracoid, run directly backwards parallel to the major body-axis.
It is formed by the entire lower section of the mesio-ventral portion of the lateral
muscle, which remains undifferentiated right up to its point of insertion.

(3) The Anterior Muscles.

In the anterior region we again encounter in the Holocephali an advanced state
of evolution of the muscles which work the pectoral girdle. The muscles will be
described in the following order, homologies with those of Selachians being implied
by the use of identical terms : —

{a) Protractor dorsalis pectoralis.

(6) Muscles derived from the mesio-ventral portion of the lateral muscle :

(i) Levator pectoralis.

(ii) Coraco-branchiales.

(iii) Coraco-hyoideus.

(iv) Coraco-mandibularis.

(a) Protractor dorsalis 'pectoralis ( p.d.p .) — It has already been seen that a large
sheet of muscle derived from the entire latero-dorsal portion and to a certain extent
from the mesio-dorsal portion of the lateral muscle is inserted on the posterior
border of the uppermost section of the scapula ; this was designated the retractor
dorsalis pectoralis. Another muscle, of precisely similar extent, bestrides the

550

CAPTAIN E. W. SHANN ON

anterior border of the scapula, and its fibres, passing forwards and slightly down-
wards, take origin on the posterior aspect of the strong post-orbital ridge of the
cranium ; to this muscle the name protractor dorsalis pectoralis is given.

The p.d.p. is covered to a certain extent by the l.p. ; this is particularly marked
in Callorhynchus, where a tract of the latter muscle is excised in the drawing in
order to expose the insertion of the former.

(b) Muscles derived from the mesio-ventral portion of the lateral muscle.

(i) Levator pectoralis (l.p.). Trapezius superficialis, Vetter; Trapezius,

Jaquet.

Arising from the post-orbital ridge of the cranium, opposite to the lower half of
the orbit in Chimsera, opposite the entire orbit and extending above its level in
Callorhynchus, the fibres of this muscle run backwards and downwards to their
insertion on the outer aspect of the scapula. The insertion extends in an arc in
Chimsera, while in Callorhynchus it runs in a straight line, parallel with that of r.l.v.p.
The similarity of the insertion levels of the muscle under consideration and of the
r.l.v.p., and the similar direction taken by their fibres, led me at first to believe
that the former was also a derivative, of the latero-ventral portion of the lateral
muscle — in fact, that it represented the p.l.v.p. of Selachians. The origin of the
muscle high up upon the cranium seemed also to agree with such a conclusion.
Vetter’s observation that this muscle in Chimsera is innervated in a precisely
similar manner to the “trapezius” [i.e. levator pectoralis) of Acanthias, namely, by
certain twigs from the ramus intestinalis of the vagus, seems to me a weightier
argument than one based on mere topographical grounds. Moreover, there is a
small muscle, figured and described by Vetter as trapezius profundus, which lies
beneath the foregoing mass. The origins of these muscles are conterminous, they
run parallel to one another, but whereas the larger outer one is inserted upon the
scapula, the smaller inner one is inserted upon the last pharyngo-branchial. Now,
this last recalls the small muscle which was mentioned under the Selachii (see p. 540)-
as a part of the so-called trapezius. It appears, then, that the protractor latero-
ventralis pectoralis is entirely wanting in Holocephali.

(ii) Coraco-branchiales (c.br.). Coraco-branchiales, Vetter, Jaquet.

A muscular sheet of uniform thickness, arising from the anterior border of the
girdle (both scapular and coracoid portions), is inserted by very short branches on
the lowermost portions of the gill arches, thus indicating its composite structure.
From the nature of their insertions it follows that the fibres are short dorsally and
increase in length as the ventral border of the muscle is approached. The muscle
serves, as in the Selachians, to form the lateral wall of the pericardium.

Vetter has described this muscle in considerable detail for Chimsera (88,
Taf. xii, fig. 5).

The general relations of the muscle are similar in Callorhynchus.

THE SHOULDER GIRDLE AND PECTORAL FIN OF FTSHES.

551

The nervous supply is derived from the anterior branch of the united ventral
roots of spinal nerves I and II.

(iii) Coraco-hyoideus (c.hy.). Coraco-hyoideus, Vetter, Jaquet.

There is a marked dissimilarity in the two types under consideration in respect
of this muscle.

In Chimsera it arises from the anterior face of the V-shaped septum described
above, its outermost fibres being attached further back and more superficially than
the median ones, since the latter are displaced by the origin of the deepest portion
of the coraco-mandibularis. The median fibres meet their fellows of the opposite
side, but do not fuse with them ; from this median line of contact the muscle masses
of either side slope downwards and outwards, thus forming the furrow in which runs
the keel of the c.m. The insertions of the c.hy. of both sides are contiguous, and
occur upon the postero-dorsal aspect of the basihyal cartilage.

There is another muscle mass, which Vetter regarded as the posterior portion of
the c.hy., and which must probably be regarded in part, at least, in that light. Arising
from the anterior border of the girdle, from just above the origin of the c.m. to the
level of the glenoid border (where it is covered by the c.br.), the fibres of this muscle
run downwards and slightly outwards to their insertion on the posterior face of the
V-shaped septum. Since not only the fibres of the c.hy. are attached to the anterior
face of this septum, but also those of the c.m. 2, the muscle may be compared to the
coraco-arcualis communis of Selachians. Pending further evidence, I do not propose
to alter Vetter’s description of the coraco-hyoideus of Chimsera.

In Callorhynchus the c.hy. is a relatively slender paired muscle which, arising
from the dorsal aspect of the c.m., runs forward to its insertion on the basihyal
cartilage. There is no manifestation of a posterior portion of the muscle connecting
it with the coracoid, such as has been described for Chimsera. It is not implied that
no fibres of the' c.hy. reach the coracoid, but that any such are so completely blended
with the huge c.m. that they cannot be distinguished from it. In fact, the neck
muscles of Callorhynchus have attained a state of fusion which would render com-
parison with the Selachian type obscure, were it not for the intermediate condition
which has been observed in Chimsera.

(iv) Coraco-mandibularis (c.m.). Coraco-mandibularis, Vetter, Jaquet.

The coraco-mandibulares of the Holocephali expand anteriorly, and their tendinous

insertions cover nearly the entire ventral aspect of the mandible. There is a fairly
distinct division of the insertion into an outer, superficial portion (c.m. 1 of Vetter)
and a median, deeper portion (c.m. 2 of Vetter). The latter is an azygous muscle,
composed of the united median fibres of the c.m. of either side. The central portion
is not severed completely from the lateral portions, for the deeper fibres of both
portions are in contact ; the fascial fold which serves to separate the muscles super-
ficially is sufficiently diagnostic to be of descriptive value.

552

CAPTAIN E. W. SHANN ON

The fibres from the outer line of insertion (c.m. l) run inwards, and those of
either side eventually meet at a point (35 mm. Chimsera, 25 mm. Callorhynchus)
behind the exposed portion of the mandible, thus hiding the fibres which run from
the main line of insertion (c.m. 2) nearly parallel to the long axis of the body.
After coalescing, the c.m. portions of either side unite and are only barely dis-
tinguishable superficially, while internally their fibres become intermingled. The
united c.m. portions gain in depth at the expense of their width, and the depth is
further increased by the addition of fibres from the azygous (c.m. 2) portion.

So far the two types under consideration are uniform, but from this point onwards
they require separate consideration. In Chimsera the main mass of the c.m., derived
from the entire c.m. 1 portion and from the superficial fibres of the c.m. 2 portion,
is found to take origin from the deep groove on the anterior face of the coracoid
symphysis. The remaining fibres of the c.m. 2 portion, shortly behind their insertion,
become piled upon one another so as to form a vertical keel of muscle. This keel
runs in a deep groove between the coraco-hyoideus muscles, from which it is
separated by a fascia. At a point about 45 mm. behind their insertion the fibres
which compose the keel are intercepted by a Y-shaped septum, whose apex is
directed posteriorly. The deepest fibres of the c.m. may thus be described as arising
from a wedge-shaped septum between the c.hy. muscles. Through a circular opening
in either face of this septum passes the branch of the combined ventral roots of
spinal nerves I and II which innervates the deep portion of the c.m. The above
account differs from that of Vetter in that the superficial fibres of the c.m. 2 portion
are described as running directly from their insertions on the mandible to join the
deep fibres of the c.m. 1 portion near their origin, whereas Vetter describes the
entire c.m. 2 portion as arising from the septum in question. Vetter, however,
remarks that the c.m. 1 portion is augmented in the region of its origin by certain
fibres from the septum, and my observations corroborate this. Briefly, the question
at stake is : Are all the fibres, which arise from the anterior rim of .the coracoid and
are inserted on the median (c.m. 2) area of the mandible, intercepted in their course
by a septum ? Vetter’s answer is in the affirmative, mine in the negative. It is
perfectly possible that there is variability in this respect, though my observations
are drawn from an examination of four adult specimens ; but, in any case, the fact
remains that the distinction between the two portions of the coraco-mandibularis
(c.m. 1 and c.m. 2) is purely superficial. The condition recalls that observed in
Rhina among Selachians.

In Callorhynchus the c.m. runs without interruption throughout its depth from
the mandible to the pectoral girdle ; the mass is characterised by its enormous bulk
and by the peculiar distribution of its component fibres. The divisions of the c.m. in
the area of insertion have been indicated above ; the superficial fibres of the c.m. 1
portion, running towards the ventral middle line, soon cover the original ventral
superficial fibres of the c.m. 2 portion, while their outer border is defined by a sharp

THE SHOULDER OIRDLE AND PECTORAL FIN OF FISHES.

553

ridge, above which is a deep trough, gouged, as it were, from the lateral aspect of the
muscle. This ridge and trough are most marked anteriorly, and gradually disappear
posteriorly, the ridge being rounded off and the trough filled out ; this effect is
attained at a quarter of the length of the muscle from its origin on the coracoid.
The deeper fibres of the c.m. 1. portion, instead of making for the middle line, sweep
outwards and upwards, forming the floor of the trough which was mentioned above,
and find origin on the anterior face of the girdle. The origin of this portion is, in its
ventral area, exposed on the lateral aspect (after removal of the gills and operculum) ;
dorsally, however, it is covered by the coraco-branchiales, beneath which it extends
upwards to the scapular region of the girdle, slightly above the level of the adductor
muscles of the fin. The c.m. 2 portion is not interrupted, as in Chimsera, by a septum,
but extends as a solid mass of muscle within the outer coating of the c.m. I portion.
The mesial fibres of the c.m. 2 portion blend with the mesial fibres of the c.m. 1
portion, where the faces of these muscles are in contact, and accompany them to their
origin. The lateral fibres sweep outwards and upwards, running beneath and parallel
to the lateral fibres of the c.m. 1 portion. This powerful mass of muscle, whjtfh
represents the fused coraco-mandibulares of either side, is, then, near its origin on the
pectoral girdle, a homogeneous structure, whose section is U-shaped, the concavity
being directed dorsally.

2. Pectoral Fin.

The morphology of the pectoral fin skeleton of the Chimseroids has been discussed
by Howes (48, p. 22). In this discussion, Mivart’s interpretation of the relationship
of the parts is advocated, and subsequent authors (e.g. Goodrich, Lankesters Treatise
Zook, ix, p. 81, Oallorhynchus) appear to accept this view. The fin is supported
upon two basal elements, both of which articulate with the glenoid border of the
pectoral girdle ; the postaxial is held to represent the metapterygium, the preaxial the
proptergyium (the mesopterygium is wanting). A diagnostic triangular cartilage,
whose base articulates proximally with the propterygium, represents the fused
portions of the four propterygial radiales ; this will be designated in brief the
propterygial radial.

(l) Adductor Musculature.

The adductor musculature closely resembles that of Sharks ; that is to say, a series
of muscle-bundles radiate fanwise from the coraco-scapular region of the girdle to the
blade of the fin. The two preaxial bundles, inserted on the propterygium and
propterygial radial respectively, are specialised and will receive separate treatment.
The remaining thirty-one bundles are divided, though indistinctly as in Selachians,
into a superficial and a deep layer. Davidoff (20) recognises a layering in the
adductor musculature of the pelvic fin of Chimsera.

The following account is based on a study of China sera alone, a certain striking
point of difference from Oallorhynchus being noted in passing.

554

CAPTAIN E. W. SHANN ON

(i) Adductor superficialis (add.s.).

The main mass of this muscle arises on the outer border of the scapula and is
inserted into the underlying adductor profundus. The first two adductor bundles
(preaxial) are not divided into layers and may be described, as in Selachians, as
belonging entirely to the superficial series. The first, arising from the postero-external
border of the scapula, opposite to the upper third of the coraco-branchialis, runs
downwards and slightly outwards' to its insertion on the inner upper border of the
propterygium ; as it fills the axil of the fin its shape conforms approximately to that
of a wedge. The second in Callorhynchus follows a parallel course and finds insertion
on the propterygial radial ; but in Chimeera we meet with a very striking modification.
The insertion is strap-like, partly tendinous, and occurs on the proximal fifth only
of the propterygial radial ; from here the fibres run almost vertically to take origin
not only from the outer surface of the scapula (opposite to the insertion of the l.p.),
but also from the fascia covering the r.l.v.p. as far up as the lateral line. (See
fig. 6.)

The third superficial bundle has a distinct origin in the cavity of the scapula near
the preaxial end of the glenoid border ; it is very slender and soon fuses with the
much stouter first bundle of the deep series, which in turn is inserted into the distal
portion of the propterygial radial. The remaining thirty bundles are divided into
two groups, of which the preaxial, as in Selachians, is overlapped near its origin by
the postaxial. The preaxial group consists of twenty-one bundles in the example
figured ; these arise from a bowl-like cavity in the coraco-scapular region of the
girdle opposite to the middle of the glenoid border. The bundles are by no means
distinct in the region of origin, and the muscle is characterised by the great depth
in comparison to its width. As the fibres pass outwards they spread laterally, and,
as a consequence, the mass becomes shallower. They begin now to be collected into
bundles, which become more and more distinct as the distal extremity of the muscle
is reached. The deepest fibres, when they reach the level of the basal cartilages of
the fin, blend with the underlying adductor profundus. In fig. 10 the superficial
bundles appear to be inserted entirely at the distal extremity of the corresponding
deep bundles ; it should be pointed out that this is a diagrammatic representation,
the remainder of the insertion (extending almost the whole length of the add.p.)
being omitted for the sake of clearness. The remaining nine bundles comprise the
postaxial group. These arise from the outer border of the scapula, a short distance
above the origin of the preaxial group, and in their outward passage obscure the
latter.

(ii) Adductor profundus (add.p.).

This muscle is composed of bundles which correspond in number and relative
position to those of the superficial series described above, with the exception of the
two preaxial ones, which, as already stated, have no representatives in the lower

THE SHOULDER CURDLE AND PECTORAL FIN OF FISHES.

555

series. There are, then, thirty-one bundles, of which the first takes origin from the
superior face of the propterygium and is inserted into the propterygial radial.

(2) Abductor Musculature.

The abductor musculature is in the form of a fan-shaped mass which radiates from
a considerable area of the coracoid, and, to a slight extent, from the coraco-scapular
region, to the outer surface of the fin. The muscle mass is exceedingly dense ventrally ;
the insertion here extends from the anterior border of the coracoid right round to the
posterior border, where it lies parallel with the ventralmost bundle of the adductor
series. The result of this formation is that the postaxialmost adductor fibres (fig. 10)
actually act as abductors ; so that, when the muscle as a whole contracts, the lower
(postaxial) border of the fin is curled inwards, while the upper and middle portions
are drawn outwards. Moreover, the obliquity of the muscle, owing to the relative
lowness of its origin, tends to draw the whole fin somewhat downwards. From the
foregoing description of the adductor musculature it is apparent that the action of the
latter is to rotate the fin back into its position of rest at the side of the body ; this
upward and backward sweep of the fin must be largely assisted by the strongly
developed second adductor bundle.

(i) Abductor superficialis (abd.s.).

The muscle takes origin dorsally from the antero-lateral face of the girdle in the
coraco-scapular region, rather below the upper level of the glenoid border ; as it is
traced ventrally the origin increases in depth, and finally spreads round the lower
edge of the glenoid border and upon the posterior face of the coracoid (see fig. 10).
The first two (preaxial) bundles are not divided into layers (see fig. 11), and are
inserted into the propterygium and propterygial radial. The succeeding two bundles
give their deepest fibres to the metapterygium, the more superficial fibres terminat-
ing on the propterygial radial and in the corresponding bundles of the profundus
system. The bundles thenceforward become less and less distinct.

(ii) Abductor profundus (abd.p.).

This muscle takes origin from the metapterygium and metapterygial radials, and
is inserted proximally into the palisade-like radials and distally into the actinotrichia.
The muscle bundles are very distinct, and, unlike those of the adductor system, run
parallel with the palisade cartilages.

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 21).

85

556

CAPTAIN E. W. SHANN ON

III. Chondrostei.

(Accipenser rubicundus.)

1. Shoulder Girdle.

(l) The Lateral Muscle.

The composition of the lateral muscle of Accipenser closely resembles that of the
Teleostei, consisting, namely, of four series of hollow cones, two above and two below
the lateral line. In the pectoral region four areas can still be traced ; of these, the
two dorsal are completely severed from the two ventral by the passage of the hori-
zontal septum which runs from beneath the lateral line to the transverse processes
of the vertebrae.

The dorsal moiety takes a double origin from the skull. The dorsalmost portion
arises from the occiput just below the first dorsal spine and is formed by the upper
halves of the dorsal backwardly directed cones, while the ventralmost arises from the
exoccipital region and is formed by the ventral forwardly directed cones. This
differs from the condition figured by Knauer for A. sturio, in which there appears
to be only one area of origin for the dorsal moiety of the lateral muscle.

The ventral moiety is very clearly divided into a latero-ventral and a mesio-
ventral portion, which condition also appears in Knauer’s figure of A. sturio. In
the former the fibres run parallel to the long axis of the body. This is also true for
the mid-ventral fibres of the mesio-ventral portion ; but as the sides are reached the
fibres become increasingly inclined, so that they make an acute angle with those
of the latero-ventral portion. The slope of the fibres is from below anteriorly to
above posteriorly.

(2) The Posterior Muscles.

When the lateral muscle reaches the pectoral girdle it becomes inserted to a large
extent into the various parts of that structure. The muscles which are attached in
this manner act as retractors of the girdle, and will be described in relation to the
areas of the lateral muscle from which they are derived.

(a) Retractor latero-dorsalis pectoralis ( r.l.d.p .). — Some of the superficial fibres
break away from the main mass of the latero-dorsal portion of the lateral muscle and
become inserted into the posterior border of the cartilaginous suprascapula.

(b) Retractor latero-ventralis pectoralis ( r.l.v.p .). — The entire latero-ventral
portion on reaching the girdle becomes inserted into the posterior face of the
coraco-scapular cartilage. Not a fibre reaches the cleithrum. The muscle is seen
in fig. 15.

(c) Retractor mesio-ventralis pectoralis ( r.m.v.p .). — In the Sturgeon this muscle
is divided into two portions ; the third or “ medius” portion figured and described in
Elasmobranchii and Teleostei is not apparent.

THE SHOULDER GIRDLE AND PECTORAL FIN OF FISHES.

557

(i) Superior (see fig. 15).

The markedly oblique fibres which compose the lower part of the lateral body-wall
pass into a tendinous sheet which is inserted into the lowermost part of the posterior
face of the coraco-scapula.

(ii) Inferior (see fig. 14).

The ventral body-wall near its middle line passes anteriorly into a sheet of con-
nective tissue which is inserted into the hinder border of the clavicles, forming the
ventral wall of the pericardium. Laterally some of the fibres pass forwards, between
the pericardial fascia and the coraco-scapula, and internal to the ventrally projecting
process of the latter, to be inserted on the bases of the visceral arches. The last-
named muscle mass will be described later under the heading claviculo-hyoideus.

(3) The Anterior Muscles.

A. Connections with the Skull. — The presence of a “ trapezius” muscle is denied
by Gegenbaur and Vetter for Accipenser sturio. Meissner (63) describes such a
muscle in A. nudiventris, A. stellatus, A. stenorhynchus, and Pseudoscaphirhynchus
Kaufmani, and associates its presence with the increased mobility of the shoulder-
girdle in these forms compared with that in A. sturio ; he even proposes on these
grounds to reinstate the old generic name of Sturio for the latter species. Ostrotj-
moff describes the development of a “ trapezius ” in A. giildenstadtii and A. ruthenus.
I have shown the muscles of this area in A. rubicundus in fig. 15. It is also present
in Poly don.

Ostroumoff states that the development of the “trapezius” muscle, likewise its
innervation (from the ramus accessorius of the vagus), leads to the conclusion that
this muscle belongs to the dorsal suprabranchial section of the superficial constrictor,
which is not severed into myomeres, and in the beginning of its development includes
six myomeres representing the five gill pouches and the sixth rudimentary pouch.
Its development pulls the shoulder girdle forwards against the gill apparatus. Thus
the embryonic condition in the Chondrostei recalls vividly the adult condition in
the Selachii.

Meissner also shows that the trapezius is divided into two distinct portions, but
does not name them.

( a i) Protractor pectoralis anterior. Trapezius (pars), Meissner.

Lies immediately beneath the skin. It takes origin on the skull just behind
the retractor hyomandibularis, and is inserted on the scapula and in part also on
the cleithrum.

(< a ii) Protractor pectoralis posterior. Trapezius (pars), Meissner.

A very small muscle, lying somewhat deeper and posterior to the foregoing, takes
origin on the suprascapula and is inserted on the scapula.

558

CAPTAIN E. W. SHANN ON

( b i) Levator pectoralis is wanting in the Sturgeon.

B. Connections with the Visceral Arches. — In further correlation with the
relative immobility of the shoulder girdle, these muscles are fewer in number and
smaller in extent than in other groups of fishes.

( b ii) Cleithro-branchialis.*

This is a compact, strap-like muscle which, inserted on the base of the fifth
branchial arch, runs outwards and downwards, spreading slightly fan- wise to its
origin on the inner aspect of the lower third of the cleithrum.

(■ b iii) Claviculo-hyoideus. Coraco-arcualis posterior, Vetter, Ostroumoff.

This muscle arises from the ventral anterior border of the clavicle and is inserted
at the base of the hyoid arch. Unlike the condition in most Elasmobranchs, this
muscle is not divided by myocommata ; according to Ostroumoff, it is formed by a
ventral outgrowth of the third myotome.

The very minute branchio-mandibularis (fig. 14, Br.mn.) of A. rubicundus appears
to be a vestige of the coraco-mandibularis so characteristic of Elasmobranchs.

2. The Pectoral Fin.

The details of the fin musculature have not been worked out, but the general
composition is diagnostic. Both the adductor and abductor muscles are divided into
two layers, a superficialis and a profundus. The fibres of each of these muscles are
collected into distinct bundles for insertion on the fin rays.

Davidoff (20) found a similar condition in the pelvic fin of A. ruthenus, except
that the abductor was not divisible into two superimposed layers.

IV. Teleostei.

(See List of Types, p. 532.)

1. Shoulder Girdle.

(l) The Lateral Muscle.

A considerable time was devoted to the study of this muscle, and the results are
recorded in the Proceedings of the Zoological Society of London , “ On the Nature of
the Lateral Muscle in Teleostei,” 1914, p. 319.

(2) The Posterior Muscles.

(a) Retractor latero-dorsalis pectoralis. Retractor scapulae, Owen. — The latero-
dorsal portion of the lateral muscle is inserted at its anterior extremity on the side

* It is interesting to note that the Cleithro-brancliialis is single in Accipenser, while in Polypterus it is divided
into three distinct parts (fig. 24). Both these types differ in respect of this muscle from the condition whieh is so
constant.throughout the Teleostei.

Key to the Musculature of the Pectoral Arch and Fin in Teleostei.

THE SHOULDER GIRDLE AND PECTORAL FIN OF FISHES.

559

«4H

o

8

8

SP

«

B

-g

c

2

J2

Ph

S

M

1

ft

p

w

o

J w

"o ^

03

O ^

03

O ^

Dt al-
able,

apz.

2 2
SK

>> «
^JtS3

03

"3 N

£

H

m

o

o

e d

d

T3

_g

.a

r2

Ph

S

g

&

>

c3

o o

Ph

Ph

O

ffl

’>

6

o

8

>

ranch.

ant.

ranch.

post.

’yo-pjayic

l

Ph i

ri

1

2

Ph

ft

o

p

ft ft

Ph

W

w

d d

w

"> ’>

d

rjd

'g

s

g

§

CO

'g

03

4-3

re3

1

rapezius.

haryng. i
ext.

haryng. <

o

-d ^

■S 2,

Ph

P

CQ

nd

nd

Ph

«d

O

ft

ft

ft

£H

Ph Ph

" a

<1

6

d d

£

"?> *£>

o

"O

'd

P

r^2 r2

8

03

J d

cp

nd

c6

EH

03

o

6

‘pH

'o

fcb fcu

’o

d

Ph O

o ^

Ph

o

rt

t>

O

H

s s

f5

PhP

4

ft

S

03-

|

CO

2

‘3

,t3

Owen.

4-3

Do.

'c

©

c8

Sh

3) fig.
No. 26.

p

2

w

I

c«

cp

Ph

^ :
c8

03

Cp

d

pP

Ph

03

03

Xfl

£

Ph

_^ft

Zfl

ft

m

P

d

s

’o

o

O

2 o d

6

:

£

o

HQ

§«o

Q

ft

§

m

Ph

03

O

5zi

p

n 3

s

P

<<

w

Spy

CO P-

co co

d

03 .

J :

Jt

6

d ^

CD 0) rg

u

03 .

03 P

D

s

r§ p

'ce

o •

2 S «2

P g

o

p

? 2
. P

t>

w

.1

■ i

.

1 . .

. rP

4-3

1 2

«3 03

nd

_

I

at. pect.-

srior

ills

o g

03

Ph

O' r. ^

+3 o .g
S'fH g

i-tj §

■g ft ^

03 r-< o

>~a o h

03 03

■+j}

ft m

*o

6

1

J |

§ o

cp

1

S

-g

CD p/O

s >

. ^ 2 -g

'g > *■§

's' 05.

S

i &•

*§ J J

2

Ph

=3

|

"8

6 c/2 S

^ <4

£p O

O

2 C/2

£5 2

P4

r“]

4-3

s 88

8

88 rp

§ o

§

2

r2

2 ^

8

2

■® w 8

8

r& ^

88 2

3-

:P

! P3

tf

P3

4+

g

w <1

S'

8

1 8

8

■i

560

CAPTAIN E. W. SHANN ON

of the skull (exoccipital, otic recess, and post-temporal). In some Teleosteans a
small superficial branch of this muscle is attached to the supraclavicle ( Blennius
pholis), or to the cleithrum ( Cottus scorpio), where it is inserted on the posterior
border of the backwardly directed spine in which that bone ends.

In Periophthalmus the mesio-dorsal portion of the lateral muscle gives a small
branch to the upper border of the post-temporal ; but this is exceptional in the
Teleostei.

Where this muscle is present the r.l.v.p., next to be described, is wanting,
and vice versa.

( b ) Retractor later o-ventralis pectoralis. Retractor scapulae, Owen. — M‘Murrich
(60) states that the latero-ventral portion of the lateral muscle (his third portion)
disappears in Amiurus before reaching the shoulder girdle ; such is the case in some
other Teleosteans, but in the majority a small bundle of fibres, undoubtedly belonging
to this portion, is attached to the upper extremity of the cleithrum, or to the supra-
clavicle ( Trigla gurnardus), or to both ( Lophius piscatorius). The muscle is
frequently flat and fan-shaped.

In Lophius this muscle has undergone most remarkable modifications, prob-
ably in accordance with the extraordinary mobility of the limbs in that form.
Hamburger (42) has already given an account of the musculature of the fore limb
of Lophius ; but, as his description is incomplete, and he has neither named nor
illustrated the individual muscles, and as the case is of peculiar interest, I shall deal
with it here in detail. Fig. 22 shows the dorsal aspect of the right shoulder after
removal of the skin, the latero-dorsal portion of the lateral muscle, and the retractor
mesio-ventralis pectoralis superior : the various parts of the r.l.v.p. are numbered i-v.

Part i is a strap-like muscle inserted mainly on the ventral surface of the supra-
clavicle, but in part also on the dorsalmost extremity of the cleithrum (ventral aspect).

Part ii is a strap-like muscle inserted on the dorsalmost portion of the cleithrum
above the origin of i. It is also distinguished by its darker colour.

Part iii is a flattened cylindrical mass inserted in a slight hollow on the ventral
aspect of the cleithrum immediately behind the glenoid border. It passes inwards
and downwards, between parts ii and v, so as to cover the peritoneum.

Part iv is a strap-like muscle given off by part v at a level defined by producing
the long axis of the supraclavicle, and inserted into part iii at a level slightly adaxial
to the backwardly directed process of the cleithrum.

Part v is a powerful strap of muscle inserted on the ventral aspect of the abaxial
third of the horizontal stem of the cleithrum. It curves abruptly inwards and
then upwards.

Parts i, ii, and v do not go to form part of the latero-ventral body muscle, but
take origin by a fascia along the lateral line. It might be supposed that these
muscles belonged tc the latero-dorsal portion, but the fact that part v is blended
with part iii (undoubtedly ventro-lateral) through the medium of part iv, indicates

THE SHOULDER GIRDLE AND PECTORAL FIN OF FISHES.

561

that part v certainly, and parts i and ii probably, should be regarded as derivatives
of the latero-ventral portion. Embryological evidence would be of interest here, as
in other problems connected with Lophius.

(c) Retractor mesio-ventralis pectoralis. — The latero-ventral portion has been
seen to decrease in width as it approaches the anterior end, and in some cases
completely to disappear. With this decrease of the latero-ventral, the mesio-ventral
portion increases in proportion, so that it becomes attached to almost the whole
length of the cleithrum. It is divided into two distinct portions, the superior (i)
and the inferior ’(iii), while the medius (ii) portion, axillary, so constant in Elasmo-
branchs, is vestigial.

(i) R.m.v.p. superior. Subcoracoideus, Owen; Lateralis 4 (pars), M‘Murrich.

A broad thin sheet of muscle, usually traversed by myocommata, inserted along

the posterior border of the cleithrum and, in part, on the coracoid. The insertion is
entirely dorsal to that of the fin muscles, and extends nearly to the dorsal end of the
cleithrum. In this muscle the post-clavicle, when present, is embedded.

(ii) R.m.v.p. medius.

There is a triangle of connective tissue immediately behind the pectoral fin, and
this is traversed by a few muscle fibres, which are inserted on the adductor muscle
of the fin. These fibres are the only trace of a “ medius” in most Teleostei, but in
Trigla (fig. 21) the muscle shows considerable development.

(iii) R.m.v.p. inferior. Protractor ischii, Owen; Lateralis 5, M‘Murrich ;

Pelvico-clavicularis, Hamburger.

This muscle is attached to the cleithrum from just below the fin muscles to the
ventral symphysis, where it meets its fellow of the opposite side, with which it
is continuous. In those forms whose pelvic fins are thoracic or jugular in position
this muscle becomes considerably specialised, and ceases to show the myocommata
which are characteristic in lower types ; it may then be described as a pelvico-
cleithrale (fig. 21).

(3) The Anterior Muscles.

A. Connections with the Skull.

(a) Protractor pectoralis, Lambeau, Cuvier ; Protractor scapulae, Owen ;

Occipito-clavicularis, Vogt and Yung, Hamburger ; Trapezius,

Humphry, M‘Murrich.

This muscle is constantly present in Teleosteans. It takes the form of a fan-
shaped mass which, arising from the otic region of the skull, becomes inserted partly
into the forward border of the supraclavicle and partly into the cleithrum. The
muscle is superficial in position, and forms the posterior border of the upper portion
of the gill cavity.

562

CAPTAIN E. W. SHANN ON

The protractor pectoralis is invariably divided into two distinct portions, as
described below.

(i) P.p. anterior.

This part of the muscle is usually slightly internal to the succeeding part. In
point of size it may be larger, as in I Zeus faber (fig. 16), or smaller, as in Lophius
piscatorius (fig. 22), than the succeeding portion. The insertion is entirely upon the
cleithrum, but may take place upon the inner aspect (Zeus), or upon the outer aspect
(Lophius).

(ii) P.p. posterior.

This part of the muscle arises behind the former and is inserted partly upon the
supraclavicle, and, to a greater or less extent, upon the cleithrum.

(b i) Levator pectoralis. Diaphragmamuskel, Cuvier ; Trapezoide,
Vogt and Yung ; Levator, Hamburger ; Sternocleido-mastoideus,
Humphry.

A flattened columnar muscle underlies the foregoing muscles. It arises from the
basioccipital or parasphenoid and is inserted on the inner aspect of the cleithrum.
This muscle usually forms the lateral wall of the pericardium, and for that reason
was named by Cuvier “ diaphragmamuskel.”

It should be recalled that this muscle is absent in the Sturgeon ; in many
Teleosteans it is exceedingly minute, and in Motella it is absent, unless it is represented
in that form by the ligament (figs. 19, 20) which runs from the inner aspect of the
cleithrum to the basioccipital. I am inclined to believe that this is not the case,
for both in Trigla and Lophius, which possess .a normal levator pectoralis, the liga-
ment is present in addition. It is curious that in Lophius this ligament, instead of
being inserted on the side of the basioccipital or of the vertebral centra (as in the
majority of Teleosteans), unites with its fellow of the opposite side in the space
between the kidney and the vertebral column.

B. Connections with the Visceral Arches.

(b ii) Cleithro - branchialis. Branchiretractors, Owen ; Pharyngo-
clavicularis, Vogt and Yung, M‘Murrich ; Branchio-clavicularis,
Hamburger.

Two compact little bands of muscle invariably run from the anterior border of
the cleithrum to the fourth ceratobranchial. The two muscles run approximately at
right angles to one another, and their insertions on the ceratobranchial are usually
close together, or even contiguous (Zeus).

a. Cli. br. externus.

This muscle arises from the lower third of the cleithrum, and follows an approxi-
mately vertical course (Zeus, fig. 16). In Motella (fig. 20) the muscle runs obliquely
upwards. In Trigla the muscle is traversed by two transverse septa.

THE SHOULDER GIRDLE AND PECTORAL FIN OF FISHES.

563

ft. Cli. br. interims.

This muscle arises, often through the medium of a tendon (Trigla, Zeus, fig. 16),
from the upper third of the cleithrum and follows an approximately horizontal course.
It is usually considerably broader than the cli.br. ext. Hamburger states that in
Lophius this muscle is inserted upon the hyo-branchialis muscle, but my dissections
do not bear this out ; I find that the insertion occurs upon the fourth ceratobranchial
in the normal manner ; the roots of the two muscles in question are, however,
contiguous.

(b iii). Cleithro-hyoideus. Sternohyoid, Stannius, Vogt and Yung ;
Retractor hyoidei, Owen ; Hyo-pectoralis, M‘Murrich ; Hyo-
clavicularis, Hamburger.

A powerful muscle which forms the ventral border of the gill cavity. It has a
double origin on the cleithrum ; the larger portion arises from the outer aspect of
the cleithrum, usually in its lower third, while a smaller portion arises from the
region of the clavicular symphysis, where it comes in contact with its fellow of the
opposite side. The insertion occurs upon the base of the hyoid arch. The muscle
is usually much reduced in bulk as it reaches the point of insertion, but Zeus (fig. 16)
forms an exception.

2. Pectoral Fin.

(1) Adductor.

The adductor system, as in the foregoing groups of fishes, is divided into two
layers— -an outer, superficialis, and an inner, profundus. Each of these is divided
distally into very distinct bundles which become inserted, usually by long tendons,
upon the bases of the lepidotrichia. There is in addition in certain Teleosteans (Zeus,
fig. 16) a third, deeper and more ventral, layer, which serves to dilate the fin. The
third layer has been described by Cuvier, Yogt and Yung, and others, and is named
the dilator posterior. These muscles take origin in part from the cleithrum and in
part from the fin skeleton proper (coracoid, scapula, and radials).

(2) Abductor.

The abductor system conforms closely to the adductor in any given type of
Teleostean. The third layer, when present, is named the dilator anterior.

(3) Modifications.

Extraordinary specialisation of the fin muscles occurs in forms whose pectoral
fins have become modified for progression on a solid surface. These modifications
have been fully described for Trigla by Williamson (92).

The most interesting development of the muscles of the pectoral fin in the
Teleostei is exhibited by Lophius. Hamburger (42) has given a complete account

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 21). 86

564

CAPTAIN E. W. SHANN ON

of this musculature ; but, through some mischance, the descriptions of his drawings
are so inaccurate that they would be very difficult to follow without the corrections
which I have ventured to make.

The description of figs. 19, 20 is correct ; but in Hamburger’s “ Figurenerklarung ”
the descriptions of these two drawings are inverted.

On comparing fig. 17 with the statement in the text no agreement is found. The
one muscle shown is marked M.b.r. 1, which the summarised description of this figure
tells us is the basi-radialis internus ; moreover, that muscle takes origin on a bone
denoted Bs. II. Examination of my own material shows that fig. 17 represents
faithfully the basi-radiale internus — the distal portion of the adductor profundus ;
it is thus a view of the dorsal (inner) aspect, not of the ventral as stated. The
lettering of the radials (basalia) has been inverted in this particular figure, which
accounts for the second discrepancy.

The two muscles described as lying on the inner aspect of the fin in reality form
the distal portion of the deeper abductor, or outer system, and are situated as shown
in fig. 18. Here the text agrees with the summarised description, but both are at
variance with fact.

Hamburger divides the fin muscles, for practical purposes, into three systems
based on their origins and insertions, but adds that they are divisible also, as in other
Teleosteans, into an external and an internal system. The complicated condition of
the fin musculature is fairly clearly derivable from the primary Teleostean condition ;
they may be classified as follows : —

Adductor (internal) System.

I. Superficialis.

1. Claviculo-radiale internus.

2. Basi-claviculare internus superf.

II. Profundus.

3. Basi-claviculare int. prof.

4. Basi-radiale internus.

Abductor (external) System.

I. Superficialis.

5. Claviculo-radiale externus.

6. Basi-claviculare ext. superf. i.

7. Basi-claviculare ext. superf. ii.

II. Profundus.

8. Basi-claviculare ext. prof.

9. Basi-claviculare transversus.

10. Basi-radiale transversus.

11. Opponens.

To avoid confusion, Hamburger’s nomenclature has been retained ; though it
must be remembered that here “ basi ” implies connection with the radials, and
“radiale” with the fin rays or lepidotrichia, while references to the “ clavicle” are
better stated in terms of the cleithrum.

A fold of the preaxial radial grips and covers slightly superficially the claviculo-
radiale internus, but this is not indicated in Hamburger’s illustration (fig. 16). The
basi-claviculare internus superficialis, as Hamburger states, serves to extend the

THE SHOULDER GIRDLE AND PECTORAL FIN OF FISHES.

565

566

CAPTAIN E. W. SHANN ON

limb ; it is not thus functionally an adductor, though, from its origin on the dorsal
aspect of the cleithrum, it should probably be regarded morphologically as a part of
the adductor system. The claviculo-radiale externus does not supply the posterior
rays, Hamburger’s “ opposable portion” of the fin. So far I agree, but it is surely
a slip which leads him to say that this muscle draws the fin rays upwards, while the
claviculo-radiale internus draws them downwards.

The writer is indebted to the Carnegie Trust for a grant to cover the greater
part of the cost of the illustrations.

LITERATURE.

(1) Agar, YV. E., “The Development of the Anterior Mesoderm, and Paired Fins, with their Nerves

in Lepidosiren and Protopterus,” Trans. Roy. Soc. Edin., vol. xlv, 1907, pp. 611-39.

(2) Anthony, R., “Adaptation a la vie arboricole chez les Vertebres,” Ann. des Sci. Nat. ( Zool. ), xv,

2-6, 1912.

(3) Allis, E. P,, “ The Skull, and the Cranial and First Spinal Muscles and Nerves, in Scomber

Scomber," J. Morph., v, 1903, pp. 45-328.

(4) Balfour, F. M., “On the Development of the Skeleton of the Paired Fins of Elasmobranchs,”

Proc. Zool. Soc. London, 1881.-

(5) Balfour, F. M., and Parker, W. H., “On the Structure and Development of Lepidosteus,” Phil.

Trans., 1882.

(6) Bolk, L., “Beziehungen zwischen Skelet, Muskulatur u. Nerven d. Extremitaten, dargelegt am

Beckengiirtel, an dessen Musk, sowie am Plexus lumbosacralis,” Morph. Jahrb., xxi, Leipzig,
1894.

(7) Bolk, L., “Kekonst. der Segmentierung d. Gliedmassenmuskulatur, dargelegt an den Musk, des

Oberschenkels u. des Schultergiirtels,” op. cit., xxii, 1895.

(8) Bolk, L., “Die Sklerogonie des Humerus, zugleich ein Beitrag zur Bildungsgeschichte dieses

Skeletteiles,” op. cit., xxii.

(9) Braus, H., “ Innerv. paar. Extremit. bei Selach.,” Jena. Zeit. f. Naturw., xxix, 1895, and xxxi,

1898.

(10) Braus, H., “ Extremitatenskelettes,” Haeckelfest. Denkschr., Med.-Nat. Ges., Jena, 1904.

(11) Braus, H., “Die Musk. u. Nerv. der Ceratodusflosse,” Zool. Forsch., op. cit., 1901.

(12) Braus, H., “Extremit. u. d. Ex.-skelettes Hertwig’s Handb. Wirbelt., 1904.

(13) Braus, H., “ Ueber neue Funde versteinertes Gliedmassenknorpel u. Musk, von Selach.,” Verh. d.

Phys.-Med. Ges., Wurzburg, xxxiv, 1901.

(14) Braus, H., “Die Nervengellechte d. Haie u. Rochen,” Jena. Zeit., xlvii, 1911.

(15) Bruch, C., Osteol. des Rheinlachses, 2te Ausg., Mainz, 1875.

(16) Budgett, J. S., “On the Structure of Larval Polypterus,” Trans. Zool. Soc., xvi, p. 315, 1902.

(1.7) Buist, T. P., “Development of the Pectoral Girdle in the Pipe-fish (Syngnathus acus)," Ann. and

Mag. Nat. Hist., Dec. 1912.

(18) Cope, E. D., “The Homologies of the Fins of Fishes,” Amer. Nat., 1890, p. 418.

(19) Corning, H. K., “Ueber d. vent. Urwirbelknospen in der Brustflosse d. Teleostier,” Morph. Jahrb.,

xxii, Leipzig, 1895.

(20) Davidoff, Y. M., “Hint. Gliedmasse der Fische,” op. cit., 1879, 1880, 1883.

(21) Dean, B., “Historical Evidence of the Origin of the Paired Limbs of Vertebrates,” Amer. Nat.,

1902.

THE SHOULDER GIRDLE AND PECTORAL FIN OF FISHES.

567

(22) Dean, B., “The Fin-fold Origin of Paired Limbs, Ptychopterygia of Palaeozoic Sharks,” Anat.

Am., xi, 1896.

(23) Dean, B., “Biometrical Evidence in the Problem of the Paired Limbs,” Amer. Nat., No. 431, 1902.

(24) Dean, B., “Chimseroid Fishes,” Carnegie Inst., Washington, 1906.

(25) Debjughn, K., “Die Entwickelung der Brustflossen und des Schultergiirtels bei Exocectus volitans,”

Zeit. Wiss. Zool., xci, 1908.

(26) Derjugin, K., “Der Bau u. d. Entw. des Schultergiirtels u. d. Brustflossen bei den Teleostiern,”

op. cit., xci, 1909-10.

(27) Dietz, P. A., Vergleijkende Anat. van de Kaak- en Kieuw-boogspieren der Teleostei, Leiden, 1912.

(28) Dohrn, A., “ Studien Wirbelthierkorpers : VI. Die paar. u. unpaar. Flossen der Selachier,” Mitth.

Zool. Stax. Neap., v, 1884.

(29) Druner, L., “Ueber d. Musk, des Visceralskelettes d. Urodelen,” Anat. Anz., xxiii, pp. 545-71,

1903.

(30) Ducret, E., Gontr. a Vetude du Devel. des Memb. pairs et impairs des Poissons teleost., Lausanne,

1894.

(31) Furbringer, M., “ Zur vergl. Anat. d. Schultermuskeln,” Jena. Zeit., 1873, p. 237.
j(32) Furbringer, M., op. cit., 1874, p. 175.

(33) Furbringer, M., “Ueber d. rnit dem Visceralskelet verbund. spinal. Muskeln bei Selachiern,”

op. cit., 1895.

(34) Gegenbaur, C., “Zur Gliedmassenfrage,” Morph. Jahrb., v, 1879.

(35) Gegenbaur, C., “Das Flossenskelett der Crossopt. u. das Archipterygium der Fische,” op. cit.,

xxii, 1895.

(36) Gegenbaur, C., “Clavicula u. Cleithrum,” op. cit., xxiii, 1895.

(37) Gegenbaur, C., “ Ueber das Skelett der Gliedmassen der Wirbelthiere im allgem. u. d Hinterglied-

massen d. Selacbier insbesondere,” Jena. Zeit., v, 1870.

(38) Goodrich, E. S., “Notes on the Development, Structure, and Origin of Median and Paired Fins

of Fishes,” Q.J.M.S., June 1906.

(39) Goodrich, E. S., “Metameric Segmentation and Homology,” op. cit., lix, pt. 2, 1913.

(40) Guitel, F., “Devel. des nageoires prs. du Gyclopterus lumpus,” Arch. Zool. exp. et. gener., vol. iv,

3me ser., 1896.

(41) Haller, B., “Ueber d. Schultergtirtel der Teleostier,” Arch. Mikr. Anat., lxvii, 1905.

(42) Hamburger, R., “Ueber d. paarigen Extremitaten von Squalius, Trigla, Periophthalmus u.

Lophius,” Rev. Suisse de Zool., xii, 1904.

(43) Harrison, K. G., “Unpaar. u. paar. Flossen d. Teleost.,” Arch. Mikr. Anat., 1895.

(44) Hartmann, R., “Ueber d. Brustflossenmuskeln einiger Fische,” Sitzungsber. Ges. Nat. Fr. Berlin,

No. 9, 1881, pp. 150-154.

(45) Haswell, W. A., “On the Structure of the Paired Fins of Ceratodus, with Remarks on the

General Theory of the Vertebrate Limb,” Proc. Linn. Soc. N.S.W., vii, 1883.

(46) Haswell, W. A., “Studies on the Elasmobranch Skeleton,” op. cit., ix, 1884.

(47) Hill, J. P., “A Fiddler ( Trygonorhina fasciata) with Abnormal Pectoral Fins,” op. cit., x,

2nd set., 1895.

(48) Howes, G. B., “ On the Skeleton and Affinities of the Paired Fins of Ceratodus, with Observations

upon those of the Elasmobranchii,” P.Z.S., 1887.

(49) Howes, G. B., “ Observations on the Pectoral Fin Skeleton of the living Batoid Fishes and of the

extinct genus Squalo raja, with especial reference to the affinities of the same,” op. cit., 1890.

(50) Humphry, G. M., “ On the Homological Relations to one another of the Mesial and Lateral Fins

of Osseous Fishes,” J. Anat. and Phys., v, 1871.

(51) Humphry, G. M., “The Muscles of the Smooth Dog-fish ( Mustelus levis),” op. cit., vi, 1872.

(52) Humphry, G. M., “The Muscles of Ceratodus,” op. cit., vi, 1872.

(53) Huxley, T. H., “On Some Parts of the Skeleton of Fishes,” A.J.M.S. , 1859.

(54) Huxley, T. H., and Hawkins, Atlas of Comparative Osteology, 1864.

(55) Jaquet, M., “ Contr. a l’anatomie comp, des syst. squelet. et muse. (Chimsera, Callorhynchus, Spinax,

Protopterus, Ceratodius, Axolotl),” Arch. Sci. Med. Bucharest, ii, iii, iv.

568

CAPTAIN E. W. SHANN ON

(56) Jungersen, H. F. E., “Structure of Amphisile and Centriscus,” Mem,. Acad. Roy. des Sci. Zett. de

Danemark, Copenhagen, ser. 7, Sc. Section, vi, No. 2.

(57) Kerr, J. G., “The Embryology of certain of the Lower Fishes, and its hearing upon Vertebrate

Morphology,” Proc. Roy. Phys. Soc. Edinburgh, xvi, No. 5, 1906.

(58) Klaatsch, H., “Die Brustflosse der Crossopterygier, etc.,” Festschr. Carl Gegenbaur, Leipzig, 1896.

(59) Luther, A., “Beitrage zur Kenntniss von Muskulatur u. Skelett des Kopfes des Haies u. d.

Holocephalen,” Acta Soc. Sc. Fenn., xxxvii, No. 6, 1909.

(60) M‘Murrich, J. P., “The Osteology of Amiurus catus,” Proc. Canad. Inst., ii, No. 3, 1884.

(61) Marey, M,, “Des mouvements de natation de la Raie,” Gompt. Rend. Acad! Sci., cxvi, Paris, 1893.

(62) Marion, G. E., “ Mandib. and Phar. Muscles of Acanthias and Raja,” Contr. Biol. Lab., ™ uft’s Coll.,

No. 43.

(63) Meissner, von W., “Neue Beit, zur vergleich. Anat. des Schultergiirtels der Accipenseriden,” Zool.

Anz., xxxii, 1907-1908, p. 465.

(64) Mivart, St G., “Fins of Elasmobranchs, etc.,” T.Z.S., London, x, 1879.

(65) Mollier, S., “Paar. Extrem. Wirb. : I, Das Ichthyopterygium,” Anat. Hefte, i, 1893.

(66) Mollier, S., II, “Das Cheiropterygiuni,” op. cit., v, 1895.

(67) Mollier, S., Ill, “Die Entw. der paar. Floss, des Stors,” 1909, op. cit., viii, 1898.

(68) Mueller, E., “Die Brustflosse der Selachier,” op. cit., xxxix, 1909, p. 118.

(69) Mueller, E., “ Untersuch. ii. d. Musk. u. Nerv. der Brustflosse u. d. Korperwand bei Acanthias

vulgaris,” op. cit., xliii, 1911, p. 129.

(70) Osburn, R. C., “ New Evidence from Primitive Sharks on the Origin of the Limhs of Vertebrates,”

Science, N.S., xxv, 1907.

(71) Osburn, R. C., “ The Origin of Vertebrate Limbs : Recent Evidence upon the Problem from Studies

on Primitive Sharks,” Ann. N.Y. Acad. Sc., xvii, 1907, pp. 415-436.

(72) Osburn, R. C., “Observations on the Paired Limbs of Vertebrates,” Amer. J. Anat., vii, 1907,

pp. 171-194.

(73) Owen, R., Anatomy of Vertebrates, vol. i.

(74) Parker, W. K., “Monograph on Shoulder Girdle and Sternum,” Ray Soc., 1868.

(75) Rauteneeld, von E., “ Morph. Untersuch. ii. d. Skelett d. Gliedmassen v. Ganoiden u. Teleostiern,”

Inaug. Diss. Dorpat, 1882.

(76) Rennie, J., “Accessory Fins in Raia batis Anat. Anz., xxviii, 1906.

(77) Ruge, E., “Vord. Extrem. von Spinax,” Morph. Jahrb., xxx, 1902.

(78) Ryder, J. A., “ Evolution of the Fins of Fishes,” Rep. Comm. Fish and Fisheries, Washington, 1886.

(79) Semon, R., “Paar. Floss. Ceratodus,” Zool. Forscli. u. Jena. Denh., 1898.

(80) Siebenrock, F., “ Ueber die Verbindungsweise des Schultergiirtels mit den Schadel bei den

Teleosteern,” Ann. K. K. Nat. Hofmuseums, Wien, xvi, 1901.

(81) Starks, E. C., “ The Shoulder Girdle and Characteristic Osteology of the Hemibranchiate Fishes,”

Proc. U.S. Mus., xxv, 1902, p. 619.

(82) Swinnerton, H. H., “A Contribution to the Morphology and Development of the Pectoral Skeleton

of Teleosteans,” Q.J.M.S., xlix, pt. ii, 1905.

(83) Swinnerton, H. H., “Evolution of the Structure and Formation of the Fin in Fishes,” Sc. Progress,

No. 19, Jan. 1911.

(84) Swirski, G., “Untersuch. ii. d. Entw. des Schultergiirtels u. d. Skelets der Brustflosse des Hechts,”

Inaug. Diss. Dorpat, 1880.

(85) Taylor, M., “The Development of Symbranchus marm.,” Q.J.M.S., 1913.

(86) Thacher, J. K., “Median and Paired Fins,” Trans. Conned. Acad., 1877.

(87) Tiesing, B., “ Ein Beitrag zur Kenntniss der Augen-, Kiefer- u. Kiemenmuskeln der Haie u. Rochen,”

Jena. Zeit., xxx, 1895.

(88) Vetter, B., “Verg. Anat. der Kiemen- u. Kiefermuskulatur der Fische,” op. cit., viii, xii, N.F. i.

(89) Wiedersheim, R., Das Gliedmassenskelet der Wirbelthiere, Jena, 1892.

(90) Wiedersheim, R., “Ueber d. Entw. des Schulter- u. Beckengiirtels,” Anat. Anz., iv, p. 428, 1889 ;

v, p. 13, 1890.

THE SHOULDER GIRDLE AND PECTORAL PIN OF FISHES.

569

(91) Wiedersheim, R'., “ Das Skelet . . . Lepidosiren annectens,” Morph. Stud. Jena, i, 1880.

(92) Williamson, H. C., “On the Anatomy of the Pectoral Arch of the Grey Gurnard,” S.F.B., xii,

pt. iii, p. 322.

(93) Woodward, A. S., “The Evolution of Fins,” Nat. Sci ., 1892,

INDEX TO FIGURES.

I. Cartilages and Bones.

Br.c. 1-

>5 = Ceratobranchials 1st to 5th.

Mpt.

= Metapterygium.

Br.hyp.

= Hypobranchial.

Ppt.

= Propterygium.

Gla.

= Clavicle.

Ppt.r.

= Propterygial radial.

Gli.

= Cleithrum.

Pt.cla.

, = Post-clavicle.

Cor.

= Coracoid.

Pt.t.

= Post-temporal.

F.r.

= Fin ray.

S.cla.

= Supra-clavicle.

Hy.b.

= Basihyal.

S.sc.

= Supra-scapula.

Hy.c.

= Ceratohyal.

Sc.

= Scapula.

Mn. .

= Mandible.

II. Muscles.

Abd.sup. = Abductor superficialis (Nos. = muscle-bundles, 1 preaxial)
Abd.prof. = Abductor profundus do. do. do

Add.sup. = Abductor superficialis do. do. , do.

Add. prof. = Adductor profundus do. do. do.

Br.mn. = Branchio-mandibularis.

G.a.c. = Coraco-arcualis communis.

C.a.s. = Coraco-arcual septum.

C.br. 1 to 5 = Coraco-branchi.ales of 1st to 5th branchial arches.

C.hy. = Coraco-hyoideus.

C.m. = Coraco-mandibularis.

Gla.hy. = Claviculo-hyoideus.

Cli.br. ext. = Cleithro-branchialis externus.

GU.br.int, = Do. internus.

Gli.br.ter. £=5 Do. tertius.

Gli.hy. = Cleithro-hyoideus.

DU. a. = Dilator anterior.

Dil.p. = Dilator posterior.

L.p. = Levator pectoralis.

P.cli. =Pelvico-cleithrale = R.m.v.p.inf.

P.d.p. = Protractor dorsalis pectoralis.

P.p.ant. = Protractor (latero-ventralis) pectoralis anterior.

P.p.pst. = Do. * do. do. posterior.

R.d.p. = Retractor dorsalis pectoralis.

R.l.v.p.ext. = Retractor latero-ventralis pectoralis externus.

R.l.v.p.int. = Do. do. do. internus.

R.m.v.p.sup. = Retractor medio-ventralis pectoralis superior.

R.m.v.p.med. = Do. do. do. medius.

R.m.v.p.inf. = Do. do. . do. inferior.

Ins. = Insertion of.

Or. = Origin of.

Yol. LII.

j Trans. Roy. Soc. Edin.

Captain E. W. Shann on

The Comparative Myology of the Shoulder Girdle and Pectoral Fin of Fishes.” — Plate II.

Fig. 8.

Add.sup.(2)
Or. (pars).
Add.sup. Or.

Add.prof.

Mpt.

Cor. Abdisup. Abd.prof.

Abd.sup. { J
Abd.sup.ro
Ins. (pars

Fig. 10.

Trans. Roy. Soc. Klin.

Yoh. LI I.

Captain E. W. Shann on “The Comparative Myology of the Shoulder Cirdle and Pectoral Fin of Fishes.”— Plate II

Fig. 13.

Vol. LII.

Trans. Roy. Soc. Edin.

Captain E. W* Shank on “ The Comparative Myology of the Shoulder Girdle and Pectoral Fin of Fishes. —Plate IY.

S.cla.

-Ligaftient

Abd.prof. -

( 571 )

XXII. — The Morphology of the Stele of Platyzoma microphyllum, R. Br. By
John M‘Lean Thompson, M.A., D.Sc., F.L.S., Lecturer on Plant Morphology,
Glasgow University. (With Three Plates and Three Figures in the Text.)

(MS. received February 3, 1919. Read February 3, 1919. Issued separately September 29, 1919.)

That the stelar problem considered in this paper may be more clearly visualised,
it will be well to divest of purely theoretical considerations the facts of stelar
structure already known for Platyzoma, and to summarise such structural features
as bear on the subject in hand.

In the first account of Platyzoma given by Robert Brown in 1810 (l), the habit
of the stem and heterophyllic leaves, and the general form and position of the
sporangia, were alone described. But in 1832 (2) a rough analysis of the structure of
the stem was added, and the tubular nature of its stele was recognised. No material
contribution to the knowledge of the stelar structure was made until 1893 (3), when
Dr Poirault described the histology of the stele as shown by a small fragment
of stem. In this material there was a sclerotic pith completely surrounded by an
endodermis, outside which lay a zone of parenchyma surrounded by a broad ring
of parenchymatous xylem. At the periphery of the xylem there was a narrow
zone of phloem surrounded by a large-celled pericycle and an outer endodermis.
Neither leaf-gaps nor perforations were found in the stele, and accordingly the
complete isolation of the pith from the cortex was recognised.

Fuller information regarding the stele was provided in 1901 by Mr Boodle (4).
In particular, he described the origin of a leaf-trace by the nipping off of a group of
phloem- and xylem-elements from the stele. The leaf-trace thus initiated departed
from the stele without the formation of a leaf-gap affecting the endodermis;, and
it was noted that frequently the exit of a leaf-trace scarcely affected the stelar-
xylem beyond forming in it a small external pit. “ The nearest approach to a true
leaf-gap ” which was observed was a parenchymatous ray running almost radially
athwart the xylem-ring, but “ longitudinal sections through the insertion of a
leaf- trace did not show any passage of leaf- trace -parenchyma towards the centre
of the stele.”

In 1916 (5) I described at some length the stelar structure as shown in part of
a single specimen of Platyzoma microphyllum. From the central stele there departed
a close succession of leaf- and root-traces. The leaf-traces varied considerably in
size. The largest originated mainly from the upper surface of the stele, the smaller
sprang from the sides and lower surface. The majority of the root-traces were
attached to the lower surface of the stele. The pith was bulky and sclerenchy-
matous, and — as in the preparations of Poiratjlt and Boodle — was surrounded by

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 22). 87

572

DR JOHN M'LEAN THOMPSON ON

an uninterrupted endodermis which was immediately followed by a few layers of
large-celled parenchyma. This parenchymatous zone was followed on its outer
surface by a ring of xylem which varied in thickness as one passed from the
upper to the lower surface of the stele. On its upper side the xylem-ring
was thin ; on its lower side it was relatively thick. The xylem was composed
of scalariform tracheides together with chains and groups of parenchymatous cells
which followed irregular courses between the tracheides, and which fused and
branched repeatedly. The xylem was roughly divisible into two concentric cylinders
according to the dimensions of the tracheides and the distribution of the
parenchyma. The inner cylinder was composed of wide tracheides associated with
numerous chains and masses of parenchyma, and formed the essentially parenchy-
matous part of the xylem. The outer cylinder was composed mainly of small
tracheides which were closely packed together. Parenchyma did not bulk largely
in the organisation of this outer cylinder, the chains and groups being
usually short and narrow, and their component elements were few. This outer
cylinder supplied the xylem to the traces of the leaves and roots. The difference
in xylem-thickness between the lower and upper sides of the stele was due mainly
to a massing of parenchymatous inner xylem towards the lower surface of the
stele. The inner xylem fluctuated from point to point either in the direction
of increase or of decrease. The greatest increases were in those regions where
the departing traces were most crowded ; on the other hand, the inner xylem
was usually less prominent in those portions of the stele where the traces were
large and scattered.

The phloem followed almost immediately upon the outer surface of the xylem,
and was separated from the xylem locally only by a single layer of small parenchy-
matous cells. It varied in bulk from point to point, but seldom exceeded five
layers in breadth. It was composed entirely of typical fern sieve-tubes, and was
interrupted only at those points where leaf- and root-traces were departing. The
phloem was followed externally by a large-celled pericycle of three to four
parenchymatous layers, and beyond this was a continuous outer endodermis. These
features are in complete harmony with those described by Poiratjlt and Boodle.

In the formation and departure of both large and small leaf-traces one general
method was adopted. The first indication of the formation of a leaf-trace was the
marking-out of a peripheral tangential band of tracheides from the outer xylem-
cylinder. Beneath this tracheidal band a definite mound of parenchyma appeared,
while outside it the phloem, pericycle and outer endodermis were gently curved.
As the band of xylem became more distinct and arched, the phloem upon its convex
face became thinner. The xylem-band then freed itself from the stelar-xylem,
and passed out as the arched xylem of a leaf-trace. In some cases it severed its
connections with the stelar-xylem simultaneously on both sides ; in others it freed
itself more rapidly on one side than on the other. In either case, while the phloem

THE MORPHOLOGY OF THE STELE OF PLATYZOMA MICROPHYLLUM, R. BR. 573

clothing the leaf-trace-xylem on its outer surface had become thinner, no diminution
of phloem took place at the points where the leaf-trace-xylem had freed itself. In
point of fact, no sooner had an edge of the leaf-trace-xylem become free than
phloem began to appear in the gap which separated it from the stelar'-xylem, and
as the gap widened a sheet of phloem appeared in the parenchymatous mass beneath
the leaf-trace-xylem. This sheet of phloem was an inward continuation of the
stelar-phloem beneath the departing leaf-trace-xylem. It was destined to meet a
similar phloem-sheet projecting through the gap which separated the leaf-trace-
xylem from the stelar-xylem at the other edge, and when these phloem-sheets had
met and fused, a continuous sheet of phloem was spread once more over this
portion of the stelar-xylem. Meanwhile the local depletion of the outer xylem
had been corrected, and the “external pit” had been obliterated.

While these structural changes were proceeding, the leaf-trace-xylem and its
accompanying crescent of phloem had continued their outward course, and the leaf-
trace-phloem had become free from the stelar-phloem. The pericycle and outer
endodermis had been moved outwards to accommodate the xylem and phloem of
the leaf-trace, and when the limit of this accommodation was reached, the leaf-
trace possessed a crescent of xylem with phloem on its convex surface and around
its margins.

When the gap in the peripheral cylinder of stelar-xylem had been repaired, and
phloem again covered the xylem, a new endodermal formation appeared on both sides
of the departing leaf-trace. Endodermal cells spread obliquely inwards from two
points upon the arched outer endodermis, and having spread inwards, they united
to establish a continuous endodermal wall which completely excluded the leaf-trace
from the stele. No sooner had this endodermal formation been completed, than the
outward arched portion of the outer endodermis which had accompanied the other
components of the leaf-trace on their outward course became free as a crescentic
band following the general outline of the trace. As a consequence of these modifica-
tions the endodermis of the free leaf-trace was at this stage incomplete, being
maintained upon the outer side of the trace, but absent from the inner side. This
condition was not long maintained. An inner band of endodermal cells soon appeared
completing the leaf-trace-endodermis upon its inner side. It was then evident that
in the process of leaf -trace-formation and -departure, the inner endodermis was in
no way involved ; and further, there existed no point where the stele was not entirely
enclosed by an outer endodermis. Not only has it been demonstrated that in this
plant — so far as the materials examined are concerned — there existed no ground-
tissue-connection between the cortex and pith, but also that no connection existed
between the cortex and the pericycle at the point of final liberation of the leaf -trace
into the cortex, or at any other point. The possibility of such a connection had,
in fact, been removed by the formation of the fused endodermal sheets as
described above.

574

DR JOHN M‘LEAN THOMPSON ON

It frequently occurred that when the liberation of a leaf-trace made a heavy
drain on the tracheides of the outer cylinder, so that the parenchymatous inner
xylem was locally in direct parenchymatous connection with the parenchyma beneath
the leaf-trace-xylem, a parenchymatous bridge appeared in transverse section passing
athwart the xylem-ring. But many chains of parenchyma traversed the inner xylem
from side to side at points other than the insertions of leaf-traces, and in not a few
instances were continued almost to the periphery of the outer xylem-cylinder.
Where the latter was locally parenchymatous or thin these rays were sometimes seen
in transverse section to be continuous from the inner to the outer surface of the
xylem. The rays were usually narrow and followed varying courses, and, in branch-
ing, ran in obliquely longitudinal directions, both backwards and forwards in the
stem. In no case observed did they disturb , or threaten , the unbroken continuity of
the inner and outer endodermal bands.

From the above account it will be apparent that no leaf-gaps were found in the
length of stem examined, and that the local depletion of the outer xylem-cylinder,
or a parenchymatous development within it, led to the formation of rays which
traversed the xylem from side to side, and that since such local depletions of the
outer xylem-cylinder were unavoidable at the points of departure of leaf-traces, it
was at these points that continuous rays were most frequently seen in transverse
sections of the stele.

No special interest was presented by the origins of the root-traces. The latter
were restricted to the lower surface and lower portions of the sides of the stele.
They originated like the leaf-traces by a protrusion of the outer xylem-cylinder, and
frequently coincided with xylem-rays, but in departing they created no gaps in the
stelar ring by which direct continuity of pith and cortex could be established.

The facts thus summarised have been fully illustrated in Plates I, III, and IY of
the memoir mentioned above (5). With the facts recorded by Dr Poiratjlt and
Mr Boodle, they provide the only record of stelar structure which has been made
public for Platyzoma. They bear only one interpretation, namely, that in the
materials so far investigated the stele comprises a conductive cylinder with both
outer and inner endodermis. It is- devoid of leaf-gaps and perforations , and
includes centrally a pith completely isolated from the cortex. The phloem is present
only on the outer face of the xylem. The stelar structure is of the so-called
ectophloic-siphonostelic type.

This curious stele has been the subject of some discussion, and it will be well to
draw together and briefly consider the theoretical statements which have been made
regarding it. In the memoir already referred to (3), Dr Poiratjlt considered that
the stelar structure of Platyzoma might perhaps be brought in line with that of a
Eugleichenia. In explanation of this view he suggested that the pith is the result
of confluence and final isolation of a series of foliar sclerotic pockets. This con-
ception arose from the fact that in the leaf-trace of certain Gleichenias a mass of

THE MORPHOLOGY OF THE STELE OF PLATYZOMA MICRQPH YLLUM, R. BR. 575

decurrent sclerenchyma is gradually surrounded by tbe leaf- trace as the latter is
followed downwards, so that a deep sclerotic pocket is enclosed by the tubular portion
of the trace and appears, in transverse section, completely isolated from the cortex.
“ If now,” he continues, “ we imagine that, the length of the internodes diminishing,
the leaves of a Eugleichenia approach each other closely, the base of each one of
these being the seat of formation of a pocket, one can conceive the coalescence of
these closely meeting' formations leading to a general inclusion which will be enclosed
in the stem like that seen in Gleichenia polypoides, and the resulting structure will
be very near to that seen in the stem of Platyzoma.” *

It will be apparent that Dr Poirault imagined certain non-existing conditions
as fulfilled in a Eugleichenia , and that he merely contemplated a possible explana-
tion for Platyzoma , for which no adequate supporting evidence was provided.

This idea of the origin of the pith of Platyzoma was taken up in 1897 by
Dr Jeffrey (6). On page 869 of the report of the British Association meeting it
is stated : — “ Among the Gleicheniacese we have in Martensia the cortex sending
parenchymatous strands into the vascular axis of the stem down through the
channelled leaf-traces. In Gleichenia and Platyzoma these are completely cut off
from the outside cortex, and we have a completely included pith similar to that
present in Osmunda. The pith in these forms is in reality extrastelar, but no longer
communicates with the peripheral cortex.” Thus a possible explanation advanced
by Dr Poirault for the stele of Platyzoma was transformed by Dr Jeffrey into
the true and actual explanation, even although no structural evidence was advanced
which would justify dogmatism. For it will be noted that a direct ground-tissue-
connection between the cortex and stelar pith of Platyzoma has not been recorded
by any investigator, and accordingly the cutting of such a connection has not been
demonstrated even in the phyletic sense.

In 1900 (7) the extremity of this position was modified, though indeed no further
facts had emerged regarding this plant. Thus on page 7 of this paper it was stated :
“ It has further been rendered probable by the interesting investigations of Poirault
that the apparently medullated monostelic central cylinder of the stem of the
Gleicheniaceous genus Platyzoma also possesses a pith derived from the extrastelar
fundamental tissue.”

In 1901 (4) Mr Boodle, after personal observation of the stelar structure, wrote
as follows : — “ Platyzoma appears to be a xerophytically reduced form in which the
leaf-traces have become small and crowded ; it is perhaps probable that it may have
been derived from a solenostelic form, by obliteration of the leaf-gaps and disappear-
ance of the internal phloem of the stele. This view is, I believe, held by Dr E. C.
Jeffrey. It is possible, on the other hand, that Platyzoma may have been derived
from a protostelic Gleichenia, and its structure might then be due to the formation
of a pith and internal endodermis.” An extrastelar origin for the pith of Platyzoma

* The italics throughout this paper are those of the writer of this memoir.

576

DR JOHN M‘LEAN THOMPSON ON

was thus contemplated by Mr Boodle as 'perhaps probable, but an intrastelar origin
within an original protostele seemed also possible.

In thus stating the views which seemed alternative, Mr Boodle clearly showed
that the structural facts — to which he was the chief contributor — were inadequate
foundations for a confident opinion as to the true origin and nature of the medullated
stele of Platyzoma.

Mr Boodle’s work on the anatomy of the Gleicheniacese was followed in 1902 by
a paper by Dr Jeffrey on “The Structure and Development of the Stem in the
Pteridophyta and Gymnosperms” (8). It included no new observations on Platy-
zoma, but the stele was said to closely resemble that of Osmunda cinnamomea,
since it is a collateral stelar tube with an internal endodermis. The author
further remarked: “Through the kindness of Dr D. H. Scott he# has had the
opportunity of examining some sections of a dried stem of this species, f The foliar
gaps are nearly degenerate in this small extremely xerophytic fern. It would be
interesting to investigate branching specimens of this species if such occur, to
discover whether the internal endodermis is continuous with the outer one through
the ramular gaps. It seems not improbable that this may turn out to be the case.”
At a later point in the present paper it is shown that although the stem of
Platyzoma is frequently branched, no ramular gaps exist in the materials examined ;
the pith has always been found to be completely isolated from the cortex, and no
connection has been seen between the outer and inner endodermal sheaths. The
suggestion contained in the concluding remarks of the above quotation may therefore
be set aside as disproved. But the statement that “ the foliar gaps are nearly
degenerate ” remains to be considered. It clearly implies that some observation has
been made from the sections referred to suggesting the presence of reduced or
degenerate foliar gaps, and, if justified, might aid materially in the interpretation of
the stele. For it will be evident that the presence of nearly degenerate foliar gaps
would be strong evidence in support of a belief that the so-called ectophloic
siphonostele of Platyzoma is the result of reduction, and was at least preceded at
some point in the ancestry by a siphonostele in which foliar gaps were well defined,
so that the pith and cortex were in direct connection through them. Dr Scott and
Mr Boodle have kindly put at my disposal the sections already examined by
Dr Jeffrey. Each one possesses parenchymatous xylem-rays like those already
described in the preface of this paper. At several points leaf-traces are departing
from the stele, and in some instances a parenchymatous ray links the parenchyma
beneath the leaf-trace-xylem with the parenchymatous zone immediately inside the
xylem-cylinder.

It was from such sections as these that Mr Boodle’s analysis of the vascular
construction was made, and it was probably from one of them that his illustra-
tions of a departing leaf-trace were taken (figs. 33, 34, pi. xxxix), “ to show the
* Dr Jeffrey. + i.e. Platyzoma.

THE MORPHOLOGY OF THE STELE OF PLATYZOMA MICROPHYLLUM, R. BR. 577

nearest approach to a true leaf-gap met with.” In them no leaf-gap affecting the
endodermis is present, and the only structures which might possibly be interpreted
as nearly degenerate leaf-gaps are the parenchymatous xylem-rays immediately
internal to the leaf-trace-xylem. The most obvious ray of this kind which I have
found in these sections is represented in fig. 26 of the present paper. By them-
selves such rays can scarcely be held to prove the prior existence of foliar gaps of
which they are the degenerate relics, and accordingly they cannot reasonably be
brought into court as proof of the present reduced state of the stele of Platyzoma.
If such xylem-rays are to be considered the almost degenerate relics of leaf-gaps,
all prominent rays which completely or almost completely traverse the xylem-ring
might similarly be considered the degenerate, relics of stelar gaps of one kind or
another. However true it may prove to be that “ there is ground for the belief that
the solenostele may give rise by reduction to a modification not distinguishable in
any way from the medullated monostele of Van Tieghem” (8), the stelar reduction
indicated above has not been proved for Platyzoma. And further, there will be a
natural hesitation in accepting as justifiable Dr Jeffrey’s comprehensive statement
that “ the simplest view, and that most in harmony with all the facts, seems to be
that the medullated monostelic central cylinder has been derived from an ancestral
siphonostelic condition with internal phloem by reduction” (8). For the stele of
Platyzoma is a medullated monostelic cylinder which has not been shown to have
possessed at any time internal phloem and foliar gaps.

The reduction theory was supported by Mr Tansley in 1908 (9), and it may be
well to quote at length his remarks. On page 44 of his Lectures he wrote : “ Platy-
zoma has a curious and rare type of stele containing a pith surrounded by an
endodermis but no internal phloem, and with quite small collateral leaf-traces whose
departure does not break the continuity of the vascular ring as it does in the normal
solenostelic type.” And on page 106, while discussing the interpretation of the
Osmundaceous stele, he stated that “ferns which have a continuous internal
endodermis and no internal phloem are decidedly rare. We have such a case,
however, in the Grleicheniaceous genus Platyzoma, and the stele of this plant also
resembles that of the Osmundacese in having no leaf-gaps putting the cortex in
connection with the pith, the leaf-traces departing from the outer surface of the
xylem-cylinder somewhat as in the simpler Osmundaceous types. There is good reason
for considering Platyzoma as an instance of reduction.” On page 108 it was stated :
“It is no doubt conceivable that Osmunda skidegatensis is, reduced from a normal
dictyostelic type in which the leaf-traces have secondarily come to depart from the
outer face of the xylem-cylinder, their internal phloem has ceased to be connected
with the internal phloem of the stem, and the leaf-gaps have become very narrow.
Some such process of reduction as this has very likely occurred in Platyzoma.”
And lastly, on page 134- he wrote : “If the reduction of leaf- traces relatively to the
stele is great enough, a siphonostelic form like Platyzoma may have no leaf-gaps.”

578

DR JOHN M'LEAN THOMPSON ON

I have quoted these remarks in full so that the reader may see more clearly that
they are not the bearers of new morphological facts, but are merely the restatement
of the reduction theory already advocated. But they entail an implication which
may be briefly considered, namely, that in an organism under xerophytic conditions
simplification or reduction of the axile stele may be expected to take place, and in
the case of Platyzoma has actually occurred in the manner indicated above. Such
a suggestion may gather strength from the facts that the leaves of Platyzoma are
densely crowded upon the axis, and many of the leaves are of reduced and remark-
ably modified type (5). The axis is then probably of a condensed and slow-growing
type, as might be expected under xerophytic conditions. But from this it does not
seem to follow reasonably that telescoping of the stele has taken place in the
phyletic sense with the obliteration of foliar gaps. And further, even though it
were proved that stelar modifications had arisen in Platyzoma under xerophytic
conditions, no structural evidence has been advanced to show that the stele of
Platyzoma was at any time solenostelic. Much less has it been shown that in any
solenostelic fern under xerophytic conditions stelar reduction has taken place and
has followed the lines indicated, namely, that leaf-gaps have been obliterated and
inner phloem lost.

But there is another aspect of this question which must be considered, for it
lies behind Mr Boodle’s alternative view that “ Platyzoma may have been derived
from, a protostelic Gleichenia, and its structure might then be due to the formation
of a pith and internal endodermis.” This suggestion may justifiably be transformed
so as to be read that the present medullated monostelic condition of Platyzoma is the
direct result of medullation of an original proto stele which has not acquired leaf -gaps
during transformation. It is to be granted at once that this alternative upgrade
view stands, as does the reduction theory, not only “ not proven,” but also so far
unsupported by facts of observation. It presumes that the pith and internal
endodermis were created de novo, whereas the reduction theory presumes intrusion
of the cortex and outer endodermis into the stele through leaf-gaps, the formation
of a pith of purely cortical origin, and other modifications which culminated in the
establishment of a typical solenostele ; and when this had been accomplished —
according to the reduction theory — the leaf-gaps were obliterated and the inner
phloem completely disappeared. The implication of Mr Boodle’s view of the stele
of Platyzoma is that the present medullated monostelic state is the high-water
mark of elaboration from an original protostele reached either under xerophytic
conditions or independently of them. If this view could be reasonably supported
by structural facts, it would merit the same consideration as would be due to the
reduction theory if structural evidence of degenerate internal phloem and of actual
leaf-gaps were forthcoming.

The matter stood thus in suspense when in 1908 Dr Jeffrey’s paper entitled
“ Are there Foliar gaps in the Lycopsida ? ” appeared (10). From it I am extracting

THE MORPHOLOGY OF THE STELE OF PLATYZOMA MICROPHYLLUM, R. BR. 579

certain statements which, seem to have a bearing on the subject in hand. On page
241 it is stated : “ There are two phylogenetic types of tubular central cylinder,
namely, that in which only ramular gaps are present, and that in which both
ramular and foliar gaps occur.” “ The use of these constant and characteristic
anatomical features results in the division of the Vasculares into two great primitive
stocks — the Lycopsida which are cladosiphonic and palingenetically microphyllous,
and the Pteropsida which are phyllosiphonic and palingenetically megaphyllous.”
“ The Lycopsida include the Lycopodiales and Equisetales ; the Pteropsida include
the Filicales, Gymnosperms, and Angiosperms” On page 242, in amplification of
the above statement, he continued : “ Pteropsida. Palingenetically megaphyllous
vascular plants with dorsi-sporangiate sporophylls, the tubular central cylinder when
present characterised by foliar gaps or interruptions in the fbro-vascular tissue
immediately above the outgoing foliar traces .” In this way it is implied that a
fern possessing a tubular central cylinder like that of Platyzoma, devoid of foliar
gaps or other interruptions in the vascular tube, is uncharacteristic for ferns, and
can reasonably be regarded only as of downgrade and reduced type. The alternative
upgrade view was not even considered.

This attitude was still more clearly shown in 1910 when Dr Jeffrey’s paper
on “The Pteropsida ” appeared (ll). In it the “morphological status” of pith
was regarded as “ external tissue included by the stele in the course of evolution.”
On page 408 it is stated : “ It appears evident, since the stelar system of the stem
in the case of certain of the lower Pteropsida is able to include tissues and sub-
stances which are beyond question extrastelar in their origin, that no difficulty
arises in regarding the pith present within the siphonostelic central cylinder of the
lower as well as the higher Pteropsida as morphologically equivalent with the
fundamental tissues of the cortex, with which it is often continuous through the
gaps in the stelar walls resulting from the exit of the vascular supply of leaves and
branches.” “ A further argument in favour of this view is the frequent textural
similarity between pith and cortex, even where the former as the result of upward
evolutionary tendencies is no longer continuous with the fundamental tissues
without the stele.” On page 412 he concludes : “ It now in fact appears much more
clearly than formerly that the primitive condition of the vascular system in both
stem and leaf in the Vasculares as a whole was what the present writer has
designated protostelic, i.e. a condition in which the fibro-vascular tissue harboured
no pith. Following this condition was one in which the fibro-vascular system
became transformed especially in the more progressive stem into a stelar tube lined
both externally and internally with phloem and endodermis. In the process of time
the internal phloem became degenerate, probably on account both of the absence
of direct relation to the leaves and of the appearance of secondary growth, advan-
tageously localised ultimately on the outer surface of the stele. The internal
endodermis more slowly followed the internal phloem into oblivion, and is often

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 22). 88

580

DR JOHN M‘LEAN THOMPSON ON

found at the present time in the young individual when it is absent in the adult.
( Ophioglossaceee , Equisetacese, Ranunculacese.) The pith must in all cases he
regarded as a derivative of the cortex, which has become more or less completely
segregated within the stele." These remarks contain no actual reference to
Platyzoma, but according to their conclusion the pith of Platyzoma must be
regarded by Dr Jeffrey as of extrastelar origin, and the stele itself as reduced so
as to have lost internal phloem and leaf-gaps, but still to have retained its inner
endodermis.

I have dwelt on the published statements of those who have supported the
reduction theory for the stele of Platyzoma so that the arguments involved might
be fully understood. By reiteration rather than the support of cogent facts they
have been strengthened, but they have not solved the stelar problem. I am driven
to the conclusion that the reduction theory as so far stated stands merely upon
unwarranted and comprehensive assumptions, and that there is no justification in
the recorded facts of fern-anatomy for regarding the reduction hypothesis for
Platyzoma as anything more than an undemonstrated possibility. Having regard
to the recorded facts, I am constrained to maintain this position with regard to the
stelar morphology of Platyzoma in direct opposition to Dr Jeffrey’s recent state-
ment that, “ so far as the plentiful evidence in the case of the Filicales is concerned,
it seems beyond reasonable doubt that the median parenchyma of the tubular or
siphonostelic central cylinder has come from the outside, and is not the result of
internal differentiation within the stele” (12).

But the object of this paper is not mere criticism of the theoretical statements
which have grown around a limited body of fact. It is rather the recognition
of the problems of stelar structure which await solution for Platyzoma, and the
contribution and employment of certain new facts which have emerged from recent
investigation. It was evident from the outset that the facts obtained from a few
sections of the stem and from the general habit of this plant were totally inadequate
to allow of a confident opinion as to the true nature and origin of the tubular stelar
structure. Ample material embracing “sporeling” stages and mature plants could
alone provide the fuller ontogenetic facts which are necessary. The “sporeling”
stages have not yet been obtained, but through the kindness of Dr Bailey of
Brisbane Botanic Gardens a number of plants of various ages have been available
for study. In most cases the stem was robust, and, though basally incomplete,
measured from three to five inches in length. In all of them the apparent zonation
of leaves already described was seen, and all were heterophyllic (5), (13). The
majority of the larger stems were dichotomously branched, and in one instance the
branching had occurred at least twice. The materials included one small and
apparently relatively young plant. Its stem was more slender than that of any of
the larger specimens. Basally it was incomplete, but otherwise it was a perfect
specimen. It showed well the leaf-zonation, there being three groups of pinnate

THE MORPHOLOGY OF THE STELE OF PLATYZOMA MICROPHYLLUM, R. BR. 581

leaves alternating with zones in which filiform leaves predominated. The stem of
this small plant and those of a number of the larger specimens have been sectioned
throughout their entire length, so that no opportunity might be lost of discovering
in them a solenostelic condition, foliar gaps, or evidence of the loss of gaps and
internal phloem. The discovery of any one of these would materially affect any
interpretation of the stelar state. Degenerate gaps and phloem would undoubtedly
raise the reduction theory to the plane of probability, but an actual solenostely
without evidence of degeneration or reduction would be open to interpretation as
the result of stelar amplification, provided an upward progression from protostely
to this hind of tubular stele could be indicated in the ontogeny. But in any case,
neither the presence of degenerate leaf-gaps and phloem, nor the existence of the
so-called ectophloic-siphonostelic state, would provide any grounds for a belief in
the cortical origin of the pith of Platyzoma. At most they would have demon-
strated merely a connection between pith and cortex which would seem to have
been broken frequently in the phyletic sense in this plant.

On the other hand, the discovery of a protostelic state giving place in the
ontogeny to a well-defined medullated protostele with isolated internal endodermis
and entire absence of foliar gaps and internal phloem would lend support to the
theory of upgrade development, unless it be further assumed, without evidence
from fern-anatomy , that such a protostele merely demonstrates the extent to which
reduction may go. Such a view of widespread reduction would necessitate that the
internal endodermis had followed the inner phloem and the leaf-gaps “ into
oblivion,” with the result that a stelar structure in no way distinguishable from
an original protostele was established. The supporter of such a reduction theory
would be compelled to view with suspicion most fern-protosteles, for each would be
open to interpretation as the result of drastic reduction from a solenostelic state
or from some other type of tubular stele. The “ sporeling ’’-structure of modern
ferns — in which protostely so commonly figures — would likewise be placed under
suspicion as a possible reduction phenomenon. The logical conclusions to such an
attitude would be, first, to suspect everything with a pure protostele or a medullated
protostele of being reduced ; second, to negative ontogeny unless it shows through-
out a solenostele or some upgrade modification thereof ; * and third, to consider
solenostely and not protostely as primitive in the sense recently advocated by Di-
Jeffrey, “that the tubular condition is both typical and primitive for the ferns
in general, with the sole exception of those forms in which the organisation of the
stele maintains the original protostelic state ” (14). This would imply that protostely
must persist throughout the entire ontogeny if it is to be considered an original
protostely, and that once solenostely appeared, the original protostely was lost, and
that where protostely and solenostely exist in the same individual, the latter is the

* In discussing the stele of Angiopteris evecta in 1902 (8), Dr Jeffrey stated that he had examined some young
plants, but “ unfortunately these proved rather too young for the present purpose, since the central cylinder had as yet
hardly passed into the tubular condition .”

582

DR JOHN M‘LEAN THOMPSON ON

result of reduction from the former. Behind the reduction theory lies the idea that
all normal ferns with tubular steles must have stelar gaps, for unless this be so, a
fern such as Platyzoma with an isolated pith, if considered normal and unreduced,
would be a stumbling-block to the generalisation that “ the pith must in all cases
be regarded as a derivative of the cortex which has become more or less completely
sequestered within the stele” (ll). Such a view as that advocated by the theory
of the cortical origin of pith would seek to exclude the alternative view that ferns
with pithed tubular steles may in certain cases illustrate a condition in which stelar
gaps have not yet been formed either in the ontogenetic or phyletic sense. But if
in any fern a pith has arisen without the creation of stelar gaps , it cannot have been
formed by cortical intrusion. It may be noted further that on the upgrade theory
the presence of ill-defined or scattered phloem, or of incipient leaf-gaps, might con-
ceivably be considered evidences of further amplification of an original protostele,
provided it be granted that tissues of one kind or another can be formed where
required and are not necessarily referable in origin to one original and continuous
tract which has been extended and disintegrated in the phylogeny.

The stelar structure in the large specimens of Platyzoma will be referred to first.
In each case there was throughout a well-defined pith enclosed by a continuous
tubular endodermis which was at no point in connection with the outer endodermis.
Neither inner phloem nor any structure suggestive of inner phloem was found.
There were neither leaf-gaps nor evidences of the prior existence of leaf-gaps. The
only points which these plants added to the facts already acquired were the following.
The pith fluctuated in bulk from point to point. In some plants it was a relatively
thick strand where the traces of large leaves were most numerous, and more slender
where the stem bore mainly small leaves. In other plants no such fluctuations were
observed, even although the leaf-zonations were well shown. In some instances the
pith was sclerotic throughout ; in others it was in part sclerotic, in part parenchy-
matous. A number of dichotomies were cut in serial sections, and in each instance
the division of the stele into two followed the plan illustrated in text-figure 1. The
stele of the parent stem became expanded laterally (a, b), and as preparation for
division of the xylem proceeded the inner endodermis became separated into two
distinct and complete tubes (c). Between these a bridge of xylem was established
( d ), so that the two parenchymatous cores now separated from each other were
not placed in open parenchymatous connection with the pericycle. The stele
then divided into two similar steles (e) by division of the xylem-bridge, the com-
pletion of phloem and pericycle around the now isolated tubes of xylem, and the
separation of the outer endodermis into two complete cylinders.

It will be seen that neither ramular gaps nor other breaks in the continuity of
the stele existed in the materials examined, and that the pith of the parent stem,
in dividing to form the pith of the branches, was not brought into contact with the
cortex. The investigation of these stems has added nothing to our knowledge which

THE MORPHOLOGY OF THE STELE OF PLATYZOMA MICROPHYLLUM, R. BR. 583

materially alters or helps to solve the problem of the origin of the pith and the
nature of the stele of Platyzoma. On the other hand, the small specimen referred
to above revealed some interesting features. These will now be described in
advancing order as they presented themselves in the series of sections into which
the stem was cut from its incomplete base to the apex. The series of illustrations
of this plant given in the accompanying plates have been prepared by camera-lucida
drawings, and are made to a uniform magnification.

At the broken base of the stem the stelar structure was as in fig. 1. The
outer endodermis (O.E.) was continuous, and root- and leaf-traces were being
liberated (R.T. and L.T.). There was a narrow band of phloem (Ph.) around the
bulky xylem. The latter was divisible into a narrow peripheral zone of small
elements, and a mass of inner parenchymatous xylem much more bulky towards

a t> c

Text-fig. 1. — Diagrams of the structure of the stele of Platyzoma at successive points in a
bifurcating stem, as seen in transverse sections. Both outer and inner endodermis are
represented by black lines ; phloem is hatched and xylem is solid black. Pericycle,
inner parenchyma, and pith are white. ( x 12.)

the lower than the upper side of the stele. The xylem enclosed a mass of parenchyma,
the central part of which was surrounded by a small continuous endodermis (I.E.).
This condition was followed by that shown in fig. 2. The bulky xylem was main-
tained, and apparently isolated tracheides were found in the mass of parenchyma
which it surrounded. The inner endodermis had closed down to form a narrow
tube enclosing only a few rows of parenchymatous cells. A little in advance of the
point where this latter condition was seen the stelar structure was as in fig. 3.
The inner endodermis had been reduced to a vanishing-point, so that it appeared
in section as a small group of endodermal cells enclosing no parenchyma (I.E.).
In the immediately succeeding sections the stele appeared as in fig. 4. The inner
endodermis had ceased to exist at this point, and a stelar structure in no way
distinguishable from that of a medullated protostele had been established. This
condition was followed almost immediately by that shown in fig. 5. The pith
was no longer bulky, and a stelar structure not far removed from an almost solid
protostele was seen. As the series was followed forward, the condition illustrated
in fig. 6 was reached ; an irregular but continuous medulla had been created by a

584

DR JOHN M‘LEAN THOMPSON ON

continuation of parenchymatous tissue from the immediate neighbourhood of the
slender persistent medulla into the mass of parenchymatous xylem. Within this
expanded pith there then appeared an irregular and incomplete endodermal tube
and an isolated chain of endodermal cells (I.E.). The resulting structure is illus-
trated in fig. 7. This condition led on to that shown in fig. 8 ; the inner endodermal
tube had become closed so as to isolate a body of parenchyma from the main
medullary mass. The independent chain of endodermal cells was still present in
the sections. Further forward the inner endodermal tube became open once more,
and new endodermal formations arose within the medulla. The resulting structure
appeared in section as in fig. 9. By the opening of the inner endodermal tube the
parenchyma which it surrounded at the point immediately preceding was again
placed in open communication with the main medullary mass. An independent row
of endodermal cells lay within the partially enclosed parenchyma, and a sheet and
slender tube of endodermal nature had arisen towards the opposite side of the pith.
A number of fluctuations in the disposition of the inner endodermal formations
followed, but these soon gave place to the arrangement shown in fig. 10. A wide
cylinder of inner endodermis was now clearly defined, but locally it was imperfectly
differentiated, while isolated endodermal cells appeared sporadically in various
situations. One such cell is shown within the parenchyma enclosed by the inner
endodermal tube. As the series was followed still further forward the sporadic
endodermal formation ceased. The wide and continuous inner endodermal tube was
maintained for some distance, though still locally imperfectly differentiated (fig. ll).
But a return to an ill-defined condition was soon made, the continuous cylindrical
inner endodermis giving place to an interrupted tube within which detached chains
of endodermal cells were enclosed. The resulting structure appeared in section as
in fig. 12. The inner endodermal formation had been fundamentally disrupted.
These transformations prefigured the condition shown in fig. 13. The inner endo-
dermal formation had been resolved as the sections were followed forward into a
narrow and almost complete tube enclosing a large body of the medulla, and into
an independent irregular group of cells. This latter group was found to diminish
and finally to end abruptly in the sections immediately succeeding the one figured
(fig. 13). The condition depicted in fig. 14 was subsequently established, the
persisting inner endodermal tube having been completely formed though narrowed
considerably. Around the inner endodermis a general development of parenchy-
matous inner xylem had been organised. It will be apparent that at this point
the medulla was again diminishing, and the inner endodermis was in fact once more
in process of being eliminated. The chief steps in this elimination as shown in the
sections are represented in figs. 15, 16. The inner endodermis became a slender
tube within a very restricted mass of pith (fig. 15), and when the vanishing point
for the endodermis was passed, a stelar structure in no way distinguishable from a
medullated protostele was again established (fig. 16). The purely parenchymatous

THE MORPHOLOGY OF THE STELE OF PLATYZOMA M1CROPHYLLUM, R. BR. 585

pith thus formed expanded rapidly as the advancing series was followed, and within
it a new endodermal formation arose (I.E., fig. 17). It passed quickly from this
irregular and ill-defined condition to that seen in fig. 18. At all points the inner
endodermal cylinder had become complete and fully differentiated. It embraced
a large body of the medulla, part of which had become sclerotic. Further forward
the stele as a whole underwent diminution (fig. 19), but later it expanded to the
condition shown in fig. 20, having a wide and complete inner endodermal tube
enclosing a somewhat sclerotic portion of the pith. At this point the stele showed
the general structure which has hitherto been known for Platyzoma.

It will be seen that so far as the description of this stem has gone, the inner
endodermis has twice been narrowed down to a vanishing point, so that a stelar
structure not distinguishable from a medullated protostele has been established
locally. In the second place, it will be clear that with the increase in the medullation
of the two protostelic zones thus formed an inner endodermis has arisen de novo within
the pith itself. In both instances the diminishing endodermal tube has retained
an unbroken identity to the end, but in the process of endodermal-re-creation there
has been at first a marked indefiniteness. The new endodermis has not appeared
within the growing pith as its predecessor vanished from the diminishing pith, for
there is no single point of endodermal-re-creation from which an immediately tubular
endodermis has expanded, whereas there was in both instances a single point to
which the decreasing tubular endodermis was narrowed. It will be further noted
that, in the first instance of re-establishment of the inner endodermis, the fluctuations
preceding the establishment of a clearly tubular endodermis were prolonged and
varied, but in the second instance they were few and comparatively simple. The
two zones in which the protostelic state existed, and that portion of the stele where
considerable constriction occurred (fig. 19), were the main points of origin of the
small leaf-traces, while those zones which possessed a large pith gave rise to the
majority of the large leaf- traces. It was difficult to determine the destination of
each leaf-trace even in a careful reconstruction of the stem, but the opinion was
formed that the protostelic zones coincided roughly with two of the groups of
small leaves.

The general construction of the most interesting portion of the stele of this plant
is illustrated in text-fig. 2. It purports to represent in slightly simplified form the
general proportions and relationships of the component tissues as seen in median
section. The outer and inner endodermal layers are indicated by black lines, both
pericycle and pith are white, phloem is hatched, and xylem is solid black. It will
be seen that throughout the greater part of the stele thus represented, the protostelic
zones show conspicuous increase of the xylem towards the lower surface of the stele.
An examination of the figures given in the plates of this paper will show that this
increase is due to marked local increase of the parenchymatous inner xylem.
Throughout the entire length of stem so far described neither foliar gaps nor

586

DR JOHN M‘LEAN THOMPSON ON

indications of stelar gaps of any kind were found. The same remark applies to the
apical portion of the stem which remains to be considered.

This younger portion of the stem retained to its apex the well-defined tubular
stele, and accordingly only -a small part of its stele is represented in text-fig. 2.
With a number of other apical portions of stems it was examined throughout so as
to demonstrate the steps of stelar differentiation from the apical meristem. In all
essential points the differentiation which it showed was in agreement with what
was seen in the other materials. The description given for it will suffice for all.
The stelar structure may conveniently be read backwards from the apex to the point
of complete stelar differentiation. In this way the entire stelar structure of the
plant under consideration will have been described.

Close to the apex of the stem the stele appeared in section as in fig. 25. “Both

Text-fig. 2. — Diagram of the structure of a portion of the stele of the small plant of Platyzoma described in the text and
illustrated in the plates. Both outer and inner endodermis are represented by black lines ; phloem is hatched and xylem is
solid black. Pericycle, inner parenchyma, and pith are white. The diagram represents the general arrangement of tissues
as seen in what would have been a median longitudinal section of the stele. ( x 16.)

outer and inner endodermal sheaths were clearly defined. Between them lay the
still imperfectly differentiated pericycle, phloem, xylem, and inner parenchyma.
Xylem-differentiation had commenced intermittently at the periphery of the xylem
and in the nascent traces, but the inner xylem was not yet clearly distinguishable.
No stelar protoxylems were seen. Further back the condition represented in fig. 24
was found. The outer xylem was almost completely differentiated, but though the
large tracheides of the parenchymatous inner xylem were clearly indicated, their
bonification had only just commenced. The differentiation of the inner xylem
progressed mainly centripetally as the lignification of the outer xylem approached
completion (fig. 23), and soon reached the stage represented in fig. 22. From this
point it was a mere step to the fully differentiated state depicted in fig. 21. The
general centripetal differentiation of the inner xylem and the ill-defined sequence
in the outer cylinder were shown in all the materials examined. Although no
well-defined stelar protoxylems were recognised, typical spiral protoxylems were
occasionally observed in the departing leaf-traces. An example of this is represented
in fig. 27. It shows an almost median and two intra-marginal protoxylem-groups
which, if continued down as the leaf-trace-xylem becomes completely concurrent with

THE MORPHOLOGY OP THE STELE OF PLATYZOMA MICROPHYLLUM, R. BR. 587

the outer stelar xylem, would occupy the positions indicated by the three crosses
in text-fig. 3, and if projected into the stele itself, could not fail to have a
mesarch position. The theoretical position of the stem-protoxylems would therefore
be mesarch.

It will be evident that the tubular portions of the stele of Platyzoma are con-
structed on the plan indicated in text-fig. 3, with outer and inner endodermis
(O.E. and I.E.), between which lie in centripetal sequence pericycle (P.), phloem (Ph.),
outer xylem (O.X.), inner xylem (I.X.), and inner parenchyma. Internal to the
inner endodermis is what is generally recognised as the pith (M.), but it will be clear

ax,

Text-fig. 3. — Diagram of structure of medullated stele of Platyzoma as seen
in transverse section. O.E., outer endodermis ; I.E., inner endodermis ;
P., pericycle; Ph., phloem; O.X., outer xylem; I.X., inner xylem;
■ M., medulla. ( x 50. ) The three crosses indicate the positions of the
leaf- trace protoxylems.

that if the 'pith as a whole is in this case of purely intrastelar origin , the inner
parenchyma immediately within the parenchymatous inner xylem is a portion of the
original pith which has not been included within the inner endodermis. On the
other hand, if the pith were of cortical origin, then only the ground-tissue enclosed
by the inner endodermis could reasonably be regarded as pith, and the inner paren-
chyma above referred to would be open to interpretation as the intrusive extension of
the pericycle. There seems no ground for the latter belief in the facts now recorded.
The outer and inner endodermal cylinders have been shown to be entirely inde-
pendent of each other, the former being invariably maintained inviolate while the
latter has fluctuated and locally vanished, and neither fully developed nor degenerate
stelar gaps have been found. Throughout the entire plant which has just been con-
sidered the tissues were apparently normal in form and differentiation, and no grounds
were found for considering the plant abnormal or perverted in any of its parts.

TRANS. ROY. SOC. EDIN., VOL. LIT, PART III (NO. 22). 89

588

DR JOHN M‘LEAN THOMPSON ON

The fuller facts of stelar structure have now been recorded. It remains to con-
sider the effect of their recognition upon the stelar problem. It is at once admitted
that they are inadequate to allow of a solution of the problem, for they lack the
support of the evidence of the “ sporeling ’’-structure, and of the earlier ontogenetic
stages generally. But the following points have been proved : —

(i) Platyzoma is at points protostelic.

(ii) It has been .shown to greatly augment an attenuated pith from within

the stele ; on the other hand, the pith may be drastically diminished.

(iii) It has been shown to develop entirely from within the stele an inner

endodermis around a portion of the pith.

(iv) It has neither foliar nor ramular gaps ; nor does it show any structural

evidence of having possessed at any time either stelar gaps or
internal phloem. Thus an original solenostele has not been
demonstrated.

Among the results of this investigation there is not one fact which could reason-
ably be considered evidence in favour of the reduction theory of the stelar structure,
unless it be that protostely recurs in those zones which are apparently the bearers of
the majority of the small leaf -traces. This alone could be considered evidence of
reduction if it be granted that the stelar structure is liable to be reduced under
xerophytic conditions. It might be held to demonstrate the extent to which
reduction may go. But no evidence of the original solenostely upon which the
reduction theory hangs has emerged ; much less have the materials shown “ pocket-
ing ” in an original protostele by which the cortical ground-tissue has “ invaded ” the
protostele and established a pith. It will be evident that if the course of stelar
transformation and reduction proposed for Platyzoma has actually been followed in
descent, this is the first recorded example of a fern which has reached solenostely
by amplification of an original protostele, and has returned to protostely by reduc-
tion. On the other hand, it will be clear that, even supposing the so-called ectophloic-
siphonostely of Platyzoma be the result of reduction from solenostely, no evidence
has been advanced to show that this solenostely arose by a process of cortical
“ intrusion.” If, as seems the case in the materials discussed above, both diminution
and augmentation of pith and inner endodermis are purely intrastelar phenomena,
there is no initial reason why inner phloem should not arise within a medullated
protostele not only where required, but also independently of the outer phloem.
But whatever be the possibilities, the fact remains that in promulgating their view
of the stelar structure of Platyzoma the advocates of the reduction theory have seen
neither the solenostelic stage they presume existed nor the protostelic stage which
has been described above, and from which arises a purely intrastelar pith. It will
be of interest to see what bearing the fluctuations of the inner endodermis above
recorded may have on the problem raised by the isolated endodermal spindles which
Mr Boodle has recorded from the stele of Schizsea dichotoma (15).

THE MORPHOLOGY OF THE STELE OF PLATYZOMA MICROPHYLLUM, R. BR. 589

But if the reduction theory has received little support from the recorded facts,
on the other hand the theory of upgrade internal differentiation has not been proved,
though the general trend of the stelar fluctuations which have been described in
these pages might reasonably be considered in its favour. Thus the fact that within
a protostele a pith has been shown to grow and to become more or less included in
an inner endodermis created de novo, points directly to the course which a supporter
of the theory of intrastelar origin of the pith in Platyzoma would expect to be
followed in the ontogeny. In the case described such an intrastelar development
would be considered to have occurred at least three times, and on all but the last
occasion had not been maintained, but had reverted directly to the original proto-
stelic state. The inference would be that any one of three definite states may exist
in the stele of a fern, and that these states are continuous in the evolutionary sense.
First, the original protostely may persist undisturbed throughout the entire ontogeny.
Primitive protostelic ferns would be examples of this. Second, the original protostely
may be transformed at an early stage in the ontogeny once and for all into one or
other of the more elaborate stelar states. Third, the original protostely may recur
in the ontogeny in ferns in which the advanced stelar state is not yet permanently
established. Whatever be the actual course of stelar amplification followed in
individual cases, the second state appears to have been widely exemplified in modern
ferns. Platyzoma may provide an illustration of the third and apparently excep-
tional state in which the stelar structure of the ancestry has not yet been definitely
restricted to the “ sporeling ” stages, but tends to persist in the mature organism.
If this be so, then the structural evidence dealt with above may indicate a remini-
scence of the steps taken in the initial transformation of the ancestral protostele.
Such steps would be, first, the growth of the pith within the stele itself, and second,
the inclusion of the bulk of the pith within an independent internal endodermis
formed de novo. It has been shown by Dr Lang that endodermal cells may be
formed where required in the stem of Helminthostachys (16). And further, Flas-
kaemper has shown that pith may be locally present and locally absent in one and
the same root according to the general conditions provided for the organism (17).
He found that after a seedling of Vicia had been grown until the radicle had begun to
elongate, if the cotyledons were removed, the growth of the plant was checked,
and the root, pithed in its older part, was devoid of pith in the part subsequently
formed. But when the plant had become strong under continual exposure to
favourable conditions, a pith was reorganised in the later parts of the same root. In
like manner the stelar pith of the stem of Platyzoma may be viewed as open to fluctua-
tion either by decrease or increase under the varying conditions of life, and need not
be regarded as the relic of a cortical intrusion which has persisted under xerophytic
conditions. A parallel may possibly be found in the case of Lepidophloios Scottii,
Gordon (18), in the stem of which the pith may come and go, though indeed no
evidence has been advanced to show that in it this fluctuation is a sign of reduction.

590

DR JOHN M'LEAN THOMPSON ON

It seems more reasonable to raise ontogenetic evidence — even though incomplete
— to the place of honour in the interpretations of adult structure, than to reject it
in favour of distant lateral comparisons which necessitate mere assumptions. But
it may be objected that the creation of an inner endodermis entirely independent
of its outer correlative is a most exceptional occurrence, and that there is no
apparent and reasonable need for it in a stele which is completely included in an
unbroken outer endodermis. The assumption would then be that in the case of
Platyzoma there is no reasonable explanation of an inner endodermis unless the
stele has been at some time possessed of stelar gaps by which the pith and cortex
have been directly connected. In thus assuming that the stele has been at some
stage in its history on a more elaborate footing than at present, the supporters of
the reduction theory would consider that the pericycle, phloem, xylem, and inner
parenchyma should normally be delimited from the ground-tissue of both cortex
and pith by endodermal barriers. ' To this it might reasonably be replied that the
inner endodermis may have arisen when required in a growing intrastelar pith for
purposes of intrastelar physiological control, the creation of the pith having rendered
such a new endodermal formation necessary or advantageous. It may be asked,
“ From whence arose the outer endodermis ? ” The presumption seems to be that
it was independently evolved when required for the delimitation of a primitive
stele. And further, if the formation of outer endodermis antedated the evolution
of inner endodermis, may it not be because the delimitation of the stele from the
general ground-tissue antedated medullation? These are questions which cannot
yet be definitely answered, but their bearing upon the present problem is obvious.

The conclusion to which this paper may be drawn is that in the case of Platyzoma
the recorded ontogenetic facts would clearly allow of the tubular stele with inde-
pendent outer and inner endodermis and only outer phloem being the upgrade
development from within of an original protostele. At the same time the reduction
hypothesis would meantime suffer exclusion merely from lack of evidence in its favour.
The stelar problem presented by Platyzoma cannot be solved until the “sporeling ”
stages have been fully investigated, and solenostely is proved or disproved.

Summary.

The introduction to this paper consists of a summary of the facts of stelar
structure already recorded for Platyzoma. These have been provided in succession
by Dr Poirault, Mr Boodle, and the present author. It is shown that in the
materials so far examined the stele was an unbroken medullated cylinder devoid of
either leaf-gaps or perforations. It was possessed of independent outer and inner
endodermal sheaths, between which lay in centripetal succession pericycle, outer
phloem, xylem, and inner parenchyma. The included pith was entirely isolated
within the conductive cylinder, having no ground-tissue connection with the cortex.

THE MORPHOLOGY OF THE STELE OF PLATYZOMA MICROPHYLLUM, R. BR. 591

In leaf-trace-formation and -departure the inner endodermis was in no way involved,
and no point existed where the stele was not entirely enclosed by an outer
endodermis.

The theoretical statements which have been made regarding this stelar structure
are then considered. They have been advanced by Dr Poirault, Dr Jeffrey, Mr
Boodle, and Mr Tansley. It is shown that in first suggesting that the pith of
Platyzoma is the result of confluence and final isolation of a series of foliar sclerotic
pockets, Dr Poirault imagined certain non-existing conditions as fulfilled in a
Grleicheniaceous plant and merely contemplated through them a possible explanation
for Platyzoma. He did not profess to support his interpretation by actual structural
evidence from the plant itself. To Dr Poirault’s conception of the stelar structure
support has been given by Dr Jeffrey and Mr Tansley. The former has affirmed
that the pith of Platyzoma is in reality of extrastelar origin, or is at least most
reasonably interpreted as such, though as a result of stelar reduction it no longer
communicates with the peripheral cortex. Mr Tansley has held that there is reason
for considering the stele of Platyzoma as reduced from an original solenostele by
loss of internal phloem and leaf-gaps. Dr Jeffrey has further claimed to have
observed nearly degenerate leaf-gaps in the stele. It is shown that in none of the
materials so far examined is there structural evidence which can be reasonably
advanced to support a belief in the present reduced state of the stele of Platyzoma
or of its present or prior possession of foliar gaps. As an alternative interpretation
of the stelar structure Mr Boodle has suggested that the medullated stele of
Platyzoma may possibly have been derived from a protostelic origin by the direct
formation of an independent intrastelar pith and inner endodermis.

It is held by the present author that the evidence so far advanced is inadequate
to allow of a confident opinion regarding the nature and origin of the medullated
stele of Platyzoma. For while no evidence of the intrastelar origin of the pith
and inner endodermis has emerged to support Mr Boodle’s suggestion, the reduction
theory is not supported by the demonstration of cortical “ intrusion ” into the stele
through stelar gaps, nor by the presence of stelar gaps and a solenostelic condition.
The reduction hypothesis seems to stand upon the more comprehensive assumptions,
but neither view can claim to be considered to indicate more than the undemon-
strated possibilities.

A description of the stelar structure of a number of recently acquired incomplete
specimens of Platyzoma follows. It is shown that in the majority of these the
stelar-structure was identical in plan with that already described. Neither a
solenostelic condition nor evidence of degenerate stelar gaps or inner phloem was
found. The investigation of these plants added nothing to our knowledge which
materially alters or helps to solve the problem of the origin of the pith and the
nature of the stele. But on the reduction hypothesis they might be considered
evidence of the firm establishment of the reduced condition, while on the theory of

592

DR JOHN M‘LEAN THOMPSON ON

upgrade intrastelar differentiation they might be held to demonstrate how entirely
intrastelar in nature as in location are the pith and inner endodermis.

On the other hand, a small — and possibly young — but incomplete specimen
revealed some interesting features which are illustrated in the accompanying plates.
It is shown that as its stele was followed forward from the broken base the pith and
inner endodermis decreased until the latter was_ narrowed to a vanishing point and
a medullated protostele was locally established (figs. 1-5). The pith again expanded
and within it there arose de novo an inner endodermis, which, at first of irregular
and indefinite construction, was subsequently definitely tubular (figs. 6-11). By
disintegration of this wide tubular inner endodermis an irregular formation again
arose within the pith, but when the latter had once more diminished the inner
endodermis was an unbroken tube narrowing on a vanishing point (figs. 12-15).
For at least the second time in the history of this plant a medullated protostele
was established locally (fig. 16), and on subsequent re-expansion of the pith inner
endodermis was again created de novo, and was subsequently maintained to the
apex of the stem as an unbroken cylinder (figs. 17-25).

A study of the stelar differentiation showed that the xylem was roughly divisible
into two concentric cylinders. The outer of these was differentiated first and was
alone involved in supplying xylem to the traces. In it the order of differentiation
was indefinite and no spiral stelar protoxylems were recognised. The inner cylinder
showed more or less clearly a centripetal order of differentiation. Spiral protoxylems
were occasionally found in the departing leaf-traces but were not continuous into the
stele. The theoretical position of the stem-protoxylems is mesarch.

The new facts of stelar-structure having been recorded, their bearing on the
stelar problem is considered. It is recognised that Platyzoma is at points protostelic,
and may augment from within the stele itself an attenuated pith. In like manner
it has been seen to develop an intrastelar inner endodermis. Among the results
secured there is nothing which can reasonably be considered evidence in favour
of the reduction theory unless it be that protostely recurs where small leaf-
traces predominate.

On the other hand, the theory of upgrade internal differentiation has not been
proved, though the general trend of the structural evidence might reasonably be
considered in its favour.

It is considered that any one of three definite states may exist in the stele of a fern,

(i) The original protostely may persist undisturbed throughout the entire
ontogeny.

(ii) The original protostely may be transformed at an early stage in the
ontogeny once and for all into one or other of the more elaborate
stelar states.

(iii) The original protostely may recur in the ontogeny in ferns in which
the advanced stelar state is not yet permanently established.

THE MORPHOLOGY OF THE STELE OF PLATYZOMA MICROPHYLLUM, R. BR, 593

It is suggested that Platyzoma may provide an illustration of this third and
apparently exceptional state in which the protostelic structure of the ancestry has
not yet been definitely restricted to the “ sporeling ” stages but tends to persist in
the mature organism. If this be so, the structural evidence discussed in these pages
may indicate a reminiscence of the steps taken in the initial transformation of the
ancestral protostele, namely, the growth of a pith within the stele itself, and the
inclusion of the bulk of this intrastelar pith within an independent internal endo-
dermis created de novo. If this view should prove to be correct when “ sporeling ”
plants have been examined, and if parallels to it be found in the ontogeny of other
primitive ferns, the establishment of foliar gaps either late or early in the ontogeny,
but subsequent to medullation , may reasonably be regarded as merely a further step
in stelar amplification by which the initially distinct cortex and pith are sooner
or later directly connected. On such a view the present medullated conductive
cylinder of Platyzoma would be regarded as the high-water mark of stelar
amplification so far reached for this plant, neither inner phloem nor stelar gaps
having been evolved.

In conclusion, it is held that the recorded facts would clearly allow of the tubular
medullated stele of Platyzoma being the result of upgrade development from within
of an original protostele. Nevertheless, the stelar problem cannot be solved until
the “sporeling” stages have been investigated and solenostely and reduction are
proved or disproved.

The author wishes to express his indebtedness to the Carnegie Trust for a grant
in aid of the illustration of this memoir.

REFERENCES.

(1) Prodr omus Florae, Novas Hollandias et Insulae Van Diemen.

(2) Plantas Javanicas Rariores.

(3) “Recherches sur les Cryptogames Vasculaires,” Ann. des Sci. Nat., Bot., 7e s4r., tom. xviii.

(4) “Anatomy of the Glcicheniaceas,” Annals of Botany, xv, 1901.

(5) “The Anatomy and Affinity of Platyzoma microphyllum, R. Br.,” Trans. Roy. Soc. Edin., vol. li,

part iii, No. 20.

(6) Report of the British Association for the Advancement of Science, Toronto , 1897 : Transactions of

Section K, “ The Morphology of the Central Cylinder in Vascular Plants,” by E. C. Jeffrey.

(7) “The Morphology of the Central Cylinder in the Angiosperms,” by E. C. Jeffrey, Transactions of

the Canadian Institute.

(8) Phil. Trans. Roy. Soc. London, Series B, vol. cxcv, pp. 119-146.

(9) “Lectures on the Evolution of the Filicinean Vascular System,” New Phytologist reprint, No. 21.

(10) Botanical Gazette, October 1908.

(11) Botanical Gazette, 1910.

(12) The Anatomy of Woody Plants, 1917, p. 283.

(13) “A Further Contribution to the Knowledge of Platyzoma microphyllum, R. Br.,” Trans. Roy. Soc.

Edin., vol. Iii, part i, No. 7, 1917.

(14) The Anatomy of Woody Plants, 1917, p. 281..

594

DR JOHN M'LEAN THOMPSON ON

(15) “ Comparative Anatomy of the Hymenophyllacex, Schizxacex, and Gleicheniace ee,” Annals of Botany,

xvii, 1903, p. 516.

(16) “ Studies in the Morphology and Anatomy of the Ophioglossacese,” I, Annals of Botany, April 1913.

(17) Flora, 1910.

(18) “On Lepidophloios Scottii," Trans. Roy. Soc. Edin., vol. xlvi, 1908.

DESCRIPTION OF FIGURES IN THE PLATES.

(All the figures are from the stem of Platyzoma.)

Plate I.

Fig. 1. Transverse section of the stele at the broken base of the small plant described in the text. The
inner endodermis is a narrow though uninterrupted tube. ( x 50.)

Fig. 2. Transverse section of the stele at a point in advance of that represented in fig. 1. The inner
endodermis has been narrowed down to a very slender tube, and scattered tracheides are in the surrounding
parenchyma. ( x 50.)

Fig. 3. Transverse section of the stele at a point in advance of that represented in fig. 2. The inner
endodermis has diminished until it is merely a small group of cells including no pith. ( x 50.)

Fig. 4. Transverse section of the stele at a point in advance of that represented in fig. 3. The inner
endodermis has now vanished and a medullated protostele has been formed locally. ( x 50.)

Fig. 5. Transverse section of the stele at a point in advance of that represented in fig. 4. An almost
solid protostele has been formed. ( x 50.)

Fig. 6. Transverse section of the stele at a point in advance of that represented in fig. 5. The pith has
expanded within' the parenchymatous xylem. ( x 50.)

Fig. 7. Transverse section of the stele at a point in advance of that represented in fig. 6. An irregular
endodermal formation has arisen in the growing pith. ( x 50.)

Fig. 8. Transverse section of the stele at a point in advance of that represented in fig. 7. The irregular
inner endodermal formation has been resolved into an irregular tube and an isolated group of endodermal
cells. ( x 50.)

Fig. 9. Transverse section of the stele at a point in advance of that represented in fig. 8. The inner
endodermal tube has been ruptured and involves a single chain of endodermal cells. A new tube and sheet
of endodermal cells have arisen within the pith. ( x 50.)

Plate II.

Fig. 10. Transverse section of the stele at a point in advance of that represented in fig. 9. A wide
though still incomplete inner endodermal tube is indicated and includes an isolated endodermal cell. ( x 50.)

Fig. 11. Transverse section of the stele at a point in advance of that represented in fig. 10. The inner
endodermal-cylinder is almost completely differentiated. ( x 50.)

Fig. 12. Transverse section of the stele at a point in advance of that represented in fig. 11. The inner
endodermal-cylinder has been disrupted and an irregular endodermal-formation has arisen. ( x 50.)

Fig. 1 3. Transverse section of the stele at a point in advance of that represented in fig. 1 2. The inner
endodermal-formation has been resolved into an incomplete tube around a limited body of pith and an
irregular group of endodermal cells. ( x 50.)

Fig. 14. Transverse section of the stele at a point in advance of that represented in fig. 13. The pith
has been greatly diminished and the irregular inner endodermis has become a narrow though incomplete
tube. ( x 50.)

Fig. 15. Transverse section of the stele at a point in advance of that represented in fig. 14. The pith
has been drastically diminished and the inner endodermal tube has been tapered towards a vanishing
point. ( x 50.)

THE MORPHOLOGY OF THE STELE OF PLATYZOMA MICROPHYLLUM, R. BR. 595

Fig. 16. Transverse section of the stele at a point in advance of that represented in fig. 15. A medul-
lated protostele has been established locally. ( x 50.)

Fig. 17. Transverse section of the stele at a point in advance of that represented in fig. 16. The pith
has been augmented and within it a new endodermal-formation has arisen. ( x 50.)

Fig. 18. Transverse section of the stele at a point in advance of that represented in fig. 17. A complete
inner endodermal tube has been formed, and part of the pith which it includes is sclerotic. ( x 50.)

Plate III.

Fig. 19. Transverse section of the stele at a point in advance of that represented in fig. 18. The inner
endodermal-cylinder has become locally narrow. ( x 50.)

Fig. 20. Transverse section of the stele at a point in advance of that represented in fig. 19. The inner
endodermal-cylinder has again expanded. ( x 50.)

Figs. 21-25. A series of transverse sections of the stele at successive points in advance of that represented
in fig. 20. Fig. 25 represents the undifferentiated stele just behind the growing apex of the stem, while
figs. 24, 23, 22, 21 show successive steps in the stelar differentiation. ( x 50.)

Fig. 26. Transverse section of a portion of the vascular cylinder in a section lent by Dr Scott and
Mr Boodle. It represents a xylem-ray at the point of departure of a small leaf-trace. ( x 100.)

Fig. 27. Transverse section of a portion of the stele of Platyzoma to show the spiral protoxylem groups
seen occasionally in the leaf-trace. ( x 100.)

TRANS. ROY. SOC. EDIN., VOL. LII, PART III. (NO. 22).

90

Vol. LII.

Trans .Roy. S o c . E dirf

DR John M9L. Thompson; Platyzoma microphyllum.R.Br.- Plate I.

J.M9L. Thompson del,

!

Vol. LII.

Trans . Roy. S oc . E du£

D^John M9L. Thompson; Platyzoma microphyllum.R.Br— Plate II.

16. 17. 18.

J. M?L. Thompson del.

Vol. L1I.

Trans. Roy. Soc. EdinT

D R John M?L. Thompson: Platyzoma microphyllum.R.Br -Plate III.

J.M^L. Thompson del.

( 597 )

XXIII. — Amphicheiral Knots. By Mary Gertrude Haseman, Ph.D. Communicated
by Dr C. G. Knott, General Secretary. (With One Plate.)

(MS. received May 20, 1918. Read November 4, 1918. Issued separately December 22, 1919.)

§ 1. The Intrinsic Symbol of an Amphicheiral.

The intrinsic symbol* of an amphicheiral knot is based on the idea of the
sequence of the crossings ; it replaces each letter of the alphabetical symbol by a
number equal to one-half of the number of crossings intervening before the next
occurrence of that letter as the knot is traversed in a definite direction. Hence it
is seen that two knots, which have the same intrinsic symbol, are identical. Since
the same number of pairs of crossings must elapse before the next occurrence of
corresponding crossings of two identical knots when the knots are traversed in a
given direction, it is seen that the converse is true also. It may be necessary to
consider the complementary intrinsic symbol in order to detect identical knots. For
example, the two knots

(1) 10 5599993444488 10 5599993444488

a g b l c in d h i j f k g b h n e i j f It a l c m d n e ,

(2) 3559 999 10 444488 3559999 10 44448. 8
a g b l cm d aejfkgbhniejfk li l cm d n i,

are found to be identical since the intrinsic symbol of (l) coincides with the com-
plementary symbol of (2). That they are identical may be verified by the fact that
their compartment symbol is

By an interchange of the two crossings, a and h in (2), it is seen that the two
symbols may be made to coincide. So two amphicheiral knots are identical when their
intrinsic symbols agree except for an interchange of complementary numbers.

The intrinsic symbol of an amphicheiral knot of the first order offers certain
points of interest. An amphicheiral centre of an amphicheiral knot of order 1 is

* M. G. Haseman, Trans. Roy. Soc. Edin., vol. lii, p. 235.

TRANS. ROY. SOC. EDIN., YOL. LII, PART III (NO. 23).

91

598

MARY GERTRUDE HASEMAN ON AMPHICHEIRAL KNOTS.

defined as a mid-point of a lap of the thread so located that corresponding crossings
occur at equal arcual distances when the knot is transversed along this thread in
opposite directions from the point. Let the amphicheiral centre (p1 be the mid-point
of the lap of thread between the two corresponding crossings y and q of a knot
with n crossings, and denote by a1; f31 the number of pairs of crossings which elapse
before the next occurrence of p, q respectively as the knot is traversed from p to (pi
through q. Hence n— 1 — cq, n — 1 — Pi will be the number of pairs of crossings

which elapse before the next occurrence of the crossings p, q respectively as the

knot is traversed from q to (pi through p. By the definition of an amphicheiral
centre, oq = n— 1 — fii or a1-\-fi1 = n — 1. Similarly + $ = n — 1 , where and ft, are
two crossings of the intrinsic symbol at equal arcual distances from ^i- Thus an
amphicheiral knot of the first order, as well as its pairs of amphicheiral centres, may
be detected very easily from its intrinsic symbol. For example, the intrinsic symbol

444848 - 5 9 5 9 9 9 4 • 9 4 4 4 8 4 8 ■ 5 9 5 9 9 9 4 • 9

a f b k c l d n e a f b g c h m i d j e k g l h m i n j

exhibits one pair of amphicheiral centres between the crossings l, d and e,k, as well
as a second pair between g, c and n, j.

On the other hand, from Tait’s # definition of an amphicheiral of the second
order, it is seen that every crossing is separated from its correspondent by n — 1 pairs
of crossings, and hence the first n numbers of the intrinsic symbol are identically
equal to the remaining numbers, the sequence of numbers being the same. Accord-
ingly the knot whose intrinsic symbol is shown in (1) on page 1 is an amphicheiral
of the second order.

§ 2. A New Construction for the Amphicheirals of Order 1.

The intrinsic symbol for all of the amphicheiral knots of the first class of
orders 1 and 2, as constructed by Tait f is arranged in one of the two sequences
stated above, whereas this is not the case for all of the amphicheirals of the second
class of orders 1 and 2. However, in a census (M. G. Haseman, Trans. Roy. Soc.
Edin., vol. lii, p. 253) of the amphicheiral knots with twelve crossings, it is found
that the form shown in fig. 3, which is obtained by the unsymmetrical distortion
D]" of the form shown in fig. 2, and belongs therefore to the second class,
possesses the intrinsic symbol,

177 10 44177 10 44177 10 44177 10 44;

this classes it among the amphicheirals of the second order. But it is impossible to
put this knot on the sphere so that corresponding compartments are opposite ; that
is to say, it cannot be constructed in the plane by means of a great circle and a pair
of twin circuits. Hence it cannot belong to order 2 as defined by Tait.J It is
* Tait, Trans. Boy. Soc. Edin., vol. xxxii, p. 497. f Ibid., Plate LXXIX. J Ibid., p. 497.

MARY GERTRUDE HASEMAN ON AMPHICHEIRAL KNOTS.

599

possible, however, to construct on the sphere one of the amphicheiral forms of
the knot given in fig. 2 by means of a curve Ci in contact with a great circle
at two diametrically opposite points, Pi, P2, and of a curve C2 which is obtained
as a reflection of C! in the plane ttx of the great circle, followed by a second
reflection in a plane tt2 passed through the points Pi, P2 perpendicularly to the
plane 7 iq. To secure this construction in the plane suppose cx, which is the

projection of the curve Ci in the plane tt1} to be the broken curve in fig. 4,
with contacts at the diametral points px, The curve C2, represented by the
dotted curve, is the same curve as cx, but drawn on the outside of the circle and
reflected in the line pxp2- Now, imagine the curve c2 to be rotated through an angle

of - to the right or left ; the resulting knot, where the contacts are regarded as

4 & ’ O’ o _

crossings, is found to be the knot shown in fig. 3.

The foregoing construction led me to seek for a similar construction in the plane
of the amphicheirals of the first order with any number of crossings. Let the curve

Fig. 2. Fig. 3.

Ci have k contacts, 2 < intersections with the circle and <r self sections ; and denote by
c2 the same curve on the outside of the circle but reflected in a diameter passing

through one of the contacts. Now, imagine c2 rotated through angles — , etc., and

the resulting curve is found to be an amphicheiral of order 1 with 2(/c + 2* + cr) cross-
ings. The curve c2, obtained by the reflection in the plane ttx of the great circle,
ensures the desired correspondence of compartments ; the reflection of the curve c2
in the plane tt2 does not alter the number of compartments, nor the number of laps
of thread bounding each compartment ; instead it interchanges corresponding
adjacent compartments. For instance, suppose the two regions px, p2 by the first
reflection to go into the two adjacent regions px, p2 respectively. If, now, by the
second reflection px becomes adjacent to p2, then p2 must become adjacent to px by
the same reflection. Since rotation through an angle merely adds one crossing to
each of the regions which are in the relation of px, p2 , then the desired correspond-
ence of compartments remains unaltered. Further corresponding crossings occur at
equal arcual distances from the mid-points of the lap of thread common to two
corresponding compartments. Hence the knot is an amphicheiral of the first order.

600

MARY GERTRUDE HASEMAN ON AMPHICHEIRAL KNOTS.

It is. possible to have the two curves and C2 in their various positions intersect
in or' pairs of points, although it is not always possible to make corresponding laps of
the thread intersect without introducing extra crossings. By this process I have
succeeded in constructing all of the amphicheirals of order 1 with four, six, eight,
and ten crossings (Nos. 1-21 in the Plate*), and find the tenfold knot No. 21 in
the Plate, omitted by Tait in his census (Trans. Roy. Soc. Edin ., vol. xxxii,
Plate LXXIX).

Likewise the knot No. 22 in the Plate has been omitted from my census (see
Trans. Roy. Soc. Edin., vol. lii, pp. 253-4) of the amphicheirals with twelve crossings.
It is possible that this method will reveal other omissions. It is to be noted that the
maximum number of contacts were used in the constructions of these amphicheirals,

Fig. 4.

Fig. 5.

Fig. 6.

but 1 cannot say whether this is necessary in the construction of the knots with
a greater number of crossings.

§ 3. Skew Amphicheirals.

If, however, the curve c2/, obtained by a single reflection in the plane is used
instead of the curve C2, there results upon rotation a knot which can be distorted
into an amphicheiral of the first order— that is to say, it belongs to Tait’s second
class. When the curve Cj is symmetrical about the line pip%, the resulting knot
is an amphicheiral of the first class of order 1, since then the curve is identical
with the curve C2.

In this construction there arise certain knots, called by me skew amphicheirals
of the second order, which exhibit the amphicheiral symmetry in spite of the fact
that they belong, by the above statement, to the second class of order 1. The
intrinsic symbol of all such knots, as I have found, classes them with the amphi-
cheirals of the first class of order 2, although corresponding regions are not opposite
on the sphere. An example of a skew amphicheiral is shown in fig. 5 ; it is found
to be identical with the knot in fig. 3. It is to be noted that in the knot shown in
fig. 5 the curve possesses symmetry of such a nature that its relation to curve cx
is the same whether c2' be rotated to the right or left.

■* The numbers at the lower right-hand corners are Tait’s numbers.

MARY GERTRUDE HASEMAN ON AMPHICHEIRAL KNOTS.

601

Another very good example (see fig. 6) of a skew amphicheiral is given by
the symbol

1 3 3 22 1 8 8 22 20 20 15 15 1 3 3 22 1 8 8 22 20 20 15 15

a e b a c l d c e b fvgkligirj i k h Id

1 3 3 22 1 8 8 22 '20 20 15 15 1 3 3 22 1 8 8 22 20 20 15 15

m q n m ox p o q n r j s w t s u f v u w t x p.

Because of the relation of the curve c2r to the curve cx the crossing a may
correspond to either g or s ; likewise the crossing m may correspond to either g
or s. Therefore we may expect not only the numbers in the last half of the
sequence to be a repetition of those in the first half, but also the first half to
consist of a repetition of a certain group of numbers ; the number of repetitions
will probably depend on the number of contacts.

The only skew amphicheirals, which I have found, may be obtained as the
unsymmetrical distortions of an amphicheiral of the first class of order 1, and
in view of the fact that the curve for knots with 4, 6, 8, 10, 12, 14 crossings
seems to lead always to a knot which can be distorted into an amphicheiral of
the first class of order 1, I am of the opinion that they do not constitute a
■distinct class, although it may be that they will form another class in the case
of the knots with a greater number of crossings.

If, therefore, an amphicheiral knot is defined as one whose primary and secondary
symbols are identical — that is to say, one whose intrinsic symbol belongs to one of
the two arrangements mentioned on p. 598 — it is seen that Tait’s classification given
in Trans. Roy. Soc. Edin., vol. xxxii, p. 499, is sufficient provided that it be
admitted that an amphicheiral knot can belong to the first class of one order and
to the second class of the other order.

§ 4. Amphicheiral Knots of Order 2 with Fourteen Crossings.

As has been shown by Tait, Trans. Roy. Soc. Edin., vol. xxxii, there are no
amphicheiral knots of order 2 with 4, 6, 8, or 10 crossings. There are two knots *
of the second order with twelve crossings, both of which may be constructed on
models involving one pair of contacts, although they appear among the knots
which required a greater number of contacts. In a consideration of the maximum
number of contacts necessary to construct the amphicheirals with n crossings I
was led to construct the amphicheirals of the second order with fourteen crossings,
of which there are ten in number, as shown in Nos. 23-32 in the Plate. Nine of
these were constructed on models with one pair of contacts, whereas the tenth
one, No. 32, required two or more pairs of contacts. Hence it will be necessary
to pass to the amphicheirals with a greater number of crossings in order to
determine the maximum number of contacts required.

An interesting amphicheiral is the form which is obtained by the single distor-
* M. G. Haseman, Trans. Boy. Soc. Edin., vol. lii, p. 254.

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 23).

92

602

MARY GERTRUDE HASEMAN ON AMPHTCHEIRAL KNOTS.

tion Di of the amphicheiral of the second order with fourteen crossings, No. 25 in
the Plate. Its intrinsic symbol

11 10 3 3 11 10 4 3 1 10 10 12 4 3 10 9 1 3 3 12 10 9 3 2 10 10 3 2

shows that it is an amphicheiral of the first class of order 1 with one pair of
amphicheiral centres. This is an example of Tait’s supposed third class (Trans.
Roy. Soc. Edin., vol. xxxii, p. 499), which has the property of being changed into
its own perversion by a single distortion, but, contrary to his idea, it must belong
to the first class of order 1 and to the second class of order 2.

It is of interest to note that all of those amphicheirals which belong simul-
taneously to the first class of orders 1 and 2 exhibit two pairs of amphicheiral
centres with the exception of one, which has fourteen pairs.

Trans. Roy. Soc. Edin.

Vol. LI I.

Mary G. Haseman: Aaphicheiral Knots.

2 ''

~S 8 Am

8 Bt 5

8 By

6 8B 7 -Y -"8Ap 8 y' 35

9 lO

11

A. RITCHIE & SON, EDIN*

( 603 )

XXIV. — On Old Red Sandstone Plants showing Structure, from the Rhynie Chert
Bed, Aberdeenshire. Part II. Additional Notes on Rhynia Gwynne-Vaughani,
Kidston and Lang; with Descriptions of Rhynia major, n.sp., and Hornea
Lignieri, n.g., n.sp. By R. Kidston, LL.D., F.R.S., and W. H. Lang, D.Sc.,
F.R.S., Barker Professor of Cryptogamic Botany in the University of
Manchester. (With Ten Plates.)

(Read July 8, 1918. MS. received May 17, 1919. Issued separately January 20, 1920.)

Introduction..

In Part I * a general account was given of the silicified peat-bed found at
Rhynie, and one vascular plant was described in detail under the name of Rhynia
Gwynne-Vaughani. Further study has shown that there are two species of Rhynia
which we now distinguish as Rhynia Gwynne-Vaughani and Rhynia major. The
account in Part I applies to both these species. Along with them there occurs a
plant of similar grade of organisation to Rhynia , though quite distinct from that
genus: this we name Hornea Lignieri. Asteroxylon Mackiei, on the other hand,
which will be described in the next part, was a plant of larger size and much more
complex morphology.

We are thus now able to establish the existence and main features of four archaic
Vascular Cryptogams from the Rhynie bed. In the present paper additional notes
on R. Gwynne-Vaughani are given, and R. major and Hornea Lignieri described
in detail. In conclusion, the morphological bearings of the facts will be briefly
considered.

Our former paper was based mainly on sections prepared from loose blocks of the
chert, and partly from specimens collected from the chert bed in situ. The material
for the present paper was mostly obtained in position in the chert bed, though a few
loose blocks have yielded valuable information.

Rhynia.

As mentioned above, two species of Rhynia have to be distinguished. Their simi-
larity in organisation is so great that they are not always readily separated from one
another, and in our former paper they were described together under the name Rhynia
Gwynne-Vaughani. The second species, which we now separate as Rhynia major,
is larger in all its parts, and differs from R. Givynne-V aughani in the absence of
hemispherical projections and adventitious branches from the stems, in the greater
size of the stele and xylem strand, and in the much larger size of the sporangia
and spores.

* Trans. Boy. Soc. Edin., vol. li, 1917, p. 761.

TRANS. ROY. SOC. EDIN., YOL. L1I, PART III (NO. 24).

93

604 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

The difficulty of sharply distinguishing particular fragments of these two plants
is in no way inconsistent with the opinion we have been led to form, that they are
distinct but closely allied species. Their comparative study is assisted by certain
beds or loose blocks containing the remains of only one or the other species. The
difference in size of the two plants will be evident on comparing the block of chert
represented of natural size in PI. I, fig. 1, with the figs. 2 and 5 on PI. II of our
former paper. The peat in fig. 1 is composed of stems of R. major, many of which
have a diameter of 5 mm., while the stems of R. Gwynne-Vaughani in figs. 2 and 5
are about 2 mm. in diameter.

Although we were dealing with mixed material of two species, which we now know
separately, the description of the plant and its differentiation into rhizome, stems,
and sporangia given in Part I holds good for both species of the genus Rhynia. It
has only to be qualified by recognising the specific differences to be dealt with below.
It will be convenient at this stage, however, to review the illustrations to our former
paper and to indicate which belong to R. Givynne-Vaughani and which to R. major.

AIL the figures on PI. II are of R. Gwynne-Vaughani, as are also figs. 6-10 on
PI. Ill, fig. 20 and figs. 23-30 on PI. V, all the figures on PL VI except fig. 37, and
all the figures on Pis. VII and VIII. The sporangium in figs. 63 and 63a on PI. IX
is the only one figured that belongs to this species. Pigs. 72-74 on PI. X are small
stems of this species. A survey of the figures named shows that the general habit
and size of R. Gwynne-Vaughani, the external features of the stems, the details , of
structure of the stems, the hemispherical projections and the occurrence in their
place of adventitious branches, and the sporangium were represented. The trans-
verse section of a stem in Dr Mackie’s paper,* which was the first figure of Rhynia
published, also belongs to this species.

To R. major, on the other hand, we now refer fig. 1 on PI. I, showing a block of
the chert with the stems of natural size ; all the figures of rhizomes on PI. IV (figs.
13-19), with the details in figs. 11 and 12 of the preceding plate; the large stems
in figs. 21 and 22 t of PI. V ; the outer cortex and stoma in fig. 37 on PI. VI ; the
large sporangia in figs. 62 and 64-69 on PI. IX ; the spores in figs. 70 and 71 of PI. X ;
and the partially decayed stems in figs. 76-78 of the same plate. J Consideration
of these figures will ' show that the rhizomes, stems, and sporangia of R. major
were represented, though the detailed account of the structure of the stem is
based on R. Gwynne-Vaughani. On the other hand, all the rhizomes described
and figured in Part I are of R. major, and the description of the sporangium is
mainly based on this species.

The rhizomes figured on PI. IV of Part I all occurred in one block of silicified

* Trans. Edin. Geol. Soc., vol. x, 1913, pi. xxii, fig. 5.

t The magnification of figs. 21 and 22 on PI. V of Pt. I was erroneously given as 20 diameters. These stems are
magnified about 14 diameters.

| To complete this survey of the illustrations to Part I, it should be stated that the poorly preserved specimen in
fig. 75 of PI. X does not belong to Rhynia, hut to Asteroxylon.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 605

peat, composed of partly decayed stems of R. major , with sporangia of that species,
and were evidently preserved as they grew at one level of the peaty mass. The
presumption is that they belong to R. major and not to R. Gwynne-Vaughani, no
stems of which are found in this block. This conclusion, though confirmed by some
details, would be difficult without the evidence afforded by association. Rhizomes
of R. major have, however, now been met with in other blocks, and rhizomes of
similar type, though smaller, which undoubtedly belong to R . Gwynne-Vaughani
will be described below.

Additional Notes on Rhynia Gwynne-Vaughani, Kidston and Lang.

(PL I.)

1917. Rhynia Gwynne-Vaughani, Kidston and Lang (pars), Trans. Roy. Soc. Edin., vol. li, p. 7 G ♦ ,
pi. ii, figs. 2-5 ; pi. iii, figs. 6-10; pi. v, figs. 20, 23-30; pi. vi, figs. 31-36, 38-40; pi. vii,
tigs. 41-51 ; pi. viii, figs. 52-61 ; pi. ix, figs. 63-63a; pi. x, figs. 72-74.

Many of the features of R. Gwynne-Vaughani were sufficiently described and
illustrated in Part I, and only some supplementary notes are required.

Since there is reason to regard all the rhizomes described in Part I as of R. major,
a rhizome belonging to R. Gwynne-Vaughani is figured here (PI. I, fig, l). The
rhizome, which is cut transversely, has given off laterally an ascending stem. This
rhizome has a thin-walled epidermis, a slight but evident distinction between outer
and inner cortex, and a small stele, the xylem of which consists of about four
tracheides. Rhizoids extend downwards into the peat from the epidermis of the
lower side of the rhizome. The ascending branch shows epidermis, cortex, phloem,
and xylem ; it agrees in size and structure with the typical stems of R. Gwynne-
Vaughani making up the peat of this region of the bed. Comparison with the
rhizomes of R. major in Part I will show the essential similarity in structure of this
region of the plant in the two species, and the difference in size.

We have little to add to the description of the stem in Part I, which was based
almost entirely on this species. It may, however, be pointed out that, while the
small strand of xylem is often uniform and composed of similar tracheides, this is
not always the case. In larger- strands a distinction of smaller central tracheides
surrounded by outer tracheides of greater diameter is evident. This distinction can
be seen in figs. 41 and 45 on PI. VII of Part I. Attention is directed to this char-
acter in R. Gwynne-Vaughani, since it will be found to hold regularly and strikingly
for the steles of R. major and Hornea Lignieri.

In the peat, composed of numerous stems of R. Gwynne-Vaughani, a number of
cases have been noticed in which stems have been cut in the neighbourhood of their
tips. The most striking example is shown in PI. I, figs. 2-4, and demonstrates the
appearance of the growing point. Fig. 2 shows at a, among a group of typical stems
of this species, one with its tip cut longitudinally. This tip is more highly magnified

606 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

in fig. 3. The meristematic tissue is composed of small cells filled with dense
contents, which evidently represent the slightly contracted protoplasts. The elon-
gation of the cells in the central region of the stem is apparent a short distance
behind the actual apex. The cells of the growing point itself are isodiametric, and
those of the outermost layers were evidently undergoing periclinal divisions. The
whole appearance suggests a small celled meristem, but it is of course impossible to
say whether an apical cell was present or not. The cuticle can be traced from the
older contracted region of the branch over the growing point ; it has become separated
from the outermost layer of cells by a clear zone.

By the side of this undoubted apex, at b in fig. 2, is seen what we can only
interpret as the extreme tip of another growing point which has been removed by a
horizontal section and is thus viewed from above. The arrangement of the meri-
stematic cells in this transverse section is shown more highly magnified in fig. 4.
If this interpretation is correct, the specimen gives additional evidence as to the
absence of any prominent initial cell.

In connection with the apical meristem, fig. 5 may be referred to. This longi-
tudinal section, while not including the apex, shows a region situated close to it with
the protoplasts almost filling the small cell cavities.

The only sporangium of R. Givynne-Vaughani figured in Part I (PI. IX, figs.
63, 63a) lay amongst characteristic stems of that species, and terminated a stem
showing the distinctive vascular strand and tracheides. It contrasted with what
we now know to be the sporangia of R. major by its much smaller size. It measured
about 3 mm. in length by 1 mm. across. At the junction of stalk and sporangium
in this specimen there is a constriction (PI. I, fig. 6) to which we do not attach
importance as indicating a natural feature. Above the constriction the epidermal
cells assume the character of the sporangial wall. They are deeper than on the
vegetative stem, but as seen in surface view at the top of the sporangium have the
fusiform outline (PI. I, fig. 7). The cell walls are moderately thickened and brown.
Within the well-marked epidermis come some layers of thin- walled tissue, mostly
perished, and only indicated towards the base of this sporangium. Bounding the
empty cavity of the sporangium is the persistent layer which we have spoken of
in Part I as the “ tapetum.” The nature of this layer in the sporangia of Rhynia
will be further considered below under R. major.

A transverse section of a slightly larger sporangium is shown in fig. 8. The
well-developed epidermal layer, the perished middle zone of the sporangial wall, and
the persistent tapetal layer can be distinguished. The sporangial wall is about '2 mm.
in thickness, and the width of the sporangium must have been slightly over 1'5 mm.
There were no spores remaining in this sporangium, but another of similar size, and
with an equally well differentiated wall (fig. 9), was filled with well-preserved spores.
The tapetal layer in this specimen appears a number of cells deep, but this is probably
to be explained by the section being oblique.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 607

The sporangium represented in fig. 10 is peculiar in that the epidermal layer of
its wall is not sharply distinguished from that of the stalk. A firm tapetal
layer seems to be completely wanting, though the soft tissues of the wall
are well preserved. This peculiar sporangium widened out gradually from the
stalk without any constriction at the base. The sporangial cavity is partly filled
with mature spores.

Another small sporangium (fig. ll) was even more remarkable. The group of
spores, still arranged in tetrads, is surrounded by a thick wall, which has neither a
definite tapetal layer nor a characteristic thick-walled epidermis. The ill-preserved
epidermis is bounded by a cuticle which is ridged like that of adjacent vegetative
stems of R. Gwynne-Vaughani ( cf. Part I, PI. VI, figs. 31, 36, p. 769).

The two sporangia last described, which clearly belong to R. Gwynne-Vaughani ,
suggest a less differentiated condition in which spores were formed within the end of
a stem without this being modified into a definite sporangiun. It is of interest to
find this type of sporangium co-existing in the same species with a more specialised
type, provided with a tapetal layer and a specially constructed and thickened
epidermal layer. There is no evidence, however, that any of the sporangia of* this
plant had a definite dehiscence.

The spores of R. Gwynne-Vaughani in the sporangium in fig. 11 were still
associated in tetrads (fig. 12). Other mature spores from the sporangium in fig. 9
are represented in fig. 13. The fully mature but not shed spores of this species
measure about 40 p in diameter.

Phynia major, n.sp. (Pis. II and III.)

1917. Rhynia Gwynne-Vaughani , Kidston and Lang (pars), Trans. Roy. Soc. Edin., vol. li, pi. i, fig. 1 ;

pi. iii, figs. 11-12; pi. iv, figs. 13-19; pi. v, figs. 21-22; pi. vi, fig. 37; pi. ix, figs. 62,
64-69 ; pi. x, figs. 70, 71, 76-78.

The figures quoted above, and the descriptions in Part I based upon them, give
the structure of all the regions of the plant — rhizomes, stems, sporangia — of R.
major. It has been already pointed out that the differences distinguishing this
species from R. Givynne- Vaughani are mainly those of size and of minor details. It
is necessary to consider the new species a little more fully, however, in order to bring
out clearly the nature of these differences.

The remains of R. major so far found do not give any such picture of the habit
of the plant as a whole as the specimen figured in Part I, PL II, fig. 5 did for
R. Gwynne-Vaughani. The greater diameter of the stems, as shown in PI. I of
Part I, suggests that R. major was a larger plant throughout, and this is supported
by the large size of the sporangia. We have, however, no direct evidence of the
height this species attained.

We have nothing to add to the description of the rhizomes of R. major given in
Part I, except to record the occurrence of similar though less perfect examples,

608 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

also associated with the typical stems of this species, in some specimens of the
silicified peat collected from the section exposed in situ.

The stems of R. major must be more fully described, since the description of the
stem of Rliynia in Part I was almost entirely based on specimens of R. Gwynne-
Vaughani.

As regards size, the examination of large numbers of sections of stems has shown
that in R. major a diameter of 5-6 mm. was often attained (PI. II, figs. 14, 19).
Finer branches, probably from the upper region of the plant, are met with down
to a diameter of 1‘5 mm. or less (fig. 15). These measurements contrast with those
of R. Givynne-Vaughani, in which the usual diameter of larger stems is about 2 mm.,
though a thickness of over 3 mm. is sometimes attained, while the finer stems or
branches may be less than 1 mm. in diameter.

On PI. Y of Part I two transverse sections of R. major were figured in figs. 21
and 22. Large as they appear in comparison with the other stems on the same
plate, the difference is in reality greater, since their magnification should have been
given as x 14 instead of x 20. A true comparison with the stems of R. Gwynne-
Vaughani, magnified 20 diameters, in Part I, PI. V, figs. 23-29, is afforded by the
accompanying figs. 14, 19, 17, and 15 on PI. II of this memoir.

A general view of a portion of the silicified peat composed of large stems and
sporangia of R. major is given at a lower magnification ( x 7) in PI. Ill, fig. 25. The
stems shown in figs. 18 and 21 are magnified 14 diameters.

All these figures show how closely the stem of R. major agrees in general
construction with that of R. Givynne- Vaughani* As is best shown in the well-
preserved stem in fig. 14, there is a clear differentiation into epidermis ( ep .), outer
cortex (o.c.), inner cortex ( i.c .), phloem (ph.), and xylem ( x .). A difference that
may be at once emphasised is the absence of the hemispherical projections which
are such a striking feature of many stems' of R. Gwynne-Vaughani. In the large
number of sections of R. major examined no trace either of hemispherical pro-
jections or of the adventitious branches corresponding in place of origin to them
has been seen.

The epidermis had the outer walls thickened and covered with a well-marked
cuticle. It appears smooth, and no ridges have been observed on the epidermal cells.
Stomata were sparingly present as in R. Gwynne-Vaughani. The outer cortex, two
or three cells deep, was evidently a mechanically important hypoderma in the larger
stems. Within it came the wide inner cortex, the outer cells of which, when well
preserved, often have the dark contents common in the same position in R. Gwynne-
Vaughani (fig. 14). The outer cortex was interrupted beneath the stomata, as is
shown in figs. 14 and 18 and in Part I, PI. VI, fig. 37. In the large stems the outer
cortex was very resistant to decay, while the wide inner cortex often broke down
(figs. 25, 19, 18).

* This is also brought out in comparing figs. 21 and 22 with the other figures on PI. V of Pt. I.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 609

The stele, which is clearly marked off from the inner cortex, though no special
boundary line is present, has a wide clear zone of phloem surrounding the central
solid strand of xylem. Even in the smallest stems of R. major (fig. 15) the stele
is relatively large as compared with R. Gwynne-Vaughani.

The cells of the phloem are thin-walled, and as seen in transverse section (fig. 16)
fit closely together. The intercellular spaces present in the inner cortex, when this
is well preserved, cease on passing to the phloem. In longitudinal sections (figs: 17
and 20) the cells of the phloem are seen to be thin-walled and elongated, resembling
the same tissue in R. Gwynne-Vaughani.

The xylem consists of far more numerous tracheides than in most steles of R-
Gwynne-Vaughani. A distinction, such as was occasionally found in the latter
plant, is here regularly present between the central and peripheral tracheides. The
inner tracheides are much smaller than the outer ones, and have thinner walls
(figs. 22, 16a). Although the xylem appears beautifully clear in many transverse
sections, no satisfactory indication of the thickening of the tracheide walls is shown
in any of the longitudinal or oblique sections as yet examined. Everything points
to the walls having been thickened, but the thickening seems to have readily perished,
and practically all the material of the plant had undergone decay to this extent at
least. There is sometimes an appearance of porose thickening, but this often fills up
the cavity, and from the examination of many specimens we are convinced that this
appearance is due to an alteration of the original thickening, and that it would not
be wise at present to attach any weight to it.#

While, as stated above, no adventitious branches have been met with in R. major,
the stems of this plant showed equal or dichotomous branching. An example is seen
in Pl. III, fig. 21, at a, where two steles of the same size are present in the one cortex.

A number of sporangia of R. major were figured in our previous paper (Part I,
PI. IX, figs. 62, 64—69), and the description of the sporangium there given is based
on them, and applies to this species. The figures on this plate show how much larger
the sporangia of R. major were than those of R. Gwynne-Vaughani. There is some
range in size of the sporangia of R. major, and still larger examples than any described
in Part I have since been met with. Thus the sporangium on PI. Ill, fig. 24, was
nearly 4 mm. in diameter. Its wall was well preserved, and showed the vertically
extended, thickened epidermal cells (ep.), the zone of thin-walled tissue ( ml .), and
the “ tapetal ” layer (tap.) around the large cavity filled with spores. t The sporangium
in PI. Ill, fig. 23, may have been slightly larger, since the wall is broken and over-
lapped at a ; only the epidermal layer enclosing the spores is preserved. It seems
clear that the sporangia of R. major attained a length of more than 12 mm., a
diameter of more than 4 mm., and a thickness of wall of ’3-'4 mm.

* The decayed steles in Pt. I (PI. X, figs. 76-78) belong to II. major, and the comparison there made with
Dawson’s figure of Psilophyton must be modified in so far as the outer dark zone in our specimens is derived from
the outer xylem and not as suggested from the altered phloem.

t Gf. Pt. I, PI. IX, fig. 67.

610 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

The spores (PI. Ill, fig. 26) were about 65 y in diameter. The difference in size
from the spores of R. Givynne-Vaughani was constant in all the sporangia of the
two species so far studied.

The remarkable and very persistent layer of cells which lines the cavity of the
sporangium in both species of Rhynia has been referred to throughout as the
“ tapetum.” Its position corresponds to that of a persistent tapetal layer, but it
shows peculiarities which call for special remark. When best preserved it is seen
to be a layer of somewhat flattened, elongated cells, the walls of which remain
bounding the empty-looking cell cavities. It suggests to our minds a relatively
rigid tissue, and is perhaps best compared with the sheath of tracheidal cells which
has been described in some other fossil sporangia.* The appearance of the walls
of these “tapetal” cells in Rhynia is quite consistent with their being of tracheidal
nature, though the preservation does not allow of any special thickening being
recognised.

Diagnosis,

The recognition of two species of Rhynia necessitates some modification of the
generic characters and a differential description of the species.

Rhynia.

Plant gregarious, rootless and leafless, consisting of subterranean rhizomes
attached by unicellular rhizoids, and erect, dichotomously branched, cylindrical aerial
stems. Stomata present. Stele consisting of a zone of phloem surrounding a strand
of tracheides. Sporangia cylindrical, without columella, terminal on aerial stems.
Homosporous. Spores with cuticularised wall, developed in tetrads.

Rhynia Gwynne- Vaughani.

Aerial stems tapering upwards, probably about 20 cm. high, and ranging in
thickness from 3 mm. to under 1 mm. Small hemispherical protuberances of super-
ficial tissues of the stem occur, and sometimes, in place of them, adventitious branches,
the stele of the branch not being continuous with that of the main stem. Xylem
strand of stele slender, only sometimes showing a distinction of smaller central and
larger peripheral tracheides. Tracheides with broad annular thickening. Sporangia
about 3 mm. long and 1'5 mm. in diameter. Sporangial wall about '2 mm. thick.
Spores about 40 y in diameter.

Locality. — Muir of Rhynie, Aberdeenshire.

Horizon. — Old Red Sandstone (not younger than the Middle Division of the Old
Red Sandstone of Scotland).

F. W. Olivee, “ On a Vascular Sporangium from the Stephanian of Grand Croix,” New Phytologist, vol. i, p. 60.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 611

Rhynia major.

Plant larger in all its parts than R. Gwynne-Vaughani. Aerial stems tapering
upwards, and ranging in thickness from 6 mm. to 1'5 mm. or less. No hemispherical
projections or adventitious lateral branches. Stele large, xylem strand of numerous
tracheides differentiated as smaller central and larger peripheral tracheides.
Sporangia reaching a length of 12 mm. and a diameter of 4 mm. Sporangial wall
about *3-'4 mm. thick. Spores about 65 g in diameter.

Locality. — Muir of Rhynie, Aberdeenshire.

Horizon. — Old Red Sandstone (not younger than the Middle Division of the Old
Red Sandstone of Scotland).

Hornea Lignieri, n.sp. (Plates IY-X.)

Another vascular plant from the Rhynie peat-bed agreed in general build with
Rhynia, but differed in the sporangia having a columella of sterile tissue, around and
above which the dome-shaped spore-sac extends. We regard it as a distinct, though
allied, genus, and have named it after Dr Horne, to whose energy and interest the
successful discovery of the Rhynie peat-bed in situ is largely due. The specific name
has been given in reference to the acute morphological speculations of the late
Professor Lignier, some of which gain confirmation and reality from the discovery
of these simple types of Vascular Cryptogams.

The remains of Hornea occur in considerable quantity, at places forming an
almost pure peat and elsewhere mixed with the other plants of the Rhynie deposit.
The preservation is not quite so good as in the case of Rhynia and Asteroxylon, but
shows all the essential structural features.

Hornea Lignieri resembled Rhynia in being a rootless and leafless plant,
differentiated into rhizome, dichotomously branched cylindrical stems, and terminal
sporangia. The continuity of these parts has been established. From the lobed
tuberous rhizome, which was probably subterranean but might have been only partly
embedded in the soil, a number of cylindrical stems arose separately. These bore no
appendages, nor did they have hemispherical projections or lateral branches. They
were from about 2 mm. to slightly under 1 mm. in diameter, and branched dichoto-
mously. This occurs in the larger stems, but was probably more frequently repeated
in the upper region of the plant, the stems diminishing in thickness as they
subdivided. The sporangia terminated stems of various diameters, and show a
corresponding range in size. The transformation of the end of a stem into a
sporangium often affected a nascent dichotomy, the branching being then evident
in the construction of the columellate sporangium.

The rhizome, stems, and sporangia may now be described in detail.

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 24).

94

612 DR R. K1DST0N AND PROF. W. H. LAND ON OLD RED SANDSTONE PLANTS

Rhizome. (Plates IY-YI.)

The rhizome of Hornea was of a different type from that of Rhynia, as will be
evident from figs. 27 and 28 on PI. IY. It had no vascular strand of its own, but
was a lobed parenchymatous body of considerable size, from which arose stems, each
provided with a central stele. The stele of each stem ended separately and blindly
below in the parenchymatous tissue of the rhizome (figs. 35 and 38). Numerous
long rhizoids attached the rhizome to the peat. From its resemblance to the
protocorm of the young plants of some species of Lycopodium , this type of rhizome
may be distinguished as protocormous from the stem-like rhizome of Rhynia. The
largest rhizome met with is represented at a low magnification ( x 7%) in fig. 31. It
is branched, and at first sight appears of a different type, not being evidently lobed.
It is, however, characteristic that no vascular system proper to the rhizome is seen.
Comparison with the next section of the series showed that a number of stems arose
from this rhizome, the plane of section in fig. 31 passing horizontally through the
rhizome below the bases of the steles of these stems.

The general structure of the rhizome is shown in the vertical section in fig. 29.
The plane of section just misses the stem, and the rhizome appears as a massive,
somewhat flattened body, about 8 mm. across by 2 mm. thick. It is composed of
thin-walled parenchymatous tissue, between the cells of which are small intercellular
spaces.* Towards the periphery the cells are somewhat smaller, and tend to be
arranged in rows vertical to the surface, giving the appearance of an ill-defined
outer cortex. The superficial layer constitutes a rather ill-defined epidermis, the
small cells of which on the lower side of the rhizome bear the long rhizoids.

The rhizoids, which are visible in a number of the illustrations (figs. 29, 30, and 38),
are non-septate, and each is a protrusion of the middle portion of the outer wall of
an epidermal . cell, from the cavity of which the rhizoid is not separated by a wall.
Fig. 30 shows the rhizoid-bearing surface in vertical section, while in fig. 32 the
surface is cut tangentially. Fig. 33 is a more general view of the latter section,
and shows the long rhizoids extending from the small-celled epidermis.

The structure of all the rhizomes shown on Pis. IV, Y, VI corresponds to that
described above, some showing particular features more clearly. A number of them
show the lobed form of the rhizome (figs. 27, 28, 35). In this connection the
specimen in fig. 34 is of interest. A rhizome with a stem has apparently been cut
obliquely, and two adventitious lateral growths had evidently developed from the
superficial tissues as small protocormous rhizomes. The one shown in the figure
bears rhizoids at rh.

No vascular elements were present in the section of the rhizome in fig. 29, though
some of the internal cells have peculiar brown walls, and in their neighbourhood the

* The question whether the fungal hyplise which occur in the intercellular spaces of the rhizome are to be
regarded as saprophytic or mycorhizic will be considered in a later paper when dealing with the fungi occurring in
the Rhynie peat. Their frequent presence must, however, be mentioned here.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 613

section was just touching the edge of the base of a stem-stele. The' stems attached
to the protocormous rhizomes in most of the specimens figured show their steles at
some point or other.

The characteristic base of the vascular strand of a stem springing from a rhizome
is best seen when the section is truly vertical to the rhizome and follows the median
plane of the stem (figs. 35-38). The best specimen is that shown in figs. 35 and 36,
the other two specimens serving to confirm it. As fig. 35 shows, one lobe of a
rhizome has been cut vertically, and the section follows in an accurately longitudinal
direction the stele of the stem arising from the lobe ; the cortex has decayed from
the further part of the stele as it lies in the peat. The rhizome shows the structure
already described, and bore rhizoids below. While there is no vascular tissue in the
rhizome itself, the stele of the stem when traced downwards into the upper part
of the rhizome (figs. 35, 36) ends in connection with a brown-celled tissue without
tracheidal thickenings. The inverted cup-shape of this mass of tissue is characteristic,
though not often so favourably shown as in this specimen ; it is seen also in figs. 37
and 38. It explains the frequent occurrence of patches of brown tissue in sections
of rhizomes (cf figs. 29 and 31).

Stem. (Plates VII and VIII.)

The basal regions of the stems attached to the rhizomes described above were
often more or less decayed, but were sometimes well preserved. Portions of both
the upper and lower regions of similar stems in various stages of decay make up the
peat. A group of the best-preserved specimens is shown in fig. 39, some being cut
in transverse and one in longitudinal section. The general similarity in structure
to the stem of Rhynia will be evident, though the preservation is not so good in the
material of Hornea yet examined. The well-defined cuticle and thickened epidermal
wall, the broad cortex, and the stele with a zone of phloem surrounding the rather
stout central strand of xylem, are distinguishable in fig. 39. The transverse section
of a small stem from the upper region of the plant in fig. 40 may be compared with
this. The corresponding regions are shown in longitudinal section in fig. 41.

The tissues may now be described in greater detail, starting from the outside
(figs. 40 and 4l). There was a well-marked cuticle. The epidermis had its outer
walls somewhat thickened, but otherwise did not differ much from the underlying
cortical cells. No stomata have been observed, but owing to the state of preservation
of most of the material much weight cannot be attached to this negative result.
The cortical tissues cannot be sharply distinguished into outer and inner cortex,
though the cells diminish in diameter on passing inwards. The cortical cells are
somewhat elongated, and have transverse end walls (fig. 41).

Within the cortex and immediately around the xylem was a zone of thin-walled
elements which readily collapsed. This zone represents the phloem (figs. 42 and 43).
It often has a peculiar appearance, as if the cells were thickened at the angles or small

614 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

intercellular spaces were present there (figs. 42, 43, 50, 51). In the more decayed
stems the phloem may or may not be recognisable. It seems to have been less
characteristically developed at the lower part of the stele near its insertion into the
rhizome. All that can be said regarding the phloem in longitudinal section is that
it consisted of narrower and more elongated elements than the cortex.

The xylem formed a centrally situated, solid strand of tracheides. It exhibits a
considerable range in diameter, probably in relation to the regions of the stem-system
of the plant. A xylem strand of moderate size is shown in transverse section in
fig. 47, and others of larger size at the same magnification in figs. 45 and 48, with
which fig. 46 can be compared. The xylem is shown more highly magnified in
figs. 50 and 51. A distinction between narrower central, and wider peripheral
tracheides is a constant feature. This holds for the steles of moderate diameter near
to the rhizome, for the steles of large diameter, which we presume came from a region
between this and the finer branches, and for the finer branches themselves. There are
no thin-walled cells mixed with the tracheides, but the central xylem often appears
broken down within a ring of peripheral xylem composed of wider and intact elements
(figs. 43, 45, 46, 48). When such steles are cut longitudinally this appearance is
seen to be due to repeated transverse breaks or interruptions of the core of central
xylem, the tracheides of the peripheral xylem remaining continuous (fig. 49).

The thickenings on the tracheide walls have in most cases disappeared owing to
the decayed condition of the stems. They are often best shown in the region of the
stele near to the rhizome. In favourable specimens there is a distinct thickening
of narrow bands forming irregularly connected rings or a spiral (figs. 52 and 53).
As the extreme base of the stele in the rhizome is reached, the zone of phloem
disappears and the tracheides become shorter and wider (fig. 54) and pass gradually
into the brown-celled tissue at the base of the stele.

The stems of Hornea branched dichotomously, and sometimes, show the stele
dividing (figs. 55 and 56), or two equal steles enclosed in the same cortex (figs. 42
and 57). From the frequency with which it is met with in smaller stems, this
dichotomous branching was probably most marked in the upper region of the plant.

Sporangium. (Plates IX-X.)

Sporangia of a remarkable type occur associated with the remains of the vegetative
organs of Hornea, especially with the finer branching stems. The proof that they
belong to this plant is afforded by sporangia terminating stems with the structure
described above (figs. 58, 60, 61).

The general construction of the sporangium will be evident from the longitudinal
sections in figs. 58-60, and the transverse sections in figs. 63 and 64. As these
show, the sporangial cavity was enclosed by a fairly thick wall, and had a sterile
column of tissue projecting so far from its base that the actual cavity is dome-shaped.
The resemblance to the columella of some bryophytic sporogonia is so great as to

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 615

make it convenient to use the name columella for the sterile projection in Hornea .
The cavity is filled by the numerous spores with cuticularised walls.

The sporangia evidently arose by the transformation of the tips of certain
branches of the plant. When the apex was simple, a single sporangium (such as that
shown in fig. 58 to the right, and in fig. 64) resulted. When, however, the apex was
in a more or less advanced stage of division, this is reflected in the subdivision of
the sporangium and the lobing or branching of the columella, as shown in fig. 58
to the left, and in fig. 63.

The stalk of the sporangium is thus simply an ordinary branch of the plant. It
shows the same structure as the stem, with epidermis, cortex, and stele more or less
well preserved. Stalks bearing sporangia and showing the position of the stele are
represented in 'figs. 58 and 60. The size of the sporangium differs according to the
size of the stem bearing it. Thus the sporangium in fig. 58 is about 2 mm. long
by slightly over 1 mm. broad, while that in fig. 64 is about 2 mm. in diameter.

The wall of the sporangium was of considerable thickness (about '25 mm.), and
its tissues, like those of the stems, are usually imperfectly preserved. An epidermal
layer, a middle zone consisting of a number of layers of thin-walled cells, and a more
resistant tapetal layer can be distinguished (figs. 65 and 67), as in the case of
the sporangium of Rhynia. The wall shows irregularities in outline or projecting
processes which do not seem to be wholly accounted for by accidental contraction.
A common appearance is the flattening or broadening of the tip of the sporangium,
as shown in figs. 58 and 59.

The epidermal layer of the sporangium is, at least in some cases, more marked
than that of the vegetative . stems. It is bounded by a cuticle, and has the outer
walls of the cells thickened, the thickening extending inwards on the lateral walls.
No indication of any place of dehiscence has been seen even in complete transverse
sections. The structure of the epidermis is seen in figs. 67 and 68, while that of the
wall as a whole is well shown in fig. 65. Below the epidermis comes a zone of some
six layers of small thin-walled cells ; this is usually badly preserved. The innermost
layer of the wall is the persistent “ tapetum,” which is continuous over the columella
and thus forms a complete lining to the sporangial cavity. The tapetal cells are
well shown in fig. 70. Their walls are dark, and give the impression of having been
rigid, as in the case of the corresponding layer in the sporangium of Rhynia. No
indication of tracheidal thickening of these cells has been met with.

The columella (fig. 69) is composed of narrow, elongated, thin-walled cells which
give the tissue, as usually preserved, a peculiar fibrous appearance. Though corre-
sponding in position to a continuation of the stele of the stalk, it exhibits no
agreement in histological structure with the central region of this, but resembles
rather the phloem of the stele. The specimen shown in fig. 62 shows that the
columella is directly continuous with the phloem of the stalk.

The spores are often met with still associated in tetrads (fig. 7 1 ), while in other

616 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

cases they have become isolated in the sporangial cavity (fig. 72). They measure
about 50 n in greatest diameter. Spores are also found distributed through the
peaty matrix in which the plant is embedded.

Diagnosis.

Hornea, Kidston and Lang, n.g.

Plant rootless and leafless. Stems arising from protocorm-like rhizomes,
dichotomously branched. Sporangia terminal on ultimate branches, with a sterile
columella projecting from the base into the sporangial cavity, and cuticularised
spores developed in tetrads.

Hornea Lignieri, Kidston and Lang, n.sp.

Plant small, consisting of a lobed rhizome from which arise stems which branch
dichotomously and range from 2 mm. in diameter downwards. Stele of stem with
a zone of phloem surrounding the xylem composed of small central and wider peri-
pheral tracheides. Sporangia cylindrical, terminal on branches, indehiscent, with
thick wall composed of thickened epidermis, thin-walled tissue, and persistent tapetal
layer. Sterile columella composed of thin-walled elongated cells extending from base
to near top of sporangium. Homosporous. Spores about 50 p in diameter.

Locality. — Muir of Rhynie, Aberdeenshire.

Horizon. — Old Red Sandstone (not younger than the Middle Division of the Old
Red Sandstone of Scotland).

Classification of Rhynia and Hornea.

Rhynia and Hornea, while distinguished generically, agree so closely in the
simplicity of their organisation that they must be regarded as genera of the same
Family. For this Family we suggest the name Rhyniacese.

It is characterised by the plants being rootless and leafless and composed of
rhizomes which bear rhizoids, branched aerial stems, and terminal sporangia. The
vascular system is correspondingly simple, the central stele having a cylindrical
strand of xylem either composed of similar tracheides or with a distinction of central
and peripheral xylem.

The Family Rhyniacese comes into the class of Vascular Cryptogams to which
we have given the name of Psilophytales (Part I, p. 780), since “the sporangia are
borne at the ends of branches of the stem without any relation to leaves or leaf-
like organs.”

The consideration of more complicated types of the Psilophytales can be deferred
until Asteroxylon is described.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 617

The classification of the plants so far described from the Rhyme peat-bed is
therefore as follows : —

PTERID OPHYT A.

Class. PSILOPHYTALES.

Family, rhyniacea;.

Genera. Rhynia (2 species).

Hornea (l species).

Comparative Discussion.

The Vascular Cryptogams preserved in the Rhynie peat-bed are the most ancient
plants of which the internal structure and external morphology are adequately
known. They facilitate our understanding of those plants of early Devonian age
which are only known as impressions, while these assist in forming a conception of
the form and habit of the plants of this remarkable flora. The consideration of this
flora as a whole is beyond the scope of the present paper. In this discussion we are
concerned only with the simpler types of the Psilophytales which have been placed
above in the Family Rhyniacese.

The great interest of these plants ( Rhynia , Hornea ) is that they acquaint us
with a type of construction fundamentally more primitive, not only than that
of existing land plants but of most plants composing the Upper Devonian and
Carboniferous floras.

Between the Lower and Middle Old Red Sandstone and the Upper Old Red
Sandstone or Devonian, as represented by the flora of the Kiltorkan beds, there is
a most remarkable difference in the type of plant life. It is doubtful if a single
species which occurs in the Lower and Middle Old Red Sandstone of Scotland is
present in the Upper Old Red as represented by the Kiltorkan beds. In fact, the
botanical affinity of. the plants of the latter is with the Lower Carboniferous flora
rather than with the Middle and Lower Old Red Sandstone plants.

The type of plant exhibited by Hornea and Rhynia is the simplest known in
undoubted Pteridophyta. We know nothing of the sexual generation, the existence
of which there is no reason to doubt, and of the early stages of the sporophyte. The
full-grown sporophyte is differentiated into rhizome, stems, and sporangia. It has
no leaves or roots. The rhizome, which was probably subterranean, was intimately
connected to the substratum by long non-septate . rhizoids. In Hornea it was a
tuberous parenchymatous structure from which the stems arose. In Rhynia it had
a central vascular strand, and is like a peculiar region of the stem. The cylindrical
stems branched dichotomously, diminishing in diameter. They had a simple
vascular system, the stele consisting of a cylindrical strand of xylem surrounded
by a zone of phloem. Stomata occurred in the epidermis of Rhynia, and there

618 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

was a well-developed system of small intercellular spaces. The peculiar hemi-
spherical projections of Rliynia Gwynne-Vaughani were the seat of adventitious
lateral branches, and, whatever their other functions, may provisionally be
regarded as dormant branches. The sporangia correspond to transformed and
more or less specialised ends of stems of various sizes. Their thick wall was
indehiscent, and the epidermis was usually specially thickened. There was a
peculiar (and possibly tracheidal) tapetum. The spores were developed in
tetrads and had cuticularised walls. No sterile cells were mixed with the
spores. In Hornea the central tissue formed a sterile columella projecting into
the sporangial cavity.

This is a simpler type of plant than is known in any existing Vascular Cryptogam,
or in any extinct forms from the Carboniferous period onwards. In the Devonian
period, however, there are indications of plants of correspondingly simple
morphology. It will be sufficient to mention here Sporogonites as possibly
related to the Rhyniacese.

Sporogonites exuberans, Halle,* from the Lower Devonian rocks of Roragen,
Norway, consists of a long unbranched axis terminating in a sporogonial or sporangial
structure which is similar to the sporangium of Hornea in a number of characters.
There are distinct differences, however, but the two agree in the presence of a
columella rendering the spore-containing cavity dome-shaped. Halle regards
Sporogonites as a sporogonium and as finding its “ nearest analogy in the sporogonia
of the Bryophyta.”t Our fuller knowledge of Hornea , however, lends weight to
another suggestion of Halle that “ the possibility must be faced that it may
represent only the upper part of a more highly developed sporophyte, perhaps on
the line of descent of the pteridophytes.” J So far as the evidence afforded by the
remains of Sporogonites goes, it does not establish the bryophytic nature of the
plant. In showing a general resemblance to sporogonia of Bryophytes the sporangia
of Hornea and Rliynia agree with Sporogonites , and in them the relations of the
sporangia to the whole simple vascular plant is known.

While the comparisons with Sporogonites and with some undescribed im-
pressions which we have seen are of great interest, our fairly complete knowledge
of the structure as well as the external morphology of Rliynia and Hornea makes
them more suitable for comparison with other plants. In dealing with such
simply organised Pteridophyta comparisons must be widely extended, but at present
are best only indicated in outline.

The simple organisation of the sporophytes of Rliynia and Hornea may there-
fore be briefly compared with some other Pteridophyta, with Bryophyta, and
with Algae.

* Halle, T. G., “ Lower Devonian Plants from Roragen in Norway,” Kongl. Svensk. Vetenskapsaked. Handlingar,
Bd. 57, 1916, p. 27, pi. iii, figs. 10-32.

t Loc. tit., p. 31.

| Loc. tit., p. 40.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 619

Comparisons with Pteridophyta.

In comparing the Rhyniacese with other Pteridophyta we are limited by the
fact that we are dealing with the simplest types of Vascular Cryptogams known.
All the other Vascular Cryptogams, including such archaic forms as Asteroxylon
and Psilophyton, are more highly organised than these rootless and leafless plants,
the body of which might as well be termed a cylindrical, branched, vascular thallus
as a stem. Thus no detailed comparison with the organisation of the whole plant
in the higher Pteridophyta is possible, and what we have to say on this question
will find its proper place in dealing with Asteroxylon.

We can, however, recognise correspondences with particular structures in some
other Pteridophyta, and it is not without interest that some of these affect features
which have always appeared aberrant in the plants possessing them. This applies
especially to the rootless rhizome, which may be first briefly discussed.

The only rootless Pteridophyta at present known that need be considered
are the Psilotacese.* Both Psilotum and Tmesipteris have rhizomes embedded

in the substratum and attached to this by non-septate rhizoids. It is not at
present known whether a root is absent from the young sporophyte of Psilotum
in its development from an embryo, but none is present in plants arising
from the minute bulbils t formed on the rhizome. No root is present in the
young plant of Tmesipteris \ borne on the prothallus. The resemblance between
the rhizomes of Rliynia and of the Psilotacese is close, and in neither case is
there anything suggesting that the rootless condition is other than primitive.
Although more complicated in the aerial shoots, the Psilotacese appear to have
retained the simplicity of the subterranean region characteristic of the most
primitive Vascular Cryptogams.

The rhizome of Hornea, on the other hand, appears to correspond most
closely to that remarkable and much discussed region of the young plants of
certain species of Lycopodium and of Phylloglossum known as the protocorm.
As is well known since the investigations of Tretjb,§ the young plant of Lyco-
podium cernuum does not at once initiate a shoot and root, but enlarges outside
the prothallus (to which it is attached by a small foot) to form a tuberous
parenchymatous body. This protocorm is attached to the soil by numerous
rhizoids, and bears on its upper side a number of leaves in no apparent relation
to a stem-apex ; these leaves are termed protophylls. Later a stem bearing
spirally arranged leaves develops, and the first root forms at the base of this
shoot. The corresponding stage in L. later ale and L. ramulosum, two New

• Salvinia and some of the smallest Hymenophyllacese may be mentioned, but the rootless condition in them
appears to be due to reduction.

+ Solms, Annales du Jardin Bot. Buitenzorg, vol. iv, 1884, pp. 139-190.

f Holloway, Trans. New Zealand Institute, vol. 1, 1917, pp. 1-44.

§ Treub, Annales du Jardin Bot. Buitenzorg, vols. iv and viii.

TRANS. ROY. SOC. EDIN., YOL. LII, PART III (NO. 24).

95

620 DR R. K1DSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

Zealand species, the life-history of which has been described by Holloway,*
attains a much larger size before a stem-apex is differentiated. “ This protocorm-
ous rhizome plays a much larger and more important part than does the simple
protocorm of L. cernuum. It constitutes the plant body for a whole season and
may even branch. It appears also that in both L. laterale and L. ramulosum
it is able to reproduce itself vegetatively.” t The vascular bundles of the more
or less numerous protophylls borne on the protocorms of these three species of
Lycopodium end blindly below in the parenchymatous tissue, the protocorm itself
having no vascular system.

The protocorm of Phylloglossum, which in the embryo plant resembles that of
L. cernuum, appears to be repeated as the specialised resting tuber by means of
which this plant perennates. This region in Phylloglossum can be regarded as both
primitive and specialised.

The rootless plant of Horned appears to retain in the adult condition an organisa-
tion comparable to the protocorm stage in the species of Lycopodium mentioned
above. The relation of the aerial stems of Hornea to the rhizome is similar to that
of the protophylls to the protocorm in Lycopodium.

The resemblance of the protocormous rhizome of the ancient Vascular Cryptogam
Hornea to the protocorm present as a transient embryonic phase in some existing
species of Lycopodium supports the suggestion made by Treub as to the nature
of the protocorm. He was led to assume “ que chez les ancetres des crypto-
games vasculaires actuels un organe a pris naissance qui a du avoir beaucoup de
ressemblance avec le tubercule embryonnaire des Lycopodes.” His discussion is
summed up thus : —

“ Le tubercule embryonnaire chez les Lycopodes est un organe rudimentaire.”

“ L’organe admis theoriquement chez les ancetres des Cryptogames vasculaires
actuels, et designe ci-dessus comme ‘ predecesseur de la pousse feuilMe telle quelle
se presente maintenant chez les plantes vasculaires ’ existe encore aujourd’hui
a l’etat passager, dans le genre Lycopodium. Cet organe nest autre que le
tubercule embryonnaire .”

“ Aussi je propose de donner au tubercule embryonnaire des Lycopodes le nom
de protocorme .” |

In view of the morphology of Hornea it may be said that the existence of this
organ does not require to be theoretically assumed in archaic Vascular Cryptogams,
but is demonstrated.

So far as the anatomical construction of the Rhyniacese is concerned, we can only
recognise its simplicity and its direct relation to the primary vegetative functions
of land plants, absorption, conduction, aeration, and protection. It is only necessary

* Holloway, Trans. New Zealand Institute, vol. xlviii, 1915, pp. 253-303.

t Ibid., vol. xlix, 1916, p. 90.

t Treub, Annales du Jardin Bot. Buitenr.org , vol. viii, 1890, p. 30.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 621

to recall the rhizoids, the simple vascular strand, the intercellular spaces and stomata,
and the thick cuticle to indicate how these needs are met.

The simplicity of the stelar anatomy limits comparison with other vascular plants,
but may be compared with the rhizome of the Psilotacese. In the xylem of the
central cylinder of the Rhyniacese we find two grades of construction illustrated :
(a) the strand of equivalent and similar tracheides in many stems of Rhynia Gwynne-
Vaughani ; ( b ) a differentiation of the strand of xylem into smaller inner tracheides
and larger peripheral tracheides, as in some stems of Rliynia Gwynne-Vaughani
and regularly in R. major and in Hornea. This differentiation suggests comparison
with the xylem of the stem-stele and leaf-traces of a number of Pteridophyta in
which the protoxylem is surrounded by metaxylem (centrarch).

The sporangia of Rhynia and Hornea will be compared below with the sporogonia
of Bryophyta and the stichidia of the Red Sea-weeds. What in other Pteridophyta
may correspond to their simple terminal position will be better discussed in a later
paper when the more complicated types of Psilophytales have been considered.
A peculiarity of the sporangia of Rhynia and Hornea is that they are relatively
large and indehiscent. The thickened epidermis seems an advance in specialisation,
though apparently useless for dehiscence. It looks as if the thickened epidermis,
serving for the protection of the spores, had preceded any arrangement for the
opening of the sporangium in these plants. The sterile columella of the sporangium
of Hornea may correspond to the sterile projection from the base of the sporangium
in some species of Lycopodium.

Comparisons with Bryophyta.

The difficulty in comparing Rhyniacese with Bryophyta lies in our ignorance of
the sexual generation and the relations of the young sporophyte to it in the ancient
Pteridophyta. The simple morphology of the vegetative region of the plants of
Rhynia and Hornea as well as their sporangial structure appear, however, to bring
us a step nearer to a comparative morphology of the sporophyte of Vascular Crypto-
gams and the bryophytic sporogonium. It is sufficient at present to indicate the
general resemblances that can be traced between the fertile regions of these Old Red
Sandstone plants and the sporogonia of Hepaticse, of Anthoceros, and of certain
Mosses. Whether the sporogonium is a simpler parallel development to the
sporophyte or is a reduced equivalent of such a plant as that of the Rhyniacese must
remain an open question. The possibility that these questions, which have hitherto
had to be treated in a purely speculative fashion, may be at least partially answered
by the discovery of further plants of early Devonian age,~is strengthened by the
experience of the last few years.

622 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

Comparisons with Alg^e.

Without implying direct derivation of the Pteridophytes from any more complex
Algae or actual relationship with any known Algae, it is impossible to overlook the
resemblance in habit between the plant of Rhynia and Hornea and some Algae.
The comparison is closest with some Red Algae, which show a distinction of a rhizome-
like basal region and cylindrical dichotomously branched axes and have their
tetraspores developed within the more or less altered ends of some of the branches.
As examples, Furcellaria fastigiata and Cordylecladia ( Gracilaria ) erecta may be
mentioned. Such resemblances, though they may be superficial, indicate the need of
caution in deciding whether impressions of early Devonian plants of this habit were
Pteridophyta or Algae. Formerly a number of these plants, which we now know
to be Vascular Cryptogams, were classed as Algae or Fucoids. We anticipate that
others will prove really to be Algae, and that the comparison between them and
the Pteridophyta of the same early geological age will be of peculiar interest.

A more special question concerns the equivalent of the sporangium of the
Rhyniaceae in the Algae. The tetrads of spores in certain Algae and the tetrads in
the land plants appear to correspond strictly. In some Red Algae we find the spore-
mother-cells (or tetrasporangia) confined to certain special branches, or to the tips of
ordinary branches and there situated below the surface. Such special tetraspore-
bearing branches are known as stichidia. It is one possible view of the sporangia
of Pteridophyta to regard them as the equivalents of such stichidia. The general
resemblance and the correspondence in regional anatomy which can be traced
between the sporangium of Hornea and the less differentiated sporangia of Rhynia,
on the one hand, and the tetraspore-containing branches of such Algae as Cordyle-
cladia, Furcellaria, Gigartina, and some Rhodomelaceae on the other, supports
this way of regarding the sporangia of these early Pteridophyta.

We are here probably dealing with a homoplastic similarity which is an expression
of general morphology rather than of relationship. The whole question is too wide
to be appropriately treated in this place, but the preceding remarks will suffice to
indicate the bearing of the new facts upon it. In our ignorance of the complete life-
histories and the forms of the sexual and asexual generations, both of the ancient
Algae and of the simplest Pteridophytes, any suggestion of relationship would be
dangerous and is not implied in the above comparisons. The facts are, however,
consistent with the Rhyniaceae finding their place near the beginning of a current of
change from an Alga-like type of plant to the type of the simpler Vascular Cryptogams.

Conclusion.

If we review the comparisons we have been led to make, it will be evident that
all the features of Rhynia and Hornea point to the Rhyniaceae illustrating an early

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 623

stage in the organisation of the sporophyte of Vascular Cryptogams. On grounds of
their general structure it may be presumed that these simple plants were land plants.
They are somewhat advanced along the line of progression of the land-growing
sporophyte, but not so far as to obliterate clues to the origin of this.

Speculations as to the origin of the sporophyte of the Pteridophyta when broadly
considered fall into two groups. On the one hand, it has been suggested that the
vegetative organs of the sporophyte have arisen as the result of a process of pro-
gressive sterilisation within an interpolated phase of the life-history ; this phase in
its simplest form was composed wholly of reproductive spores. On the other hand,
the origin of the sporophyte of the Pteridophyta has been looked for in the modi-
fication of a plant body such as we see in the asexual stage of a number of Algae ;
this would have a more or less differentiated vegetative portion bearing asexual
reproductive organs.

It appears to us that, while not inconsistent with the first of these two hypo-
theses, the morphology of the Rhyniaceae is much more in line with the second.
The organisation of Hornea and Rhynia make the transition assumed by this second
hypothesis readily conceivable.

We do not, however, desire to enter at present into a discussion of the general
problem or to draw conclusions that would be premature. The interest of Rhynia
and Hornea lies in their providing new and definite historical data on early steps
in the evolution of land plants. Since we may hopefully expect further dis-
coveries to make the story more complete, any hypothetical construction of the
course of evolution would be superfluous. It is sufficient, without entering into
particular theories, to have indicated the bearing of the new facts on some of
the chief hypotheses as to the early steps in the origin of land plants that have
been entertained by morphologists. That these archaic and simple Pteridophyta
should compel us to institute comparisons, not only with some existing Vascular
Cryptogams but with bryophytic sporogonia and the organisation of the plant
in some Algse, appears to us pregnant with morphological interest whatever the
actual lines of descent may have been.

EXPLANATION OF PLATES.

(All the figures are from untouched photographs.)

Plate I.

Rhynia Gwynne- Vaughani, Kidston and Lang.

Fig. 1. A small rhizome ( r .) in transverse section attached by rhizoids (rh.) and giving off laterally an
ascending stem ( br .). x 14. (Slide No. 2423.)

Fig. 2. Longitudinal (a) and transverse ( b ) sections of apical regions of stems. x 14. (Slide No. 2425.)
Fig. 3. The longitudinal section of the apex of a stem shown in fig. 2 (a), more highly magnified, x 105.
(Slide No. 2425.)

624 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

Fig. 4. The transverse section of a stem apex in fig. 2 ( b ). x 105. (Slide No. 2425.)

Fig. 5. Longitudinal section of stem showing incompletely developed tissue with cell contents, x 33.
(Slide No. 2426.)

Fig. 6. Basal region of the sporangium shown in Part I, PI. IX, figs. 63 and 63 (a), ep., epidermis ;
m.l., decayed middle layers of wall ; tap., tapetum. x 60. (Slide No. 2392.)

Fig. 7. Apical region of the same sporangium. ep., epidermis ; tap., tapetum. x 60. (Slide
No. 2392.)

Fig. 8. Sporangium in transverse section, ep., epidermis; m.l., decayed middle layers of wall; tap.,
tapetum. x 14. (Slide No. 2427.)

Fig. 9. Sporangium in obliquely transverse section full of spores, x 14. (Slide No. 2424.)

Fig. 10. Obliquely longitudinal section of sporangium without tapetum and without specially thickened
epidermal layer, the cavity partially filled with fully developed spores, x 14. (Slide No. 2423.)

Fig. 11. Small sporangium in which the mass of spores is enclosed by a wall like the ordinary tissues
of the stem. x 33. (Slide No. 2428.)

Fig. 12. Spores in tetrads from the sporangium in fig. 11. x 105. (Slide No. 2428.)

Fig. 13. Mature spores from the sporangium shown in fig. 9. x 160. (Slide No. 2424.)

Plate II.

Rhynia major, Kidston and Lang.

Fig. 14. Transverse section of a moderately large and well-preserved stem, ep., epidermis; o.c., outer
cortex; i.c., inner cortex; ph., phloem; x., xylem; a, interruption in outer cortex in relation to a stoma,
x 20. (Slide No. 2429.)

Fig. 15. Transverse section of one of the smallest stems met with, showing the relatively large stele,
x 20. (Slide No. 2430.)

Fig. 16. Transverse section of stele, photographed to show the phloem. x 33. (Slide No. 2431.)

Fig. 16a. Central portion of specimen seen in fig. 16, photographed with different screen to show details
of the xylem. x.o., outer xylem ; x.i., central xylem. x 60. (Slide No. 2431.)

Fig. 17. Longitudinal section of stem with well-preserved stele, o.c., outer cortex ; i.c., inner cortex;
ph., phloem ; x., xylem. x 20. (Slide No. 2429.)

Fig. 18. Longitudinal section of stem, the inner cortex and phloem decayed, o.c., outer cortex interrupted
at a beneath a stoma ; x., xylem. x 14. (Slide No. 2395.)

Fig. 19. Transverse section of large stem showing the breaking down of the inner cortex as the result

of decay, o.c., outer cortex; i.c., inner cortex; ph., phloem; x., xylem. x 20. (Slide No. 2398.)

Plate III.

Rhynia major, Kidston and Lang.

Fig. 20. Longitudinal section of stele with the soft tissues preserved, i.c., inner cortex ; ph. phloem ;
x., xylem. x 60. (Slide No. 2429.)

Fig. 21. Group of stems compressed in the peat, the one at a showing two steles in preparation
for dichotomous branching, x 14. (Slide No. 2432.)

Fig. 22. Stele from a stem photographed to show the xylem. i.c., inner cortex ; ph., phloem ; x.o.,

outer xylem ; x.i., central xylem. x 60. (Slide No. 2432.)

Fig. 23. Transverse section of a large decayed sporangium filled with spores. Only the thickened
epidermis is preserved, and this is broken and overlapped at a. x 14. (Slide No. 2433.)

Fig. 24. Transverse section of sporangium with well-preserved wall, ep., thickened epidermis; m.l.,
middle layers of wall; tap., tapetum; sp. spores. x 14. (Slide No. 2434.)

Fig. 25. Portion of the silicified peat composed of decayed stems of Rhynia major with two sporangia
of this species, x 7. (Slide No. 2435.)

Fig. 26. Spores from sporangium in fig. 23. x 160. (Slide No. 2433.)

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 625

Plate IV.

Hornea Lignieri, Kidston and Lang.

Fig. 27. Large protocormous rhizome in vertical section with three lobes, a stem arising from each lobe,
x 14. (Slide No. 2436.)

Fig. 28. Portion of the silicified peat, the matrix of which is composed of decayed stems of Hornea with
spores scattered throughout. Portions of three rhizomes are shown. The largest is cut vertically and shows
two tuberous lobes from each of which a stem arises. A stele is present in each stem. x 14. (Slide
No. 2437.)

Plate V.

Hornea Lignieri, Kidston and Lang.

Fig. 29. Vertical section through a rhizome missing the attachment of any stem, but showing the general
parenchymatous structure and the distribution of the rhizoids. ox., outer zone ; i.c., inner tissue ; b.t., brown-
walled tissue in the neighbourhood of the base of a stele, x 14. (Slide No. 2438.)

Fig. 30. Portion of the lower surface of a rhizome in vertical section showing the rhizoids springing as
protrusions of all the epidermal cells, x 33. (Slide No. 2442.)

Fig. 31. Section of portion of the largest rhizome met with. This is a horizontal section of a rhizome
below the bases of the steles of a number of stems which arose from it. The brown parenchymatous tissue
in relation to some of these is seen at the points marked b.t. Lying beside the rhizome at st. is an obliquely
transverse section of a stem. x 7|. (Slide No. 2441.)

Fig. 32. Tangential section showing the rhizoid-bearing surface of a rhizome. The bases of the rhizoids
are cut across at a, where they spring from the central portion of each epidermal cell. x 60. (Slide
No. 2443.)

Fig. 33. The tangential section of the rhizoid-bearing surface from which fig. 32 was taken, showing the
long rhizoids extending into the peat, x 14. (Slide No. 2443.)

Fig. 34. Small protocormous rhizome (r.') with rhizoids (rh.) springing as an adventitious growth of the
superficial tissues of the main rhizome (r.). x 33. (Slide No. 2441.)

Plate VI.

Hornea Lignieri, Kidston and Lang.

Fig. 35. Accurately vertical section of a lobe of a rhizome showing the attachment of a stem. The stele
of this on being followed down into the rhizome shows the characteristic base in longitudinal section. The
peat around contains spores and decayed stems of Hornea. x 14. (Slide No. 2444.)

Fig. 36. Base of the stele shown in fig. 35. x., xylem composed of elongated tracheides ; x.' , region with

short tracheides passing into an inverted bowl-shaped mass of brown-walled parenchymatous tissue (b.t.).
x 33. (Slide No. 2444.)

Fig. 37. Base of the stem-stele of another specimen showing the same features as fig. 36. x., xylem

composed of elongated tracheides; x.', region with short tracheides passing into an inverted bowl-shaped
mass of brown-walled parenchymatous tissue (b.t.). x 33. (Slide No. 2445.)

Fig. 38. Vertical section of a rhizome showing good rhizoids and with the characteristic base of a stem-
stele displaced downwards through the decay of the inner tissue, x 20. (Slide No. 2446.)

Fig. 39. Portion of silicified peat containing fairly well-preserved stems of Hornea. These stems, one
of which is in longitudinal section, show the structure of the lower region close to the rhizome, x 14.
(Slide No. 2447.)

Plate VII.

Hornea Lignieri, -Kidston and Lang.

Fig. 40. Slightly oblique transverse section of a small stem, probably from the upper region of the plant.
cut., cuticle ; c., cortex with indications of outer and inner zones ; ph., phloem ; x., xylem. x 33. (Slide
No. 2448.)

626 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

Fig. 41. Longitudinal section of stem about 5 mm. above its origin from a rhizome, ep., epidermis ;
e., cortex with indications of outer and inner zones ; ph., phloem ; x., xylem. x 33. (Slide No. 2440.)

Fig. 42. Transverse section of stem with stele just divided in preparation for dichotomy, cut., cuticle ;
c., perished cortex; ph., well-preserved phloem; x,, xylem. x 33. (Slide No. 2446.)

Fig. 43. One of the steles in fig. 42 more highly magnified, pli., phloem ; x.o., outer xylem ; x.i., central
xylem. x 105. (Slide No. 2446.)

Fig. 44. Small stem with relatively large stele. x 20. (Slide No. 2449.)

Fig. 45. Stele of stem in fig. 44. x.o., outer xylem ; x.i., central xylem partially broken down, x 60.

(Slide No. 2449.)

Fig. 46. Another large stele with broken-down central xylem. ph., phloem ; x.o., outer xylem ; x.i.,
position of central xylem. x 33. (Slide No. 2450.)

Fig. 47. Transverse section of stem-stele close to the rhizome, x.o., outer xylem; x.i., central xylem.
x 60. (Slide No. 2447.)

Fig. 48. Transverse section of large stele, x.o., outer xylem ; x.i., perished central xylem. x 60. (Slide
No. 2447.)

Fig. 49. Longitudinal section of xylem of stem-stele showing the continuous outer xylem (a;.o.) and the
central xylem (x.i.) interrupted by breaks. x 60. (Slide No. 2441.)

Plate VIII.

Hornea Lignieri, Kidston and Lang.

Fig. 50. Transverse section of stele near its origin from the rhizome, ph., phloem ; x.o., outer xylem ;
x.i., central xylem. x210. (Slide No. 2447.)

Fig. 51. Portion of the stele shown in fig. 47. ph., phloem; x.o., outer xylem; x.i., central xylem.
x 210. (Slide No. 2447.)

Figs. 52, 53. Tracheides in longitudinal section showing the fine irregularly spiral or reticulate thickening,
x 210. Fig. 52 (Slide No. 2452); Fig. 53 (Slide No. 2451).

Fig. 54. Short tracheides from the basal region of a stele showing the fine irregularly spiral or reticulate
thickening, x 210. (Slide No. 2442.)

Fig. 55. Stele showing dichotomy, x 60. (Slide No. 2442.)

Fig. 56. Small stem with stele showing dichotomy, x 20. (Slide No. 2452.)

Fig. 57. Small stem a showing dichotomously divided stele. x 20. (Slide No. 2453.)

Plate IX.

Hornea Lignieri, Kidston and Lang.

Fig. 58. Portion of peat containing, besides more or less decayed stems of Hornea, two slender stems
terminating in well-preserved sporangia cut in accurate longitudinal section. The sporangium to the left
is bifurcate, that to the right is simple. A portion of the stele (st.) is seen in the stalk of each, x 14.
(Slide No. 2454.)

Fig. 59. Imperfect sporangium in longitudinal section showing flattened top. w., wall; sp., spores;
col., columella. x 20. (Slide No. 2424.)

Fig. 60. Broad divided sporangium terminating a stalk with two steles (st.). w., wall ; sp., spores ; col,

columella, x 20. (Slide No. 2456.)

Fig. 6 1 . Portion of peat with spores of Hornea distributed through it, and containing on the left a small
stem (a) and on the right an obliquely- cut sporangium (b) showing the characteristic stele of Hornea. x 14.
(Slide No. 2455.)

Fig. 62. Oblique section of the base of a sporangium showing the continuity of the tissue of the
columella with the phloem of the stele, tv., wall ; sp., spores ; col., columella ; ph., phloem ; x., xylem.
x 60. (Slide No. 2456.)

Fig. 63. Transverse section of a large partially divided sporangium with a single cavity, but the columella

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 627

showing two lobes, one of which is again subdividing, to., wall ; tap , tapetum ; col. a ; col. b ; col. b', the
three divisions of the columella. x 20. (Slide No. 2457.)

Fig. 64. Sporangium in transverse section lying beside a longitudinal section of a stem, to., wall; sp.,
spores; col., columella. x 20. (Slide No. 2457.)

Plate X.

Hornea Lignieri, Kidston and Lang.

Fig. 65. Portion of transverse section of a sporangium of Hornea. ep., epidermis ; m.l., middle layers
of wall; tap., tapetum; sp., spores; col., columella. x 60. (Slide No. 2458.)

Fig. 66. Transverse section of a sporangium showing the thickened epidermis, ep., epidermis ; m.l.,
middle layers of wall; tap., tapetum; col., columella. x 20. (Slide No. 2456.)

Fig. 67. Portion of wall of sporangium in fig. 66 more highly magnified, ep., epidermis with thickened
outer walls; m.l., perished middle layers of wall; tap., tapetum. x 60. (Slide No. 2456.)

Fig. 68. Fragment of the epidermis of a broken sporangium showing the thickening of the outer walls
extending on to the lateral walls. x 60. (Slide No. 2439.)

Fig. 69. Portion of one of the sporangia in fig. 58 showing the columella in longitudinal section, w.,
wall ; sp., spores in the dome-shaped cavity lined by the tapetum (tap.). x 60. (Slide No. 2454.)

Fig. 70. Portion of a tangential section of a sporangium to show the persistent tapetum. w., wall ;
tap., tapetum. x 60. (Slide 2451.)

Fig. 71. Spores in tetrads around the columella (col.). x 105. (Slide No. 2459.)

Fig. 72. Mature spores in sporangium, col., columella. x 160. (Slide No. 2456 )

We again gratefully acknowledge our indebtedness to the Executive Committee of the Carnegie Trust
for a grant to defray the expense of the plates illustrating this memoir.

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 24)

Trans. Roy: Soc. Edin^

KidstonakdLang.

Plate I. Vol. Lll.

*111111

Zinco-Colloty p« Co? Edinburgh.

R.Kidston ,Photc

Rhynia Gwynne -Vaughan i, K.&L

K i d sto n and Lang.

Plate II. VoI.LIl.

Trans. Poy: Soc.

Edinr

Zinco-Collotype Co, Edinburgh.

R.Kids>toi\ .Phot:-

Rhynia Major, K&L.

Trans. Roy; Soc. Edmr

Kl D ST O N and L AN G .

Plate III.

Vol.LIl.

TKidston, Photo.

26.

Zinco-Colloty pe Co, Edinburgh.

i.c.

24.

Rhynia Major, K&L.

Trans. Rov: Soc. Edinr

Kl D 5TON ante) Lan g

Plate IV Vol.Lll.

Hornea Lignieri, K&L.

Plate V Yol. Lll

Kidston and Lang

Edmr

Trans. Poy; Soc

Co, Edinburgh.

Zinco-Colioty p»

Hohnea Ligniebi, K &L

Trans. Roy: So c. Edinr

Ki D STO N and L AN G

Plate VI. Vo I. LI1

■R.Kidston , Photo.

Zinco-Coliotype Co, Edinburgh.

Hornea Lignieri, K &I

Plate VII. Vol.LIl.

Trans. Roy: Soc. fdinr

Kidston and Lang.

Hornea Lignijeri, K &L.

Trans. Roy; Soc. Edin?*

Kidstona^dLang.

Plate VII I. Yol.LIl

55.

P.Kidston .Photo.

-Collotype Coj Edin burgh.

Hornea Lignieri, K&L

Trans. Koy; Soc. Edin1' Kidstonanx>Lang.

Plate IX. Yol.LIL

ThKidstcm, Photo.

63.

Zi nCo-C o Holy pe Co, Edinburgh

tap.

:ol.a.

Hornea Lignieri, K.&L.

Trails. Roy; Soc. Fdinr

Kidston and Lang.

Plate X. VoI.Lll.

67

PKidsfon ,PKoto.

Zinco-Colloty pe Co, Edinburgh.

Hornea Lignieri, K &L.

( 629 )

XXV. — Theoretical Determination of the Longitudinal Seiches of Lake Geneva.
By A. T. Doodson, R. M. Carey, and R. Baldwin, Tidal Institute, University
of Liverpool. Communicated by Dr E. M. Wedderburn.

(MS. received November 11, 1919. Read January 12, 1920. Issued separately April 2, 1920.)

1. The theoretical determination of the longitudinal seiches of a lake was reduced
by Professor Chrystal # to the finding of those solutions of the differential equation

*Y+A_v-o

dor? p(x)

(1)

which vanish at x = 0 and x = a.

Here x denotes the area of the surface of the lake from one end up to any trans-
verse section, and ranges from 0 to a, a being the total area of the surface of the
lake ; p(x) denotes the product of the area of the transverse section at x and its
breadth at the free surface ; ^ = <r2/g, & being the “ speed ” of the periodic motion and
g the acceleration due to gravity ; V denotes the total volume of water which has
passed the section up to the time t.

If £ denotes the elevation of the free surface at the transverse section corre-
sponding to x, we have

£= -

3V

dx

(2)

An inverted graph of the function p(x) was called by Chrystal the “ normal
curve” of the lake, and it can be constructed from the results of a bathymetrical
survey. By approximating to the normal curve by geometrically simple curves,
Chrystal showed how to find the required solutions of (l), and he and E. M.
Wedderburn applied the method to Lochs Earn and Treig.f The normal curves of
these lakes are fairly regular, and results were obtained closely in agreement with
observation.

In a recent paper by J. Proudman, \ a general solution of the problem has been
given which does not involve approximations similar to those of Chrystal, and the
object of the present paper is to give details and results of the application of this
solution to Lake Geneva. The normal curve of Lake Geneva had been constructed
some years ago by Wedderburn ; he suggested that the present method should first
be applied to this lake, and very kindly supplied his data and calculations. A glance
at the diagrams illustrating the normal curve will show how complex the curve is ;
the approximate methods introduced by Chrystal were not easily or satisfactorily

• “ Hydrodynamical Theory of Seiches,” Trans. Roy. Soc. Edin., vol. xli, p. 599.

f “Calculation of the Periods and Nodes ot' Lochs Earn and Treig,” ibid., p. 823.

\ “ Free ami Forced Longitudinal Tidal Motion in a Lake,” Proc. Lond. Math. Soc., series 2, vol. xiv, p. 240.

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 25). 97

630 MESSRS A. T. DOODSON, R. M. CAREY, AND R. BALDWIN, THEORETICAL

applied. A noteworthy feature of Proudman’s method is that it can take account
of all the natural irregularities of the body of water, and its application has been
found to give very satisfactory results. In §§ 9-11 will be found a comparison
between observations and the results of the theoretical determination of the longi-
tudinal seiches of Geneva ; better agreement could hardly be expected.

General Account of the Method.
2. Mathematical Basis.

The solution we are using involves the function

where

and

for rC> 0.

R(4

A) = 2(-A)h*G> v)

I0(£ v) = y-£

u(,V)-r r

J = s = ( p(s)

= r r

J S'=( J s=s' p{s)

dsds'

-dsds

(3)

(4)

(5)

(6)

The functions In(£, y) depend only on the transverse sections corresponding to
x = £, x = y, and the variable x may be measured from either end of the lake, provided
that £ always refers to that section which is the nearer to that end ; that is, £ is less
than y. This is obviously true for n = Q, and its general truth can then be readily
seen from (5) and (6).

The conditions of the problem are such that the solution is given by

R(0, a, A) = 0 ..... (7)

A value of X- which satisfies the above equation will then determine the period of
a free oscillation; if possible values of X are denoted by X1; X2) . . . . . ., then

the periods are given by

T = 2ir/(g\s)i (8)

and the corresponding mode of motion is given by

V = R(0, x, \s) . . . . (9)

on omitting a factor which is a simple harmonic function of the time. W e shall then
have

£= • . • • • (10)

rn(o, x) = ^-i„(o, x).

where

DETERMINATION OF THE LONGITUDINAL SEICHES OE LAKE GENEVA. 631

Thus the problem resolves itself into finding the functions Ira(0, x), I'ra(0, x), which
are determined from (5) as

l'*(0, 4- (11)

Jo p(x)

l»(0,x)= [\(0,s)ds (12)

Jo

with the initial form

l0(0,x) = x ...... (13)

3. Arrangements for Computation.

It is clear that if p(a) is zero the integrands in (5) and (6) become infinite
for s = a. These infinities, however, can be avoided in a very elegant manner as
follows.

Equation (35) of the paper by J. Proudman, already quoted, gives the relation

R(x,v,\)^,x,X)-n(^,x,\)l-R(x,v,X) = R^,v,X) . . . (14)

which is a direct consequence of the fact that R(£, x, X) and R(a:, n, are two
solutions of the fundamental differential equation. If these functions be expressed
as power series in X by (3), and the coefficients of equated, we have

2, | b(f, *)“ L(£ x)^fn-r(x, ??)] = hz(£ v) ■ ■ (15)

Hence, if the functions I„(0, x) and L(x, a) are known, together with their first
differential coefficients with respect to x, the value of I„(0, a) can be at once
determined, and no infinities occur in any of the integrands.

We have already pointed out that the variable x may be measured from either
end of the lake, but we have assumed that £ is less than »?. It is now convenient to
split the lake into two portions at some arbitrary point x = X. Let measurements
from the west end of the lake yield the functions ln(0, X) and I'n(0, X), and let
measurements from the east end yield the functions In(X, a), I,n(X, a). Wherever
necessary, the system will be denoted by w or e to indicate the manner of measuring
the variable, so that we have

I»(£’7) = I n(a-v>a~£) ■ (16)

and, if x = a— x,

w e e

In(x, rj) = In(a- r], a- x) = ln(a — 7j, x) .... (17)

0 w 0 e 0 e

^fn(x>v) ~ — V’ a ~ *)' ~ — dx' L(a — V> x ) • • • (1®)

In particular, our problem requires the functions for which £=0 and *1 = a. We
shall choose X, the arbitrary value of x, to be such that the corresponding section
divides the lake into two equal surface areas ; that is, X = \a. Let dashes denote
differentiation with respect, to x or to x\ according to the system ; then we shall

632 MESSRS A. T. DOODSON, R. M. CAREY, AND R. BALDWIN, THEORETICAL

require In(0, x ) and I'M(0, x ) in both systems for values of x from 0 to \a , and equation
(15) yields

^o(InJ'r + L'n Jr) = K(0,«) .... (19)

where, on the left-hand side, the arguments are (0, \ a ).

4. Determination of Periods and Elevations.

The unit of length will be chosen so that the surface of the lake has twTo units of
area, and this unit of length will be called a “ lake unit.” Thus we have a = 2, and
all the functions on the left-hand side of (19) will have the arguments (0, l).

It is therefore necessary to evaluate the functions I„(0, x), I'n(0, x) from * = 0
to x= 1 in both systems, and substitution in (19) will give the values of I„(0, 2) for
the whole lake. These determine the value of A for which R(0, 2, A) is zero, or, by
(3), for which

J)(-A)»Ire(0,2) = <> (20)

The elevation £ is then obtained as in equation (10) ; use is made of the values
of r„(0, x) in both systems, and the two sets of values are combined as follows.
Reference to (14) shows that when R(0, 2, As) is zero we have

R(#, 2, As)^-R(0, x, As) = RfO, x, Xs)^-R(x, 2, A) . . . . (21)

ox . ox

whence

R(0, x, \S)/R(x, 2, As) is independent of x . . . . (22)

Consequently

-^-R(0, x, As)/— R(.r, 2, As) is independent of x . . . (23)

OX OX

and the values of the elevation in the w-system are a constant multiple of those in
the e-system. The two systems can thus be fitted together at the point x=l.
The relation (21) also affords a verification of the value of Ag obtained from (20).

Application to Lake Geneva.

5. Construction of Normal Curve.

The chart used by Wedderburn was that of Hornlimann and Delebecque,
Atlas des lacs frangais, with a horizontal scale of 1/50,000, and with the depths
stated in metres. He had drawn, across it thirty-one lines as the surface traces
of transverse sections. The lengths of these lines were measured, and the area of
the corresponding transverse section calculated. The product of these two quantities
gave the corresponding value of p(x). The area of the lake from one end up to a
transverse section was measured by means of a planimeter, thus giving x, and the
value of a lake unit was found to be u= 17052 x 106 centimetres. The original data
are given in Table I, and from them were deduced Tables II and III, which give
the values of p(x) at intervals of 0'01 in x. The normal curve for Geneva is
illustrated in fig. 1, and fig. 2 shows on a larger scale the western portion of the

DETERMINATION OF THE LONGITUDINAL SEICHES OF LAKE GENEVA. 633

curve. It will be seen that a large number of peaks exist, and it was necessary,
therefore, to provide a very smooth curve in order to avoid excessive numerical
errors due to the use of integration formulae. This was obtained by plotting from
the original data, and the nominal accuracy of the interpolated values was

increased by redrawing on a greater scale. Great care was taken in smoothing the
curve. For the small values of x , in both systems, the graph of x/p(x ) was used in
this smoothing process.

6. Computation of I„(0, x) and Vn(0,x).

Let w(x) and e{x) denote the reciprocal of 10 ~15u3p(x) in the w-system and
e-system respectively ; these were calculated at intervals of 0 ‘01 in x. We now have
to apply formula (ll) and (12). The integration was performed by means of the
well-known Simpson formula

3 jo ydx = %0 + 4yh + yi7l] .... (24)

where A = 0'01 in the present instance.

Successive application of this formula gives the values of three times the integral

for x = 0, 2 h, 4 h , and, by a second series of calculations, for x = h, 3 h, 5 h,

The odd and even series are thus calculated independently, and a check on the end
values is practically sufficient ; this check was performed by means of the inter-
polation formula

1 6 y3h = - y0 + 9 yA + 9 yth - yrh . . . . (25)

Thus any term in the odd series can be calculated by interpolation from the even
series and compared with the value obtained from the Simpson integration.

In the actual calculations powers of three were retained and powers of ten
ignored until all the integrations were completed.

Since l0(0,cc)=:r, the first procedure was to evaluate xw(x), and the accuracy of

634 MESSRS A. T. DOODSON, R. M. CAREY, AND R. BALDWIN, THEORETICAL

the multiplications was checked by differencing. Then the Simpson integration
was performed, giving 3l,1(0, x), the work being checked at three points by
the interpolation formula. A second integration gave 32I1(0, x), and was checked
as before.

The next procedure was to multiply 32I;i_(0, x) by w(x), and this was checked
by differencing ; errors might have arisen either in the multiplication or in the
copying of 32I1(0, x) from the calculating machine. This differencing was essential,
as the interpolation formula only checks the summation process arid does not
check clerical errors.

Successive application of the integration yielded ultimately 32wI„(0, x) and
32n_ir„(0, x), each multiplied by powers of 10.

Similar processes were carried out in the e-system.

The values of these functions for x = 1 in both systems were then extracted,
and since in all cases interpolation showed slight differences between the odd and
even series, the final values were corrected so as to distribute this error equally
between the two series. The values so obtained can be written as k-,T„(0, l) and
K~nI'n(0, l). The factor 10 15tT3 in w(x) has also to be accounted for, and also the
factor 32 in the integration process. The values obtained for I'n(0, l) should be
multiplied by 3 if they are to be written in the form 0, l). Allowing for

powers of 10 introduced in the integration formula, and also for any powers of
10 introduced to avoid an undue number of decimal places in the integration
processes, we have ultimately

/c=9~x x 10-16 x w3 = 55-096 .... (26)

The corresponding values of K_nIm(0, x) and *"nI'n(0, x) at intervals of 0T in x
are given in Tables IV to VII for the western and eastern systems.

7. Extrapolation Methods for the Larger Values of n.

It was found, when the period equation (20) was formed, that too few values
of In(0, l) had been calculated, and as the extra terms required were of no great
importance, extrapolation was used so as to avoid further integrations. The most
useful procedure was to construct the ratio of «:"”In(0, l) to ic 1}I}1_ x(0, l) for suc-
cessive values of n. When the ratios were differenced, extrapolation was easily
carried out to the required degree of accuracy, and the resulting values of «~n I„(0, l)
are indicated by square brackets in the table of the next paragraph. In addition,
this method of differencing provided a very useful check on the accuracy of the
values of k“5T„(0, l).

8. Computation of k_?T„(0, 2).

The evaluation of the components «-_nI„(0, l) and k_7T„(0, l) in both systems
having been satisfactorily performed, they were arranged as in the table below.

DETERMINATION OF THE LONGITUDINAL SEICHES OF LAKE GENEVA. 635

The extreme columns give the values of <~nln and k~wI„, and are written with the
values of n increasing up the column. The central portion of the table is supposed
to be written on a separate piece of paper, so that it can slide up and down between

the extreme columns. The values of k ~nl'n are written, with n increasing down the

w

column, on the left-hand side of this central piece and in proximity to * ~nln ;

w e

similarly K~nl'n is written in proximity to ^nln- The middle column contains the
values of /c_rTn(0, 2) as calculated from equation (2l), and the calculation is easily
accomplished. Suppose the value of k‘6I6(0, 2) be required ; then the central piece
is moved down until the value n = 6 on it is opposite to n = 0 on the fixed sheet ;
the terms of the various products of (19) are then contiguous, and the whole process
can be rapidly performed on a calculating machine. The table given shows the
arrangement for n= 6, and the central movable piece is enclosed by double lines.

w

K-nln.

n.

n.

K-»rn.

K-nL(0, 2).

K-n’,,.

n.

n.

l

[-00000004

13

13

[•00000072"

12

12

[-00001065'

11

11

| -00013349

10

10

■00140154

9

9

•012096

8

8

[-000000

00660]

•083874

7

7

[-000000

2578]

•45376

6

0

1-00000

2-0000

1 0000

0

6

■000007

9728

1-8445

5

1

1-84296

18-5144

9-4035

i

5

■000186

982

5-3519

4

2

■82141

55-5745

22-9944

2

4

•003154

1

10-3037

3

3

•163192

84-9787

25-8675

3

3

•035646

11-7405

2

4

•018270

79-3360

16-5171

4

2

•24352

6-4240

1

5

•0013098

50-0488

6-7554

5

1

■84389

1 0000

0

6

■00006522

22-7913

1-9182

6

0

1-00000

7

[-000002387]

7-8521

•40085

7

8

[-000000067]

2-1201

•064320

8

9

•46111

•008187

9

10

•08258

•0008485]

10

11

•012399

•00007313

11

12

•0015844

"00000533’

12

13

•00017452

’•00000034’

13

9. Determination of the Periods.

The solution of the period equation (20) is best effected by the use of Horner’s
method. Let the notation Jn = /c_”I71(0, 2) be used ; then we have to solve the
equation

J0-(kA)Ji + OcA)2J2- . . . +(-K\y\Jr = 0,

where Jr is the last term considered. The accuracy of the solution will be affected

636 MESSRS A. T. DOODSON, R. M. CAREY, AND R. BALDWIN, THEORETICAL

by the omission of the subsequent terms, but the relative importance of the various
terms, and the accuracy of the solution, can be tested afterwards. Since the most
important terms are those for which n is small, it is best to deal with the series

where

+ . . . + (-l)'J, = 0 . . ' (27)

j*=l/(*A) O28)

The result of Horner’s method is to give the first three roots as follows

^ = 5-2764, k\= -18952 \

^2 = 1-1725, k\2= -8529 l (29)

-7459, k\3 = 1-3407 1

We have now to estimate the accuracy of these results. In the first place, the
accuracy of the work, regarding J0, Ji, . . . as absolutely accurate, may be tested
by substituting the above values of kX on the right-hand side of the equation
(/c\)J1 = J0 + (kA.)2J2-(kX)3J3 +

whence division by J\ should give the value of ; the effect of terms beyond J,. is
easily estimated by this process, and it was shown that the values of k\ as given above
were not affected, to the order given, by taking terms beyond J13 into account. But
the J’s are only known to a limited degree of accuracy, and this is not great enough
for the third decimal of to be considered as more than approximately correct, and
the second decimal of >c\3 is only an approximation. The value of k\ 1} however, is
probably correct to at least four decimals.

The actual periods are given by the relation

T = 2tt /(g\,y, ..... (30)

where g, expressed in lake units, has the value 9807 divided by 17052 x 106.
Introducing the factor k once for all, we have

T = 2TTKiff~h/isi =--- 32-417/^s minutes .... (31)

Therefore, the values of the first three periods are respectively

Tx = 74'45 minutes, -i

T2 = 35T „ i . (32)

T3 = 28 „ J

and all the figures here given may be regarded as significant.

10. Determination of the Elevation £ and of the Nodes.

The elevation £ may be calculated by the formula

£ = - j^(- *),

and this function is given in Table VIII at intervals of 07 in x. The value of x is
measured from the western end. Application of the method discussed in § 4 was

DETERMINATION OF THE LONGITUDINAL SEICHES OF LAKE GENEVA. 637

made, and the values obtained from the eastern system were multiplied by appro-
priate constants, these being the ratios of the values of £ at5c=l in the two
systems.

The calculation of the nodes was accomplished by taking a few intermediate
values of I'n(0, x), and the positions of the zero values of £ obtained by interpolation
from them. The following results were obtained : —

Uninodal oscillations : node at x= '377.

Binodal oscillations : western node at x = 'll 3.

eastern node at X — 1'067.

By interpolation in the table of latitude and longitude of the mid-points of the
transversals (see Table I), we get the nodes defined by the transversals whose mid-
points have the following latitude and longitude : —

Uninodal oscillations : lat. = 46° 23,-9, long. =4° 0,-2 E. of Paris.

Binodal oscillations : lat. = 46° 17,-9, long. = 3° 52,-4 E. of Paris.

lat. = 46° 27'*3, long. = 4° 12'-8 E. of Paris.

These sufficiently define the nodes, and reference to any map of Geneva will at once
determine the exact positions.

11. Comparison of Results with Observations.

The zero of the chart used in the present calculations is given as “ lm772 au
dessous de repere de la pierre a Niton de Geneve.”

Most of the observations on Lake Geneva are recorded by Forel in Le Leman.
His zero plane is taken as “ un plan passant a 3'0 m. en contrebas du repere de la
Pierre du Niton de Geneve” (vol. i, p. 455). His zero plane is therefore 1'228 m.
below the zero plane of the chart, and as most of his observations ostensibly refer
to this mean level of the lake, an exact comparison between calculations and obser-
vations cannot be made. Forel’s mean value for the period of a uninodal oscillation
is 73‘5 minutes ; our theoretical determination gives 74'45 minutes. On p. 121 of
vol. ii of Le Leman is given a set of observations when the lake was 30 cm.
above his zero level ; on this occasion the period was found to be 74 ‘0 minutes.
Forel attempts to correct for the height of the lake, but his method is not very
satisfactory ; indeed, it would be difficult to find an easy way of reducing observa-
tions to standard height. The agreement between calculation and observation is,
therefore, as good as can be expected when the conditions of observation are not
identical with those of calculation.

The value of the period of the binodal oscillations, according to Forel, is
35 '5 minutes, but the mean level of the lake on the occasions of observation is not
given. Our calculations give 35 7 minutes. No comparison is possible for trinodal
oscillations.

The positions of the nodes are not easily determined experimentally, and Forel

TRANS. ROY. SOC. EDIN., YOL. LII, PART III (NO. 25).' 98

638 . MESSRS A. T. DOODSON, R. M. CAREY, AND R. BALDWIN, THEORETICAL

was only able to place them approximately. He places the node of the uninodal
oscillations somewhere to the west of the line Rolle-Tbonon ; but he also records
that no uninodal oscillation was observed at Sechex or at Fleur D’Eau. Our
calculations place the nodal line about 2 km. west of the latter line.

The western node of the binodal oscillations was not observed by Forel, but he
places the eastern node as being approximately on the transversal through Morges ;
the calculated value for this node gives it as the transversal about 2 km. east of the
transversal through Morges.

When it is considered that the exact determination of nodes by observation is very
difficult, and when allowance is made for differences in mean lake level, the results
that have been obtained by calculation may be regarded as extremely good, and they
indicate that the method gives a most satisfactory solution of the problem of free
longitudinal motion in a lake.

12. Remarks on Methods of Calculation.

The processes of successive integration require a high degree of nominal accuracy
throughout the work of calculation ; otherwise the values that are obtained for
In(0, 2) will diverge rapidly from the true values as n increases. Whatever be
the degree of accuracy of p(x), it is essential that there should be as exact a corre-
spondence as possible between the adopted function and the calculations based on it.
The difference between theoretical results and observations will then not be due to the
method of calculation. The experience acquired in the present instance shows that
when there are considerable irregularities in p(x) a great many values should be
used in the numerical representation of the function. For values of x between
0 and 0*3, measured from the west end of the lake, as usual, it would have led to
better results if considerably smaller intervals in x had been chosen. Values of
p(x) at intervals of 0*002 in this region would have led to a large increase in

accuracy. As it is, the values of I„(0, l) are certainly correct to 1 in 10,000,

but this degree of accuracy is not sufficient for the higher roots of the
period equation. For the first few values of w a much greater degree of accuracy
is desirable.

There is nothing in the theoretical basis for the partition of the lake into western
and eastern systems (see § 4) that requires the partition at x= 1. It can be at any
suitable value of x, and in the present case the division at a; = 0'3 would have been
better, as the value of p(x) is most irregular between £C = 0 and x = 0'3, the remain-
ing portion being a smooth curve.

The method of solution of the problem, as expounded in §§ 2—4, is due to Professor
J. Proudman, and this paper is simply an application of his method. The calcula-
tions were begun by R. Baldwin, and a graphical method of integration was used :

his results indicated the need for a much greater degree of accuracy. The work

DETERMINATION OF THE LONGITUDINAL SEICHES OF LAKE GENEVA. 639

was then undertaken by A. T. Doodson, who prepared the table of p(x) and initiated
the method of calculation as described in the paper ; the integrations were carried
out chiefly by R. M. Carey, and the rest of the work, together with the exposition
as given here, was performed by A. T. Doodson.

Throughout, the authors have experienced much assistance from Professor
Proudman, and they desire to express their thanks to him. They also desire to
record their thanks to Dr E. M. Wedderbtjrn for the data provided by him, and
for the help and suggestions received from him.

Table I.

Lake Geneva Measurements, u = 1‘7052 x 106.

Line.

South End.

North End.

Mid-point.

ub(x)

u2A(x)

u3p(x)
1015 '

x.

Lat.

Long.

Lat.

Long.

Lat.

Long.

5 x 105'

5 x 109’

1

46° 12-9

3°50-4

46° 13*3

3° 49-1

46° 13-1

3°49-7

•370

•0197

•0183

•0071

2

13-9

513

14-5

49-0

14-2

50-1

•617

•1493

•230

■0273

3

15-2

51 -5

157

49-5

15-5

50-5

•546

•1742

•238

•0507

4

15 6

51-6

16-3

50-1

16-0

50-8

•478

•1455

T74

•0583

5

16-1

52-6

17-2

50-0

16-6

51-3

•776

•2224

•432

■0730

6

16-4

52-9

18-0

50-3

17 2

51-6

•885

•3101

•686

•0870

7

17-3

53-9

18-8

51 2

18-0

525

•895

•4072

■911

•1175

8

18-2

54-4

19-2

51-7

18-7

53-0

•777

•3527

•685

•1366

9

19-5

55-3

20-4

52-3

20-0

53-8

■835

•3923

•819

•1752

10

20-7

56-5

21-8

53-2

21-2

548

•931

. -3401

•792

•2189

11

21-5

57-1

23-2

548

223

53-0

•899

•4219

•948

■2554

12

22T

58-5

23-8

565

230

57-5

•791

■3422

•677

•2875

13

22-3

4° 0'0

25-3

57-3

23'8

58-7

1-310

•5750

1-883

•3235

14

22-2

0-8

25-8

58 2

24-0

59-5

1-492

•8211

3-065

•3517

15

20-5

3 0

26-8

593

237

4° l'-l

2-520

1-6927

10-668

•4198

16

21-8

5-8

27-9

4° 2'*6

24-8

4-2

2-456

2-8707

17-633

■6081

17

226

8-4

28-1

5-0

25-3

6-7

2-228

3-4481

19-207

•7374

18

24-3

10-2

28-7

7-6

26-5

8-9

1-800

3-3114

14-901

•8601

19

23-7

12-7

31-0

10-3

27-3

115

2-768

4-9460

34-225

•9968

20

23-9

| 13-8

30-5

12-6

27-2

13-2

2-464

5-2897

32-58

1-0903

21

24-2

16-2

31-1

15-3

27*6

15-8

2-596

5-5438

35-98

1-2348

22

24-4

19-3

303

19-9

27-3

19-6

2-210

5-5256

30-53

1-4360

23

24-7

235

29-4

24T

27-1

238

1-830

4-3035

19-69

1-6187

24

24-1

25-6

28-5

27-4

26-3

26-5

1-695

3-6954

15-66

1-7317

25

23-7

28-2

27-5

30'5

25-6

29-3

1-539

2-7796

10-70

1-8353

26

23-3

29-4

27-2

31-2

25-2

303

1-505

2 2963

8-64

1 8848

27

23-3

30-6

26'9

31-9

25T

31-2

1-386

1-4072

4-88

1-9204

28

23-6

31-9

26-5

331

25-0

32-5

1-104

•9257

2-555

1-9457

29

23-8

32-9

25-8

34-5

24-8

337

•860

•5643

1-212

1-9603

30

23-8

34-2

24-8

35-7

24-3

35-0

•522

■2134

•278

1-9948

31

23-7

35-2

23-7

35-2

23-7

35-2

•000

•0000

•000

2-0000

The latitude and longitude (east of Paris) of the terminal points and of the mid-points of the
transversals drawn by 'Wedderburn on the map of Geneva are given above. The value of b(x) is the
breadth of the lakes, A(.r) is the area of the transverse section, and p(x) is the product ; these quantities
are expressed in lake units.

640 MESSRS A. T. DOODSON, R. M. CAREY, AND R. BALDWIN, THEORETICAL

Table II.

Values of p(x) for a?=0 to x=\, Western System.

,

u3p(x)

1016

X.

Isjo

X.

u3p{x)

To15 '

X.

1

uzV(x)

101*

X.

v?p(x)

1015

■01

•075

•21

•800

•41

9-60

•61

17-70

•81

16-98

•02

•148

•22

•792

•42

10-65

•62

17-90

•82

16-47

03

•208

■23

•820

•43

11-38

■63

18-08

•83

15-97

■04

•230

•24

•875

•44

12-00

•64

18-25

■84

15-50

■05

■205

•25

•945

•45

12-55

■65

18-42

•85

15-10

•06

•243

•26

■945

46

13-06

•66

18-59

■86

14-90

•07

•380

•27

•850

•47

13-50

•67

18-74

■87

1510

■08

•562

•28

•708

•48

13-91

•68

18-87

•88

15-S8

•09

•730

•29

•680

•49

14-29

•69

18-98

•89

17-50

•10

•860

•30

•750

•50

14-65

•70

19-08

•90

19-65

■11

•918

•31

1-000

•51

15-00

•71

19-15

•91

21-80

T2

•895

•32

1-575

•52

1534

•72

19-20

■92

23-85

•13

•770

•33

2T50

•53

15-67

•73

19-22

•93

25-78

•14

■662

•34

2-550

•54

15-97

•74

19-20

•94

27'55

•15

•648

•35

2-975

•55

16'25

•75

19T2

•95

29-18

•16

•690

•36

3-500

■56

16-52

•76

18-98

■96

30 63

T7

•795

•37

4-250

■57

16-78

■77

18-73

•97

31-88

•18

•878

•38

5-300

■58

17-02

•78

18-38

•98

32-90

•19

■880

•39

6-750

•59

17-25

•79

17-95

•99

33-70

•20

•845

•40

8-250

•60

17-48

•80

17-48

LOO

34 25

Table III.

Values of p(x) for x = 0 to x = 1 , Eastern System.

(The value of x iu the western system is the excess of 2 over the eastern value of x.)

X.

u3p(x)

1015

X.

u3p(x)

To^~ ‘

X.

u3p(x) 1
1015

X.

u3p(x)

"10“' ‘

X.

u3p(x)

1Q15

•01

•42

•21

12-92

•41

20-88

•61

30-42

•81

35-50

•02

•87

•22

13-44

•42

21-29

•62

30-93

■82

35-22

03

L35

•23

13-97

■43

21-72

•63

31-43

•83

34-88

•04

1-85

•24

14-47

•44

22T5

•64

31-91

■84

34-51

•05

2-39

•25

14-93

•45

22-58

■65

32-38

•85

34-12

•06

3-05

■26

15-36

•46

23-02

■66

32-84

•86

3373

•07

3-85

•27

1575

•47

23-47

■67

33-29

•87

33-38

•08

4-82

•28

16T0

•48

23-93

•68

33 72

■88

33 05

.09

5-98

•29

16-45

•49

24-40

■69

34T2

•89

32-80

•10

7-24

■30

16-80

■50

24-87

•70

34-48

•90

32-65

•11

8-28

•31

17T5

•51

25 35

■71

34-82

•91

32-60

T2

8-89

•32

17-49

•52

25-83

•72

35T2

•92

32-67

•13

9-35

•33

17-83

•53

26-32

•73

35-39

•93

32-87

T4

9-73

•34

18T8

•54

26-82

•74

35-61

•94

3317

•15

10-13

•35

18-55

•55

27-33

■75

35-79

•95

33-57

T6

10-54

•36

18-93

•56

27-84

1 '76

35-90

■96

34-02

•17

1097

•37

19-31

■57

2836

*•77

35-95

■97

34 42

•18

11-43

•38

19-69

•58

28-88

•78

35-94

•98

34-59

T9

11-91

•39

20-08

■59

29 40 1

•79

35-87

•99

34-55

■20

12-41

•40

20-48

■60

29-91

•80

35-72

1-00

34-25

DETERMINATION OF THE LONGITUDINAL SEICHES OF LAKE GENEVA. 641

Table IV.

Values of * nIn( 0, x), Western System.

X.

n— 1.

2.

3.

4.

5.

6.

7.

8.

9.

o-i

•0734

•0174

■0020

•0001

0-2

•2990

•1460

•0356

•0052

■0005

•00004

•000002

0-3

•7454

■6219

•2582

•0642

■0107

■00128

•000116

■000008

•0000005

0-4

1-4316

1-6690

•9475

•3162

•0696

01090

•001281

•000118

•0000087

05

2-1795

2-9210

1-8636

•6919

•1686

02915

•003778

•000382

•0000311

0-6

2-9586

43083

2-9605

1-1831

•3106

05791

•008101

•000883

•0000776

0-7

3-7688

5-8484

4-2802

1-8276

•5140

•10287

•015459

•001813

•0001712

0-8

4-6123

7-5683

5-8794

2-6773

•8054

,-17267

■027815

•003499

•0003540

0-9

5-4978

9-5307

7-8780

3-8387

1-2395

•28572

•049522

•006703

•0007298

10

6-4240

11-7405

10-3037

5-3519

1-8445

•45376

•083874

■012096

•0014015

Table V.

Values of K~nI'n( 0, x), Western System.

X.

n = l.

2.

3.

,

5.

6.

7.

8.

9.

0-1

1-4870

•5055

•0764

•0066

•0004

0-2

3-2041

2-4616

•8140

•1497

•0176

•0015

■00010

0 3

5-9889

8-0730

4-6600

1-4880

•3019

•0424

•00441

•000356

•000023

0-4

7-3034

11-8457

8-3358

3'2639

•8183

•1431

■01854

•001856

•000149

0-5

7-6381

13-1760

10-0068

4-2804

1-1788

■2273

•03250

•003592

•000317

0 6

7-9438

14-5998

12-003.0

5-6065

1-6910

■3573

•05592

•006759

•000651

0-7

8-2620

16-2439

14-4794

7-3659

2-4191

•5556

•09435

•012349

•001286

0-8

8-6203

18-2498

17-6866

9-7886

3-4900

■8679

•15926

•022486

•002522

0-9

9-0967

21-0726

22-4506

136091

5-2917

1-4303 j

■28456

•043470

■005269

1-0

9-4035

22-9944

25-8675

16-5171

6-7554

1-9182

•40085

■064320

•008187

Table VI.

Values of k nIn( 0, x), Eastern System.

X.

n= 1.

2.

3.

4.

5.

6.

o-i

•00979

•00032

•000005

•0000001

0-2

•03415

•00195

■000056

0000010

•00000001

0-3

•07284

•00605

•000254

•0000064

■00000010

•0000000010

0-4

T2756

01407

•000785

0000264

•00000059

■0000000088

0-5

T9979

02769

•001942

■0000821

•00000231

•0000000452

0-6

■29006

•04854

•004108

•0002096

• -00000713

•0000001714

0-7

•39841

07812

•007743

•0004626

•00001846

•0000005231

0-8

•51516

•11819

•013446

•0009223

•00004228

•0000013814

0-9

•67238

•17175

•022206

•0017334

•00009053

•0000033780

DO

•84389

•24352

•035646

•0031541

•00018698

•0000079277

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 25). 99

642 DETERMINATION OF THE LONGITUDINAL SEICHES OF LAKE GENEVA.

Table VII.

Values of * nl'n( 0, sc), Eastern System.

X.

n= 1.

2.

3.

4.

5.

6.

o-i

•17988

•00842

•000178

•000002

0-2

•31238

•02649

•001015

•000022

•0000003

0-3

•46390

•05789

•003242

•000102

•0000020

•00000003

0-4

•63316

•10544

•007874

•000332

0000089

■00000016

0-5

•87219

16974

•015912

•000842

•0000286

•00000066

0-6

•99317 •

•24961

•028165

001795

•0000735

•00000208

07

M7414

•34481

•045509

•003396

•0001627

•00000540

0-8

1-36394

•46121

•070083

•006022

■0003321

■00001272

0-9

1-58847

•61907

•108107

•010662

■0006741

•00002960

1-0

1-84296

•82141

•163192

•018270

0013098

•00006522

1

Table VIII.

Values of Elevation.

(These values have to be multiplied by a simple harmonic
function of the time.)

X.

Uninodal oscillations.

Binodal oscillations.

00

-1-0000

. -1-000

0-05

- -8732

- -490

0-1

- -7358

- -055

0-2

- -4758

+ -375

0-3

- 1251

+ -460

0-4

+ -0114

+ -377

0-5

+ -0372

+ -328

0-6

+ -0560

+ -277

0-7

+ 0720

+ -22,2

0-8

+ -0868

+ -158

0-9

+ -1036

+ -077

1-0

+ 1126

+ -028

1-1

+ -1194

- -014

1-2

+ -1255

- -055

1-3

+ -1308

- -093

1-4

+ -1360

- -132

1-5

+ -1393

- '154

1-6

+ ’1464

- '221

1-7

+ '1515

- -268

1-8

+ -1560

- 313

1-9

+ -1601

- -355

2-0

+ -1657

- -416

( 643 )

XXVI. — On Old Red Sandstone Plants showing Structure, from the Rhynie Chert
Bed, Aberdeenshire. Part III. Asteroxylon Mackiei, Kidston and Lang.
By R. Kidston, LL.D., F.R.S., and W. H. Lang, D.Sc., F.R.S., Barker
Professor of Cryptogamic Botany in the University of Manchester. (With
Seventeen Plates.)

(Read December 1, 1919. MS. received December 1, 1919. Issued separately May 11, 1920.)

Introduction.

Asteroxylon Mackiei was a plant of more complicated organisation and larger
size than either Rhynia or Hornea, which have been described from the silicified
peat-bed at Rhynie in the two preceding papers of this series.* The generic name
refers to the stellate outline of the xylem of the stem as seen in cross section,
while the specific name commemorates the original discovery of the plant remains
by Dr Mackie. t

The remains of Asteroxylon, though abundant, are unfortunately fragmentary,
and caution is therefore necessary in mentally reconstructing the plant as it grew.
The descriptions in this paper, though grouped as naturally as possible, are founded
on isolated portions of a plant of considerable size. A large number of sections of
the chert containing Asteroxylon have been examined, and there is reason to believe
that the chief facts have been ascertained. While, however, we have a fuller know-
ledge of Asteroxylon than of most extinct plants, any fortunate block of the chert
may supply data to complete or qualify the account given here. In anticipation
of this, care has been taken to keep the statement of our observations as objective as
possible. The mutual relations of the parts, and the probable habit of the plant, will
be discussed at the close of the descriptive portion of the paper.

Asteroxylon was first seen in a few loose blocks of the chert discovered by Dr
Mackie. It has since been found in specimens taken from the section of the bed
exposed in situ by Mr David Tait, and represented on p. 762 of Part I. Remains of
Asteroxylon have been met with in the upper portion of the bed labelled A" 1 , in bed
A" 2, and in the upper six inches of bed B. The other blocks of the chert containing
Asteroxylon have been found loose.

The most abundant and generally distributed parts are shoots of various sizes,
the stem bearing numerous leaves. The central cylinder of these stems had a solid
xylem, which was star-shaped in cross section, and gave off small leaf-traces from

* Trans. Roy. Soc. Edin., vol. li, p. 761, and vol. lii, p. 603.

f Mackie, W., “ The Rock Series of Craigbeg and Ord Hill, Rhynie, Aberdeenshire,” Trans. Edin. Geol. Soc.,
vol. x, pp. 205-236, pi. xxiii.

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 26).

100

644 DR R. KIDSTON AND PROP. W. H. LANG ON OLD RED SANDSTONE PLANTS

the enlarged ends of the rays. This construction was shown in the single figure
published, without description, in Dr Mackie’s paper.* The structure of the
branched shoot-system was primary throughout, no indication of secondary thicken-
ing having been found even in the largest stems. In addition to the leafy shoots,
leafless cylindrical axes with a simpler type of central cylinder are met with. These
rhizomes constituted the subterranean region of Aster oxylon. Their finer branches
behaved like roots, while the transition between the larger rhizomes and the leafy
shoots can be clearly followed. There is conclusive evidence, both from continuity
and from the histology, that the rhizomes, with their fine root-like branches and the
leafy shoots, were parts of the one plant, Asteroxylon Mackiei , the vegetative organs
of which are thus pretty fully known.'!

In two loose blocks of the chert small axes of a peculiar type have been found
closely associated with the typical remains of the vegetative organs of Asteroxylon.
In one of the blocks there were in addition dehiscent sporangia containing cuticu-
larised spores. We regard these structures as possibly, or probably, constituting the
fertile portion of Asteroxylon. Their description will therefore be given in this
paper before the general discussion of the morphology of the plant, although in the
absence of evidence of continuity or histological identity these interesting remains
must be treated to some extent independently.

The various parts mentioned above will now be described. It will be convenient
to deal in order with the leafless rhizomes, the transition from these to the leafy
shoots, the shoots of various sizes, and lastly with the peculiar axes and the sporangia
associated with them.

Rhizomes.

The leafless axes of various sizes which are met with among the other remains of
Asteroxylon have a broad cortex and a central cylinder consisting of a simple central
strand of xylem surrounded by a wide zone of phloem. While the behaviour,
especially of the finer branches, is very root-like, the axes with this type of con-
struction, whatever their size, will be spoken of as rhizomes. The main reason for
this is that, as will be seen in the next section, a gradual transition can be traced
between certain of the leafless axes and the leafy shoots. Since some of the leafless
axes were thus undoubtedly rhizomes, and no clear distinction can be drawn between
these and the more root-like specimens, they seem best regarded as parts of a
rhizome-system something like that of the existing Psilotales.

The mode of occurrence of the rhizomes at places in the upper part of bed A" 1

* Mackie, W., “The Rock Series of Oraigbeg and Ord Hill, Rhynie, Aberdeenshire,” Trans. Edin. Geol. Soc. ,
vol. x, pp. 205-236, pi. xxiii, fig. 6.

t It may assist the reader in forming an idea of the vegetative organs of Asteroxylon if we note at the outset the
parallel afforded by the Psilotacese, in which the plant has a leafless, cylindrical, subterranean rhizome of simple
structure, certain branches of which pass into the more complex aerial leafy shoots. There are also resemblances in
structure between the leafy shoot of Asteroxylon and that of the Psilotacese, but the most helpful parallel as regards
the anatomy is found in some species of Lycopodium.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 645

is particularly interesting. In some blocks of the chert from this region a dark
sandy band enclosing large stems and rhizomes of Asteroxylon overlies a clear peat
composed of Hornea and Rhynia major. Rhizomes of Asteroxylon extend down-
wards in a root-like fashion into the upper portion of the clear peat, sometimes
penetrating the portions of plants in it. It appears that, in this region at any rate,
we have Asteroxylon preserved where it grew in the peat. It is clear that the
rhizomes were the underground . parts of the plant, and behaved physiologically like
a root-system. This is further shown by the frequent intrusion of these axes into
other rhizomes and into stems of Asteroxylon itself.

A general view of a section showing large stems of Asteroxylon embedded in the
dark sandy matrix overlying clear peat is given in PI. I, fig. 1. In the corresponding
section given in fig. 2, the sandy band above contains stems of Asteroxylon and a
rhizome a cut transversely. Two more slender rhizomes, at b, extend downwards
through the underlying peat, which is composed of more or less decayed Hornea and
Rhynia major. One of these rhizomes of Asteroxylon has grown through a rhizome
of Hornea in the peat. The larger rhizome of Asteroxylon at a in the layer above
contains a similar intrusive rhizome. This behaviour, which is also shown in the
specimen in fig. 4, recalls a common appearance in rootlets of Stigmaria.

Intrusive rhizomes of small size have been frequently met with in the stems of
Asteroxylon , and at first suggested comparison with the endogenous roots enclosed
in the cortex of some species of Lycopodium. This interpretation was, however,
negatived by their obvious intrusion in some cases, while in others they were situated
in the centre of the stele. In fig. 3, two stems of Asteroxylon are shown enclosed in
a sandy matrix. In the cortex of the large stem above, a small rhizome is cut
transversely at a. Another rhizome of similar size ( b ) is cut longitudinally as it
lies free in the matrix. This rhizome passes right across the cortex of the stem,

, which occupies the lower portion of the figure.

The largest rhizomes were more than 5 mm. in diameter (fig. 4), and all sizes
have been met with down to below 1 mm. (fig. 14). A remarkable feature, which
holds for the rhizomes of all sizes, is the complete absence of absorbent hairs. In
well-preserved specimens the surface is bounded by a smooth and continuous
epidermis, but this has frequently broken down and disappeared.

A general idea of the range in size and structure of the rhizomes will be gathered
from the examples figured on Pis. I, II, and IV. In the most clearly diflerentiated
specimens it is possible to distinguish the epidermis (ep.), a narrow zone of outer
cortex (o.c.), a zone of inner cortex (i.c.), wrhich in the larger rhizomes is relatively
broad, a clear zone of phloem (ph.), and the central strand of xylem (x.).

These regions are all shown in a transverse section of a fairly large rhizome
which was figured in Part I (PI. X, fig. 75), under the mistaken idea that it belonged
to Rhynia. The general correspondence in construction with the stem of Rhynia is
brought out by the mistake, which is now corrected,

646 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

The examples figured may now be briefly described.

The large rhizome in fig. 4 is slightly flattened, but must have been about 6 mm.
in diameter. Growing within it is another about 3 '5 mm. thick, and in the cortex of
this are two small specimens under 1 mm. thick. The xylem is evident in all the
rhizomes, and in the two larger specimens the zone of phloem around it is clearly
marked. The inner and outer cortex can be distinguished in the largest rhizome,
but the epidermis had almost completely broken down.

A specimen of medium size, with the stele divided preparatory to branching, is
represented in fig. 5, and a similar dichotomy in obliquely longitudinal section in
fig. 7. The longitudinal section in fig. 6 is of a rhizome of about the same size.
The strand of xylem is only followed for a short distance, but the section is otherwise
truly longitudinal. The surface is smooth, the epidermis is intact and well preserved,
the outer and inner cortex can be distinguished, and the phloem is fairly well
marked.

In the small rhizomes (figs. 8, 9, 14, 26-29) the same regions can be distinguished
as in the larger specimens, but this is not always evident without close examination.
Thus in fig. 29 the epidermis is wanting, but the regions are fairly well marked,
and they can also be readily distinguished in fig. 26. In fig. 9 the epidermis is
wanting, the outer cortex is about two cells deep ; there is a very narrow zone
of inner cortex and a fairly wide zone of phloem, but the limit between these
tissues is difficult to distinguish. The junction of inner cortex and phloem is clearer
in fig. 14. The epidermis is present on the small specimen in fig. 27. The transverse
■ section on the left in fig. 28 is interesting, since on part of its circumference the
epidermis is persistent and the surface smooth, while on the other side the
epidermis has broken down.

Rhizomes of all sizes evidently branched repeatedly. The most usual method
seems to have been equal or dichotomous branching, as is shown in figs. 5 and 7.
In other specimens there are indications that the branching was unequal or
lateral. It seems always to have been exogenous, and in favourable cases the
epidermis has been traced from a main axis to a branch.

The examples of rhizomes described above were all cylindrical. It should be
mentioned that some other sections have shown a more irregular outline, and
included several steles. There is nothing to suggest, however, that they were
anything but cases of irregular subdivision of the rhizome, such as are met with
in Psilotum.

The details of structure of the rhizomes can be dealt with briefly, since the
histology of some of the more important tissues will be more fully considered in
relation to other parts of the plant. There are, however, some features peculiar
to these underground parts.

The epidermal cells are smaller and shallower than those of the outer cortex
(figs. 6, 27, 28). They are well shown in the longitudinal sections of the outer

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 647

tissues of large rhizomes in figs. 12 and 13. In many specimens, as already
mentioned, the epidermis has broken down, the surface thus losing the smooth
appearance (fig. 29).

The zone of outer cortex is usually well marked in the larger rhizomes, where
it consists of some five or fewer layers of rather large cells which form a
relatively persistent zone usually free from fungal hyphse (figs. 12 and 13). In the
smaller and more root-like rhizomes the outer cortex is still clearly recognisable,
but may be reduced to one or two layers.

The inner cortex (figs. 13 and 29) is composed of narrower and more delicate
cells. It is often more or less broken down, and frequently contains a particular
fungus, which is usually wanting in the outer cortex on the one hand and the
phloem on the other. The same features characterise this region in the small
rhizomes, where it may be reduced to one or two layers of cells, and be difficult to
distinguish from the phloem unless the fungus be present.*

The phloem persists, and is fairly well preserved in many of the rhizomes.
In the larger examples (fig. 4) it forms a broad, clear zone of closely associated
elements, all of one kind. Even in the smallest rhizomes it is usually distinguish-
able (fig. 14). The elements composing it, which are elongated tubes with pointed
or transverse ends, will be described in the next section in connection with the
transition region which has supplied the best-preserved examples (cf. figs.
20 and 22).

The xylem in rhizomes of all sizes forms a simple, more or less cylindrical,
strand of tracheides, with no indication of a distinction between protoxylem and
metaxylem. All the tracheides have the peculiar type of spiral thickening
characteristic of Asteroxylon. This feature is shown in figs. 10 and 11, which are
details from the rhizomes in figs. 8 and 6 respectively. It will be described in
connection with the leafy stems.

Transition Region from Rhizomes to Leafy Shoots.

A series of intermediate stages has been traced between the characteristic
structure of the rhizome, as described in the preceding section, and that of the
leafy shoot of Asteroxylon. It has not as a rule been found possible to follow
the changes in a series of sections of one specimen, but the intermediate grades
are so completely represented as to leave no doubt as to the reality of the
transition.

The transition is indicated by a number of changes shown in different degrees
by the various specimens. As compared with that of the rhizome, the epidermis
becomes more distinct and definite, and has thicker external walls. The cortex,

* The question as to the saprophytic or possible mycorhizal nature of this fungus is deferred to Part 1Y, in
which the fungi present in the peat-bed will be considered.

648 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

while retaining the simple contrast of a narrow outer and a broad inner zone,
often appears more definite and better preserved. The central xylem becomes
stellate in cross section, thus intruding into the wide zone of phloem. Scale-
leaves appear on the surface. These sometimes have leaf-traces in relation to them,
but may be borne on a region of the stem with no leaf-traces.

The general features will be evident from the specimens represented in figs. 15-25
and fig. 95. Some of these examples will now be referred to and described.

In the portion of clear peat included by fig. 15 there are to be seen a
rhizome of medium size cut in obliquely longitudinal section, and below this
a transverse section of an axis with the characters of the transition region. The
general correspondence in structure between them will be evident at a glance.
The- xylem strand of the transition region is damaged, but is surrounded by a
broad zone of phloem, and is not giving off any leaf-traces, and none are present
in the cortex. This is differentiated into a broad inner cortex and a narrow zone
of outer cortex. The epidermis has broken down around part of the circumference,
but is very distinct on the other part. Here the outline is irregular, owing to
projections which comparison with other specimens shows to be scale-leaves. A
small intrusive rhizome is to be seen in the cortex on the lower side of the figure.

The longitudinal section shown in fig. 16 has also essentially the same structure
as the rhizome, and may be compared with fig. 6. It differs in the strongly
developed epidermis and cuticle, the indications of scale-leaves (sc. 1.), and the
presence of a few leaf-traces, one of which is shown at the lower part of the figure
at l.t. The cortex, which is rather broken down, shows outer and inner zones.
The stele is of the type characteristic of the rhizome, with a broad zone of
exceptionally well-preserved phloem (cf. fig. 22) around the large but simple
strand of xylem.

The two axes, still connected as they have arisen by branching, in fig. 25 show
a similarly simple structure in transverse section, while the smaller axis above
is in a more advanced stage of transition, with an angled xylem giving off leaf-
traces.

Successive stages of the transition from different specimens are seen in figs. 17,
18, 19, and 24. The specimen in fig. 17 hardly differs from a rhizome, except in
the distinctness of the epidermis. There are no leaf-traces or indications of scale-
leaves. The stele of this axis, which is to be regarded as in an early transition
.stage, has just given off a small lateral branch. The transverse section in fig. 18
shows the small scale-leaves very well. No leaf-traces are passing out to them, but
the xylem of the stele is distinctly angled. This is still more marked in the small
stem in fig. 25, in which a leaf-trace is departing from one angle of the quadrangular
xylem. The stem represented in fig. 19 is preparing to branch, and contains two
steles. The xylem of these has projecting rays, which extend nearly to the limit
between the phloem and inner cortex. This stem bears scale-leaves, and, in relation

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED. ABERDEENSHIRE. 649

to these, small leaf-traces are given off from the angles of the xylem. A scale-
leaf from another specimen is more highly magnified in fig. 23.

In the stem represented in fig. 24 the stele has all the features to be described
below for the large leafy stems, but the cortex is of the simpler type, and the leaves
appear to have been small and scale-like. The large specimen on PI. XIII, fig. 95,
is better preserved, but its construction is at about the same grade of complexity.

In the details of its histology the transition region is intermediate between the
rhizomes and the leafy shoots, but more like the former. Only special features,
or those tissues that are particularly well shown in this region, need be mentioned.
The tissues are brought out most clearly in fig. 20, which is a transverse section
of an early transition region from the same axis as fig. 17 ; in fig. 22, which is a
portion of the longitudinal section in fig. 16, and is also of the early transition
region; and in fig. 21, which is a portion of the late transition region shown in
fig. 19.

The epidermis is usually well preserved, and its small cells covered with cuticle
form a more definite boundary than on the rhizome (fig. 20, ep.). Stomata were
present in it (fig. 73). The outer and inner cortex resemble the corresponding
tissues of the large rhizomes, and, as fig. 20 shows, may be very well preserved;
The inner cortex is often infested with the same fungus as that occurring in the
rhizomes (figs. 20 and 22).

The xylem calls for no special description, except to note that when it becomes
stellate in the more advanced transition region the rays from which the leaf-traces
depart are composed, in some cases at least, of narrower tracheides.

The phloem has tended to persist, and some of the best-preserved examples
of this tissue are supplied by the transition region. It consists of closely associated
elements, all of one kind (figs. 20 and 21). In longitudinal section (fig. 22) these
appear as elongated tubular cells, which sometimes fit together with pointed ends,
and at other times have transverse end-walls.

In the less advanced specimens of the transition region the phloem, as in the
rhizome, forms a broad, clear zone around the more or less cylindrical xylem. As
the latter becomes stellate its projecting rays divide up the zone of phloem, but do
not reach to the periphery of this. There is thus phloem between the rays of xylem
and also forming a narrow zone outside the rays (fig. 21). The xylem of a leaf-trace
passing out to the scale-leaves is surrounded by a sheath of this outer zone of phloem
(fig. 21, It.)

The Leafy Shoot.

Leafy shoots ranging in diameter from 1 centimetre to 1 millimetre ( cf. figs. 96
and 30) form the most abundant material of Asteroxylon met with in those portions
of the chert that have been examined. The structure in the large as in the small
specimens is primary, and was determined once for all in development. The

650 DR R, KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

differences in size are best explained by the fragments coming from different
positions in a branch-system. Along with the difference in size there is naturally
some in the details of the structure, but a common plan of construction holds
throughout. This will be evident in comparing the specimens figured on
Pis. VI and VII.

Since the shoots in the material examined are represented by short fragments,
no general picture of the branch-system can be obtained. Examples of the
branching, which in most cases was lateral, will be described later.

The stem bore numerous leaves of relatively small size, so that the general form
of the plant must have depended on the branching. The first leaves on the
transition region from the rhizome have been seen to be small scales (figs. 18, 19, 23,
and 95), sometimes without any leaf- traces in connection with them, and at other
times with a small trace extending to the base of the leaf. The leaves on the
aerial shoots are larger, but their leaf-traces also never extend into the free portions,
ending in the enlarged leaf-bases (fig. 35). The long simple leaves are seen attached
to the stem in the longitudinal sections represented in figs. 31 and 32. They are
met with cut across, around many of the transverse sections of stems (figs. 96
and 46). These and other sections also show the projecting leaf-bases, while in
fig. 33 the leaves are just separating from the stem. While the leaf-bases on the
stem include the ends of the leaf-traces, the sections of the free portions of the leaves
have never shown any indication of the presence of a vascular bundle, however
well preserved the tissues may be (cf. figs. 96 and 37). Fig. 36 shows the leaves
closely crowded, and so arranged as to suggest that the section has passed across a
bud. On the stems the leaves were more or less closely placed, and, as the arrange-
ment of the leaf-traces shows, were borne spirally. A leaf lying free in the matrix
is shown in fig. 38, while others are cut so as to show the stomata in fig. 40. Fig. 30
is a cross section of a very small twig, with the closely placed leaves around it.
In cross section the leaves were oval, being thus slightly dorsiventral.

The arrangement of the tissues in the stem is most clearly brought out in
transverse section. A survey of the figures on Pis. VI, VII, and XIII will show
the features mentioned in the followingfgeneral description, which is based on the
better-preserved examples of large or medium size.

The surface of the stem and leaf-bases is bounded by an epidermis (ep.). This
has a well-marked cuticle, which is often traceable as the limit of the stem when the
epidermal cells are unrecognisable. Within the epidermis is a narrow zone of outer
cortex ( o.c .). The inner cortex ( i.c .) is sometimes uniform, but is usually differentiated
into outer, middle, and inner zones ( i.c.o ., i.c.m., i.c.i.). Only in the case of one
stem has an endodermis-like layer ( en .) been recognisable limiting the stele, which
in all cases, however, is clearly contrasted with the cortex. The arrangement of the
tissues in the stele has already been traced in the more advanced specimens of the
transition region. The xylem ( xy .) is stellate in cross section, and groups of

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 651

protoxylem elements (px .) are sometimes recognisable in the enlarged ends of the
rays. The phloem (ph.) occupies the bays of the stellate xylem, and forms a narrow
zone outside the ends of the rays of xylem (fig. 72) ; it is often more or less
completely decayed.

Leaf-traces ( l.t .), consisting of a small centrarch strand of xylem surrounded
by a zone of phloem, can be seen departing from the stele or on their way out
to the leaf-bases. The much larger traces for the lateral branches (br.tr.) are
occasionally met with, but, unlike the leaf-traces, pass out from the cortex of
the stem into the branches.

Attention may now be directed to some of the figures in which particular
features mentioned in the above general account are shown especially clearly.

The complete differentiation of a large stem is well shown in fig. 96, which will
be referred to later in relation to the preparation for equal division exhibited
by the stele. The corresponding tissues are labelled in the portions of other large
stems in figs. 41 and 43. The epidermis is not evident, though the cuticle marks
the limit of the stem. In the outer cortex ( o.c .), which is light coloured owing to
the disappearance of the cell-contents, some leaf-traces (l.t.) are present. Within
it. the outer zone of the inner cortex ( i.c.o .) is recognisable as a narrow and not
- very well characterised layer. The broad middle zone of the inner cortex (i.c.m.) is
prominent owing to the trabecular arrangement of the cells with wide intercellular
spaces between the rows. The coherent tissue around the stele, dark in fig. 41, is
the inner zone of the inner cortex (i.c.i.). The stellate xylem (x y.), with the
widened ends of the rays divided and enlarged, is well preserved. Leaf-traces (l.t.)
are present in relation to the rays. The phloem (ph.) between the rays of xylem
has decayed and disappeared, while the narrow zone of phloem within the cortex
and around the xylem is compressed, but recognisable.

The somewhat smaller stem in fig. 42 does not show the zoning of the decayed
inner cortex, but the dark remains of the cell-contents have persisted in the outer
-cortex and render it prominent. The tissues of the leaf-bases and free portions of
the leaves have a similar appearance. The stem in fig. 45 shows the zones of the
cortex clearly. The stele has the massive four-rayed xylem well preserved, but the
phloem has perished from between the rays.

Fig. 44 is of a stem of fair size, but with a less massive, rayed xylem. This and
the smaller stem in fig. 46 have a similar zoning of the cortex to the larger stems,
and have the outline irregular from the projecting leaf-bases. The still smaller stem
in fig. 47 shows the trabecular middle zone of the inner cortex clearly, and has the
xylem of the stele three-rayed. The cortex of the three small twigs associated
together (fig. 48) has so completely decayed, that their outlines are only dimly
traceable.

All the transverse sections of stems on PI. VI are magnified 14 diameters, and
are thus strictly comparable.

TRANS. ROY. SOC. EDIN., YOL. LII, PART III (NO. 26).

101

652 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

Two small twigs magnified 20 diameters are represented in fig. 54, and show
the proportional size of the solid triangular or quadrangular strand of xylem and
the stem itself better than those in fig. 48. The soft tissues are, however, very
poorly preserved. The bays in the angled xylem are here shallow, and there is some
reason to think that in some ultimate branches the xylem formed a cylindrical
strand, though no well-preserved specimens of this have been seen (cf the
specimens in fig. 47, on the right).

The stem in fig. 49 is a good example of one in which the inner cortex is
composed of uniform tissue with large intercellular spaces, and is not differentiated
into zones. The narrow outer cortex is distinct from this. The obliquely cut stele
is giving off large traces, one of which is probably for a branch and the other a
leaf-trace.

The arrangement of the tissues in the stele has been briefly described above, and
was evident in the figures of complete transverse sections of stems on PL VI. A
number of steles are represented somewhat more highly magnified on PI. VII, and
show the details better. Thus fig. 50 is the stele of a small stem, with the xylem
three-rayed, and the ends of the rays gradually widening. The phloem is fairly well
preserved in the bays. Fig. 53 is a typical large stele as usually preserved. The
ends of the four rays of the xylem are greatly expanded, and leaf-traces are
departing from them. The phloem in the bays is only partially decayed, but that
around the xylem has collapsed. This stele may be compared with the specimen
in fig. 72, in which the phloem is exceptionally well preserved. The stele in fig. 57
shows only the xylem well, but this is of interest on account of the more slender
three-rayed central portion which connects the widened and divided ends of the
rays from which the small leaves-traces depart. In the stele in fig. 52 the central
connecting portion of the xylem is imperfectly developed, and the numerous widened
ends are partially isolated. In fig. 51 the well-preserved xylem of the stele is
separated into distinct masses by the phloem extending across.

Before passing to consider histological details, mention must be made of some
sections of the shoot in which the imperfect differentiation of the tissues suggests
nearness to the apical region. The specimen represented in fig. 58 can only be
interpreted as of this nature. It is an accurately transverse section of a twig, the
outline of which is irregular owing to the projecting leaves. The tissues of these
and of the epidermis and outer cortex, with which they are continuous, appear almost
mature and of full thickness. The inner cortex, on the other hand, is narrow and
immature, though it shows the distinction of zones. The stele is represented by the
central clear area, with no indication of developed xylem, a tissue which in Astero-
xylon is very persistent. The stele thus appears completely undifferentiated, and,
unlike the cortex, does not show the cells clearly.

Fig. 59 is of a specimen of somewhat different character, and appears to belong
to a leafy shoot of fair size. It bears leaf-bases, and the outer and inner cortex have

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 653

numerous leaf-traces cut across. The leaf-traces close to the periphery of the stele
have well-developed xylem. The only other thickened elements are the poles of
the stellate xylem, from which differentiation in the thin-walled tissue of the immature
stele will proceed centripetally. This is more apparent in the next lower section of
the same stele, which is more highly magnified in fig. 60. In this the poles of the
imperfectly differentiated xylem, marked x., are readily distinguishable from the
leaf-traces. Figs. 59 and 60 are apparently of sections just below the apex, where
the lignified leaf-traces are crowded and pursuing a more vertical course, while the
stem-xylem is only beginning to be lignified.

The various tissues, the arrangement of which in transverse sections has been
described, are of course found in longitudinal sections (fig. 63), but few good
examples have been available for study. The longitudinal sections tend to be less
satisfactory, since the preservation of the tissues in these shoots, though good up to
a point, is by no means perfect. Owing, in part at least, to the effect of saprophytic
fungi the cell-walls are often hardly to be traced, although the zones of tissue may
be obvious. The difficulties arising from this imperfect preservation are especially
felt in studying longitudinal sections of Asteroxylon (figs. 63 and 31).

In proceeding to deal with the various tissues in detail, the information obtained
from longitudinal or oblique sections will be combined with that from the transverse
sections.

The epidermis, as seen on the stem, leaf-bases, and leaves, is a single layer of
cells, the outer walls of which are thickened, and covered with a distinct cuticle
(figs. 61 and 62). The cuticle often persists when the rest of the cell-walls has
disappeared. The outer walls of the cells may be flat or papillate (figs. 61 and 39).
In surface view or tangential section the epidermal cells are somewhat longer than
broad (fig. 65)» and in some well-preserved specimens their outer walls have shown
a median line similar to that in Rhynia Gwynne-Vaughani (Part I, PI. VI, fig. 31).
Stomata of ordinary form have been seen on the leaves, though even there they
seem to be relatively few (fig. 40), and they occurred on the aerial stems also, as
they did in the epidermis of the transition region (cf. fig. 73). A well-preserved
stoma from the aerial shoot is represented in fig. 65, and another in vertical section
in fig. 66. As the latter figure shows, the guard-cells were distinctly depressed
below the general surface. The pore of the stoma on the aerial shoot was much
smaller than on the transition region (cf figs. 65 and 73).

The outer cortex is a narrow zone, some six cells deep. Its cells were
tangentially extended, and are usually compressed (figs. 62 and 58). The contents of
the cells may persist as a dark substance, making the region prominent (figs. 42, 62,
and 58), or the contents may have disappeared and the zone appear clear (figs. 43
and 45). This applies to the contents of the epidermal cells also (figs. 61, 62,
and 65).

The tissues of the leaves are continuous with the epidermis and outer cortex of

654 DR H. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

the stem. When the free portions of the leaves are well preserved (fig. 37) the
thick-walled epidermis encloses a uniform tissue, with dark contents similar to the
cortex. As already mentioned, no vascular tissue is present in the free portion
of the leaf.

The wider inner cortex may be a uniform tissue (fig. 49), but it is usually
differentiated into outer, middle, and inner zones (figs. 41 and 43). In the cases in
which it is uniform the tissue has large intercellular spaces. When three zones are
distinguishable, it is the middle zone that has the prominent intercellular spaces, and
often shows a definitely trabecular arrangement of the rows of cells (fig. 67). The
narrow collapsed cells between the intercellular spaces come out more clearly in
fig. 71, though the trabecular arrangement is not shown in this specimen. Longi-
tudinal sections show that the trabeculae were vertically extended plates, and not
rows of cells. The outer zone beneath the outer cortex, and especially the wider
inner zone around the stele, are composed of more compact tissue. The former is
often ill-marked, while the inner zone is well characterised, and contrasts with the
middle zone by reason of its closely associated small elements (figs. 67 and 68).
While the cortical tissues show clearly the differences of the various zones, the cells
have been so far altered in the process of decay as to make detailed description
inadvisable. Longitudinal sections show that all the cells were somewhat longer
than broad, and that those of the inner zone of the inner cortex were closely
packed and more elongated. There is no evidence, however, to show that these
latter elements were developed as fibres.

As a rule there is no prominent layer marking off the cortex from the stele, but
in one large stem * a well-marked layer in the position of an endodermis was so
evident, that it cannot be passed over as without significance, although found in a
single example only (fig. 69 ; cf. figs. 100 and 102). The soft tissues of the inner
cortex and phloem of this stem were rather more than usually indistinct, though
the zonation of the tissues was clear, and the layer of cells in question stood out
between these two regions. Its cell-walls were prominent owing to their dark
colour. The whole appearance suggests comparison with such an ill-defined endo-
dermis as is met with in the stems of some species of Lycopodium.

The soft tissues of the stele, which have been termed collectively the phloem,
occupied the bays and formed a narrow zone outside the arms of the xylem. As has
been seen, they are usually badly decayed or compressed, and* have often in great
part disappeared, or been replaced by saprophytic fungi. In a few specimens of
the leafy shoot they have been found well preserved. The more decayed examples
must evidently be interpreted in the light of those better-preserved ones and of the
sections of the transition region. It has been described above how, with the change
from the circular to the stellate cross section of the xylem, the phloem of the stem
can be clearly related to that of the rhizome.

* This particular stem was undergoing branching of a peculiar endogenous type, as will be described below.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 655

The large, thin- walled elements of the phloem were preserved in the small stele
represented in fig. 50. The large stele in fig. 72 is the best example that has been
met with. In this the phloem is complete, both in the bays and around the
stele. No clear distinction can be drawn between the thin- walled tissue in these
two regions. The elements in the centre of the bays are wider than those abutting
on the xylem, and suggest a possible distinction of phloem proper and conjunctive
parenchyma. It does not appear advisable, however, to attempt such distinctions
of tissues on the specimens so far examined. The elements of the phloem with
dark contents in fig. 72 have probably been cut across not far from the transverse
end-wall.

The phloem in the stem is as a rule so much more decayed than in the rhizomes
or transition region that it is better to rely on these ( of fig. 22) for showing the
structure in longitudinal section. An example from a stem is represented in fig. 70.
The elements of phloem within the cortex are elongated and compressed. The wider
elements from one of the bays are partially decayed. From this and other specimens
it was clear that they were wide elongated tubes, which sometimes had the almost
transverse end-walls to which attention has been directed in the case of the better-
preserved phloem of the transition region.

It is convenient to select for the more detailed description of the xylem and of
the structure of the tracheides examples in which the protoxylem was recognisable,
but the fact that this was not always possible must be borne in mind. This is the
case, for example, with the steles in figs. 50, 53, and 57, which may be contrasted with
fig. 56, where the position of the immersed protoxylems comes out prominently.

Figs. 74, 75, and 76 represent three rays of the xylem of one large stele. There
is here no difficulty in distinguishing from the larger tracheides making up the bulk
of the xylem the groups of small and compressed tracheides ( px .) within the ends of
the rays but close to the surface. These small tracheides are the protoxylem, and
they are here enclosed by the larger elements of the metaxylem. Some-
times a protoxylem. group appears divided (fig. 75), and when a division is extending
to the surface there is the preparation for the separation of a leaf-trace. The
tracheides in figs. 74 to 76 are to some extent separated by the decay of the middle
lamella. The apparently uniform thickening of the wall proves on careful examination
to be the optical expression of a thin wall from which the localised thickening projects
inwards, thus diminishing the cavity. There are none of the characteristics of a pitted
wall in the transverse sections of the tracheides, either of the protoxylem or metaxylem.

The character of the thickening is evident when the tracheides are seen in longi-
tudinal section (figs. 77 to 79). The wall of the tracheide is moderately thin, but
has a spiral or irregularly spiral thickening. This projects into the cavity as rather thin
flat ledges, which, when seen from above in the transverse section, gives the appearance
of a uniformly thickened wall. The tracheides of the protoxylem (figs. 77 and 78,
px.) are much narrower, and have spiral or annular thickening. The clear distinction
of protoxylem and metaxylem, while the thickening of all the tracheides is spiral, is

656 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

characteristic. No trace of any scalariform or pitted tracheides has been seen. The
tracheides were of considerable length, and had pointed ends.

The leaf-traces arise from the arms of the xylem, more than one vertical series
coming from each arm, at least in the case, of the large and medium-sized stems.
This is shown in a number of figures, but the reader may be especially referred to the
steles in figs. 53 and 57. The arrangement and relations of the leaf-traces to the stem-
xylem are also well shown in figs. 55 and 56, in which the tissues other than the xylem
are partially or wholly omitted owing to the screen used in taking the photograph.

The xylem of the larger leaf-traces, which often arose from steles in which the
protoxylem was recognisable, shows a clear differentiation into small central
tracheides (protoxylem) surrounded by larger metaxylem elements (figs. 80 and 8l).
These in suitable specimens could be seen to be respectively continuous with the
protoxylem and metaxylem of the stem. The differentiation may be evident in the
leaf-trace even when it cannot be traced into the stem xylem (fig. 82). It can be
seen even in small traces, although in them the metaxylem elements often form
an incomplete layer around the protoxylem.

The leaf-traces exhibit a considerable range in size which does not go strictly
parallel with the size of the stems or steles. They are often larger, and sometimes
hollow near the periphery of the stem, an appearance that is probably in part to be
accounted for by changes occurring with age (fig. 43).

• The xylem of the leaf- trace acquires a sheath of the soft tissues or phloem which
surrounds the xylem of the stele. This appears according to the preservation
as composed of more or less collapsed or decayed, elongated cells. The nature of
this tissue in the leaf-trace and its continuity with the phloem of the stem is most
clearly seen when the latter tissue is well preserved, as in the specimen from the
transition region already described, which is represented in figs. 19 and 21.

The leaf-traces thus arising from the stele seem to have pursued an almost vertical
course for some distance, and are therefore commonly met with in the immediate
neighbourhood of the stele, and especially in the outer zone of phloem. This is seen
in many of the transverse sections, and also in fig. 63. The leaf-trace seems to have
been more oblique in its passage through the inner cortex, but on reaching the outer
cortex it is again almost vertical, till it ends in the leaf-base. This is best shown in
the obliquely longitudinal section in fig. 64.

Branching of the Shoot.

The shoot of Asteroxylon was evidently repeatedly branched, and a number of
examples must be referred to and several types of branching distinguished. This is
all that can be done, since the fragmentary nature of the remains is a disadvantage
in the investigation of this feature of the plant. The type of branching most
frequently met with is exogenous, and, as shown by the nature of the vascular

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 657

supply, clearly lateral. In a single remarkable example the lateral branch appeared
to be endogenous. In addition to lateral branching, indications of dichotomy, both
• of large and small shoots, have been found.

The large stem in fig. 96 appears from the disposition of the tissues in the stele,
where two distinct stellate xylems are enclosed in a common mass of phloem, to be
preparing for dichotomous branching. It is the only specimen of the kind so
far met with. There are other indications of the occurrence of dichotomy, however.
Thus in the case of a lateral branch, to be described below, the series showed the
continuation of the branch into two equal branches, although the actual division
of the stele was not followed. The division had taken place at a level where the
stele of the primary branch had not acquired the characteristically stellate xylem.
It is significant that in some other examples, in which two equal steles have been
found within the same cortex, the outlines of their xylems are suggestive of an
early stage in the progression to the stellate type. There is, however, no direct
evidence in these cases, one of which is represented in fig. 89, that we are dealing
with an early dichotomy of a lateral branch.

Another region in which the branching may have been dichotomous was in the
fine ultimate branches. These, however, are so ill preserved, that little can be
said beyond drawing attention to the fact that their association is suggestive of such
a mode of subdivision. The group of small twigs in fig. 48 may be compared with
the smaller but equally decayed twigs on the right side of fig. 47.

The relations of lateral exogenous branches to the main stem have been more
clearly followed in a number of examples. The lateral nature of the branch is
shown not merely by its small size relatively to the main axis (figs. 86, 90, and 95),
but by its vascular supply being derived from a cylindrical branch-trace separated
from one arm of the stellate xylem. This trace is considerably larger than a leaf- trace,
but, like it, takes a sheath of phloem as it separates from the stele. It passes more
or less obliquely through the cortex of the main stem and enters the branch. In
this it gradually assumes the characters of a stem-stele, the xylem becoming
angled and stellate, and giving off leaf-traces to the leaves of the branch.

Examples of different stages in the process thus summarised may now be
described and illustrated. In fig. 83 the separation of a branch-trace from one arm
of the stelar xylem is seen in a moderately small stem. In fig. 84 a branch-trace
is present well out in the cortex. Figs. 85, 86, and 87 form a series following a
branch- trace from the outer cortex of a large stem (fig. 85) into a small branch, the
stele of which rapidly assumes the characters of that of a leafy shoot, the xylem
becoming stellate (figs. 86 and 87). The base of a small lateral branch still
connected with the parent stem is shown in fig. 95, both axes being cut transversely.
In figs. 90 and 91 the branch is standing at right angles to the parent stem, so that
while the latter is cut transversely the branch is followed in longitudinal section.
The further portion of the branch in fig. 91 is shown more highly magnified in

658 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

fig. 92. An interesting feature in connection with the lateral branching is the way-
in which the branch may remain connected to its main stem by a web-like
flange of cortical tissue, as is shown by the example in fig. 88.

The general character of the branch-trace and its size relatively to the leaf-
trace can be gathered from figs. 83 and 84. No protoxylem elements are dis-
tinguishable in any of the branch-traces that have been observed. As is more
clearly shown by fig. 94, the xylem of the trace is surrounded by a zone of
collapsed phloem.

A point of some interest is the more or less close parallel between the assumption
by the branch-trace of the features of the stem-stele and the changes in the
transition from the rhizome to the leafy shoot. In the examples of lateral
branching figured the change is effected rapidly. A specimen has been met with,
however, in which a branch on a leafy shoot maintained for more than a centimetre
a broad, clear zone of phloem around the cylindrical strand of xylem of its stele.
The base of another branch cut transversely in the tangential section of its parent
stem is shown in fig. 93. The structure of the branch at this level resembles that
of a rhizome, but soon passed, as succeeding sections of the series showed, into that
characteristic of the transition region. This section at first suggested the lateral
origin of a rhizome from a leafy shoot, but cannot by itself be taken as affording
evidence of this.

A special account must be given of one case of branching which is remarkable for
its being apparently endogenous. The large stem exhibiting it has already been
referred to as the only specimen showing a well-marked layer in the position of an
endodermis (p. 654); it occurred in the block containing the sporangia, to be
described below. The stem can be traced unbranched through, a number of sections
of the series cut from this block. An enlargement of one arm of the xylem and a
bulging out of the endodermis indicates the approach to the place of branching
(fig. 97). As the large branch-trace separates and moves outwards it becomes evident
that the cortex of the branch is continuous with the tissue within the endodermis of
the main stem (fig. 98, and more highly magnified in fig. 100). This amounts to saying
that the branch is endogenous. This conclusion is supported by the discontinuity
between the superficial tissues of the main stem and the branch (fig. 100, X,
and fig. 101 ). Such a type of branching of the shoot is, so far as we know, unique,
both among living and extinct Vascular Cryptogams.

The cortex of this branch is continuous with that of two small stems, which in
the lower sections of the series lay in the neighbourhood of the large stem. The
whole series is consistent with the giving off of. an endogenous lateral branch that
bent back and divided equally. In the higher section figured in fig. 99 the. stele of
the curved branch is cut through twice, first at hr. tr., where the branch is emerging
from the main stem, and again where it has divided into two steles within the one
cortex \Ba, Bb ) at a distance from this. In the section at a lower level (fig. 98)

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 659

the branch-trace ( hr . tr.) is cut in the neighbourhood of the parent stele, while the
most distal part of the branch is shown divided into two ( Ba and Bb ) adjacent
to the main stem, but not connected with it.

Possible Fertile Region of Asteroxylon.

The parts so far described are shown by evidence of continuity, by the existence
of intermediate grades, and by histological agreement to be the rhizomes, stems, and
leaves of one plant. We have thus a fairly complete knowledge of the vegetative
organs of Asteroxylon MacJciei.

We owe our knowledge of the structures that have now to be described entirely
to two small blocks of chert, which were found loose. In sections of one of these a
number of small axes of peculiar type, with single or paired steles (fig. 107), had been
met with, associated with stems, leaves, and rhizomes of Asteroxylon, when Dr Gordon
kindly sent us a section he had cut from another loose block of the Rhynie. chert.
He also sent us the block, which was cut into a series of sections.

We wish to express our indebtedness to Dr Gordon for this important addition
to the material upon which this paper is based.

This block proved to be a bedded peat, free from sand, and largely made up of
portions of the peculiar type of axis mentioned above. In the peat were a number
of sporangia of a new type, and numerous scattered spores derived from them.
The peculiar axes and the sporangia were associated with stems of Asteroxylon of
various sizes, one large one showing the endogenous branching already described.
There were also structures resembling the rhizomes of Asteroxylon. Some of these,
unlike the other remains, tended to cross the bedding as if they had penetrated an
already formed peat.

The appearance of the peat in the block we received from Dr Gordon, and
especially the association with undoubted Asteroxylon of the peculiar axes that
have to be described, will be gathered from the photographs on PI. XV, figs. 102-106.
The region in fig. 102 shows a large stem of Asteroxylon and some more or less
well-preserved smaller stems. The peat in the lower portion of the figure is made
up of the peculiar axes, and the track of a perished rhizome of Asteroxylon is traceable
obliquely across the bedding. Fig. 104 shows a fairly large stem of Asteroxylon
below, while the peat above is made up of a mixture of the peculiar axes and
of rhizome-like branches of Asteroxylon, or of isolated strands of xylem composed of
the characteristic tracheides of this plant. In fig. 103 the branching structure has all
the features of a root-like rhizome of Asteroxylon. It is associated with character-
istically preserved specimens of the peculiar axes, only the epidermis and the central
cylinder of which have not perished. Another rhizome-like axis of Asteroxylon is
shown in transverse section in fig. 106, close to one of the peculiar axes, which is
exceptional in having the cortex preserved. Isolated strands of xylem are present in

TRANS. ROY. SOC. EDIN., VOL. Lll, PART ill (NO. 26). 102

660 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

the clear peaty matrix. Adjoining the peculiar axis shown in fig. 103 the peat contains
several more or less damaged specimens of sporangia still filled with the spores.

The rhizome or rhizome-like axes in this peat call for a brief reference. Some
of them are most naturally regarded as rhizomes of Asteroxylon, but others have
suggested a different interpretation. It is possible that we have in this block aerial
axes of Asteroxylon with a simple strand of xylem that are difficult to differentiate
from the rhizomes. It is unnecessary to multiply examples of such structures, the
nature of which must be left doubtful, but it seems desirable to record our impression
of the problem for future elucidation. It would be misleading to dismiss all these
remains as rhizomes, which they undoubtedly closely resemble, without raising this
question.

The association of the remains of Asteroxylon with the peculiar axes and sporangia
is such as to suggest strongly the possibility of the latter constituting the fertile
region of the same plant. We therefore describe them below, without giving them
special names. It must, however, be clearly stated at the outset that we have not
obtained evidence either of continuity or of histological identity which would con-
vert the possibility or probability of their belonging to Asteroxylon into a certainty.

The peculiar axes and the sporangia will now be described as the possible
sporangiophores and sporangia of Asteroxylon. Their reference to this plant will be
discussed later in the paper.

(a) The 'peculiar Axes or possible Sporangiophores of Asteroxylon.

The general appearance of these in the peat is well shown in figs. 102 and 105 on
PL XV. The preservation is poor, the cortical tissues having usually disappeared,
but the outline is often traceable by the more or less persistent epidermis, and the
stele also persists. In the stele the xylem, composed of narrow elements, is often
still surrounded by remains of the collapsed thin -walled tissue, which, from its posi-
tion, will be referred to as the phloem. The decay and disappearance of the cortex
render the outline of the longitudinal sections of these axes difficult to follow, but it
is evident that they were of considerable length, and branched dichotomously.

One of the specimens from the block in which they were first met with is shown
in transverse section in fig. 107. The circular outline is only disturbed by contrac-
tions during preservation. The epidermis and cortex are poorly preserved. Centrally
placed, there is a paired stele with two groups of xylem enclosed in a common invest-
ment of thin-walled phloem, which extends between them. The elements of both
xylem and phloem are much narrower than in the undoubted vegetative organs
of Asteroxylon.

The arrangement of the tissues in the steles of these axes exhibits a remarkable
variety, which is best seen in comparing the transverse sections. Some characteristic
examples are represented in figs. 108-116. In the simplest case the stele was single,
and consisted of a solid core of xylem surrounded by a zone of phloem (figs. 108 ;

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 661

cf 106, 118). There may be a small island of thin-walled phloem (fig. 109) or a
larger island (fig. 1 1 0) within the strand of xylem. This internal phloem may be
continuous with the outer phloem, making the cross section of the xylem horse-shoe
shaped (fig. 111). Larger examples of this distribution of the tissues are also met with
(fig. 114). Where the phloem extends right across, the two xylem strands are more
or less distinctly and widely separate and embedded in a common mass of phloem
(figs. 112, 113, 107). This leads on to two distinct steles lying within the same
cortex (figs. 115, 116).

The range of structure in these specimens, though they cannot be regarded as
constituting a series, is evidently consistent with a preparation for dichotomous
division of the axis. The more or less completely double steles are, however, so
frequent as compared with evidence of actual branching, that this may not be the
complete explanation. The specimens so far described show no departing lateral traces.

Other examples have been met with, though less abundantly, in which the xylem
of the stele was stellate or triangular in transverse section, and in which small traces
were departing from the angles (figs. 117 and 119). These are more like small twigs
of Asteroxylon , and it is noteworthy that in them the tracheides tend to be some-
what wider than is usual in these peculiar axes. This is well shown in fig. 119.

In yet other examples the stele was of the usual type for these axes, but was
splitting off small lateral portions (figs. 120, 126 at a). It is of interest that these
specimens were usually in the immediate neighbourhood of the sporangia, to be
described below (fig. 126).

While the general relations of the tissues in these axes is satisfactorily ascertained,
the poor preservation makes any detailed description of the tissues difficult. In only
a few cases, such as the specimen illustrated in figs. 106 and 118, were all the tissues,
including the cortex, fairly well preserved.

The epidermis is smooth, and its cells have their outer walls distinctly thickened
and covered with a cuticle. As mentioned above, this layer often remains when the
internal tissues have decayed (fig. 114). Stomata with short guard cells and a small
pore have been observed in the epidermis (fig. 121). The cortex in the examples in
which it is best preserved (fig. 118) is a uniform tissue composed of parenchymatous
cells. Little can be said of the phloem except that it consisted of narrow, elongated,
thin-walled elements. These, as preserved, tend to show in transverse section small
dark triangular markings at the junction of the cells similar to those to which
attention was directed in the corresponding tissue of Hornea Lignieri.

The tracheides composing the xylem were narrow, elongated tubes, with pointed
ends and brown walls (fig. 122). Though they remain connected and show the out-
line clearly in transverse and longitudinal sections, repeated examination of numerous
examples has failed to show the type of thickening of the tracheide walls. There was
almost certainly a thickening that has been lost or altered by decay, but we are not
prepared to place weight on such indications of its nature as we have yet observed,

662 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

In this respect, as well as in the narrow diameter of the tracheides, there is a
remarkable difference from the xylem of undoubted portions of Asteroxylon. While,
however, these differences are not to be under-estimated, it should be borne in mind
that the diameter of the tracheides exhibits such a range in different specimens of
these axes, and also in different axes of Asteroxylon in the same block of chert, that
it is not easy to draw a dividing line on account of the size of the tracheides. This
especially applies to some cylindrical strands of xylem, the tracheides of which,
though narrow, show the thickening characteristic of Asteroxylon ( cf fig. 104).

( b ) Sporangia possibly belonging to Asteroxylon.

The sporangia, which are associated with Asteroxylon and the axes just described
in the block found by Dr Gordon, differ characteristically from those of the Rhyniacese.
No confusion with the latter would be possible, and indeed no traces of Rhynia or
Hornea occur in the peat of this block. The sporangia and spores in question have
so far only been met with in this one specimen of the chert.

The spores and the remains of the sporangia occur throughout the, block, but
more abundantly in certain layers of it. The sporangia are more or less imperfect
and decayed, and nothing can be said with certainty as to their connection with
either the undoubted stems of Asteroxylon or with the peculiar axes that make up
the peat. There is a suggestive association of them with some of the small axes, the
stele of which is giving off branches (fig. 126), and the impression left by a study of
the block is, that they were most probably borne on the peculiar axes, and that the
latter were the sporangiophores of Asteroxylon.

In many cases only the thickened epidermal layer is preserved. This may be in
fragments (fig. 124), or may give a fair idea of the general size and shape of the
sporangium, and show that it dehisced at its wider free end (fig. 123). Here the
epidermal cells are thickened on their inner and lateral walls, and in the immediate
neighbourhood of the place of dehiscence they become shallower. The vertical
section of the epidermal layer thus becomes pointed towards the line of dehiscence,
as is particularly clearly seen in the two overlapping halves in fig. 125.

The specimen represented in fig. 126 gives some idea of the shape of the
sporangium and its rather thick wall. It appears to have been pear-shaped,
widening out from a thick stalk at the base, and about 1 mm. in length. There are
some indications that a fairly thick zone of thin-walled tissue had been present
within the epidermis and formed with this the sporangial wall.

Other specimens have the numerous spores still enclosed by the sporangial wall
(fig. 103). The clearest example met with is represented in fig. 123. It is cut
longitudinally, and shows the pear-shaped outline, with the epidermal layer becoming
thickened in the upper portion as the place of dehiscence (X) is approached. There are
remains of thin-walled tissue within the epidermis, and the cavity is filled with spdre§,

The mature spores are also found free in the peat. They are about 64 ^ in

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 663

diameter. The thin, smooth, cuticularised wall that alone persists has a character-
istic bright yellow colour, and shows the tri-radiate marking. A number of spores
lying free in a sporangium are shown in fig. 127. They have also been met with
still grouped in tetrads.

The Connection of the Parts Described and the Probable Habit of

Asteroxylon.

Before we can attempt to compare Asteroxylon with other plants it is necessary
to briefly summarise and discuss the connection of the various parts described in the
preceding pages, and in the light of this to see what picture can be formed of the
general habit of the plant. Our knowledge is unfortunately based on abundant
fragments only, and we lack the assistance of impressions which would give in whole
or in part the general form. The position of some of the parts and their mutual
connection can, however, be deduced on conclusive evidence, while the picture can be
extended with more or less probability to include other parts, the connection
of which is not actually established.

The relation of the root-like branches of the rhizome to the peat, and the presence
of stomata on the well-developed simple leaves, justify us in regarding Asteroxylon as
a terrestrial plant which grew in and on the peaty soil.

It is also clear that the plant of Asteroxylon had a leafless, dichotomously
branched, subterranean rhizome, some of the branches of which passed by a gradual
transition into aerial leafy shoots. The transition is marked by the appearance first
of small scale-leaves and then of the larger foliage leaves. The internal structure
shows a corresponding change from the simple stele of the rhizome without leaf-
traces to the stele of the stem with a stellate xylem giving off leaf-traces. Other
branches of the rhizome of varying degrees of slenderness have been shown by their
relations to the peat to have taken the place physiologically of a root-system. There
is, however, no evidence to justify us in drawing a morphological distinction between
roots and rhizomes.

The aerial shoots bore numerous, relatively small, simple leaves, each of which
had a leaf-trace, though this did not extend into the free part of the leaf. The leaf-
traces depart in more than one vertical series from each of the arms of the stellate
xylem of the stem-stele. The aerial shoot must have formed a system of branches,
maintaining a common plan of construction, but of successive degrees of slender-
ness. The branching in most cases observed has been lateral, but dichotomy is known
to occur both in large stems and at the base of a lateral branch. The facts on the
whole appear to indicate erect shoot-systems springing from the subterranean leafless
rhizomes. We have met with nothing pointing to a horizontal leafy stem giving oft'
erect branch systems on the one hand and rhizomes with their root-like branches
on the other.

664 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

The above general mental picture of the vegetative organs of Asteroxylon may
be said to be established in its main features. The connection of the various parts
as belonging to one plant is proved by direct evidence of continuity and by the
histological identity of important tissues.

On the other hand, conclusive evidence is wanting as to the fertile region and
sporangia of Asteroxylon. On the ground of close association we regard it as prob-
able that certain slender axes of peculiar structure and a type of dehiscent sporangia
constituted the fertile region of this plant. Our knowledge of Rhynia Gwynne-
Vaughani, Rhynia major, and Hornea Lignieri makes it certain that the structures
in question have nothing to do with them ; they either belong to Asteroxylon or to
an additional plant from the bed, of the existence of which there is no further
evidence. The weight we are prepared to lay on the peculiarly close association is
indicated by not giving separate names to these peculiar axes and sporangia. It
must be left for future investigation to confirm the assumed connection of these
structures with Asteroxylon, or to show that they belong to another plant.

Even if our assumption is correct, we have no evidence as to the mode of connec-
tion of the peculiar fertile axes and sporangia with the vegetative shoots. The
fertile axes might have replaced the leaves on a distal region of the shoot, or have
been ultimate subdivisions of a fertile region continuing certain vegetative shoots.

If we assume that these parts were the fertile region of Asteroxylon, we can
vaguely complete the picture of the plant, with the leafless rhizomes and branched
leafy shoots, by fine branched leafless axes bearing dehiscent sporangia. The
morphology and habit of the vegetative region of Asteroxylon are practically estab-
lished by evidence. The suggestions as to the fertile region are, at most, probable.

Diagnosis and Classification.

Asteroxylon, Kidston and Lang, n.g.

Diagnosis. — Plant consisting of leafless rhizomes continuing into branched aerial
stems bearing numerous small leaves. Stele of rhizome with a cylindrical strand of
xylem, while the xylem of the stem-stele is stellate and gives off leaf-traces. The
fertile region probably consisted of slender, branched, leafless axes bearing pear-shaped
sporangia of moderate size, with definite dehiscence at the wider jfree end.

Asteroxylon Mackiei, Kidston and Lang, n.sp.

Diagnosis. — Khizomes from over 5 mm. to under 1 mm. in diameter, some of the
branches evidently behaving physiologically as roots ; no absorbent hairs ; epidermis,
outer cortex and inner cortex ; stele with zone of phloem around a central strand of
xylem.

Transition region from rhizomes to shoots, with stomata in the epidermis and

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 665

small scale-leaves ; stele, with the xylem within the zone of phloem becoming stellate
and giving off leaf-traces ; cortex as in the rhizomes.

Shoots from over 1 cm. to under 1 mm. in diameter, with simple leaves about
5 mm. long, without any vascular bundle in the free portion — the leaf-trace
stopping at the leaf-base. Epidermis with thick outer walls and with stomata ;
narrow outer cortex ; wide inner cortex usually differentiated into three zones, the
middle zone having the vertical plates of tissue separated by wide intercellular
spaces, and extending between the outer and middle zones as radial trabeculae.
Stele with stellate xylem without any mixture of parenchyma ; protoxylem immersed
in the ends of the rays ; phloem between the rays .and around the stele. Elements
of phloem uniform, thin walled and elongated. Thickening of the tracheides, both
of the protoxylem and metaxylem, spiral. Leaf-traces departing from the arms of
the xylem in more than one vertical row ; protoxylem surrounded by metaxylem,
and this by a narrow zone of phloem. Trace ending in the outer cortex in the leaf-
base. Branching lateral and dichotomous. Small stems of similar structure, but
with xylem three-rayed, or even a cylindrical strand.

There is a probability, based on close association of the remains, that the fertile
region of Asteroxylon consisted of slender, branched, leafless axes with peculiar
structure and pear-shaped sporangia about 1 mm. long ; sporangia with epidermal
layer thickened towards the summit, where regular dehiscence took place. Homo-
sporous. Spores developed in tetrads about 64 m in diameter.

Locality. — Muir of Bhynie, Aberdeenshire.

Horizon. — Old Red Sandstone, not younger than the Middle Division of the Old
Red Sandstone of Scotland.

The Classification must remain to some extent obscure until the uncertainty
as to the fertile region is cleared up. We suggest for the present associating Aster-
oxylon with Psilophyton princeps in a family of the Psilophytales to be called
the AsteroxylaceEe.

The Classification of the Vascular Cryptogams or Pteridophyta associated together
in the Rhynie bed of silicified peat would thus be as follows : —

PTERIDOPHYTA.

Class. Psilophytales.*

Family, rhyniacea;.

Genus. Rhynia (R. Gwynne-Vaughani, R. major).

Hornea ( H . Lignieri).

Family, asteroxylace^e.

Genus. Asteroxylon ( A . Mackiei).

* See Part I, Trans. Roy. Soc., Edin., vol. ii, p. 780, 1917.

666 DR R. KlDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

Comparative Discussion.

The organisation of Asteroxylon, though more complex than that of Rhynia and
Hornea, appears to be in some respects simpler than that of most Vascular Crypto-
gams. The uncertainty as to the fertile region of the plant makes full comparison
difficult. On the other hand, there is the great advantage of a fairly complete
knowledge of the anatomy of the vegetative organs of Asteroxylon.

On the grounds both of external organisation and of internal structure it is
possible to make more or less close comparisons with some existing types of
Pteridophyta and with some extinct archaic forms of this group. In the present
imperfect, though growing, knowledge of the plants of Early Devonian times the
recognition of points of comparison in various directions appears more desirable than
attempts at a hypothetical completion of the facts. We propose in this discussion
to confine speculation within definite limits, though it cannot be altogether avoided.

The position arrived at in the preceding section of this paper must be borne in
mind in the following comparisons. The connection in one plant of the rhizomes, stems,
and leaves of Asteroxylon is established on satisfactory evidence. That the small axes
of peculiar structure and the sporangia constituted the fertile region of Asteroxylon
is regarded as a probable presumption on the grounds of their close association with
this plant and their absence in relation to the other vascular plants of the bed.

It should further be pointed out that, close as are some of the resemblances to
be traced, Asteroxylon stands apart in certain respects from all other Vascular
Cryptogams. One striking difference concerns the tracheides, those of the meta-
xylem as well as of the protoxylem having a peculiar type of spiral thickening.
On account of this and other differences, and because of the imperfection of our
knowledge of the plant as a whole, it is well to avoid drawing conclusions as to
actual relationships.

Comparisons.

It will be convenient to deal first with the undoubted vegetative organs of
Asteroxylon , and then to extend the comparisons with other plants to the presumed
fertile structures.

1. The organisation of the plant of Asteroxylon into leafless rhizomes and leafy
aerial shoots finds a close parallel in the existing Psilotales. The absence of hairs
even from the most root-like branches of the rhizome of Asteroxylon is a remarkable
but minor point of difference. The resemblance holds even in points of detail as
regards the absence of definite roots, the gradual passage of branches with the
structure of the rhizome into aerial shoots, and the appearance of leaves, at first as
small scales without leaf-traces, on the region of transition from the rhizome to the
shoot. The comparison is not inconsistent with the very root-like behaviour of
many of the branches of the rhizome of Asteroxylon. Something similar is shown,

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 667

though less strikingly, in the plants of Psilotum and Tmesipteris, and is brought out
very clearly in the young stage of the latter. As Holloway’s * account and figures
show, one rhizome-like branch of the young plant on the prothallus turns up and
becomes leafy, while the other extends into the soil in a root-like fashion. This
second branch, however, in some cases becomes a leafy shoot.

The close comparison thus possible with Psilotum and Tmesipteris helps us in
picturing the mode of growth of Asteroxylon , and adds weight to the view that the
Psilotales have on the whole preserved a primitive type of organisation, and do not
owe their simplicity to reduction.

In contrast to the comparison with the Psilotacese, that afforded by the species
of Lycopodium and Selaginella, which show a distinction of subterranean rhizomes
and aerial shoot-systems, is more superficial. The rhizomes in the Lycopodiacese
bear small scale-leaves, and are probably to be regarded as modified aerial shoots.
They also bear true roots.

2. As regards the external morphology of the leafy aerial shoots of Asteroxylon
and the relations of stem and leaf, the comparison is perhaps closest to Lycopodium
among existing plants, though it also holds in all essentials for the Psilotacese. The
dichotomous and lateral branching also find parallels in the shoot-systems of the
same plants.

In this connection the existence in Early Devonian times of a number of plants
known as impressions, in which the stems were more or less closely clothed with
small leaves, must be referred to here. A striking example is Thursophyton
( Lycopodites ) Milleri, from the Middle Old Red Sandstone of Scotland and Norway.
This and other ancient plants, such as Psilophyton and Arthrostigma, will be con-
sidered further below.

3. In its anatomy Asteroxylon is most closely comparable with the Psilotacese
and with Lycopodium, though some other plants must be mentioned.

The simple stele of the rhizome of Asteroxylon can be compared with that of the
rhizome of Psilotum and Tmesipteris. There is a central strand of tracheides, with-
out evident protoxylem, surrounded by phloem. The transition from this structure
to the more complex stele of the leafy shoot also finds its closest parallel in the
Psilotacese. As in these plants, it is associated with the appearance of small scale-
leaves.

The stele of the aerial shoot of Asteroxylon corresponds in a striking fashion with
that of the stem of Lycopodium. That there are also points of resemblance to the
stele of Psilotum and to the primary structure of Sphenophyllum need only be
mentioned in passing, since they concern features that are better shown in Lycopodium.
Outstanding points of comparison between the steles of Asteroxylon and of some
species of Lycopodium (e.g. L. Selago) are the circular outline of the stele as a
whole, the stellate outline of the xylem with the ends of the rays often enlarged, the
* Holloway, J. E., “The Prothallus and Young Plant of Tmesipteris ,” Trans. N.Z. Institute , vol. 1, 1917.

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 26). 103

668 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

immersed or mesarch protoxylem* in the ends of the rays, and the position of the
phloem between and around the rays of the stellate xylem. The leaf-traces in both
Lycopodium and Asteroxylon depart in more than one vertical series from each ray
of the stellate xylem. They consist of a strand of xylem, with the smaller
protoxylem elements in a central position, surrounded by a zone of soft tissue or
phloem. The anatomical parallel between Asteroxylon and Lycopodium extends to
the differentiation and zoning of the cortex, although the trabecular condition of
the inner cortex is never so marked in the latter genus.

The termination of the leaf-trace in the leaf-base of Asteroxylon, so that the
simple leaf-blade has no vascular bundle, is a noteworthy point of difference from
Lycopodium. Other differences concern the type of thickening of the tracheides
of the metaxylem, already referred to, and the absence of definitely specialised sieve-
tubes in the phloem of Asteroxylon. In both these respects the stele of Asteroxylon
appears to be less differentiated than that of Lycopodium.

A more distant comparison can be made between the stele of Asteroxylon with
its stellate xylem and, the stellate stem-steles of some Zygopteridese, such as Astero-
chlasna and Asteropteris. The points of difference as regards both stele and leaf-
traces are perhaps more important than those of agreement. The discovery in
Asteroxylon of an ancient plant with a stele so like that of Lycopodium tends
to strengthen the tentative comparisons that have been made between the
Zygopteridean and Lycopodium steles, t

Another comparison that requires mention is that between the anatomy of
Asteroxylon and Stauropteris. The vascular structure of the latter is clearly more
fern-like. It would be dangerous to go further at present than to recognise
that there are certain similarities in the arrangement of the tissues of the stele
between Stauropteris and the stems of Asteroxylon. These features in the structure
of Stauropteris have excited interest and proved a source of difficulty since the plant
was first recognised.];

4. The various comparisons made above were between the vegetative organs of
Asteroxylon and other plants. The following comparisons concern the small axes
with peculiar anatomy, and the associated sporangia which we regard as probably
the fertile region of the same plant.

{a) The type of association of the plant remains in the block of chert which
we owe to Dr Gordon is suggestively like the appearance presented by sections
across the massed remains of Stauropteris oldhamia. It has been pointed out

* The protoxylem of Lycopodium is usually exarch, as it may also he in Asteroxylon. Where larger leaf-traces
join the stem-xylem in some species of Lycopodium a mesarch structure is shown. (Cf. Sinnott; “On Mesarch
Structure in Lycopodium ,” Bot. Gazette, 48, pi. x, fig. 5.)

f Cf. Bertrand, P., Progressus Rei Botanicae, Bd. iv, p. 276.

t The vascular structure of Stauropteris will evidently be regarded differently according to whether it is assumed
that it is a specialised Zygopteridean frond (Bertrand, P., Etudes sur la Fronde des Zygopteridees , Lille, 1909), or that
it is a plant of more archaic type than the Zygopterideas, though tending in their direction (Lignier, Bull. Soc.
Bot. de France, t. 59 (1912), p. 1).

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ARERDEENSHIRE. 669

that larger and smaller stems of Asteroxylon, leafless branches of simpler structure,
and the peculiar axes and sporangia can be distinguished in this portion of the
Rhynie peat (PI. XV). The comparison is with the larger and smaller branches,
the finer ultimate subdivisions, and the sporangia of Stauropteris. The reader may
compare fig. 104 of this paper with fig. 321 of vol. ii of Seward’s Fossil Plants, or
with fig. 1 on pi. i of Bertrand’s memoir on the Zygopteridese.

( b ) No close comparisons can be made with certainty between the small axes
with peculiar structure and other plants. The general resemblances they present,
however, to ultimate fertile ramifications of Stauropteris on the one hand, and on the
other to the sporangiophores of various extinct and existing Pteridophyta, may be
indicated. We are inclined to attach considerable significance to this.

(c) The sporangia, with their thick wall, specially thickened epidermal layer
without definite annulus, and terminal dehiscence, can be compared most closely
with those of Stauropteris. Less close, but interesting, comparisons may be made
with the sporangia of the Ophioglossacese, especially Helminthostachys, of the
Psilotacese, and of Lycopodium and Selaginella.

5. Owing to the absence of conclusive evidence that the small axes with peculiar
structure are the ultimate fertile branches of Asteroxylon and bore the associated
sporangia, we enter a more speculative region in seeking for comparisons with
the whole plant of Asteroxylon as completed by combining these as its reproductive
organs.

If this assumption proves to be correct, the general habit and morphology of
Asteroxylon would be comparable with that of some Early Devonian plants, the best
known of which is Psilophyton princeps.

The plant of Psilophyton princeps, as reconstructed by Dawson in his well-known
figure,* had horizontal rhizomes bearing more slender root-like appendages, erect
branched aerial stems with small spine-like leaves, and large sporangia borne ter-
minally on the smooth ultimate branches. The known specimens of the plant do not
suffice to establish this reconstruction. There is doubt as to the appearance of
the rhizome. The characters of the aerial stems, with their spiny leaves and dicho-
tomous or lateral branching, are confirmed by specimens known from a number of
localities. The actual connection of the smooth branched axes bearing the large
sporangia with the leafy shoots is not established. They have, however, been found in
association with the stems of Psilophyton in Norway and France as well as in the
original locality at Gaspe in Canada, a circumstance that cannot be ignored. Halle f
has separated the stalked sporangia under the name Dawsonites arcuatus, on the
ground of absence of proof of connection with Psilophyton. He, however, recognises
in Psilophyton Goldschmidti a plant showing that smooth branch systems are borne

* Geological History of Plants, fig. 19.

t Haxltc, T. G., “Lower Devonian Plants from Roragen in Norway,” Kungl. Svenska Vetenskapsakademiens
Handlingcvr, Bd. 57 (1916), p. 1,

670 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

on main stems with small spine-like leaves. This, in our opinion, goes far to establish
the organic connection of his Dawsonites arcuatus with Psilophyton princeps.

We are not in a position to enter into a critical consideration of the plant of
Psilophyton princeps as a whole. In estimating the well-founded objections
mentioned above to Dawson’s reconstruction, it should be borne in mind that
he had exceptional opportunities for considering the plant as preserved at Gasp6.
Even if there is a considerable element of imagination in his reconstruction, it would
appear that it represents with remarkable justice or prescience a particular type of
plant characteristic of Early Devonian times.

The remains of Asteroxylon go far to establish the existence of another plant of
Old Red Sandstone age showing all the regions of Psilophyton as reconstructed by
Dawson. The leafless rhizome with root-like branches, and the stems with small
leaves and more or less lateral branching, are features of agreement dependent on
definitely ascertained facts. So also are the smooth leafless axes with the sporangia,
if these are rightly ascribed to Asteroxylon. The sporangia with terminal dehiscence,
though of smaller size, have their counterpart in Psilophyton.

The two reconstructions, though both admittedly imperfect, exhibit a corre-
spondence that makes them mutally supporting.

Some other Early Devonian plants may be considered in relation to the general
type indicated by Asteroxylon and Psilophyton. Thus Arthrostigma agreed in the
branched stems, with a central vascular strand, bearing pointed or spine-like leaves.
Nothing is known of the subterranean parts of this plant, or of its reproductive organs.
Pseudosporochnus, known from impressions in Bohemia, which are reterred to the
Middle Devonian, appears to have consisted of a stem-system of large size with
continued subdivision of its branches. Some of the fine ultimate branches end in
swellings, which are doubtfully referred to sporangia by Potonie and Bernard.*
There is no evidence of the existence of leaves in this plant, and the subterranean
organs are unknown.

The crowded arrangement of the small leaves on the branched stems of
Thursophyton ( Lycopodites ) Milleri, from the Middle Devonian of Scotland and
Norway, suggests comparison with Asteroxylon. Nothing is known of the repro-
ductive organs of Thursophyton. In this connection the prevalence of a number of
types of land-plants of moderate size in the Middle Devonian flora of Norway, as
described by Nathorst,| may be referred to. One of these, Broggeria Norvegica ,
seems worthy of consideration in relation to the possible fertile region of Asteroxylon.
From Nathorst’s figures * the plant may have had terminal fertile regions with
crowded sporangiophores, and not a definite cone.

In another direction comparison may be extended to certain frond-like remains
without any flattened lamina or pinnules ( Aphyllopteris , etc.) common in Middle

* Flore Devonienne de Vintage H. de Barrande, p. 32, figs. 71-72.

f Nathorst, “Zur Devonflora des westlichen Norvegens,” Bergens Museum Aarbok, 1914-1915, No. 9.

X Ibid., Taf. 3, figs. 5-7 ; Taf. 4, figs. 4-9.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 671

Devonian rocks. Some specimens in which sporangia are terminal on ultimate
ramifications are of special interest, and have been compared by other inves-
tigators to Psilophyton. These include Dimeripteris * and Rhacophyton con-
drusorum t from the Upper Devonian. Perhaps the most striking comparison on
these lines, however, is with the better, but still imperfectly, known plant Stauro-
pteris from the Carboniferous rocks, especially if, as suggested by Lignier,| it should
prove to have grown as an independent “ frond,” and not have been borne on a stem.

. On reviewing the main comparisons, which we have been led to make or suggest
with Asteroxylon , it will be seen that the only plants with which there may be
agreement in total organisation are Psilophyton and some other more or less
incompletely known archaic vascular plants. A definite conclusion cannot be reached
on this question, owing to the uncertainty as to essential facts, both for Asteroxylon
and all these other ancient plants.

The comparisons with the Psilotales as regards both general organisation and
anatomy are close and striking. This group may be regarded as representing in
a modified and specialised form the features of the Asteroxylon type of plant as
persisting to the present time. The Psilotales must, however, be maintained as a
distinct class of Pteridophyta.

The comparison with the Lycopodiales, and especially with Lycopodium, comes
out clearly in the general habit, and particularly in the anatomy, though comparison
cannot at present be extended to the reproductive organs. The antiquity of the
general type of construction of the shoot and stele exhibited by Lycopodium is
clearly established by our knowledge of Asteroxylon.

The characters of Asteroxylon suggest an advance from a more primitive type in
the direction of the Lycopodiales rather than of the Filicales. There are, however,
some interesting resemblances to archaic Ferns or plants which can be regarded
as on the line to the Filicales.

The possible fertile region of Asteroxylon, though imperfectly known, appears
to allow of comparison both with primitive Filicales and with the sporangiophoric
Pteridophyta. This suggests a general point of view from which all sporangiophores
may be regarded.

The Characters and Systematic Position of Asteroxylon considered in Relation
to Lignier’s Views as to the Evolution of Pteridophyta, and Halle’s
Application of them to the Devonian Floras.

The characters of Asteroxylon and the comparisons with other Pteridophyta
suggested above are consistent with a general view of the sporophyte in land-plants

'• ScHMALHAUSBN, J., “Ueber devonische Pflanzen aus dem Donetz-Becken,” Mem. du Comite Ge'ol., vol. viii,
No. 3, p. 29, 1894.

f Crispin, F., “ Description de quelqu'es plantes fossiles de l’etage des Psammites du Condroz (Devonien
Superieur),” Bull, de V Acad. roy. de Belgique, vol. xxxviii, p. 1, 1874. ( Psilophyton condrusorum.) Crispin, F.,

“ Observations sur quelques plantes fossiles de depots devonien, etc.,” Bull. Soc. roy. de Botan. de Belgique, vol. xiv,
p. 224, 1875. (Footnote, Rhacophyton condrusorum.) t Loc. cit., 1912.

672 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

that has emerged and been further developed of late years. On this view the plant
of the Vascular Cryptogams is comparable with, and derived from, a thalloid plant-
body such as we meet with in a number of the higher Algse. In its modern form it
was first developed by Potonie, but it is especially in the form it assumed in the
speculations of Lignier that it applies to Early Devonian plants. This investigator
indeed made special use of Dawson’s reconstruction of Psilophyton in formulating
his theory.

It will avoid unnecessary repetition if we give a free summary of those salient
features of the theoretical views of the late Professor Lignier, and of the discussion of
them in relation to the Devonian flora by Dr Halle, which appear to us to bear on
the consideration of Asteroxylon. The reader is referred to the full statements of
the views of these investigators as given in the literature cited below.*

The primitive type of plant-body in the Pteridophyta is supposed by Lignier to
have consisted of erect, dichotomously branched, cylindrical cauloids bearing small
flattened phylloids and with terminal sporangia. The later origin of roots is
supposed to have taken place by the transformation of certain subterranean cauloids.
Psilophyton is instanced as exhibiting most clearly this primitive type. Among
existing plants the Psilotales are regarded as differing chiefly in the localisation
of the sporangia on reduced branches. From such a primitive type of plant-body
the Lycopods are derived by further specialisation, their leaves being regarded as
phylloids. The leaves of the Filicales, on the other hand, are of a different origin,
being derived by the specialisation of certain cauloidal branch-systems. In the
more primitive forms ( Dimeripteris , Stauropteris) the bi-valvular sporangia were
borne terminally on the ends of cauloids, as they were in Psilophyton. The flattened
leaf-blades arose later by agglomeration of the cauloids into pinnules, and the
sporangia, at first marginal, then became situated on the lower surface.

Lignier’s view is susceptible of further simplification by Tansley’s suggestion f
that the leaves of the Lycopodia] es may have been derived by “ foliar specialisation
of short undivided branchlets of the thallus, instead of 'whole branch-systems as
in the Filicinean type.” This does away with the sharp distinction of phylloids
and cauloids.

The general view provides us with a conception of a common origin of the
divergent types of sporophyte represented by the Psilotales, Lycopodiales, and
Filicales. +

* Ligniek, M. 0., “ Equise tales et Sphenophyllales : Leur origine filicineene commune,” Bull. Soc. Linn.

Normandie, ser. 5, vol. 7, 1903, p. 93. “Essai sur l’Evolution Morphologique du regne vegetal,” Gomptes de l’ Assoc.
Frangaise pour VAvancement des Sciences, 1908, p. 580. “ Le Stauropteris Oldhamia Binney et les Coenopteridees a la

lumiere de la theorie du meriphyte,” Bull. Soc. Bot. de France, t. 59, 1912, p. 1.

Halle, T. G., “ Lower Devonian Plants from Roragen in Norway,” Kungl. Svenska Vetenskapsakad. Handlingar,
Bd. 57, p. 1.

y Lectures on the Evolution of the Filicinean Vascular System, p. 9. (Reprint, 1908.)

| For the present purpose it is unnecessary to complicate the discussion by entering into the consideration of the
Equisetales qnd Sphenophyllales. ,

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE.

673

In a recent paper Halle has discussed the Early Devonian flora on these lines. H6
surveys the assemblages of plants described from various localities, and indicates the
general distinction between those of Lower and Middle Devonian strata, and the
contrast of both with the plants of the Upper Devonian. In his general botanical
conclusions he points out the prevalence of plants with relatively small spine-like
leaves (Arthro stigma, Psilophyton) in the Lower Devonian, and suggests that from
such plants we can pass to the undoubted Lycopodiaceous forms of the Upper
Devonian. He also points out the prevalence in the Lower Devonian of remains
like the branched rachises of fern-fronds, but with no indication of pinnules or
laminae. He suggests that the points of comparison of these, on the one hand with
such a form as Psilophyton Goldschmidti, and on the other with fern-fronds, enables
us to think of the Lycopsida and Pteropsida as connected in origin from a common
form that was already vascular. Halle regards the fructification of Psilophyton as
described by Dawson (which, however, he separates under the name of Dawsonites
arcuatud) as the best evidence of the beginning of the Filicinean phylum in the
Older Devonian flora. He points out the prevalence in succeeding ancient Palaeozoic
strata of large sporangia borne on special branched fronds without a lamina. The
natural interpretation is that these fertile fronds represent the primitive state,
and that the flattened pinnules are a later development by cladodification as
suggested by Lignier, the sporangia being pre-existent in respect to the laminae
of the pinnules.

We have summarised the views of Lignier instead of discussing the question
afresh, because, without accepting his theory in all details, we find ourselves in
general agreement with his conception of a divergent progression from the primitive
land-plants towards the more specialised Lyeopodiales on the one hand and towards
the Filicinean type of plant on the other. Halle’s survey of the Devonian
floras in the light of these views also appears to us to be a fair statement of the
present bearing of the imperfectly known facts. It seems advisable to keep the
general conception on broad lines, in the hope that as our knowledge of the historical
data increases the details will emerge naturally. The interest and significance of
the more generalised plants, the external form and internal structure of which are
becoming better known, will remain even if it should prove that higher types
coexisted with them.

The characters of Asteroxylon are consistent with, and support, the conception
of the general course of the differentiation of early land-plants outlined above.
Asteroxylon appears to agree with Psilophyton in possessing in a generalised and
archaic form characters that are definitely specialised in the Psilotales, Lyeopodiales,
and Filicales. The Geological age and succession of the Early Devonian plants are
on the whole consistent with the origin of the various groups of Vascular Cryptogams
from a common source. If we think of Psilophyton as giving a fair idea of the
characters of this type, Asteroxylon shows indications of divergence in the direction

674 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

of the Lycopods, while the characters of Stauropteris (though this plant is in its own
way more advanced and specialised) may suggest the connection of the fern type.

Partly because of the insufficiency of our knowledge of the plants, and partly
because of the divergent tendencies towards more specialised groups which they
exhibit, it must be a matter of difficulty to define the characters of the group to which
Asteroxylon and Psilophyton belong. We have reason for regarding these plants as
differentiated into rhizomes, branched aerial shoot-systems with relatively small
leaves, and leafless branches bearing dehiscent sporangia. Such plants have a more
complex organisation than the Rhyniacese, with which they will be compared below.
The present state of our knowledge regarding them would appear to be fairly, though
tentatively, expressed by distinguishing a more complex family of the Psilophytales,
which may be termed the Asteroxylacese.

Comparison of Asteroxylon with the Rhyniacese. Conclusion.

Asteroxylon must, in conclusion, be compared with the still simpler type of
Vascular Cryptogams represented by the Rhyniacese.

The simple structure of the rhizome and basal region of the plant in Asteroxylon
is clearly comparable with that found throughout the plants of the Rhyniacese.
There are also indications of a similarly simple construction in some ultimate distal
divisions' of Asteroxylon. In the leafy middle region of the shoot of Asteroxylon a
complication in internal structure accompanies that in the external form.

The small dehiscent sporangia which are presumed to belong to Asteroxylon
appear to be of an advanced and specialised type as compared with those of Rhynia
and Hornea.

Asteroxylon differs from Rhynia and Hornea in the presence of undoubted leaves
on the aerial stems. The only structures in the Rhyniacese which suggest a
comparison with leaves are the peculiar hemispherical projections on the stems of
Rhynia Gwynne- Vaughani, and it is possible to regard these as affording a clue to
the leaves of Asteroxylon. On the other hand, we are acquainted in the Algse with
a number of independent examples of the integration of originally similar members
of a thalloid branch system into a new whole in which the continued axis of growth
appears as a relatively main “ stem” bearing the subordinated branches or “ leaves” ;
the relation of the shoot in Asteroxylon to the thalloid type of plant-body in the
Rhyniacese may alternatively be regarded in some such fashion.

To enter more fully into the meaning of the differences between the organisation of
Asteroxylon and that of the Rhyniacese would involve a discussion of the nature of the
leafy shoot. We shall not enter upon the consideration of the general question here.

The Vascular Plants which grew together in the Rhynie peat have fortunately
preserved for us the structure of examples of both the simple and the more complex
types of the archaic plants which we associate as the Psilophytales. They do not

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BCD, ABERDEENSHIRE. 675

demonstrate the passage of the one type to the other, and perhaps a record of this
was hardly to be looked for.

In Rliynia and Hornea we have revealed to us a much simpler type of Vascular
Cryptogam than any with which we were previously acquainted. This type suggests
the convergence of Pteridophyta and Bryophyta backwards to an Algal stock.

The knowledge of Asteroxylon confirms and enriches our conception of a more
complex but archaic type of the Vascular Cryptogams which supports the idea of the
divergence of the great classes of Pteridophyta from a common type, and links this
on to the simpler Rhyniaceae.

EXPLANATION OF PLATES.

(All the figures are from untouched photographs.)

Asteroxylon Maclciei, Kidston and Lang.

Plate I.

Fig. 1. General view of a vertical section from the upper part of bed A" 1. Large stems of Asteroxylon
are enclosed in the dark sandy matrix above. The clear peat below is composed of Hornea, but contains
rhizomes of Asteroxylon. x 2§ . (No. 2461.)

Fig. 2. Similar section from another block of the silicified peat from the same bed. In the sandy matrix
above are stems of Asteroxylon, and at (a) a large rhizome containing a smaller intrusive one. Into the peat
below, made up of Hornea and Rhynia major, rhizomes of Asteroxylon extend vertically ; the One at ( b )
perforates a rhizome of Hornea. x 2. (No. 2462.)

Fig. 3. Small rhizomes in relation to large stems of Asteroxylon enclosed in the sandy matrix, a, rhizome
cut transversely in the cortex of the large stem ; b, rhizome cut longitudinally as it lay in the peat and
passed through another stem of Asteroxylon. x 14. (No. 2463.)

Fig. 4. Large rhizome in transverse section into which another rhizome of considerable size has intruded,
while the cortex of this in turn encloses two small rhizomes, ox., outer cortex of the large rhizome ; i.c.,
inner cortex ; ph., phloem; x., xylem. x 14. (No. 2464.)

Fig. 5. Transverse section of rhizome of moderate size showing division of stele preparatory to dichotomous
branching, x 14. (No. 2465.)

Fig. 6. Slightly oblique longitudinal section of a rhizome of moderate size, ep., epidermis ; o.c., outer
cortex; i c., inner cortex; ph., phloem; x., xylem. x 14. (No. 2466.)

Fig. 7. Partially decayed specimen of rhizome in obliquely longitudinal section showing dichotomous
branching. x 14. (No. 2467.)

Plate II.

Fig. 8. Portion of the small rhizome in longitudinal section shown in fig. 3. x 33. (No. 2463.)

Fig. 9. The small rhizome in transverse section shown in fig. 3. x 33. (No. 2463.)

Fig. 10. Portion of the xylem of the small rhizome in figs. 3 and 8 showing the spirally thickened
tracheides. x 210. (No. 2463.)

Fig. 11. Portion of the xylem of the rhizome in fig. 6. The imperfectly preserved phloem is seen to the
right, x 210. (No. 2466.)

Fig. 12. Portion of a longitudinal section of the large rhizome in fig. 6 showing the superficial tissues
well preserved, ep., epidermis; o.c., outer cortex ; i.c., inner cortex. x 105. (No. 2466.)

Fig. 13. Portion of the outer tissues of another large rhizome in longitudinal section, ep., epidermis;
o.c., outer cortex; i.c., inner cortex, x 60. (No. 2468.)

Fig. 14. Two small root-like rhizomes in transverse section. x 33. (No. 2469.)

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 26).

104

676 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

Fig. 15. Portion of clear peat, including above an obliquely longitudinal section of a rhizome of moderate
size, and below a transverse section of the transition region to the shoot. The latter, on the left, has a well-
preserved epidermis, and bears small scale-leaves. A small intrusive root-like rhizome is enclosed in its
cortex below, x 14. (No. 2470.)

Fig. 16. Longitudinal section of an axis. with indications of small scale-leaves (sc. 1.) and of leaf-traces
( l.t .), but with structure closely similar to that of a large rhizome, ep., epidermis ; o.c., outer cortex ;
i.c., inner cortex; ph., phloem; x., xylem. x 14. (No. 2439.)

Plate HI.

Fig. 17 Transverse section of an axis of the transition region showing hardly any advance on the structure
of the rhizome. The stele has given off a smaller branch. x 14. (No. 2469.)

Fig. 18. Transverse section of a more advanced phase of the transition from the rhizome to the leafy
shoot. Small scale-leaves are borne on the surface, and the xylem is angled but not giving off leaf-traces,
x 14. (No. 2471.)

Fig. 19. Transverse section of a specimen of the transition region bearing small scale-leaves. Two steles
are enclosed in the common cortex, preliminary to branching. The xylem is stellate and giving off leaf-traces,
x 14. (No. 2456.)

Fig. 20. Transverse section from the same axis as fig. 17, showing all the tissues of the early transition
region, ep., epidermis ; o.c., outer cortex ; i.c., inner cortex permeated with fungus ; ph. phloem ; a:., "xylem.
x 33. (No. 2472.)

Fig. 21. Portion of fig. 19 more highly magnified, showing the stellate xylem (a;.), the distribution of the
well-preserved phloem (ph.), and a leaf-trace ( l.t .) in the cortex, with xylem surrounded;-; by phloem, x 33.
(No. 2456.)

Fig. 22. Portion of fig. 16 more highly magnified, ep. epidermis; ox., outer cortex; i.c., inner cortex
with fungus; ph., phloem ; x., xylem. x 33. (No. 2439.)

Fig. 23. Scale-leaf from the transition region. x 60. (No. 2473.)

Fig. 24. Transverse section of an advanced stage in tile transition region, with small scale-leaves and
simple cortex, but with a stele like that of the larger leafy shoots, x 14; (No. 2473.)

Plate IY.

Fig. 25. Transverse section of two axes, separating in branching, with the structure of an early^stage
of the transition region. Above is another more advanced example, the quadrangular xylem of which is giving
off leaf-traces. x 14. (No. 2474.)

Fig. 26. Transverse section of small root-like rhizome, ox., outer cortex; ix., inner cortex; ph., phloem ;
x., xylem. x 60. (No. 2475.)

Fig. 27. Transverse section of another small root-like rhizome with the epidermis (ep.) preserved,
x 60. (No. 2476.)

Fig. 28. Portion of clear peat showing two transverse sections of smaller!) izomes. The one on the left
shows the epidermis well preserved on one side, x 33. (No. 2477.)

Fig. 29. Transverse section of small rhizome, o.c., outer cortex ; ix., inner cortex ; ph., phloem ;
x., xylem. x 60. (No. 2476.)

Fig. 30. Transverse section of small leafy shoot. x 14. (No. 2479.)

Fig. 31. Longitudinal section of a large leafy shoot showing two leaves attached to the stem. The dark
bodies in the inner cortex are fungi, x 14. (No. 2476.)

Plate Y.

Fig. 32. Tangential section of leafy shoot showing two leaves in longitudinal section, that on the left
attached to the stem. x 14. (No. 2471.)

Fig. 33. Transverse section of medium-sized stem showing the leaves as they separate from the leaf-bases,
x 20. (No. 2480.)

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 677

Figs. 34, 35. Two successive tangential sections of a shoot showing the termination of a leaf-trace in the
leaf- base, the attachment of the leaf and its free portion. x 14. (Fig. 34, No. 2481 ; Fig. 35, No. 2482.)

Fig. 36. Transverse section of one-half of a bud-like grouping of leaves, the other half being lost by the
brecciation of the chert, x 33. (No. 2483.)

Fig. 37. Free portions of a number of well-preserved leaves in transverse section. x 33. (No. 2484.)

Fig. 38. Complete leaf in longitudinal section, x 20. (No. 2456.)

Fig. 39. Portion of the leaf in fig. 38 showing the papillate epidermal cells. x 60. (No. 2456.)

Fig. 40. Longitudinal view of a number of leaves in a section of the chert, showing stomata ( st .) in

surface view. x 105. (No. 2465.)

Plate VI.

Fig. 41. Transverse section of large stem, o.c., outer cortex; i.c.o., outer zone of inner cortex ; i.c.m.,

middle (trabecular) zone of inner cortex; i.c.i., inner zone of inner cortex; ph., decayed phloem; x., xylem-

x 14. (No. 2464.)

Fig. 42. Transverse section of stem, in which the outer cortex and the leaf-bases are prominent owing
to the dark cell contents. x 14. (No. 2485.)

Fig. 43. Transverse section of stem, o.c., outer cortex; i.c.m., trabecular middle zone of inner cortex;
i.c.i., inner zone of inner cortex; ph., decayed phloem ; x., xylem ; l.t., leaf-traces in the neighbourhood of the
stele and in the outer cortex. x 14. (No. 2490.)

Fig. 44. Transverse section of a moderate-sized stem showing the corresponding arrangement of tissues,
x 14. (No. 2491.)

Fig. 45. Transverse section of moderate sized stem with more massive stellate xylem than that in the
preceding figure, o.c., outer cortex; i.c.o., outer zone of inner cortex; i.c.m., middle trabecular zone of inner
cortex ; i.c.i., inner zone of inner cortex. x 14. (No. 2492.)

Fig. 46. Transverse section of small stem ; leaves in transverse section occur in the peat around. x 14.
(No. 2468.)

Fig. 47. Transverse section of small stem with well-marked trabecular zone of the inner cortex and
three-rayed xylem. In the peat, on the right, are small twigs with imperfectly-preserved cortex, giving
the appearance of dichotomous branching and of a cylindrical strand of xylem. x 14. (No. 2493.)

Fig. 48. Group of small twigs with imperfectly preserved cortex in transverse section. x 14.
(No. 2466.)

Plate VII.

Fig. 49. Obliquely transverse section of a stem with distinct outer cortex (o.c.) and uniform inner cortex
(i.c.). Large traces, one apparently for a branch and the other for a leaf, are departing from the stele,
x 14. (No. 2486.)

Fig. 50. Transverse section of the . stele of the stem shown in fig. 47. ph., well-preserved phloem ;
x., xylem. x 60. (No. 2493.)

Fig. 51. Transverse section of stele in which the phloem extends right across, separating the xylem
into distinct groups. x 60. (No. 2487.)

Fig. 52. Transverse section of stele in which the imperfection of the central region of the xylem has led
to the partial separation of the numerous enlarged ends of the rays. x 33. (No. 2472.)

Fig. 53. Transverse section of stele with massive four-rayed xylem (x.) and partially-preserved phloem
(ph.). l.t., leaf-traces. x 33. (No. 2488.)

Fig. 54. Transverse sections of two small stems with the well-preserved xylem giving off leaf-traces
enclosed by the cortex in which all trace of the tissues has disappeared. x 20. (No. 2449.)

Fig. 55. Transverse section of xylem with leaf-traces departing from the enlarged ends of the rays..
x 33. (No. 2462.)

Fig. 56. Transverse section of xylem showing the position of the groups of immersed protoxylem and the-
arrangement of the departing leaf-traces. x 33. (No. 2489.)*

Fig. 57. Transverse section of the stele of the stem in fig. 44 showing the departure and arrangement
of the small leaf-traces. x 60. (No. 2491.)

* In this and the preceding figure the soft tissues have been lost in the photograph in bringing out the details
of the xylem.

678

DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

Plate VIII.

Fig. 58. Transverse section of small shoot with the differentiation of tissues uncompleted. 1., leaf ; o.c.,
outer cortex ; middle layer of inner cortex ; i.c.i., inner layer of inner cortex ; s., undifferentiated stele,

x 60. (No. 2486.)

Fig. 59. Transverse section of stem in a region behind the. apex where the differentiation of the stelar
tissues has just begun, x 10. (No. 2454.)

Fig. 60. Transverse section of the stele of the section below that shown in fig. 59. The poles of the
developing stem-xylem are marked x. ; the other dark groups of elements are leaf-traces, x 60.
(No. 2473.)

Fig. 61. Epidermis of two adjacent stems showing the thick outer wall and cuticle, x 105.# (No. 2486.)

Fig. 62. Epidermis (ep.) and outer cortex (o.c.) from the transverse section of a small stem, x 105.
(No. 2486.)

Fig. 63. Longitudinal section of stem, o.c., outer cortex; i.c.o., outer zone of inner cortex; i.c.m.,
middle (trabecular) zone of inner cortex; i.c.i., inner zone of inner cortex; x., xylem. x 14. (No. 2494.)

Fig. 64. Oblique section of stem showing the complete course of a leaf-trace ( l.t .) from the neighbourhood
of the stele to a leaf-base. x 14. (No. 2495.)

Fig. 65. Stoma in surface view and adjoining epidermal cell% with contracted dark contents, x 210.
(No. 2491.)

Fig. 66. Stoma in vertical section across the pore, x 210. (No. 2496.)

Plate IX.

Fig. 67. The inner cortex in transverse section showing the middle (trabecular) zone well preserved.
i.c.o., outer zone ; i.c.m., middle zone ; i.c.i., inner zone, x 33. (No. 2475.)

Fig. 68. Well-preserved inner zone of the inner cortex {i.c.i.). i.c.m., middle (trabecular) zone. x 33.

(No. 2497.)

Fig. 69. Transverse section of stele showing distinct layer of cells ( en .) in the position of an
endodermis. x 33. (No. 2506.)

Fig. 70. Longitudinal section of partially-preserved phloem (ph.) of stem, x., xylem; i.c.i., inner zone
of inner cortex. x 60. (No. 2498.)

Fig. 71. Portion of middle zone of inner cortex not clearly trabecular but showing the cells and the
intercellular spaces well. x 105. (No. 2497.)

Fig. 72. Transverse section of stele of stem with exceptionally well-preserved phloem (ph.). x., xylem;

i.c.i., inner zone of inner cortex. x 33. (No. 2475.)

Fig. 73. Portion of the epidermis of stem in the transition region in surface view showing stomata with
large pores, x 210. (No. 2471.)

Plate X.

Figs. 74-76. Ends of the rays of stellate xylem from the same stele in transverse section showing the
structure of the tracheides. px., groups of small elements of the protoxylem. x 105. (No. 2499.)

Figs. 77-79. Longitudinal sections of the xylem of the stem showing the structure of the tracheides.
px., narrow elements of protoxylem. x 210. (Fig. 77, No. 2500; Figs. 78, 79, No. 2492.)

Fig. 80. Large leaf-trace in transverse section showing the Central protoxylem. x 210. (No. 2500.)

Fig. 81. Similar leaf-trace beside an arm of the stem xylem with its immersed protoxylem. x 210. -
(No. 2501.)

Fig. 82. Small leaf-trace separating from an arm of the xylem in which the protoxylem is less evident,
x 210. (No. 2501.)

Plate XI.

Fig. 83. Transverse section of stele from same stem as fig. 51 showing a branch-trace ( hr . tr.) preparing
to separate from one ray of the xylem. x 33. (No. 2483.)

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 679

Fig. 84. Transverse section of stem with a branch-trace ( br . tr.) on its way out through the cortex,
x 14. (No. 2502.)

Fig. 85. Transverse section of large stem with a branch-trace (br. tr.) just emerging from the cortex,
x 14. (No. 2459.)

Fig. 86. Next section of the same showing the branch ( B ) still connected with the stem, x 14.
(No. 2461.)

Fig. 87. Next section of the small branch, the xylem of which has now become typically stellate,
x 14. (No. 2492.)

Fig. 88. Transverse section of stem showing a small branch ( B ) connected to it by a narrow web of cortex,
x 14. (No. 2503.)

Fig. 89. Transverse section showing two steles (a and b) with imperfectly stellate xylem, enclosed in the
same cortex, x 14. (No. 2459.)

Plate XII.

Fig. 90. Transverse section of stem surrounded by the free portions of leaves and bearing a small leafy,
lateral branch cut longitudinally. x 14. (No. 2478.)

Fig. 91. Stem cut transversely and departing branch ( B ) cut longitudinally. X fI\. (No. 2470.)

Fig. 92. Further portion of the branch in fig. 91 showing the stele and leaf-traces. x 20. (No. 2470.)

Fig. 93. Tangential longitudinal section of the cortex of a stem shewing in continuity with it a branch in
transverse section with a stele like that of the larger rhizomes, ph., phloem of branch ; x., xylem of branch,
x 14. (No. 2470.)

Fig. 94. Transverse section of branch-stele in the cortex of a stem. The strand of xylem is surrounded
by the collapsed phloem. x 33. (No. 2504.)

Plate XIII.

Fig. 95. Stem with the structure of the advanced transition region with a small lateral branch still
attached, x 12. (No. 2477.)

Fig. 96. Large stem of typical structure surrounded by transverse sections of leaves. The stellate xylem
of the stele is double as if in preparation for dichotomous branching. x 14. (No. 2479.)

Plate XIY.

Fig. 97. Transverse section of a stem showing the preparation for the departure of a branch-trace (br. tr.)
from one arm of the stellate xylem. x 7|-. (No. 2505.)

Fig. 98. Transverse section of the same stem at a higher level. The branch-trace (br. tr.) has increased
in size where it lies in the neighbourhood of the parent stele. Adjacent to the main stem, but not in
connection with it, the branch is cut where it is dividing into two (Ba, Bb). x 7-|. (No. 2506.)

Fig. 99. Transverse section of the same stem at a still higher level. The stele of the branch (br. tr.) is
cut on its way out through the cortex of the main stem, and two steles (Ba, Bb) are contained in the cortex
of the further portion of the branch, which is bent back. x 7J. (No. 2507.)

Fig. 100. The stem in fig. 98 more highly magnified, c., cortex of main stem ; en., layer in position of
an endodermis limiting the stele of the main stem; c. br., cortex of branch; X, discontinuity between the
cortical tissues of the emerging branch and of the main stem. x 14. (No. 2506.)

Fig. 101. Region of discontinuity between the cortex of the main stem and of the branch in fig. 100 more
highly magnified, c., cortex of main stem; c. br., cortex of branch. x 33. (No. 2506.)

Plate XV.

Fig. 102. Portion of a section of the silicified peat of the block found by Dr Gordon. A large stem of
Asteroxylon is seen above, below there are smaller stems, while the track of a rhizome is traceable across the
peat formed of the small peculiar axes. x 14. (No. 2505.)

TRANS. ROY. SOC. EDIN., VOL. LII, PART III (NO. 26).

105

680 ON OLD RED SANDSTONE PLANTS FROM THE RHYNIE CHERT BED.

Fig. 103. Another portion showing sporangia filled with spores (s. s .) in the neighbourhood of one of the
peculiar axes with two steles, x 30. (No. 2508.) .

Fig. 104. Another portion showing a large stem of Asteroxylon (a), two cylindrical strands of xylem ( b )
composed of Asteroxylon tracheides, fragments of xylem of Asteroxylon ( c ), and a peculiar axis with its
cylindrical strand of xylem (d). x 105. (No. 2509.)

Fig. 105. Another portion showing a rhizome-like structure which is cut transversely and is giving off a
small lateral branch, embedded in the peat formed of characteristic specimens of the peculiar axes, x 33.
(No. 2509.)

Fig- 106. Another portion showing a rhizome-like axis of Asteroxylon at a, and at b one of the peculiar
axes with the cortex preserved. These are embedded in a peat formed of fragments of Asteroxylon. x 33.
(No. 2510.)

Plate XYI.

Fig. 107. Transverse section of one of the small peculiar axes, ep., epidermis; c., cortex; ph., phloem;
x. x., two strands of xylem. x 33. (No. 2511.)

Figs. 108-113. Steles of the peculiar axis in transverse section showing different distribution of the xylem
and phloem. Description in the text (p. 660). x 60. (Fig. 108, No. 2508; Fig. 109, No. 2507; Fig.
110, No. 2505; Fig. Ill, No. 2507; Fig. 112, 2509; Fig. 113, No. 2507.)

Fig. 114. Stele of peculiar axis showing horse-shoe shaped xylem {as.) and its relation to the phloem
(ph.). x 60. (No. 2512.)

Figs. 115, 116. Specimens of the peculiar axis with two distinct steles enclosed in the one epidermis.
The cortex, as usual, has decayed. x 60. (Fig. 115, No. 2505 ; Fig. 116, No. 2506.)

Fig. 117. Specimen of the peculiar axis with small traces departing from the stele. x 33.' (No. 2507,)

Fig. 118. Portion of the peculiar axis with the cortical tissues preserved shown in fig. 106. ep., epidermis ;
c., cortex; x., xylem. x 105. (No. 2510.)

Plate XYII.

Fig. 119. Transverse section of a specimen of the peculiar axis the xylem of which consists of larger
tracheides than usual and is giving off traces radially, x 60. (No. 2506.)

Fig. 120. Specimens of the peculiar axis the steles of which are giving off smaller portions laterally.
There are fragments of the sporangial walls in the adjacent matrix. x 60. (No. 2509.)

Fig. 121. Stoma in surface view from the epidermis of one of the peculiar axes. x 160. (No. 2505.)

Fig. 122. Xylem of one of the peculiar axes in longitudinal section showing the narrow tracheides without
recognisable thickening, x 105. (No. 2512.)

Fig. 123. Longitudinal section of a sporangium filled with spores. X, the place of dehiscence in the
thickened epidermis. x 50. (No. 2513.)

Fig. 124. Another sporangium in longitudinal section, only the epidermal layer persisting. X, place of
dehiscence, x 33. (No. 2508.)

Fig. 125. Overlapping portions ( a and b) of the epidermal layer from the region of dehiscence of a crushed
sporangium, x 105. (No. 2508.)

Fig. 126. Portion of the peat including at s a sporangium cut longitudinally, a, peculiar axis giying off
branches as in fig. 120 ; b, strand of Asteroxylon tracheides. x 33. (No. 2506.)

Fig. 127. Spores from the sporangium in fig. 123. x 160. (No. 2513.)

We have again to gratefully acknowledge our indebtedness to the Executive Committee of the Carnegie
Trust for a grant to defray the cost of the plates illustrating this memoir.

VoI.LIl

Plate I

Kidston and Lang.

Trans. Rov. Soc. £dinr

^ . Ki d ston , PKoto .

Zinco-CoMoty p« Co, Edinburgh.

Asteroxvlon Mackiei, K.& L

Wtojefc [Jr

"Wl

WzJjM

Trans. Roy: Soc. EdirE

Kid stonajto Lang.

Plate II

Vol.LIl

Aster oxylon

Mackie.i, K.&L.

Zinco-Collotypc Co, Edinburgh.

■R.Kidston, Photo.

KiDSTONantdLaNG.

Plate III.

V0I.LII.

Trans. 'Rov. Soc. Edin^

.Kidston,Ph©t©

Asteroxylon Mackiei, K.&L.

Zinco-CoHoty p« Co, Edinburg

Trans. Roy: Soc. Edm^

Kidston and Lang.

Plate IV Vol.LIl.

Zinco-Colloty pc Co, Edinburgh

Asteroxylon Mackiei, K.&L.

29

R.Kidston, Photo.

Trans. Rov; Soc. £din^

KidstonaktoLang

Vol.LIl

I ^ Kidston , Photo

Zinco-CoMoty pe Cot Edinburgh.

AsTEROXYLON MaCKIEI, K.&L

KiDSTONa>tdLaNG.

Plate VI

Trans. Roy: Soc. Edinr

Vol.LIl

Lom.

mm

Asteroxylon Mackiei, K.&L

p.Kid»ton,PKotc

Zinco-Collotypc Co, Edinburgh.

Trans. Roy; Soc. Edin?"

Kl D 5TO N AND L AN G .

Plate VJ I Vol.LII.

Asteroxylon Mackiei, K.&L.

Zinco-CoJIoty pt Co, Edinburgh.

Trans. Roy: Soc. Edin^

Kid ston and Lang.

Plate VIII .

VoI.LIl.

torn.

l.C.L

: R.Kidston, Photo.

Zi nco-Colloty pc Co, Edinburgh.

Asteroxylon Mackiei, K.&L

Trans. Roy; Soc. Rdin^

Kidston and Lang.

Plate IX

VoLLU.

Zmco-CoHoty pe Co, Edinburgh.

73

RKidston, Photo.

Asteroxylon Macriei, K.&L.

Trans. Roy: Soc. Edin^ KidstonamdLang.

Plate X. Vol.Lll

80

P.Kidston, Photo.

Zinco-Collotype Co, Edinburgh

Asteroxylon Mackiei, K.&L.

Kidston and Lang.

Plate X I

Vol.LIl

Trans. Roy: Soc. Edin^

K.Kidston.PKoto.

Aster

Zinco-Coliotype Co, Edinburgh.

O XYLON

Mac

ft I E I,

K.&L

Trans. Ro\: Soc. fdin!*

Kidston and Lang.

Plate XII. Yol.LII

:/■*

| ^-Kidston^PKoto.

Aster oxylon

Mackiei, K.&L

Zinco-Colloty pc Co, Edinburgh.

KiDSTONandLaNG.

Plate XIII. Vol.LII

Trans. Roy: So c. Edin^

> sr?

R.Kidston,Pkoto.

Aster oxvlon Mackiei, K.&L.

7i nco-ColIoty p e Co, Edinburgh

Kidston and Lang.

Trans. Rov: Soc. Edin^

Plate XIV. Vol.LIl

ft-Ki<Iston,Pkot©.

Asteroxylon Mackiei, K.&L.

Zinco- Collotype Co, Edinburgh.

Trans. Roy Soc. Edin^

Kidston and Lang.

Plate XV.

VolLIl

Kidston ;PKoto.

Asteroxvlon Mackiei, K.&L.

Zinco-C olio type Co, Edinburgh.

Trans. Roy; Soc. Edin?'

Kidston and Lang.

Plate XVI. Vol.Lll

p.Kfston, Photo.

117

?

ns

Asteroxylon Mackiei, K.&L.

Zi nco-Colloty p« Co, Edinburgh.

Trans. Roy: Soc. Edinr

Kidston and Lang.

"Plate XVII. Vol.LII.

127

122

a

123

124

PKidifon, Photo.

? Asteroxylon Mackiei, K.&L.

Zinco-Collofyp« Co, Edinburgh.

The Transactions of the Royal Society of Edinburgh will in future be Sold at the

following reduced Prices : —

Vol.

Price
to the
Public.

Price

to

Fellows.

Vol.

Price
to the
Public.

Price

to

Fellows.

Vol.

Price
to the
Public.

Price

to

Fellows.

£

s.

d.

£

S'

d.

£

s.

d.

£

s.

d.

£

s.

d.

£

s.

d.

VI.

0

11

6

0

9

6

XXXV.* Pt. 1

2

2

0

1

11

0

XLVI. Pt.

1

1

1

10

0

16

6

VII.

0

18

0

0

15

0

„ Pt. 2

1

11

0

1

3

6

„ Pt.

2

1

5

8

0

19

4

VIII.

0

17

0

0

14

0

„ Pt. 3

2

2

0

1

11

0

„ Pt.

3

1

7

3

1

0

11

IX.

1

0

0 1 0

17

0

„ Pt. 4

1

1

0

0

16

0

XLVI I. Pt.

1

0

19

9

0

15

0

X.

0

19

0

0

16

0

XXXVI. Pt. 1

1

1

0

0

16

0

„ Pt.

2

1

3

0

0

17

4

XI.

0

14

6

0

12

0

,, Pt. 2

1

16

6

1

7

6

„ Pt.

3

1

0

10

0

15

8

XII.

0

14

6

0

12

0

„ Pt. 3

1

0

0

0

16

0

„ Pt.

4

1

7

7

1

0

9

XIII.

0

18

0

0

15

0

XXXVII. Pt. 1

1

14

6

1

5

6

XLVIII. Pt.

1

1

2

9

0

17

2

XIV.

1

5

0

1

1

0

„ Pt. 2

1

1

0

0

16

0

„ Pt.

2

1

9

6

1

2

5

XV.

1

11

0

1

6

0

„ Pt. 3

0

16

0

0

12

0

„ Pt.

3

1

11

0

1

3

3

XX. Pt. 1

0

18

0

0

14

0

„ Pt. 4

0

7

6

0

5

8

„ Pt.

4

0

16

8

0

12

4

XXII. Pt. 2

0

10

0

0

7

6

XXXVIII. Pt. 1

2

0

0

1

10

0

XLIX. Pt.

1

0

7

6.

o-

7

6

„ Pt. 3

1

5

0

1

1

0

„ Pt. 2

1

5

0

0

19

0

„ Pt.

2

1

12

6

1

4

6

XXVII. Pt. 1

0

16

0

0

12

0

„ Pt. 3

1

10

0

1

3

0

„ Pt.

3

1

10

0

1

2

5

„ Pt. 2

0

6

0

0

4

6

„ Pt. 4

0

7

6

0

5

8

„ Pt.

4

1

1

9

0

16

0

„ Pt. 4

1

0

0

0

16

0

XXXIX. Pt. 1

1

10

0

1

3

0

L. Pt.

1

1

5

9

0

18

9

XXVIII. Pt. 1

1

5

0

1

1

0

„ Pt. 2

0

19

0

0

14

6

„ Pt.

2

1

7

0

1

0

0

„ Pt. 2

1

5

0

1

1

0

„ Pt. 3

2

3

0

1

11

0

„ Pt.

3

1

7

0

1

0

0

,. Pt. 3

0

18

0

0

13

6

,, Pt. 4

0

9

0

0

7

0

„ Pt.

4

1

2'

0

0

17

0

XXIX. Pt. 1

1

12

0

1’

6

0

XL. Pt. 1

1

5

0

0

19

0

LI. Pt.

1

Jl

5

0

0

19

0

„ Pt. 2

0

16

0

0

12

0

„ Pt. 2

1

12

6

1

5

6

;, Pt.

2

1

13

0

1

5

0

XXX. Pt. 1

1

12

0

1

6

0

„ Pt. 3

1

6

0

0

19

6

„ pt.

3

1

15

0

1

6

0

„ Pt. 2

0

16

0

0

12

0

„ Pt. 4

1

0

0

0

16

0

„ pt.

4

1

11

6

1

3

6

„ Pt. 3

0

5

0

0

4

0

XLI. Pt. 1

I

1

0

0

15

9

LII. Pt.

1

1

7

0

1

0

6

„ Pt. 4

0

7

6

0

5

8

„ Pt. 2

1

9

6

1

2

0

„ Pt.

2

1

10

0

1

2

0

XXXI.

4

4

0

3

3

0

„ Pt. 3

2

5

0

1

13

6

„ Pt.

3

2

3

0

1

11

6

XXXII. Pt. 1

1

0

0

0

16

0

.2 «

t XLII.

2

2

0

1

11

0

„ Pt. 2

0

18

0

0

13

6

| a .

) XLIII.

2

2

0

1

11

0

„ Pt. 3

2

10

0

1

17

6

XLIV.Pt.l

1

18

6

1

9

0

„ Pt. 4

0

5

0

0

4

0

«>

( „ Pt.2

1

1

0

0

15

9

XXXIII. Pt. 1

1

1

0

0

16

0

XLV. Pt. 1

1

9

0

1

2

0

„ Pt. 2

2

2

0

1

11

0

„ Pt. 2

1

7

0

1

0

0

„ Pt. 3

0

12

0

0

9

6

„ Pt. 3

1

13

9

1

5

3

XXXIV.

2

2

0

1

11

0

„ Pt. 4

0

4

6

0

3

6

* Vol. XXXV, and those which follow, may be had in Numbers, each Number containing a complete Paper.

January 1920. — Volumes or parts of volumes not mentioned in the above list are not for the present on sale
to the public. Fellows or others who may specially desire to obtain them must apply, direct to the Society. As
the Society reprints from time to time parts of its publications which have become scarce, the absolute correctness
of this list cannot be guaranteed beyond this date.

The papers published in this Part of the Transactions may be had on application to the Publishers,

at the following prices: —

Price to Public.

Price to Fellows.

Price to Public.

Price to Fellows.

No. XIX, .

2s. 3d.

Is. 9d.

No. XXIII,

Is. 3d.

Os. 7d.

No. XX, .

2s. Od.

Is. 6d.

No. XXIV,

9s. Od.

6s. 9d.

No. XXI, .

7s. Od.

5s. Od.

No. XXV,

. 2s. 4d.

Is. 9d.

No. XXII,

4s. 6d.

3s. Od.

No. XXVI,

. 15s. Od.

11s. 3d.

TRANSACTIONS

OF THE

ROYAL SOCIETY OF ED

VOLUME LII, PART IV.— SESSION

CONTENTS.

XXV [I. Scottish National Antarctic Expedition, 1902-1904: Gambrian Organic Remains from a
■Dredging in the Weddell Sea. By W. T. Gordon, D.Sc., Reader in Geology, University
of London, King’s College. (With Seven Plates), ..... 681

{Issued June 9, 1920.)

XXVIII. New Stelar Facts, and their Bearing on Stelar Theories for the Ferns. By John M ‘Lean
Thompson, M.A., D.Sc., F.L.S., Lecturer on Plant Morphology, Glasgow University.

(With Four Plates and Nine Figures in the Text), . . . . .715

{Issued July 9, 1920.)

XXIX. Isle of Wight Disease in Hive Bees .

(1) The Etiology of the Disease. By John Bennie, D.Sc. ; Philip Bruce White,

B.Sc. ; and Elsie J. Harvey. (With One Plate),. .... 737

(2) The Pathology of Isle of Wight Disease in Hive Bees. By Philip Bruce White,

B.Sc., Bacteriologist to the Bee Disease Investigation, University of Aberdeen
and N. of Scotland College of Agriculture. Communicated by Dr John Rennie.

(With One Plate), . . . . . . . .755

(3) Isle of Wight Disease in Hive Bees — Experiments on Infection ivith Tarsonemus

woodi, n. sp. By Elsie J. Harvey. Communicated by Dr John Rennie, . 765

(4) Isle of Wight Disease in Hive Bees — Acarine Disease : The Organism associated with

the Disease — Tarsonemus woodi, n. sp. By John Rennie, D.Sc. (With One
Plate and Two Figures in the Text), . . . . .768

{Issued March 25, 1921.)

XXX. Shackleton Antarctic Expedition, 1914-1917 : Depths and Deposits of the Weddell Sea.

By J. M. Wordie, M.A., F.G.S. Communicated by Professor J. W. Gregory, F.R.S., . 781

{Issued May 27, 1921.)

XXXI. Shacldeton Antarctic Expedition, 1914-1917 : The Natural History of Pack-Ice as observed
in the Weddell Sea. By J. M. Wordie, M.A., F.G.S. Communicated by Professor
J. W. Gregory, F.R.S. (With Nine Text-Figures and Four Plates), . . . 795

{Issued June 21, 1921.)

XXXII. On Old Red Sandstone Plants shoiving Structure, from the Rliynie. Chert Bed, Aberdeenshire.

Part IV. Restorations of the Vascular Cryptogams, and Discussion of their Bearing on
the General Morphology of the Pteridophyta and the Origin of the Organisation of Land
Plants. By R. Kidston, LL.D., D.Sc., F.R.S., and W. H. Lang, D.Sc., F.R.S., Barker
Professor of Cryptogamic Botany in the University of Manchester. (With Five Plates), 831
{Issued August 26, 1921.)

XXXIII. On Old Red Sandstone Plants showing Structure, from the Rhynie Chert Bed, Aberdeenshire.

Part V. The Thallophyta occurring in the Peat-Bed ; the Succession of the Plants through-
out a Vertical Section of the Bed, and the Conditions of Accumulation and Preservation
of the Deposit. By R. Kidston, LL.D., D.Sc., F.R.S. , and W. H. Lang, D.Sc., F.R.S.,

Barker Professor of Cryptogamic Botany in the University of Manchester. (With Ten
Plates and One Figure in the Text), . . . . . . 855

{Issued November 30, 1921.)

Index, ............ 903

N BU

920-1

EDINBURGH:

PUBLISHED BY ROBERT GRANT & SON, 107 PRINCES STREET,

AND WILLIAMS & NORGATE, 14 HENRIETTA STREET, COVENT GARDEN, LONDON, W.C. 2.

MDCCCCXXI.

Price Fifty-three Shillings.

( 681 )

XXVII. — Scottish National Antarctic Expedition, 1902-1904: Cambrian Organic
Remains from a Dredging in the Weddell Sea. By W. T. Gordon, D.Sc.,
Reader in Geology, University of London, King’s College. (With Seven
Plates.)

(Read March 13, 1915. MS. received December 4, 1919. Issued separately June 9, 1920.)

Introduction.

From a biological point of view considerable interest must always attend any
investigation of the earliest known organisms, and, although we may legitimately
infer that a flora and fauna existed prior to Cambrian times, the organisms preserved
for us in rocks of that age constitute, at present, the first chapter of palaeontological
history. For this reason alone the Archseocyathinse are important, since they form
part of the Lower Cambrian fauna. When, however, we consider that the genera
included in the group are very distinct from one another (indicating that the family
was probably of considerable antiquity even in those early times), that the types
have, as far as we know, a wide geographical distribution, and that to certain skeletal
characters usually associated with the Porifera they unite others more common
among the Coelenterata, interest is still further stimulated. On the other hand,
although recorded in great abundance from several widely separated localities, they
are not, as a rule, common fossils in Cambrian strata, and consequently the group
has not received much attention.

The investigation of such specimens as have been obtained has been carried on
along very different lines by the several authors who have studied them, and conse-
quently it is often a matter of no little difficulty to correlate specific diagnoses.
Thus Bornemann # based his specific determinations, to a large extent, on external
shape, and to that end employed thin sections of the specimens, not necessarily
cut in any particular direction through the organism. Taylor t employed several
features in his diagnoses, chief among which is the diameter of the pores which form
such a marked feature in all the skeletal elements. It is only fair to state that
Bornemann also refers on occasion to the pores, but in this respect the plates
which illustrate the memoir are more useful than the text. By a careful measure-
ment of his photographs, a fairly accurate idea of the porous structure may often be
obtained, even although it is not specially mentioned in the text.

The method of investigation has depended greatly on the mode and state of
preservation. Where the specimens are weathered out from the matrix, direct

* Bornemann, “ Verstein. Sardinien,” Nova Acta der Ksl. Leop. -Carol. Deut. Akad. der Nat., vol. li and vol. lvi,
Halle, 1886, 1891.

t Taylor, “ Archaeocyathinse from Cambrian of South Australia,” Mem. Roy. Hoc. S. Australia, vol. ii, pt. ii, 1910.
TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 27). 106

682

DR W. T. GORDON ON CAMBRIAN ORGANIC REMAINS

measurement was possible ; where the skeletal structures were preserved in silica
in a calcareous matrix (as in the South Australian types described by Taylor), the
details were etched out by means of acid ; where the skeletons were enclosed in the
matrix, thin sections or polished sections were employed. There is little doubt that
the preparation of thin sections is the best means to elucidate the minute details,
but there are difficulties, as pointed out by Hinde,# in obtaining a proper orientation
of such sections. This is seen in a marked degree in Bornemann’s memoir, where
many of the figures are from very oblique sections.

In the present instance the material was very scanty, and to prevent waste small
sections of individual specimens were prepared. With a certain amount of care in
the orientation of these, it has been possible in most cases to show the following
characters : — -

(a) The nature of the fixation of the skeleton to the substratum.

( b ) The appearance in cross-section at or near the top of the specimen.

(c) The size and arrangement of the pores on the walls, septa, and tabulse.

As mentioned above, the material was not abundant, and in some cases all these
characters have not been observed. In certain genera emendations have been made
on the descriptions of species already recorded, and a few new species have been
added to the group.

Before passing to a detailed description of the forms examined, it is necessary
to devote a word or two to the material investigated and the locality from which it
was obtained. All the specimens were enclosed in small pieces of a white limestone
broken from a block which was dredged from a depth of 1775 fathoms in the Weddell
Sea, in lat. 62° 10' S., long. 41° 20' W., Station 313 (text-fig. 2, X). When the fragments
were glued together, prior to being cut into sections, they formed two pieces each under
a pound in weight. Both must have belonged originally to the same mass, though
it is doubtful if they represented the whole of the original block, since they did not fit
exactly together. This is worthy of record, as erroneous statements have been made
regarding the amount of material available.! The outer crust of the fragments
showed considerable weathering effects, so that a layer of white powdery material
fully one-eighth of an inch thick encrusted the whole external surface.

The position in the north of the Weddell Sea from which the boulder was
dredged cannot be considered as marking the position of any limestone band,
for the block occurred among other debris evidently derived from a deposit of
material dropped by melting icebergs. The ocean drift near the locality is to the
north and east, so that the bed from which the material was originally torn must
lie further into the Antarctic continent. (Other forms of Archseocyathinse have been
recorded from the Beardmore Glacier moraines, and were probably carried from a

* Hinde, G. J., Q.J.G.S. , vol. xlv, p. 133, footnote,
f British Association Report : Meeting in Australia, 1914, p. 413.

FROM A DREDGING IN THE WEDDELL SEA.

683

point about 85° S.) No evidence has been obtained, however, as to how far to
the south of 62° 10r S. we might expect to find the band from which the material
was derived.

Although the specimens were dredged in 1903-4, it was not until 1913 that
Dr Bruce handed the material over to me ; meanwhile Archaeocyathinae had been
recorded from the Beardmore Glacier moraines by Professor David, Priestley,* and
T. G. Taylor.! None of their specimens, however, are specifically determinable with
certainty ; indeed, it is' open to doubt if they are generically recognisable. That they
do belong to the group is perfectly certain, and two families are represented —
Archaeocyathidae and Spirocyathidae, — according to Taylor’s investigations.

In the present instance most of the forms are well preserved, and specific
determination is perfectly easy as far as the material is concerned. The skeleton
consists of granular calcite, and the matrix of rather turbid crystalline calcite.
Occasionally the original skeleton has been leached out and subsequently replaced
by clear crystalline material. In such cases the details of the skeletal structures
may be obliterated, but in many examples the turbid matrix has remained and pre-
serves the porous structure at least. All types of preservation are often distinctly
seen in the same specimen. A section of Coscinocyathus endutus is represented
in PI. I, fig. 1. It is cut to pass vertically through the inner wall, which is shown
along the centre of the figure. The outer wall with its pores occurs at c, while
between the walls are the interseptal loculi. An examination of the inner wall at
a proves that the original wall is present, and the pores are filled in by the
cloudy matrix ; but at b the wall has disappeared, and the space is filled with clear
crystalline calcite. The pores are indicated, however, by the matrix which originally
filled them in. At various parts of the same figure the outer wall, septa, and tabulae
have been similarly destroyed, but no difficulty arises in the determination of these
structures.

In one or two examples exposed on the outside of the block, the less resistant
character of the granular skeleton is again indicated. The pores of the outer wall
are then represented by papillae (PI. I, fig. 2, along both sides of the skeleton in the
lower part of the figure). Such papillae are really the portions of the matrix which
filled in the pores, while the spaces between the papillae were originally occupied by
the granular calcite of the wall. Along the septa as at a of this figure the apparent
pores are really where the solid portion existed, but this has been dissolved out, and
the matrix is left as a continuous bridge between adjacent loculi. A similar feature
may be noted across the tabulae at b. Even in such cases, however, no difficulty
occurs in interpretation.

In addition to the remains of Archaeocyathinae, other organisms are represented
in the limestone. Fragments of shell and the carapace of trilobites occur, but-

* Compte rendu du XT Congres Geologique International, 1910, p. 774.

f British Antarctic Expedition, 1907-9, “ Geology,’' vol. i, pp. 235 et seq.

684

DR W. T. GORDON ON CAMBRIAN' ORGANIC REMAINS

clusters of bifurcating tubules referable to the Algse often occupy the greatest bulk
of the rock. Indeed, it would more properly be called an algal limestone containing
Archseocyathinae, etc. More important than the shell and trilobite fragments are
spicula probably referable to the Porifera. They are not abundant, nor are they
well preserved, but they' compare very well with similar spicules from the Beardmore
Glacier moraines, as will be seen later.

Detailed Description of the Specimens.

PLANTS.

Algse.

Epiphyton, Bornemann, 1886.

Epiphyton fasciculatum, Chapman.

1910. Solenopora? Priestley and David, “Notes on British Antarctic Expedition, 1907-9,” pp. 775-777,
tig. 7, Compte Rendu du XIe Gong. Geol. Inter. Stockholm.

1914. Epiphyton flabellatum, Chapman, David, and Priestley, Report of the British Antarctic Expedition,
“Geology,” vol. i, p. 241.

1916. Epiphyton fasciculatum, Chapman, Report of the British Antarctic Expedition, “ Geology,” vol. ii,
pp. 81-83.

The most common fossil in the material consists of clustered groups of small
radiating and bifurcating tubules (PI. I, fig. 4, a). In transverse section the tubules
are circular and sometimes closely adpressed. As a result a typical section is very
similar to that shown in PI. I, fig. 16, and the identity of this form with that figured
by Priestley and David from the Beardmore Glacier moraines is apparent when we
compare this figure with that published by these authors under the name of a
“ Solenopora-like organism.” A close examination of their figure shows that the
tubules bifurcate, and a similar phenomenon is clearly indicated in the present
material (PI. I, figs. 3, 4, and 4a). In the last two figures, at a, repeated dichotomy
of the thallus and the swollen ends impart a distinctly fucoid appearance to the
whole. Indeed, the algal form of this type is much more pronounced than in any
Solenopora. The nearest approach to the form is probably Ortonella, as described
by Professor Garwood from the Carboniferous Limestone of North-West England.

Chapman has examined the material from the Beardmore Glacier in more detail,
and in the Report of the British Antarctic Expedition, 1907-9 (“Geology,” vol. i,
1914), has correlated it with Epiphyton flabellatum, Bornemann ; but in a later
contribution (“Geology,” vol. ii, 1918), has described it under a new specific name,
E. fasciculatum, and given three figures. These figures are not very good, but the
description and measurements leave no doubt but that the form recorded here is
identical with the Beardmore Glacier type.

FROM A DREDGING IN THE WEDDELL SEA.

685

Epiphyton grande , sp. nov.

An occasional group of a much larger algal form occurs with Epiphyton fascicu-
latum. The relative size is clearly seen in PL I, fig. 3, when we compare b with a.
In the former type the tubule is about twice the diameter of that in E.fasciculatum,
and the distance between the bifurcations is also greater. Consequently a much
more lax appearance is imparted to the whole thallus. Another specimen is repre-
sented in PI. I, fig. 5, and here the branching character is particularly well exhibited.
Occurring under precisely the same conditions as E. fasciculatum, the idea that we
have here merely more robust examples cannot, I think, be admitted. If it were so,
we should expect to find intermediate forms ; but only the two sizes of tubule occur.
One is reminded in this connection of the two sizes of tubule in Micheldeania, but,
in that genus, the two forms occur intergrown with one another, and not distinct and
separate as in this case. While, therefore, one has a certain diffidence in erecting a
new species where size is the chief diagnostic character, one feels that it is justifiable
here. The new species, then, has tubules twice the diameter of those in E. fascicu-
latum , and the bifurcations are twice to three times as far apart as in that species, so
that either in transverse or longitudinal section the grouping is less compact.

Compared with E. flabellatum, Bornemann, the new form is very similar as
regards diameter of tubule. Indeed, taking Chapman’s figures, E. jiabellatum comes
between E. fasciculatum and E. grande as regards this measurement. A greater
distinction is visible when we consider the bifurcations. In E. jiabellatum and
E. fasciculatum the bifurcations are close together and only differ in minute detail,
while in E. grande the dichotomies are further apart and as a consequence the whole
tuft presents a much more lax appearance.

There are now three species of Epiphyton recorded from Cambrian rocks, all
closely similar and quite distinct from any other alga yet recorded. Von Toll’s
Confervites primordealis, Bornemann, from Siberia, seems almost certainly an
Epiphyton, but details are lacking which would assist in placing the specimens in
any of the above categories.

ANIMALIA.

Porifera.

Spicules probably referable to this group occur occasionally in the sections, but
unless the plane of section happens to pass through or near the point of junction
of the various rays it is difficult to determine the precise type of spicule. Then
again the original material has been removed in every case and casts in crystalline
calcite have been formed.

The most common type is that represented in PI. I, figs. 8 and 9. The spicule
is large and the rays seem to unite in a flat head slightly domed in the centre

686

DR W. T. GORDON ON CAMBRIAN ORGANIC REMAINS

(figs. 8 and 9, a). The number of rays present in any example is small, but the
angle of divergence between them is such that the -complete spicule was probably
heteractinellid. They are siiiiilar in size and appearance to those recorded by
Taylor from South Australia and from the moraines of the Beardmore Glacier.

Another multi-rayed example is seen in PI. I, fig. 6, and again the complete
form is rather difficult to determine. It had certainly not fewer than six rays,
but it could not be a hexactinellid spicule, since the angle of divergence of the rays
cannot be reconciled with any possible section of a true hexactinellid. It might
be said that a section through the centre of a hexactinellid type would approximate
to the figure, but, as at least two examples have been observed both almost exactly
alike, the chances are that they are sections of another type of heteractinellid
spicule rather smaller than the type of figs. 8 and 9.

We pass now to consider forms of spicule which may be classified with greater
ease than the above. The first of these is undoubtedly tetractinellid (PI. I, figs.
7 and 11). In fig. 7 the plane of section is oblique, but all four rays are indicated,
though only one is complete. The “calthrop” form, however, is very evident.
The second figure (fig. 11) presents a much more perfect example, since the plane
of section has cut through almost perpendicular to one of the rays, and the other
three diverge at the normal angle in calthrop spicules. The vertical ray is
indicated at a, where a distinct convexity appears in the angle between the
two rays.

Examples of hexactinellid spicules also occur, and two of these are figured
(PI. I, figs. 10 and 14). The great characteristic of such forms is that the rays
are set with respect to each other at angles which are multiples of 90°. Lastly,
several very characteristic specimens like those shown in PI. I, figs. 12 and 13, recall
a similar example figured by Taylor from the Beardmore Glacier locality. His
specimen certainly bears some resemblance to the spicule of the recent Lelapia,
but it is really not such a simple form. An examination of his figure * proves that
at the base of the fork an upward projection appears, and indeed that a perfectly
horizontal section through it would have looked more like that shown at a of
PI. I, fig. 12. This fortunate section is exactly in the plane of the spicule, and
several other specimens occur in the same section. A rather poor example is seen
at b of the same figure (PI. I, fig. 12), where only the forked ray is visible. Another
example is figured in the same plate at fig. 13, but again the plane of section has
passed obliquely across the spicule. The two rays passing to the right and left
of the forked one at once dissipate the resemblance to Lelapia, but the fact that
the rays are always at right angles to one another would tend to place this form
among the hexactinellids. Further, it is quite probable that rays occurred above
and below, i.e. running perpendicular to the plane of section, in which case we
may relegate the spicule to a hexactinellid type where one of the rays has become
* British Antarctic Expedition, 1 907-9. pi. Ixxix, fig. 5.

FROM A DREDGING IN THE WEDDELL SEA.

687

bifid. Similar modifications are not unknown among modern hexactinellid spicules,
but none quite like that of fig. 12 has ever been recorded.

The sponge remains, or at least the spicules which are probably sponge remains,
thus parallel in a remarkable manner the resemblances shown among the algse from
the Beardmore Glacier blocks and the mass dredged in the Weddell Sea (positions
almost opposite one another with respect to the South Pole).

Archseocyathinae, Bornemann, 1884.*

The most important organic remains in this limestone can be definitely placed
among the Archseocyathinse. Several families are represented, and the different
species may be grouped as follows —

Archaeocyathidse, Taylor, I9i0.t
Arcideocyathus, Bqllings, 1861.J
Archasocyathus pauciseptatus, sp. nov.

One exceedingly small specimen is worthy of notice, since it is the only example
in the material which approximates to the type of A. profundus , Billings. One
transverse section was obtained, and it is figured in PI. VI, fig. 63. The imper-
forate outer lamina on the outer wall, the straight septa, and the irregular but
numerous dissepiments separate the type from Spirocyathus. The septa are few in
number and perforated as seen in transverse and longitudinal section (PI. VI, fig. 63,
p, and fig. 64, p, respectively). The dissepimental platforms are very oblique and
irregular. Their inner ends are continued across the central cavity and unite in an
irregular tissue (PI. VI, fig. 64). Smaller offsets from these help to fill in the whole
central cavity. The inner wall is porous, but lack of material has prohibited a
more accurate description. The only other form which approximates to the
above is A. dissepimentalis, Taylor, but that species has more numerous septa and
no tissue in the centre of the cup.

Thalamocyathus, gen. nov.

Archseocyathidse with septa straight, perforate or imperforate. No tabulae or
dissepiments cross the interseptal loculi. Walls simple or modified. Central cavity
frequently filled by callus growth in the older part of the skeleton, i.e. near the
base. (The greater number of the species formerly in the genus Archaaocyatlius
will . fall into this new division. The reasons for erecting a new genus will be
discussed below.)

* Geol. Zeitschr., 1884, p. 706.

f Mem. Roy. Soc. South Australia, vol. ii, part ii, p. 105.

| Geology of Canada, “ New Species of Lower Silurian Fossils,” p. 3.

688

DR W. T. GORDON ON CAMBRIAN ORGANIC REMAINS

Thalamocyathus tubavallum, Taylor, sp.

1910. Arcliseocyathus tubavallum , Taylor, Mem. Roy. Soc. South Australia, vol. ii, part ii, p. 123,

text-figs. 28, 29.

One small specimen from which only three transverse sections were prepared
could probably be referred to this species. Although Taylor only obtained a
small fragment of the form, its characters are fairly well marked. The inner
wall is thick, and consequently the pores seem to be short pipes. The outer wall has
smaller pores, and is much thinner, while the septa are imperforate. Unfortunately
Taylor only published drawings of this form, and so correlation is difficult. One
section of the present specimen is given in PI. I, fig. 15. The thick inner wall is
marked, as also are the small pores on the outer wall. The septa are imperforate,
except for stirrup-pores at their inner ends. These are clearly seen at a and b of
the figure above mentioned. As no longitudinal sections could be prepared, the
diameter of the pores cannot be clearly seen. Those on the inner wall seem to
average 7 per mm., and those on the outer wall 10 to 11 per mm. The measurements
correspond fairly well with Taylor’s text-fig. 29.

Thalamocyathus trachealis, Taylor, sp.

1910. Arcliseocyathus trachealis, Taylor, Mem. Roy. Soc. South Australia, vol. ii, part ii, pp. 125, 126,

pi. i, iii, v, viii, text-fig. 22.

Three of the specimens discovered could be referred to the above species,
and, as the material allowed of examination of minute details, Taylor’s description
of the species will be emended to some extent, and consequently the new examples
must be described in considerable detail. The largest specimen was probably
2 '5 to 3 cm. in length, 1 cm. in diameter at the top, and tapering gradually to a
point. An oblique section near the top' of the cup is shown in PI. II, fig. 18. The
great number of closely set septa is at once apparent, while the large pores of the
inner wall and the small pores of the outer (notably at a) are also visible. In
fig. 17 a much smaller specimen is represented in transverse section, and a distinct
feature is the breadth of the inner wall, although in some places it has disappeared
entirely. This latter feature is still more obvious in PI. II, fig. 20, where at a
the wall is very broad, whereas at b only the inner rim is visible, and the septa
seem to stop far short of the wall, while their inner ends have become thickened.

The reason for all these features is indicated in fig. 20. The apparent inner wall
is really a broad ring of no great thickness, and arched so that a gutter is formed
on the upper surface. Should the plane of section correspond with that of the ring
a complete platform is seen, but in general the broad ring is only partly coincident
with the plane of section, and in the other parts the margin, or part of the margin,
is alone included. (See also fig. 19.) This will be even more easily understood by
reference to fig. 23, where a longitudinal section of the same specimen is represented.

FROM A DREDGING IN THE WEDDELL SEA.

689

Here, the outer wall — a continuous, porous lamina — is seen at a ; and at b the inner
wall is evidently not continuous, but a comparison of figs. 20 and 23 clearly
shows that it consists of a series of broad rings set one above the other and soldered
to the septa. As mentioned above, the rings are not thick and are grooved on their
upper surfaces.

At c (fig. 23) a septum appears cut longitudinally, and near the top of the
figure the attachment of four rings to this septum is shown. A longitudinal
section at right angles to the plane of fig. 23 demonstrates that the pores of the
inner wall are delimited by the vertical septa and the horizontal ring-platforms
(PI. II, fig. 22, a). This figure represents the vertical septa at d crossed by the
horizontal platforms b, the spaces between the intersections constituting the pores
(a). Such pores ought to be square in section, but increased growth at the fused
corners, and to a certain extent all round the pores, has given a circular shape to
the aperture. Such increase in thickness of the septa in this region accounts for
the swollen inner ends so apparent at a in fig. 19. Eeverting to fig. 22, however,
it seems difficult at first sight to determine which are the septa and which the
ring-platforms, but a closer examination soon indicates a means of distinction
even in such a section. At c in that figure, and again at a in fig. 19, a series of
small dots may be noted which are in each case continuations of the platforms.
The explanation of this row of dots is given by fig. 23, where (at b) on the under
surface of each platform a small projection is shown. Now, turning again to
fig. 19, the origin of the row of dots is quite clear. They represent a series of
small peg-like projections or denticles on the under surface of each platform. They
are again shown, more highly magnified, in PL III, fig. 26, where at a the denticles
are very evident, at s the swollen end of a septum is equally so, and aty? a septal
pore is conspicuous. The denticles are so close together that in the thickness of
a section more than one would be included, or, to put it otherwise, no section
could be cut thin enough to pass between two denticles; consequently any
longitudinal section of the ring-platforms will always show the denticle on the
lower surface of each.

Unfortunately, such a characteristic feature is not mentioned by Taylor in his
original description of the species, but this was due to the method he employed
of etching the skeleton out by dissolving the matrix in hydrochloric acid. Such
minute denticles were broken off. That such must have been the case is shown
in PI. II, fig. 24, where a longitudinal section is represented which was prepared
from material kindly given me by Mr Taylor some years ago. The identity
of this type with that of fig. 23 is undoubted. At b, in each, the denticles are
shown, at c, a porous septum, and at a, the finely porous outer wall. It is true
that the Australian specimen is distinctly more robust and larger (a point which
will be referred to later), but there is no doubt that the two are specifically
identical.

TRANS. ROY. SOC. EDIN., YOL. LII, PART IV (NO. 27).

107

690

DR W. T. GORDON ON CAMBRIAN ORGANIC REMAINS

So far the pores on skeletal members other than the inner wall have only been
mentioned incidentally : it remains to examine the porous Structures in more detail.
On the outer wall the pores are much smaller than those on the inner, there being
8 per mm. as compared with 5 per mm. On the outer wall the pores are arranged
in rows, two or three to the interseptum, but they alternate in the rows, and thus
an hexagonal outline is impressed on each individual pore (PI. II, fig. 21).

Although only a rather oblique section of a septum has been obtained, it will
suffice to give some idea of the arrangement and size of the septal pores. These
latter seem (PI. II, fig. 23) to run in vertical lines, but rather more irregularly than
the pores of the inner wall. That these apertures are not in absolutely parallel
rows may be proved by reference to transverse sections (PI. II, fig. 20, e), where
the pores on one septum are not exactly opposite those on its neighbours. As com-
pared with the Australian specimen represented in fig. 24, the arrangement of pores
is identical. In every particular except size this type agrees with Thalamocyathus
trachealis as described by Taylor or shown by the section of his material which
is figured (PI. II, fig. 24).

In one particular, however, these new forms differ from the Australian species
with which I have grouped them. That is the occurrence in the interior of the
cup of a coralloidal type of infilling material. The resulting tissue is very evidently
connected with the inner, walls (PI. II, fig. 25), and similar material has only beem
recorded in a few types. Bornemann, Hinde, and Taylor have each noted such
a feature in some type or another, but only in Taylor’s Somphocyathus does it
become prominent. In that latter genus the general form (coralloidal) is quite
comparable with that discovered in T. trachealis. The tissue has been noted in
two examples, and in each was found near the base of the cup. As will be seen later,
a similar infilling has been observed in practically all the W eddell Sea specimens ;
it only occurs near the base of the cup ; it differs in type, and so may be diagnostic
in value, but, from the position in which it occurs, I am inclined to consider it a
gerontic feature, since wherever it has been noted it is always present in a region
where the walls are thickened and the pores more or less closed by callus growth.

No evidence of rooting processes has been obtained. In the lowest section
available the diameter was -9 mm. and there were 12 septa, but it could not have
been far above the extreme point, and any processes below this level could not
have been very effective in balancing the whole cup. It may be that the form
had no special rooting processes. As regards details of measurement, the largest
specimen was 25 to 30 mm. in length, tapering to a point. The “ intervallum
coefficient,” i.e. the ratio of width of intervallum to width of central cavity, has
been found of little use. In PI. II, fig. 17, it is 1/1 ; in PI. II, figs. 20 and 25, 1/1 '5 ;
in PI. II, fig. 18, 1/4.

An emended description based on the above new facts . is as follows : — Cup, in
small specimens, a tapering cone, in larger forms more cylindrical. Constrictions

FROM A DREDGING IN THE WEDDELL SEA.

691

at intervals on larger individuals. Length up to 90 mm. Cup ends in pointed
spitz, anchoring material not preserved or . wanting. Outer wall a delicate porous
lamina ; pores 8 per mm. Wall sometimes invested by a thin layer almost closing
the pores. Septa numerous, stout, regular, with perforations in rows and about
4 per mm. in the row. Aperture of perforation small, hence they appear more
remote from one another than the pores of inner or outer walls. Inner wall of
ring-platforms firmly fixed to inner ends of septa in a vertical series, grooved on
top. Pores delimited by septa and platforms ; 5 per mm. Below each shelf a
row . of closely set denticles forming comb-like structures partially closing over
the pores. Base of central cavity, i.e. earliest part of cup, filled in by coralloidal
growth from the platforms.

Thalamocyathus infundibulum, Bornemann.

1884. Archseocyathus infundibulum , Bornemann, Geol. Zeitscbr., p. 703.

1886. „ ,, ,, “ Yersteinerung Sardinien,” Nova Acta der Ksl. Leop.-

Garol. Deut. Akad. der Nat., vol. li, p. 52, pis. ix,
x, xi, xiv.

Among the smaller forms obtained, one, represented in PI. Ill, fig. 30, can be
referred to the above species. In the only section prepared the cup is cut longi-
tudinally, and the pores of both inner and outer walls are shown in places. These
are respectively 4 per mm. and 10 per mm., numbers which correspond with those
given in Bornemann’s original description. The septa do not seem to be porous.
The shape is conical and also conforms to the original type. Some of the small
transverse sections, which occur here and there in the material, may possibly be
referred to this species, but there is no certainty in the matter until longitudinal
sections are available. No rooting structures are visible in the specimen obtained.

Thalamocyathus ichnusse, Meneghini.

1881. Archseocyathus ichnusse, Meneghini, Atti della Societa Toscana, p. 201.

1884. ,, „ Bornemann, Geol. Zeitschr., p. 704.

1886. „ „ ,, “ Yerstein. Cambs.,” Nova Acta der Ksl. Leop.-Carol. Deut.

Akad. der Nat., vol. li, p. 58, pis. ix, x, xi, xii, xiii, and xiv.

Another of the forms of which only a single example has been found is T.
ichnusse. As in the case of T. infundibulum, the sole section is a longitudinal one.
Fortunately, the walls and the septa are shown in places, so that the details can
be determined with fair certainty. In PI. Ill, fig. 31, this section is represented,
and it evidently expanded greatly towards the top, as do the Sardinian examples.
By measurement and calculation from Bornemann’s figures and statements, an
estimate of the number and arrangement of pores on the skeletal elements of his
type has been attained. The inner wall has 6 or 7 pores per mm., while the outer

692

DR W. T. GORDON ON CAMBRIAN ORGANIC REMAINS

wall lias 6 rows of pores per interseptum, which is '5 mm. wide.* Thus there are
10-12 pores per mm. on the outer wall. The septa also are porous.

The Antarctic specimen has 7 pores per mm. on the inner, and 10 pores per
mm. on the outer wall ; the septa are porous, and the general shape seems to
conform to that of the Sardinian species. All these characters confirm the reference
of the type to T. ichnusse. Since both T. infundibulum and T. ichnusae are re-
presented by single specimens, and by single sections of these examples in each
case, the specific features have been a matter of deduction rather than observation.
The clearest evidence is the size of pore, and unfortunately Bornemann laid little
stress on this point. Yet it is probably much more constant and important than
the external shape, a character often rendered useless by distortion. Indeed, some
re-examination of the Sardinian types is urgently required.

Thalamocyathus flexuosus, sp. nov.

Under the specific name of sinuosus, Bornemann! has described a Thalamo-
cyathus type with narrow intervallum and characteristic plicated outline in trans-
verse section. A well-marked feature in addition to the habit is the presence of
large pores on the inner wall and small ones on the outer, there being 3 per mm.
and 10 per mm. respectively 4 The great difference in size is also seen in other
figures illustrating the work.

A rather similar form occurs in the Weddell Sea material, as a number of small
fragments but no complete cup has been obtained. As a result of the flexuous outline
(PI. Ill, fig 27, a), it has been a matter of some difficulty to determine which is the
outer and which the inner wall. The intervallum is narrow — about '5 mm. — and as
the septa are also closely set the interseptal loculi are nearly square in transverse
section (PI. Ill, fig. 27). On account of the exceedingly narrow intervallum no
very good longitudinal sections of the wall could be made in series, but the two
figured (PI. Ill, figs. 28 and 29) are from successive sections, and illustrate in some
way the porous character of the walls. The first point to be noted is that the pores
are almost equal on each wall, a feature only recorded so far from specimens of
Australian origin — the “ equivallum ” type of Taylor. By measurement, however,
it is found that on one wall the pores number 6 per mm., and on the other 8 per mm.,
so that a slight difference does occur. Another point of difference lies in the fashion
in which the septa meet the two walls. In one case — That with the smaller pores —
the septa meet the wall between rows of pores (PI. Ill, fig. 29, s ), but they meet the
other wall along a pore-row, i.e. there are “ stirrup ’’-pores, the one perforation
allowing communication from the exterior to two adjacent loculi (PI. Ill, fig. 28, b).
The septa appear to be imperforate apart from the stirrup-pores at their ends. Since

* Bornemann, loc. cit., p. 55.

t Bornemann, loc. cit., p. 57 ; and Geol. Zeitschr., 1854, p. 704.

X Calculated by measurement from Bornemann, loc. cit., Taf. xii, figs. 6a, 66.

FROM A DREDGING IN THE WEDDELL SEA.

693

the wall with the stirrup-pores has also the larger apertures, I incline to regard it
as the inner, and that with the small pores as the outer. As mentioned above, no
complete cups have been obtained, but the fragments exceed in length any other
form present in the material. Even if the broken specimen of fig. 27 had been
complete, it would have constituted the largest specimen in the whole of this
collection.

The species may be described as follows : — Walls folded gently ; inter vallum
narrow (‘5 mm.). Outer wall simple, porous, pores 8 per mm. Inner wall simple,
pores 6 per mm., “ stirrup ’’-pores on inner ends of septa. Septa imperforate except
for stirrup-pores.

Spirocyathidse.

Spirocyathtjs, Hinde, 1889.*

Spirocyathus atlanticus (Billings, sp.).

1861. Archseocyatlius atlanticus, Billings, Lower • Sil. Fossils, p. 4, figs. 1-3.

1865. „ ,, „ Palse. Fossils, vol. i, p. 5, fig. 5.

1880. ,, „ Roemer, Letheea Geognostica.

1886. „ „ Walcott, Bull. No. 30, U.S. Geol. Surv., p. 75.

1889. Spirocyathus „ Hinde, Q.J.G.S., vol. xlv, p. 136, pi. v, figs. 8-10,

1889. „ ,, Walcott, 10 if A Annual Report, U.S. Geol. Survey.

1899. ,, sp., Yon Toll, “Sibir. Camb.,” Mem. Acad. Imp. des Sc. Petrograd, ser, viii, vol. viii,

p. 45, pi. vi, fig. 8.

1910. ,, radiatus Taylor, Mem. Roy. Soc. S. Australia, vol. xi, part ii, p. 147, text-fig. 35.

1910. „ irregularis, Taylor, Mem. Roy. Soc. S. Australia, pp. 148, 149, pi. xvi, figs. 93, 94.

Two examples of the above form were discovered in the material, and each was
about 20 mm. in length. A transverse section, fairly high in the cup, is given in
PI. Ill, fig. 32, and it compares well with the figures of the type specimen as given
by Hinde. As in that specimen, the skeletal structures have been thickened by a
coating of material, and the appearance is analogous to that produced on the septa
of many rugose corals (cf. Aulophyllum) by “ callus” growth. But the chief points
of interest in these new specimens lie in the nature of the fixing organs or processes,
and in the apparent similarity to the genus Protopharetra which is shown by the
lower parts of the cup.

A cross-section at a lower level in the skeleton is shown in PI. IV, fig. 39, and
the whole is very suggestive of a section from a Protopharetra type. In this
connection it is interesting to recall Hinde’s words when describing the new genus
Spirocyathus. He notes that he has “ considerable difficulty in determining whether
Spirocyathus could be established as a genus distinct from Protopharetra .” t
Bornemann, in erecting the new genus Protopharetra, considered that it was not
an independent type, but merely the lower part or rooting part of other forms of

* Hinde, Q.J.G.S., vol. xlv, 1889.
t Hinde, loc. cit., p. 138, footnote.

694

DR W. T. GORDON ON CAMBRIAN ORGANIC REMAINS

the Archseocyathinae.* A consideration of the specimens now figured indicates
that the early stages in Spirocyathus and Protopharetra are very similar ; but I
believe they can be distinguished, and hold with Taylor t that they may be
independent forms belonging to the same group.

In the above-mentioned figure (fig. 39) there is another important feature, namely,
the occurrence at a, b, and c of small structures not unlike the cup itself, except
that there is no central cavity. The structure at b is in continuity with the
thickened wall of the cup, but the others are separated from it. Similar continuity
between such a strand and the cup is indubitable in PI. IV, fig. 40, a. These con-
stitute the rooting processes of this form. Here (fig. 40) the Spirocyathus character
is very evident in the cup, and it is the second section- above that of fig. 39. The
lower half of this figure still shows a remarkable resemblance to Protopharetra , but
the upper part to Spirocyathus. Even in PI. IV, fig. 41, the Protopharetra affinities
are still marked, though the regular septa near the inner wall place it with Spiro-
cyathus cf. S. irregularis , Taylor. PI. IV, fig. 37, shows a section at a still higher
level in the cup, and the relation to Spirocyathus atlanticus is now evident, although,
in parts, the Protopharetra character persists. The central cavity in figs. 39, 40,
and 41 is crossed by irregular strands which are clearly outgrowths from the inner
wall and continuous with the septa.

The porous nature of the inner wall is shown in PI. IV, fig. 42, a and b. (The
figure is really on its side ; a is the top and b the bottom of the section.) The
central cavity is cut obliquely, so that the inner wall on the one side is exposed on
the right-hand side of the figure, and again at a higher level on the left-hand side.
The pores are evidently hexagonal in outline (a), but become more circular by
thickening round the edges, while rod-like synapticulse divide each pore horizontally
(PI. IV, fig. 42, a and b). The breadth' of the pore is the whole space between two
septa. The septa themselves are so porous that the wall is a meshwork of rods

rather than a porous lamina. A section through the wall thus shows a row of

circular dots, some of which may be connected when the section has coincided with
the rod-like portion of the wall between adjacent pores (PI. IV, fig. 42). This is also
indicated at a and b, fig. 38.

Another interesting feature is the occurrence of dissepiments in the central

cavity near the base of the cup. In PI. IV, figs. 39, 40, and 41, the ends of the

septa are seen to cross the inner wall, and a certain amount of dissepimental tissue
to be developed between them. Some of the longer outgrowths from the septa
pass completely across the central space (figs. 40 and 41).

The discovery of the rooting processes in the above species is of interest,
especially as these processes are not very unlike the cup in character, and it is
possible that Bornemann had obtained some such similar forms, which he named

* Nova Acta der Ksl. Leop.-Carol. Deut. Akad. der Nat., vol. li, pp. 47, 48.

t Mem. Boy. Soc. S. Australia, vol. ii, part ii, 1910, p. 113.

FROM A DREDGING IN THE WEDDELL SEA.

695

Protopharetra, and then found that higher up a more regular type was attained.
Another explanation of these processes may be suggested, namely, that they are
really reproductive stolons which gave rise to new cups under favourable circum-
stances. This may or may not be the case, but they must have functioned in the
fixation of the cup round which they occur.

It 'will be noticed that Taylor’s two new species of Spirocyathus have been
merged in the prototype, and the reasons for this will be seen by reference to the
figures now published. The chief 'diagnostic characters in S. irregularis, as given
by Taylor, are that the septa are much more radial near the inner wall and that
they are very irregular towards the outer wall. This feature is beautifully shown
in fig. 41, but, as stated already, fig. 40 is from the same specimen two sections
below the level of fig. 41, and in the upper half of that figure (fig. 40) the septa
are radial right out to the outer wall.

Again, PI. Ill, fig. 32, and PI. IV, fig. 37, are from the same specimen. In the
former, radial symmetry is almost obliterated, but in the upper part of the latter
figure this radial character of the septa is very prominent.

Similarly, S. radiata, Taylor, is merely a phase which occurs in places in each of
the Weddell Sea forms. I have therefore placed them all under the prototype. Van
Toll has figured a weathered specimen as Spirocyathus sp., and his figure compares
so well with that shown in PI. IV, fig. 38, that this specimen is also included
under S. atlanticus.

Protopharetra, Bornemann, 1883.#

The most common type present in the Scotia material is undoubtedly referable
to Protopharetra. Bornemann in his original description of the genus, considered
it as a generalised type of basal structure from which many types might ultimately
be developed, i.e. that different genera had similar basal structures. (An exact
analogy has been shown in the fossil Lycopodiales, where Stigmaria includes the
underground portions of Lepidodendron, Sigillaria, or Bothrodendron.) Taylor,
while accepting Bornemann’s conclusion as far as the Sardinian forms are concerned,
is emphatic in placing certain completed Australian types in the genus. Yet
Taylor’s genus Metaldetes bears out Bornemann’s contention, since near the base a
transverse section shows marked resemblances to Protopharetra. Both authors are
probably correct, and the genus may be a central form which became differentiated
along several lines.

Protopharetra polymorpha, Bornemann.

1884. Protopharetra polymorpha, Bornemann, Geol. Zeitschr., p. 705.

1886. ,, „ „ “ Verstein. Sardinien,” Nova Acta der Ksl. Leop.-Carol.

Deut. Akad. der Nat., vol. li, p. 46, Taf. v and vi.

A slightly oblique Section (PI. VI, fig. 66), which was obtained while making
transverse sections from another specimen, compares very well as regards dimensions

* Geol. Zeitschr., 1883.

696

DR W. T. GORDON ON CAMBRIAN ORGANIC REMAINS

with one of Bornemann’s figured types (Taf. vi, figs. 3 and 4). The vertical character
of the irregular septal plates is specially evident towards the base, where the section
is more tangential. In transverse section the septa are seen to form an irregular
meshwork (PI. VI, fig. 65), which is crossed by delicate curved plates forming
dissepimental tissue. The outer wall is well marked, as also is the inner. The septa
are plates with irregular vertical corrugations which form a series of inosculating
partitions between the two walls. Irregular emergencies near the base appear
to have fixed the whole cup (PI. VI, fig. 66). A closer examination of the inner
wall (fig. 65) shows that between the main septal plates small projections stretch
into the intervallum, so that the inner wall becomes very conspicuous. In PI. Ill,
fig. 36, an enlarged view of the cup of PI. VI, fig. 66, is given, and the inner wall
is now seen to be porous. These pores (p) are small, probably 3 or 4 to the
intersept, and this accounts for the presence, in transverse section, of the small
structures between adjacent septa. They are really the divisions between the
pores. But this figure illustrates another feature — -the central cavity is filled
by oblique cross partitions ( Boden of Bornemann) between which numerous
waved, vertical rods are clearly visible. Both the tabulae and the rods are out-
growths from the walls, and the whole completely fills in the central space. It
would be difficult, however, to call this a coralloidal growth such as has been noted
in Thalamocyatlius trachealis, Taylor, sp.

Protopharetra radiata, Bornemann.

1886. Protopharetra radiata , Bornemann, “ Yerstein. Sardinien,” Nova Acta der Ksl. Leop.-Carol. Deut.

Akad. der Nat., vol. li, p. 48, Taf. vii.

Several specimens of a rather larger form belong to this second species. The
largest cup was 1 cm. in diameter, and indeed it was the largest complete cup
obtained from the material. It is represented in PI. Ill, fig. 35, and illustrates
most of the features already described in P. polymorpha. The septa show a
certain amount of thickening, and there is no sign of infilling material in the cup.
Another specimen (PI. VII, fig. 72) presents a good example of the branching
in this species. The lowest section shows two cups joined together. Of these,
one, a, has a clearly marked inner wall and the central cavity crossed by two
tabulae, while the ends of a few vertical rods are also visible. The central cavity
of the second cup, b, is more completely filled, and consequently the inner wall
is not so prominent. In the succeeding section (fig. 73) a has disappeared, b is
cut at a higher level, and a third cup, c, with rooting process r, has joined itself to
b. In the third section (fig. 74) c is cut at a higher level (very similar to that
of a, fig. 72) and b has now disappeared. At d and e, however, sections of the
presumed fixing strands are evident. The only distinction between P. polymorpha
and P. radiata is one of size, the latter being larger in every way.

FROM A DREDGING IN THE WEDDELL SEA.

697

The question now arises, can we separate Archaeocyatlius, Spirocyathus, and
Protopharetra ? In the first place, it must be admitted that the greater number
of species placed under the heading Archaeocyathus are excluded at once from
consideration, as they contain no dissepimental tissue, the structure, which really
causes all the complexity. It seems better to institute a new generic name for
such simple forms, and the name Thalamocyathus has accordingly been adopted
for such examples. In discussing the affinities of Spirocyathus and Protopharetra
with Archaeocyathus , the particular species one has in mind is A. profundus. That
species is described more fully by Hinde than by Billings. The outer and
inner walls and septa are porous, the septa straight, and dissepiments oblique
and imperforate. Spirocyathus has septa which are much corrugated, and the
septal pores are large, a character also noticeable in Protopharetra. Dissepimental
laminae are not nearly so common in Spirocyathus as in Protopharetra , and the
inner wall is very distinctive, Protopharetra having several rows of pores between
the septa, and Spirocyathus only one row.

Thus far Protopharetra has been considered as a distinct genus. We now pass
to consider a similar type which gradually passes into quite a different form, and
hence partly confirms Bornemann’s view that Protopharetra was a type of rooting
structure rather than a genus complete in itself. Such examples are taken up
under different generic names, and Taylor’s Metaldetes is a convenient genus for
some types.

Metaldetes, Taylor, 1910.*

Metaldetes plicatus, sp. nov.

A type referable to Taylor’s genus Metaldetes is fairly common in small
irregularly folded fragments. Complete examples seem rare, but one such is
shown in PI. VI, fig. 68, The cup has been distinctly plicated, and it is attached
in part to a specimen of Spirocyathus atlantic’us, which is seen in oblique longi-
tudinal section near the top of the figure. There are certain well-marked characters
in this new species. The septa are very numerous, straight and porous (as is in-
dicated by the discontinuous course of the septa). The intervallum is very narrow,
varying from 1 mm. to 1'3 mm., but where the section is oblique it seems
much broader. Another distinct feature is the presence of numerous dissepiments,
which, however, seem better described as irregular tabulae. They are much
better displayed in longitudinal section (PI. VI, fig. 69), and each crosses
several loculi. The septa show here as vertical rods, again markedly discon-
tinuous. These discontinuities are explained by reference to PI. VI, fig. 71, a
and b. Here an oblique longitudinal radial section of the cup crosses several
septa, but the porous nature of these laminae is clearly indicated. The pores
average 4 per mm.

* Mein. Roy. Soc. South Australia, vol. ii, part ii, 1910.

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 27). 108

698

DR W. T. GORDON ON CAMBRIAN ORGANIC REMAINS

The inner wall is also porous, as may be seen in PI. VI, fig. 69, w.p., but still
more clearly, in fig. 70 of the same plate. In size these wall pores are very similar
to the septal pores, but the, wall is not so well preserved as are the septa. The
outer wall seems to be covered by a thin continuous lamina, but beneath this coating
a porous structure similar to that on the inner wall may be noted in favourable
sections. The dissepimental tissue is imperforate.

So far the specimens might quite well be placed in the genus Archaeocyathus,
but as they are cut at lower levels changes occur which place them quite outside
that genus. In PI. VII, fig. 75, there is shown what might be taken for a Proto-
pharetra with closely set septal plates, and on the upper side, a, a small rooting
process is attached. The next section is given in fig. 76, but now the septa are almost
straight from wall to wall. Rooting processes are again shown, but the cup
has become broken in outline. In fig. 77, from the section succeeding the last,
the septa are still straighter ; indeed, we now recognise the similarity to the cup
of PI. VI, fig. 68. At a of this figure a new cup is being formed from the wall
of the old one, and it becomes more noticeable at a, fig. 78. Thus the form
shown in PI. VI, fig. 68, can be traced downwards into a rooting portion which
is very similar to Protopharetra. This is precisely the character of Taylor’s
genus Metaldetes. It will be readily admitted that the form of fig. 75 cannot
be correlated with any of Bornemann’s species of Protopharetra, it is a much more
delicate type than any he has described ; but his contention that some forms of
Protopharetra are the rooting or basal parts of other genera of the Archseocyathinse
is substantiated by the above series.

Dictyocyathidse.

Dictyocyathus, Bornemann, 1891.*

Dictyocyatlius, sp. (PI. VI, fig. 67.)

Among the small cups scattered through the material and only represented by a
few sections is one shown in PI. VI, fig. 67. There seem to be large pores on the
inner wall, small ones on the outer wall, and very few septa joining the two walls.
Too few sections have been cut to admit of a distinct specific description, but it
constitutes a type quite different from any of the other forms obtained, and in all
probability is referable to Dictyocyathus. The walls and transverse septal processes
are thickened by subsequent layers of material, so that the outer wall becomes very
thick at a level three or four sections below that figured. The only longitudinal
section available shows the large pores of the inner wall ; those on the outer wall
are completely obliterated by the secondary growth.

* “ Verstein. Sardinien,” Nova Acta der Ksl. Leop.-Carol. Deut. Akad. der Nat., vol. lvi, 1891, pi. ii.

FROM A DREDGING IN THE WEDDELL SEA.

699

Syringocnemidse.

Syringocnema, Taylor. #

Syringocnema gracilis, sp. nov. (PI. V, fig. 49 ; PI. IV, figs. 43-48.)

The genus Syringocnema was instituted for certain South Australian forms of the
Archseocyathinae which were very exceptional in structure. The outer and inner
walls are not connected by septa, tabulae, or dissepiments, but by a series of almost
horizontal hexagonal tubes which had porous sides. Only one species was recorded,
and no other form has since been discovered. Among the Weddell Sea material,
however, four specimens were discovered which are indubitably referable to this
genus. It happens that the preservation is not so good as in most of the other
forms, but the essential characters can easily be demonstrated. The only rather
unsatisfactory part is connected with the porous structure of the walls of the
horizontal tubes.

A transverse section of the type (PL IV, fig. 43) is, at first sight, not unlike an
ordinary Archaeocyaihus , but should the section be slightly oblique (PI. IV, fig. 44)
a very characteristic appearance is imparted to the whole cup. The apparent septa
are seen to diverge and then converge in pairs. Where they diverge again from
each other a cross wall joins them. The explanation of this phenomenon is given by
oblique or tangential longitudinal sections (PL IV, fig. 47). Here the whole area of
the section is covered by hexagonal cells, each being a section of one of the tubes
which cross the intervallum. In longitudinal radial section it may be noted that
the tubes are not absolutely horizontal, and that they bend down at their inner
ends, and hence the apparent vesicular zone near the inner wall in some sections
(PL IV, fig. 44, a).

Referring back to fig. 43, it will be noticed that the outer wall has not the regular
outline usually seen in other Archseocyathinse, but that each tube has a curved end
convex outwards. In some cases (p, p, p) the end is more abrupt and projects
further from the general line, so that an irregular toothed outline is imparted to the
exterior. This again is best explained by longitudinal sections. In Pl. IV, fig. 45,
such a section is shown. It is oblique, so that it cuts deeper into the tubes at the
right-hand side, and is closer to the outer surface on the left. Proceeding from right
to left across the figure, we notice progressive thickening of the walls of the tubes,
and a consequent decrease in the lumen, until at the left-hand side the lumen is
rounded instead of hexagonal, and the walls seem very thick. Now in transverse
section (fig. 44) the walls are not much thickened, the apparent thickening being due
to the contraction of the mouth of the tubes towards the exterior, so that the pore
stands on the apex of a small conical projection from the general level of the wall
(text-fig. 1 b and c, o.p.). The surface, if it could be exposed, would thus be covered
by small papillae, each with a pore on its apex. In Syringocnema favus, Taylor, the
* Mem. Roy. Soc. South Australia, vol. ii, part ii, 1910.

700

DR W. T. GORDON ON CAMBRIAN ORGANIC REMAINS

surface was minutely papillate, but the papillae did not each correspond to a single
intervallar tube. In one part of one of the specimens the outside lay on the exterior
of the limestone, and a photograph is shown in PI. V, fig. 49. The hexagonal character
of sections of the tubes is well shown, but the level is too far below the actual surface
to indicate the papillate character of the outer wall. The material was too limited
to admit of sections being prepared to show the inner wall, etc. ; but such sections as

were made indicate that the
inner wall is not papillate
but smooth, and that the
pores are the inner ends of
the tubes.

The walls of the tubes
are also porous, but the pre-
servation is not very satis-
factory. The best examples
are in figs. 43 and 48, w.p.
Here the pores are seen to
be in two rows, and about
5 or 6 occur per mm. This
character alone would dif-
ferentiate S. gracilis from
S. favus, the only other
known species which has
three rows of pores on each
wall and 3 pores per mm.
The main distinctions, how-
ever, lie in the very narrow intervallum in S. gracilis , from 1 to 1-4 mm., and in
the narrow diameter of each tube, -4 to -5 mm. In S. favus the corresponding-
dimensions are 5 to 6 mm. and 1 mm. respectively.

The vesicular zone near the inner wall is not so marked in S. gracilis as in S. favus.
In the latter type it is due to the great length of the tubes so that, where they bend
down near the inner wall, several are cut through in any transverse section, and a
broad, apparently vesicular zone formed.

Another interesting point shown by the new species is a coralloidal type of infilling
towards the base of the cup, and an irregularity in the tubular arrangement. The
new species may be defined as follows : — Cup conical ; intervallum narrow (l mm.
to 1’5 mm.). Outer wall papillate, one pore on apex of each papilla. Pores lead
into tubes (maximum diameter ’5 mm.) hexagonal in section, lying horizontally
between the walls. End of tube bent down near inner wall. Inner wall smooth,
perforated, each pore the inner end of a separate tube. Apparent vesicles near inner
wall due to sections of bent part of more than one tube. Vesicular zone of one row

Text- Fig. 1.

a. Spirocyathus atlanticus. c, Syringocnema gracilis.

b, Syringocnema gracilis. d, do.

e, Coscinocyathus fultus.

PROM A DREDGING IN THE WEDDELL SEA.

701

only. Pores on tube walls in two rows, pores 5 or 6 per mm. Centre of cup filled
near the base by coralloidal tissue. All skeletal structures sometimes thickened by
secondary growth.

The specific name is due to the slender character of the walls and tubes.

Coscinocyathidse.

Coscinocyathus, Bornemann, 1884.

The genus as described by Bornemann is characterised by the presence of
horizontal tabulae which mark it off from other forms of the group. The pores
on the inner wall are defined as larger than those on the outer, but Taylor showed
that this is not a constant feature. The Australian species described by the latter
author have pores on both walls equal in size, and the Antarctic specimens conform
to the Australian type rather than to the Sardinian. These Weddell Sea specimens
fall into two categories, very distinct from one another and from all previously
recorded Coscinocyathidee. The mode of fixation of both types has been determined,
and on this character alone they are easily distinguished.

Coscinocyathus fultus* n. sp.

Although only one complete specimen of this form and the tip of a second have
been noted, the general character of the skeleton has been elucidated. The specimen
from which most of the skeletal details have been obtained lay partially exposed on
the side of the block. This weathered surface is illustrated by PI. V, fig. 50, where
the septa, the tabulae (; t ), and porous inner wall (i.w.) are all obvious. The tabulae
at once place the form among the Coscinocyathidae.

In transverse section most of the salient features become visible (PI. VI, fig. 60).
At a a small oval body, with its interior filled by convoluted partitions, is seen in
close association with the main cup, and in a lower section, represented by fig. 61,
several of these bodies are obviously connected to the surface of the cup. They
become confluent round the periphery, and form a zone of vertical partitions sur-
rounding the outer wall of the latter. Such offsets effect fixation of the whole
structure. They occur in several series, and as new series are added the outer wall
of the lower set is occluded, so that a series of walls, concentric with the outer wall
of the cup, crosses the zone of vertical partitions (PI. VI, fig. 60, p, p). These con-
centric walls are again seen in fig. 62, p, and it will be noticed that each reaches
a higher level on the cup than does the one below. The vertical partitions
are continuous across the concentric walls and are arranged perpendicularly to
the latter.

The question as to the function of these bodies is more easily answered in this
case than was possible in the case of Spirocyathus atlanticus. As far as Coscino-

* Lat. fultus, supported.

702

DR W. T. GORDON ON CAMBRIAN ORGANIC REMAINS

cyathus fultus is concerned, they appear to be organs of fixation, but they might also
act as stolons, as was suggested in connection with similar strands in the former
species. In S. atlanticus, however, there was no coalescence into exothecal lamellae.
It is interesting to note that such lamellae have been recorded in two other forms,
Archaeocyathus profundus , Billings, and A. Sellicksi, Taylor. In neither of these
cases has any confluence of small circular rooting structures been noted, but Taylor
states that new cups may arise from the exothecal mass (A. SellicJcsi), so that a
certain amount of evidence exists pointing to the possibility that these processes
function like stolons and produce new individuals.

Actual fixation has only been seen in one section, and this is shown in PI. Ill,
fig. 34. The cup is cut very close to the end, so that the central cavity c alone is
obvious. The outer wall, o.w., is also indicated, and round the whole are two
layers concentric with the walls. Attached at one side is a process, r.pr., which
is fixed to a fragment of the cup of Metaldetes plicatus at b.

On account of the mass of enveloping lamellae, and also since the walls and septa
show considerable thickening, the details of the pores are difficult to determine. In
longitudinal section (fig. 62) the pores of the outer wall are easily seen in several
places, but the pore lumen is much reduced by thickening of the walls. The
regularly placed tabulae are also well marked in this figure (fig. 62, t, t, t). The pores
run about 5 per mm. The porous structure of the inner wall is only shown on the
weathered surface (PI. V, fig. 50). There are two rows per intersept, and they number
5 per mm., as on the outer wall.

On account of the septal thickening, the porous structure on these elements has
not been ascertained, although the pores are shown perforating the primary lamella
(text-fig. 1 e, p). Tabulae are shown in both figs. 60 and 61. They are very finely
perforate ; indeed, they might be described as net-like, since the pores are quite
irregular in their distribution and closely pressed together. There are about 10
pores per mm. in any direction on the tabulae.

Lastly, in the centre of the cup a mass of coralloidal tissue similar to that
recorded by Taylor in Somphocyathus is very evident in PI. VI, fig. 60. As noted
earlier in this paper, similar coralloidal tissue has been discovered in Archaeocyathus
trachealis and Syringocnema gracilis , so that it can no longer be considered as an
abnormality.

Coscinocyathus endutus ,* n.sp.

The second species of Coscinocyathus is represented by a considerable number of
specimens cut in various directions, so that a more detailed examination has been
possible. Like C. fultus, this form is quite dissimilar to any of the species so far
recorded, the nearest being probably C. irregularis, Von Toll.

One of the chief diagnostic characters is the extraordinary mode of fixation.

Gk. endutos, clothed.

FROM A DREDGING IN THE WEDDELL SEA.

703

This will be clearly seen by reference to PI. Y, fig. 51, and PI. Ill, fig. 33. The first
of these figures represents the lowest transverse section of a cup which was obtained.
The central cavity surrounded by six loculi occupies the centre. The walls separating
the loculi are considerably thickened by callus growth. Surrounding the whole is a
thick zone of lamellae concentric with the cup, and presenting the appearance of an
outer wall thickened by subsequent deposition of calcareous material. The latter
figure, however, offers an explanation of the apparent thickening of the outer wall.
Here the cup is cut longitudinally. Inner wall, septa, tabulae, and outer wall are all
shown, while at the base the whole vase-shaped organism is supported by a lamellar
structure attached to a cluster of tubules of the alga Epiphyton fasciculatum,
Chapman. The lamellae then clothe the base of the cup in a close investment,
the outer layers spreading at their base to give a firm support to the whole. But
the lamellae often spread laterally, embracing other remains and forming a pocket
in which the base of the cup rests (PI. V, fig. 52). In other cases groups of these
plates attach themselves like struts to other organisms, expanding and flattening out
on the supporting surface, and thus helping to steady the whole cup (PI. V, fig. 53, a).
As would be expected, the pores of the outer wall are obliterated where the investing
laminae occur, but there being a sufficiency of material all the details of the porous
structure were determined, as follows — Pores on inner wall, 4 to 6 per mm. ; on outer
wall, 4 to 5 per mm. ; on septa, 5 per mm. ; on tabulae (irregular meshwork rather
than distinct pores), 8 or 9 per mm.

It will be noted in the above figures that there is considerable diversity in the
size of pore, but I am averse to creating new species on pore size alone, for in two
sections of the same specimen the measurements were 5 and 6 pores per mm.
respectively. Nor do I consider the examples giving 4 pores per mm. sufficiently
different in other respects to constitute a distinct species.

One of the best sections of the inner wall is shown in PI. I, fig. 1, and has been
already referred to. Here the pores run 4 per mm. A similar section from another
specimen (PI. Y, fig. 53) shows pores 5-5 per mm. (In addition, the fixing lamellae
a are very conspicuous in this last figure.) The outer wall in the specimen of PI. I,
fig. 1, is figured in PL V, fig. 55, and again the pores run 4 per mm. As the inner
pores and outer pores always correspond in size whether 4 per mm. or 6 per mm.,
the fossils are all of the “ equivallum” type (Taylor).

The porous character of the septa is not so easily illustrated, but one good
section is figured (PI. V, fig. 57). When we consider the tabulae the difficulty
is greater, but in PI. Y, fig. 56, a small part of one of these is illustrated. The
structure is an irregular meshwork rather than a regularly porous plate.

One of the figured sections (PI. Y, fig. 58) gives a good idea of the mode of
new septal introduction, and also the irregular character of the tabulae in many
cases. While the new septa arise practically in whorls at the same level, yet
that level does not seem to bear any relation to the tabulae.

704

DR W. T. GORDON ON CAMBRIAN ORGANIC REMAINS

As in most of the other forms recorded, the interior of the cup becomes filled
in towards the base by secondary structures. In the present instance this growth
consists of irregular oblique tabulae from which offsets pass in all directions. The
resulting tissue is very spongy and completely fills the base of the cup. In PI. V,
fig. 53, the tissue is very evident. In PI. V, fig. 58, it is again seen, though it does
not form such a marked feature. In both these examples the fixing lamellae are
very distinct.

The perfectly longitudinal section of PI. V, fig. 59, presents several interesting
features. The fixing lamellae are not well preserved at the base, though they are
indicated, and become more obvious at a higher level (Z). The tabulae are quite
irregular in distribution, and the central tissue is obviously developed from the inner
ends of the tabulae. The irregular offshoots from these tabular prolongations produce
the coralloidal structure which is so marked in this figure.

This new species may be defined as follows : —

Cup small, vase-shaped. Fixed by lamellae which wrap round the base and
spread out over the substratum. Porous structure of outer and inner walls similar
(equivallum type) ; pores 4-6 per mm. Septa regular, perforate (5 pores per mm.) ;
tabulae irregular, a mesh work rather than a porous platform ; mesh 8 or 9 per mm.
Base of cup filled by coralloidal growth, the result of proliferations from the tabulae
into the central space.

Summary.

The organic remains represented in two small pieces of limestone dredged
from a depth of 1775 fathoms in lat. 62° 10r S., long. 41° 20' W., are described.

These consist of both plant and animal fragments and skeletons. The plants
are calcareous algae and may be referred to the genus Epiphyton, Bornemann.

The animal remains are spicules, probably from sponges, complete and frag-
mentary cups of Archaeocyathinae, and sections of the carapace of trilobites.

The sponge spicules may be referred to the groups Tetractinellida, Hexactinellida,
and Heteractinellida. A peculiar forked type is also represented, but it is quite
unlike the modern Lelapia in structure.

The Archaeocyathinae belong to a number of different genera. Some of the
species have been recorded from other parts of the world, but some are new forms.
In certain of the known types, new structures have been recorded, and a more
accurate specific description than was formerly given has thus been made possible.

Taylor’s grouping of the family Archaeocyathidae has been revised and a new
genus ( Thalamocyathus ) instituted. Thalamocyathus contains groups I to IV of
Arcliaeocyathus as defined by that author. Thus in the Archaeocyathidae are
included Arcliaeocyathus, Thalamocyathus, Etlimophyllum, Archaeofungia, and
Pycnidocyathus.

FROM A DREDGING IN THE WEDDELL SEA.

705

Conclusions.

At the outset it must be admitted that this research has not led to any more
definite ideas regarding the type of organism which secreted the Archseocyathus
forms of skeleton. The method of cutting sections in definite directions has allowed,
however, of a more minute description of certain skeletal types. In several forms
the means of fixation of the skeleton have been determined, and these are not con-
stant for all species of the same genus, while closely similar methods of fixation occur
in different genera. Thus Spirocyathus atlanticus and Coscinocyathus fultus have
root-like emergencies acting as stout supports (perhaps also as stolons), while
Coscinocyathus fultus and C. endutus have very different methods of fixation.

It is suggested that the fixing strands may have functioned as stolons producing
a vegetative increase in the number of individuals, but no case has been seen where
new cups actually arise from these emergencies. Taylor, however, records a cup
arising from exothecal lamellae (fixation lamellae) in Archseocyathus Sellicksi* though
even there it is not clear whether the new cup arose from the exothecal structures
or was merely fixed to them. Apart from the fixation structures, there are one or
two marked features in this fauna from Antarctic regions. In the first place, the
specimens are small ; secondly, they show, in the majority of cases, marked thickening
of the walls near the base ; and thirdly, there are a great number of specimens (and
these belong to different genera and species) in a very small amount of material, i.e.
the fauna was abundant and varied, but the individuals small in size. The' largest
example is under 30 mm. long, and the greatest diameter noted in any cup is 10 mm.
or thereby. This compares very unfavourably with specimens from any recorded
locality, save only the other Antarctic record, from 85° S. on the Beardmore Glacier.
Despite the small size, the skeletal characteristics are constant and quite comparable
with similar forms from other regions.

Thickening of the walls and septa occurs near the base in several species, and this
is accompanied by a partial filling of the central cavity by ingrowths from the inner
wall. Such infilling may be an irregular coralloidal mass ( Thalamocyathus trachealis,
Coscinocyathus fultus, etc.) or a series of twisted tabulae (continuations of similar struc-
tures crossing the intervallum), with the interspaces filled by more irregular tissue
(Protopharetra polymorpha, PI. Ill, fig. 36). Both of these characters seem best
regarded as gerontic features, or as indicating that the specimens were at any rate
mature. In either case, when combined with the small size of the specimens, they point
to the conclusion that the conditions under which the organisms lived were far from
favourable to their growth, maturity supervening while the organism was small ; or,
what amounts to the same, growth being exceedingly slow. Such a conclusion is
confirmed rather than refuted by the fact that so many varied types occur in a very
small volume of material. It has been pointed out in various records that great

* Merri. Roy. Soc. S. Australia, vol. ii, part ii, 1910, text-fig. 9.

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 27).

109

706

DR W. T. GORDON ON CAMBRIAN ORGANIC REMAINS

abundance of certain forms in one bed or horizon (especially when the forms are small)
does not necessarily indicate that the conditions then obtaining were favourable for
that particular organism. More commonly it indicates quite the reverse, for the
small size is frequently accompanied by gerontic characters showing that the growth
was very slow, i.e. that the forms were stunted in their development.

Since all the specimens were small and relatively complete, the early stages
of growth were obtained in practically all cases. Thus in particular it was possible
to test the truth of Bornemann’s idea that Protopharetra is not a genus of the
Archseocyathinse, but represents a general type of rooting structure, i.e. that various
genera may terminate downwards in Protopharetra forms. (A direct analogy
may be seen in the so-called genus of plants — Stigmaria — which has been proved to
be a general type of rooting structure common to Lepidodendron, Sigillaria, and
Bothrodendron.) Taylor’s genus Metaldetes certainly lends colour to Bornemann’s
view, and so also does Spirocyathus, as shown above ; but, in spite of the fact that
a casual examination might cause one to group the lower parts of the skeletal
remains with Protopharetra , a closer examination shows that there are differences,
and that Protopharetra is a distinct genus.

The biological position of the whole group is still doubtful. The porous skeleton
suggests affinities with the Porifera, though not necessarily so, since certain of the
Coelenterata also have porous walls. The non-spicular character of the walls, septa,
etc., and the thickening of these structures in the lower parts of the cup, favour
relationships with corals (Anthozoa), while the peculiar internal growths can also be
paralleled in that group. The points of difference, on the other hand, are too
conspicuous to allow of any such correlation. Indeed, it is quite as likely that the
resemblances to sponges and corals are merely superficial and can be explained as
homoplastic developments.

Nor can we point to any gradual evolutionary changes within the group itself
which might give some clue to possible lines of descent. The Archseocyathinse
appear suddenly as a suite of very diverse forms at the top of the Lower Cambrian
Series, and on that geological horizon have practically a world-wide distribution.
True, a supposed organism, Atikolcania, Walcott, from Pre-Cambrian rocks was
compared with the Archseocyathinse, but the author has since proved that the
structure was really produced by inorganic means. The whole group, therefore,
is very isolated and has an extremely short geological range.

There is certainly abundant evidence that sponges coexisted with these organ-
isms, but corals are conspicuously absent from rocks of this period. Yet it must
be borne in mind that many Coelenterata secrete no skeletal elements, so the absence
of anthozoan skeletons from Lower Cambrian rocks does not necessarily imply that the
class was not represented in the fauna of that age. The only fossil with any marked
similarity to the Archseocyathinse is the calcareous sponge Barroisia from Cretaceous
rocks. Here, however, the skeleton is spicular and its sponge characters undoubted.

FROM A DREDGING IN THE WEDDELL SEA.

707

In the Archaeocyathinae we are thus confronted with a large, diverse, and widely
spread fauna, which appears suddenly at an early stage of the organic history of
the world, which produced locally great reefs of calcareous material in the oceans of
that date, and which became extinct as suddenly as it had appeared.

Having merely a series of skeletons to work from, comparison can only be made
with skeletons exhibiting similar shape and structure, and, despite the non-spicular
nature of the skeleton, it seems more natural to group these organisms with sponges
than with any other class of the animal kingdom. Of course it may be argued
that the characters justify the creation of a new phylum, though, at present, such
a course does not seem necessary.

• While the salient points in the anatomy of the skeleton in most species are clear,
the nomenclature of the group is far from satisfactory. It is impossible to group
the present forms known as Archaeocyathus under Billings’ original or emended
definition. Taylor has given a subdivision of the genus Archaeocyathus, but several
of his divisions are quite worthy of generic rank. The only forms which can be
included in the genus as defined by Billings are A. profundus, Billings, A. Sellicksi,
Taylor, A. dissepimentalis, Taylor, and A. pauciseptatus described above. All
the other species given under this generic title have no dissepimental tissue, and
it seems as logical to place them in some new genus as to place those which have
tabulae in the genus Coscinocyathus. It may be objected that in one case (A.
Sellicksi) the dissepiments only occur near the base, and that it is really intermediate
between the two types at present grouped under the name Archaeocyathus. But
there are connecting forms between other genera, and it seems preferable to keep
Billings’ definition for Archaeocyathus and erect a new genus to contain those
forms which have straight septa and neither dissepiments nor tabulae.

For this reason the genus Thalamocyathus, as defined above, has been instituted.
Forms, therefore, with straight septa and neither tabulae nor dissepiments come
into the genus Thalamocyathus ; those which have straight septa and well-marked
tabulae are grouped under the name Coscinocyathus ; specimens with straight septa
but irregular tabulae and dissepiments remain in the original genus of Billings,
Archaeocyathus.

In considering this flora and fauna from a geological point of view, it will be
convenient to summarise the whole as follows : —

Flora.

Alga.

Epiphyton fasciculatum, Chapman.

,, grande, sp. nov.

Fauna.

Porifera.

Tetractinellida, Hexactinellida, Heteractinellida ,

708

DR W. T. GORDON ON CAMBRIAN ORGANIC REMAINS

A rch'seocyathinae.

Archeeocyathus pauci.septatus, sp. nov.

Thalamocyathus flexuosus, sp. nov.

,, ichnusx, Meneghini, sp.

,, infundibulum , Bornemann, sp.

„ trachealis, Taylor, sp.

,, tubavallum, Taylor, sp.

Spirocyathus atlanticus, Billings, sp.

Syringocnema gracilis, sp. nov.

Coscinocyathus endutus, sp. nov.

„ fultus, sp. nov.

Protopharetra polymorpha, Bornemann.

,, radiata, Bornemann.

Metaldetes plicatus, sp. nov.

Didyocyathus sp.

The presence of algse in this limestone at once imposes certain limits to the
depth at which the rock must have been formed, for these plants could not
have lived in the complete absence of light. At the same time, the rock con-
sists entirely of calcareous material without admixture of sand grains. The deposit,
then, must have formed in shallow, clear water, possibly with little or no current
action.

In comparing the above list of organic remains with similar lists from other
areas, it is natural to consider, in the first place, Antarctic localities. In the various
reports of the British Antarctic Expedition, records of a Cambrian flora and fauna
are given, but unfortunately the material obtained was not very perfect. Taylor’s
report* states that two families of the Archseocyathinse were certainly present,
namely, the Archseocyathidse and Spirocyathidse. Chapman t describes the alga
Epiphyton fasciculatum from the same area, and, although the material seems
badly preserved and consequently his figures rather defective, there is no doubt
that one of the algse recorded above is identical with his species. Another fossil
common to the two localities is the forked sponge spicule which Taylor has
compared with Lelapia. Several of these forked forms have been noted in Weddell
Sea material, and the evidence now available shows that the spicule was probably
a hexactinellid type in which one ray is branched. So far, then, as the record from
the Beardmore Glacier goes, it shows a wonderful similarity with the Weddell Sea
record. These localities $ are approximately on opposite sides of the South Pole
(text-fig. 2, X, X'), but in neither case has the material been obtained in situ .

* British Antarctic Expedition, 1907-9, “ Geology,” vol. i, p. 240.

f Ibid., vol. ii, pp. 81-83.

t Bruce, Weddell Sea position, 62° 10' S., 4'>° 20' W., Scottish National Antarctic Expedition, 1902-4 ; Shackleton,
Beardmore Glacier position, 83° 42' S., 171° 30' E., British Antarctic Expedition, 1907-9, “Geology,” vol. i,
p. 235,

FROM A DREDGING IN THE WEDDELL SEA.

709

The drift in both cases has apparently been more or less northwards, but it is
impossible to assign any southern limit to the localities from which they may
severally have been drifted. All one can say is that somewhere in Antarctica there

must have been a development during Cambrian times of limestone containing
Archseocyathinse and marine algse.

A much more useful comparison, however, may be made with the Cambrian
fauna of South Australia. As described by Taylor, it constitutes the most complete
and most extensive Archdeocyathus fauna yet recorded.

710

DR W. T. GORDON ON CAMBRIAN ORGANIC REMAINS

The following genera are represented : —
Archseocyathus.

Thalamocyathus.

Archseofungia.

Archaeosycon.

Coscinocyathus.

Coscinoptycha.

Dictyocyathus.

Dokidocyathus.

Ethmophyllum.

Metaldetes.

Protopharetra

Pycnoidocyathus

Spirocyathus.

Somphocy a thus .

Syringocnema.

It will be noted that all the genera from the Weddell Sea material are repre-
sented in South Australia, and that two of these ( Metaldetes and Syringocnema)
have not been obtained elsewhere.

While the majority of the species in the two areas are distinct, yet some forms
are common to both faunas, and one of the species, viz. Thalamocyathus trachealis
(Taylor, sp.), is confined to these two localities. Even among the forms which are
not specifically identical there are sometimes similar characteristics. For example,
the species of Coscinocyathus from South Australia have, in the majority of cases,
the pores on the inner wall equal in size to those on the outer wall, i.e. they are of
the “ equivallum ” type. In other records the Coscinocyathidse have larger pores
on the inner wall. The Antarctic examples described above approximate to the
Australian “ equivallum” type.

All these cases point to a close relationship between the Antarctic and Australian
faunas of Lower Cambrian age ; and comparison with the Archseocyathus fauna in
Sardinia, Canada, or Siberia shows that the forms are not nearly so closely allied.
There is, however, one marked point of contrast, namely, the size and general con-
dition of the specimens. Taylor has recorded specimens from Australia reaching a
length of 6 inches and a diameter of 4 inches. The longest example from the Weddell
Sea is little more than 1 inch, and the greatest diameter recorded is about §- inch.
Again, as shown above, the walls and septa are in most cases thickened, while the
interior of the cup is filled with adventitious structures. These features are con-
sidered gerontic. The small size and apparently senile characters mark the Weddell
Sea types and probably indicate life under adverse circumstances. It is interesting
to note that the only complete specimens recorded from the Beardmore Glacier
moraines were embryonic forms, according to Taylor. This is explained in the
Report of the British Antarctic Expedition as due to the small size of the frag-
ments of the breccia ; but none of the fragments of larger cups — as far as one can
judge — indicate very large individuals. The small size may again be explained by
adverse conditions of growth, and in that case can be correlated with the Weddell
Sea forms.

Taking all these facts into consideration, we seem driven to the conclusion that
the life-types of Cambrian times in South Polar and Australian areas were similar,
but that conditions towards the pole were less favourable for the growth of these

FROM A DREDGING IN THE WEDDELL SEA.

711

organisms. The most obvious way in which the conditions could be adverse is by
a climatic change, since other characters seem similar. It will be interesting to
see whether subsequent discoveries confirm an apparent climatic differentiation during
Cambrian times.

In conclusion, I desire to record my thanks to Dr W. S. Bruce, leader of the Scotia
Antarctic Expedition, for giving me the opportunity of examining and describing the
Weddell Sea specimens, and for his forbearing in the long delay in publication ;
to Professor W. W. Watts, F.R.S., who has acted as supervisor, for the University
of London, and read over the manuscript ; and to Professor J. H. Ashworth, F.R.S.,
who, in my absence, read the paper to the Society. I am also indebted for
assistance in preparing the sections of the material to Mr C. H. Benson ; to
Mr J. Matheson, who has constructed the map showing the geographical points of
Antarctica referred to in the text ; and to the Carnegie Trust for the Universities
of Scotland for a generous grant to defray the cost of publication.

EXPLANATION OF PLATES.

(All figures are prints from untouched negatives.)

Plate I.

Fig. 1. Coscinocyathus endutus. Longitudinal section through inner wall, a, inner wall present ;
b, inner wall replaced by crystalline calcite, porous structure indicated by darker matrix ; c, pores on outer
wall. (Slide S. 97, x 7'5).

Fig. 2. G. endutus. External surface of weathered specimen. Outer wall practically all removed by
weathering. Porous structure still indicated near the base and on the sides. Septa and tabula? also shown
as a series of cavities arranged in lines, a, septum, the holes really represent the solid part of the original
septum, which is here dissolved away, while the solid part between the holes is the matrix filling in an
original pore ; b, tabulae. (Slide S. 156, x 5.)

Fig. 3. Epiphyton fasciculatum and E. grande. The relative sizes of the tubules are quite easily seen,
a, E. fasciculatum ; b, E. grande. (Slide S. 44, x 9.)

Fig. 4. E. fasciculatum. Slightly enlarged view of the thallus to show the tufted character, a, portion
more highly magnified in fig. 4a. (Slide S. 69, x 25.)

Fig. 4a. E. fasciculatum. Portion of a (fig. 4) more highly magnified to illustrate the equal bifurcations
of the thallus and the swollen terminal portions, a, points of bifurcation. (Slide S. 69, x 90.)

Fig. 5. E. grande. General view of a large tuft cut obliquely. Bifurcating character very apparent in
places. (Slide S. 11, x 25.)

Fig. 6. Sponge spicule ; probably heteractinellid. (Slide S. 103, x 20.)

Fig. 7. Sponge spicule ; tetractinellid type, showing all four rays. (Slide S. 33c, x 20.)

Fig. 8. Sponge spicule ; large form, probably heteractinellid. (Slide S. 113, x 20.)

Fig. 9. Spicule similar in type to that of fig. 8. (Slide S. 81, x 20.)

Fig. 10. Sponge spicule ; hexactinellid form. (Slide S. 23, x 20.)

Fig. 11. Sponge spicule tetractinellid. Three rays beautifully displayed ; the fourth is vertical but is
indicated by the knob at a. (Slide S. 76, x 20.)

Fig. 12. Sponge spicule. Hexactinellid type with one ray bifurcated. Similar to Lelapia-Yike spicule
of Taylor, a, spicule out through median plane ; b, a second example showing bifurcated ray only, as section
is oblique to median plane. (Slide S. 102, x 90.)

Fig. 13. Another spicule similar to that of fig. 12, but cut rather obliquely. (Slide S. 78, x 90.)

712

DR W. T. GORDON ON CAMBRIAN ORGANIC REMAINS

Fig. 14. Hexactinellid spicule of ordinary type. (Slide S. 36c, x 20.)

Fig. 15. Thalamocyathus tubavallum. Transverse section of cup. a, b, stirrup-pores. (SlideS. 125, x 10.)
Fig. 16. Epiphyton fasciculatum. Transverse section through thallus, showing numerous tubules filled
by a matrix of granular calcite. (Slide S. 140, x 26.)

Plate II.

Thalamocyathus trachealis, Taylor, sp.

Fig. 17. Transverse section of young cup, showing a considerable number of septa but no infilling by
callus growth into the central cavity. (Slide S. 88, x 28.)

Fig. 18. Transverse section of largest specimen obtained, a, a, pores on outer wall. (Slide S. 4, x 5'5.)

Fig. 19. Transverse section of cup, shown in last figure, at a lower level. Internal ring-platforms
prominent. Fringe of peg-like outgrowths from lower surface of shelf at a. See PI. Ill, fig. 26.
(Slide S. 22, x 22.)

Fig. 20. Transverse section of cup to show a complete ring- platform. At a the platform is complete,
at b only the outer rim is indicated. Septal pores at c. (Slide S. 23, x 22.)

Fig. 21. Longitudinal section of outer wall to illustrate the porous structure. (Slide S. 18, x 21.)

Fig. 22. Longitudinal section of inner wall, a, pores ; b, ring-platform cut vertically ; d, septa ; c, peg-
like denticles on lower surface of platform. (Slide S. 14, x 21.)

Fig. 23. Longitudinal section to show septal pores, a, pores on outer wall ; b, peg-like projection from
under side of ring-platform ; c, septum with pores. (Slide S. 15, x 22.)

Fig. 24. Longitudinal section of Australian specimen from Ajax Hill, Beltana, South Australia, a, pores
on outer wall; b, peg-like projections on lower side of ring-platform ; c, septum with pores. (Slide S. 161,
x 22.)

Fig. 25. Transverse section of a specimen showing coralloidal structures filling in the central cavity.
(SlideS 26, x 7-5.)

Plate III.

Fig. 26. Thalamocyathus trachealis. Transverse section of cup, part a of PI. II, fig. 19, highly
magnified to show the row of peg-like projections, s, septum ; p, septal pore ; a, row of pegs in action.
(Slide S. 22, x 90.)

Fig, 27. T. flexuosus. Transverse section of fragment of large cup. (Slide S. 47, x 3.)

Fig. 28. Longitudinal section of inner wall of T. flexuosus. Stirrup-pores indicated at b, (Slide

S. 46, x 9.)

Fig. 29. T. flexuosus. Longitudinal section of outer wall, s, septa indicated between pore rows, i.e. no
stirrup-pores. (Slide S. 45, x 9.)

Fig. 30. T. infundibulum. Longitudinal section through inner wall. (Slide S. 91, x 10.)

Fig. 31. T. ichnusse. Oblique longitudinal section showing pores on inner wall and septa. (Slide
S. 58, x 10.)

Fig. 32. Spirocyathus atlanticus. Transverse section of cup. (Slide S. 35c, x 11.)

Fig. 33. Coscinocyathus endutus. Longitudinal section of young cup. i.w., inner wall with pores ;

o.w. , outer wall; s, septum with pores; t, tabula. (Slide S. 29, x 26.)

Fig. 34. G. fultus. Oblique section of young cup showing fixing process ( r.pr .). b, base on which the
cup is fixed (it is part of a cup of Metaldetes plicatus) ; o.w., outer wall much thickened ; c, central cavity.

Fig. 35. Protopharetra radiata. Complete transverse section of largest cup obtained. (Slide S. 29, x 9.)

Fig. 36. P. polymorpha. Longitudinal section of central cavity. The complete specimen is figured
at PI. VI, fig. 66. Cavity filled in by proliferations from inner wall and dissepimental tissue, p, pores on
inner wall. (Slide S. 2, x 10.)

Plate IV.

Fig. 37. Spirocyathus atlanticus. Transverse section of cup near top. (Slide S. 37, x 9.)

Fig. 38. S. atlanticus. Oblique section showing that the septa are more like a meshwork of rods than
a plate perforated by pores, and thus accounting for the disconnected type of septum as seen in transverse
section. (Slide S. 48, x 2’5.)

FROM A DREDGING IN THE WEDDELL SEA.

713

Fig. 39. S. atlanticus. Transverse section at a low level, with rooting processes (a, b, c) surrounding
the cup. (Slide S. 33a (2), x 9.)

Fig. 40. S. atlanticus. Transverse section of cup at a higher level than in last figure, a, rooting
process attached to the cup. (Slide S. 35a (2), x 9.)

Fig. 41. S. atlanticus. Transverse section of same cup as figured in last two photographs, but cut at a
still higher level. (Slide S. 36a, x 9.)

Fig. 42. S. atlanticus. Longitudinal section of cup to show pores on inner wall and septa, a and b,
pores on inner w;all. See also text-fig. la. (Slide S. 41, x 9.)

Fig. 43* Syringocnema gracilis. Transverse section of small cup. p, p, p, pores on pyramidal papillae
of outer wall; w.p., pores on wall of tube. See also text-figs. 16, lc, and Id. (Slide S. 112, x 12.)

Fig. 44. S. gracilis. Transverse section of fragment of large cup. a, apparent vesicular tissue near
inner wall due to the inner ends of the tubules bending downwards. See text-fig. lc. (Slide S. 118, x 10.)

Fig. 45. Longitudinal section near outer wall. On the left the pores on outer wall are visible ; on the
right the tubules are cut further in towards the interior, and thus appear wider in the lumen. (Slide S. 120,
x 10.)

Fig. 46. S. gracilis. Oblique longitudinal section of small cup. a, tubules in longitudinal section.
(Slide S. 112, x 10.)

Fig. 47. S. gracilis. Tangential section through tubules. Section from same specimen as fig. 44, but cut
perpendicular to the direction of the latter. (Slide S. 121, x 10.)

Fig. 48. S. gracilis. Transverse section near the base of small cup Interior filled by coralloidal
growth, w.p., pores on walls of tubules. (Slide S. 108, x 10.)

Plate V.

Fig. 49. Syringocnema gracilis. Weathered surface showing tubules in section. (Slide S. 115, x 8.)

Fig. 50. Goscinocyathus fidtus. Weathered surface on outside of block. This specimen was cut up
into sections in various directions and the surface therefore destroyed, t, tabulae ; i.w., inner wall with
pores, x 5.

Fig. 51. G. endutus. Transverse section near apex of cup. The characteristic investing lamellae are
very well developed. Septa at this level six in number. (Slide S. 31a, x 26.)

Fig. 52. G. endutus. Section at a slightly higher level. The investing lamellae are spreading out and
anchoring the cup to surrounding objects. (Slide S. 66, x 9.)

Fig. 53. G. endutus. Section above that of fig. 52. It becomes oblique at the top, and the pores of
the inner and outer walls are visible, a, bundle of fixation-lamellae. Interior of cup filled by coralloidal
growth. (Slide S. 67, x 9.)

Fig. 54. C. endutus. Oblique longitudinal section of outer wall. The wall has disappeared, but the
porous character is preserved by the dark matrix. See also PI. I, fig. 1. (Slide. S. 55, x 9.)

Fig. 55. G. endutus. Longitudinal section of outer wall to show porous character. (Slide S. 98, x 9.)

Fig. 56. C. endutus. Transverse section through septum illustrating the irregular perforations ; from same
specimen as shown in PI. I, fig. 1. (Slide S. 99, x 28.)

Fig. 57. G. endutus. Longitudinal section through a septum to show the pores. (Slide S. 27, x 9.)

Fig. 58. G. endutus. Oblique longitudinal section through cup. Irregular spacing and course of septa
well shown. I , fixation-lamellae ; a, transverse section of cup of Protopharetra polymorpha. (Slide S. 53,
x 5-5.)

Fig. 59. G. endutus. Longitudinal section through cup. The internal coralloidal tissue is particularly
well exhibited. I, fixation-lamellae, (Slide S. 53, x 10.)

Plate YI.

Fig. 60. Coscinocyathus fultus. Transverse section of cup. a, 6, rooting processes ; p, p, external
boundaries of successive series of fixation offsets ; t , tabula ; c.t., coralloidal tissue in central cavity. (Slide
S. 153, x 9.)

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 27). 110

714

CAMBRIAN ORGANIC REMAINS FROM THE WEDDELL SEA.

Fig. 61. G.fultus. Section below that of last figure, a, c, rooting processes; p, boundary between two
of these processes; t, tabula. (Slide S. 152, x 9.)

Fig. 62. G. fultus. Longitudinal section of outer wall. The porous structure of the latter is prominent
in places, p, p, boundaries between rooting offsets; t, tabula. (Slide S, 154, x 9.)

Fig. 63. Archxocyathus pauciseptatus. Transverse section of cup. At p the septa are seen to be
porous. (Slide S. 145, x 9.)

Fig. 64. A. pauciseptatus. Longitudinal section. Imperforate outer lamina is clearly seen on each side.

р, septal pore ; i.w., porous inner wall ; c.t., coralloidal tissue in central cavity. (Slide S. 148, x 9.)

Fig. 65. Protopharetra polymorpha. Transverse section of cup. (Slide S. 33a (1), x 9.)

Fig. 66. P. polymorpha, Longitudinal section. Central portion at d is shown in PI. Ill, fig. 36.
(SlideS. 2, x 3-5.)

Fig. 67. Dictyocyathus sp. Transverse section illustrating very irregular rod-like septa. (Slide S. 54,
x 11.)

Fig. 68. Metaldetes plicatus. Transverse section through rather irregular cup which is attached to a
portion of Spirocyathus atlanticus along the top. (Slide S. 49, x 3.)

Fig. 69. M. plicatus. Longitudinal section showing septa, which are porous, crossed by a very irregular
set of curving tabulae, w.p., pores on outer wall. (Slide S. 90, x 9.)

Fig. 70. M. plicatus. More highly magnified figure of pores on outer wall, shown as w.p. in last figure.
(Slide S. 90, x 27.)

Fig. 71. M. plicatus. Longitudinal section of a septum to illustrate the porous character, a, h, septum
with pores. (Slide S. 94, x 9.)

Plate VII.

Fig. 72. Protopharetra radiata. Transverse section near point of bifurcation of a cup. a, one cup
cut near the top, inner wall developed and little tissue in central cavity; h, cup cut lower down than in a.
Inner wall not so clearly visible, and centre filled in by coralloidal growths. (Slide S. 65, x 5.)

Fig. 73. P. radiata. Transverse section above that of fig. 72. The section passes above the level of
cup a; h is cut at a higher level than in the last figure, and a third cup c, with rooting processes
r, is now joined to b. (Slide S. 66, x 5.)

Fig. 74. P. radiata. Transverse section above level of fig. 73. The cup 6 has now disappeared, while

с, with its rooting offsets d and e, has reached a level where the inner cavity is nearly free from
infilling material and thus the inner wall is prominent. (Slide S. 67, X 5.)

Fig. 75. Metaldetes plicatus. Transverse section at a low level, a, rooting process. (Slide S. 52, x 9.)
Fig. 76. M. plicatus. Section of the same cup at a higher level. (Slide S. 53, x 9.)

Fig. 77. M. plicatus. Section above that of fig. 76. Metaldetes form now well developed, a , new cup
forming from the wall of the older cup. (Slide S. 54, x 9.)

Fig. 78. M. plicatus. Topmost section of cup. a, the new cup at a higher level than in last figure, and
having the Protopharetra form of the early stage (fig. 75) of the present cup. (Slide S. 55, x 9.)

Trans. Roy. Soc. Kotin,

VOL. LII.

Dr W. T. Gordon: “Cambrian Organic Remains from Weddell Sea.” — Plate I.

W. T. Gordon photomic.

Zinco-Collotype Coy. Edin.

Trans. Roy. Soc. Edin.

YOL. LII.

Dr W. T. Gordon: “Cambrian Organic Remains from Weddell Sea.” — -Plate II.

Fig. 17.

Fig. 21.

22/1

Fig. IS.

Fig. 24.

Fig. 23.

»>. Fig. 22.

Fig. 25.

7-5/1

W. T. Gordon photomic.

Zinco-Collotype Coy. Edin,

Trans. Roy. S oc. Kdin.

VOL. LII.

Dr W. T. Gordon : “ Cambrian Organic Remains from Weddell Sea.” — Plate III.

Fig. 31.

Fig. 28.

Fig. 35.

Fig. 36.

Zinco-Collotype Coy. Edin.

Fig. 33.

Fig. 32.

W. T. Gordon phntomic.

Trans. Roy. Soc. Edin.

VOL. LI I.

Dr W. T. Gordon: “Cambrian Organic Kemains from Weddell Sea.” — Plate IY.

W. T. Gordon photomic.

Zinr,o-C allotype Coy. Edin .■

Trans. Roy. Soc. Edin.

VOL. LIT.

Dr W. T. Gordon' : “Cambrian Organic Remains from Weddell Sea.”— Plate V.

Fig. 56.

9/1

Fig. 53.

Fig. 49.

Fig. 59.

Fig. 54.

Fig. 51.

5-5/1

55.

W. T. Gordon photomic.

Zinco-Collotype Coy. Edin.

Trans. Roy. Soc. Edin.

YOL. LI I.

Dr W. T. Gordon: “Cambrian Organic Remains from Weddell Sea.” — Plate YI.

W. T. Gordon photomic.

Zinco-Collotype Coy. Edin.

Trans. Roy. Soc. Edin.

VOL. LIT.

Dr W. T. Gordon : “ Cambrian Organic Kemains from Weddell Sea.” — Plate VII.

Fig. 75.

Fig. 76.

b.

Fig. 77.

c.

Fig. 73.

W, T. Gordon photomic.

Zinco-Collotype Coy. Edin.

( 715 )

XXVIII. — New Stelar Facts, and their Bearing on Stelar Theories for the Ferns. By
John M‘Lean Thompson, M.A., D.Sc., F.L.S., Lecturer on Plant Morphology,
Glasgow University. (With Four Plates and Nine Figures in the Text.)

(MS. received March 1, 1920. Read March 1, 1920. Issued separately July 9, 1920.)

For over thirty years problems of stelar structure in the Ferns have claimed the
attention of anatomists. By common consent protostely has been recognised as a
primitive state. It is seen in all phyla of primitive vascular plants, except the semi-
aquatic Equisetales, it figures universally in the juvenile plants of the Filicales, and
is maintained in the adult stems of a number of their primitive genera.

But as to the origin of pith, inner phloem, and inner endodermis, and the steps
by which during descent the primitive protostely has in many instances been replaced
in the adult by other more complex stelar states, there is no general agreement.
On the one hand, the pith, inner phloem, and inner endodermis have been regarded
as purely intrastelar tissues directly referable in origin both in individual plants
and descent to the procambial mass of the growing point of the stem. On this view
medullation and solenostely are considered consecutive resultants of direct intrastelar
readjustments, the steps taken being : (i) an ontogenetic progression from the solid
xylem-core of the protostele in the juvenile axis to a medullated protostele by change
of destination of procambial elements. These have reached maturity as parenchy-
matous cells rather than as tracheides ; (ii) a further ontogenetic change from the
medullated protostele thus established to a solenostele by differentiation of procambial
cells as sieve-tubes and endodermis within the pith itself ; (iii) the establishment
of direct histological connection between the intrastelar pith and the cortex through
gaps in the stelar cylinder. On the other hand, the pith has been interpreted as
cortical tissue involved in the stele during descent. The pith is considered of purely
cortical origin, and the inner phloem and inner endodermis of the solenostele are held
to be extensions of the outer phloem and endodermis of the original protostele. On
this view the steps of progression from protostely to solenostely are : (i) the solid
protostele is followed in the ontogeny by a conductive tract into which, from the
axil of each leaf-trace., a decurrent mass of cortical tissue intrudes. These cortical
masses are lodged in endodermal pockets ; (ii) as the ontogeny advances succeeding
pockets of cortical tissue are confluent within the stele, their common product being
the pith which communicates with the outer cortex through the mouths of the
pockets or stelar gaps. By extension of phloem and pericycle of the protostele
during medullation solenostely is attained.

Both views have been held by mutual exclusion for the Filicales as a whole ; on
the other hand, pith is considered in some cases of intrastelar origin, in others as
TRANS. ROY. SOC. EDIN., YOL. LII, PART IV (NO. 28). Ill

716

DR JOHN M‘LEAN THOMPSON ON NEW STELAR FACTS,

cortical tissue. But whatever view is entertained of the origin of pith and other
stelar tissues during descent, the ontogeny holds a permanent structural record
of the stelar changes which have occurred in each individual organism. These
changes are fully revealed only by exhaustive ontogenetic study, but even when they
are known it is doubtful how far they recapitulate the stelar changes of the ancestry.
On this point light may be shed by the study of related organisms. At the present
moment the sum of available ontogenetic evidence is small, and includes fragmentary
records. But though supplemented by a large body of fact from adult structure,
it is still an inadequate basis for the superstructure of theory which it must bear.
The evidences in favour of the view of intrastelar origin of pith are only in a few
cases sufficient, while confluence of pockets, on which the theory of cortical origin
of pith is founded, has not been demonstrated.

In the following pages new ontogenetic facts are given for a number of Ferns.
In some instances these are complete, in others incomplete ; but with the structural
facts already known they may form a broader basis than has hitherto been available
on which to found a stelar theory for the Ferns. It is held that they in no way
support a theory of cortical origin of pith, but in each case medullation, and in
some cases solenostely arise in the ontogeny entirely by intrastelar readjustment in
which a change of procambial destination, as indicated above, is mainly operative.
The facts recorded are dealt with in a rough ascending scale, but are not to be
interpreted as constituting a strict phylogeny.

Schizaeacese.

Stelar Anatomy of the Sporeling of Schizsea malaccana, Bk.#

Figs. 1-6 represent successive transverse sections of the stele of the stem, and are
drawn to a uniform magnification. Fig. 1 is from the incomplete base of the stem.
In the sclerotic cortex is a leaf-trace (L) which has departed from the lower portion
of the stele. It is improbable that this is the first leaf-trace formed for this plant,
but for convenience of description it may be referred to as such. The stelar xylem is
parenchymatous, and is surrounded by unbroken bands of phloem (Ph.), pericycle (Pr.),
and endodermis (En.). At A the tracheides and wood-parenchyma are segregated,
the latter forming a definite mass to the inside of an arc of phloem and xylem (X),
which at a higher level supplies the second leaf-trace. In fig. 2 the second leaf-trace
(L) is almost free from the stele. It has departed in the protostelic manner, causing
no marked disruption of the stelar xylem, but the stelar phloem is interrupted.
An irregular bridge of endodermal cells (B) partially spans the isthmus between
the leaf-trace and the stele. This bridge is completed in sections immediately
following the one now figured, and the liberation of the leaf-trace leaves no gap in
the endodermis of either leaf-trace or stele. Local segregations of wood-parenchyma

* The sections on which the following statements are founded are Nos. 1603-1605 in Prof. Gw ynne- Vaughan’s
collection of preparations in the Botanical Department, Glasgow University. ■ *

AND THEIR BEARING ON STELAR THEORIES FOR THE FERNS.

717

are present, and one of these (A) helps to define the base of the third leaf-trace (X).
The departure of the third leaf-trace is shown in fig. 3. The steps taken in the
liberation of the second leaf-trace are here repeated ; but opposite the leaf-trace
not only is the stelar phloem interrupted, but the xylem is greatly reduced. The
chief change of stelar structure at this level is due, however, to a well-defined core
of parenchyma within the xylem (M). The sections preceding the one now figured
show this pith to be a mass of xylem-parenchyma more extensive and continuous
than the masses of similar nature already noted at lower levels. It is a purely
intraxylic tissue in origin as in location. Within it is a group of tracheides (T)
which are linked at higher and lower levels to the general xylem mass. The
departure of the third leaf-trace is thus protostelic, and pith has been formed within
the wood itself. At 0 a parenchymatous connection links the pith and pericycle.
In fig. 4 the departure of the fourth leaf-trace is shown. The stelar enclodermis
remains an uninterrupted histological barrier between the stele and cortex. Both
stelar phloem and xylem are discontinuous opposite the leaf-trace, and pericycle and
pith are in open parenchymatous connection through the intrastelar gap thus
formed (0). The closure of this gap is secured in the immediately succeeding
sections, and in the internode which follows the parenchymatous pith is increased
and is still traversed by groups of tracheides (T, fig. 5). In fig. 6 is shown the
stelar structure immediately above the point of liberation of the fifth leaf-trace.
During departure of this trace the stelar endodermis is maintained unbroken. At
the level figured the pith is composed entirely of thin-walled parenchyma (M),
though in the immediately preceding sections it contains scattered tracheides. A
parenchymatous gap in the cylinder of xylem and phloem places the pith and
pericycle in direct histological connection, but at a higher level this gap is closed.

The transition thus, traced from solid protostely without intrastelar gaps to
protostely with a purely parenchymatous pith and intrastelar gaps is shown in the
three series of sections of sporeling plants in G Wynne- Vaughan’s collection. In
them the isolation of stele from cortex is maintained complete, for until stele and
cortex are merged in the procambial mass the stelar endodermis is clearly seen as
an unbroken ring around the pericycle. But whatever interpretation may be put
upon these stelar changes in the light of adult structure and lateral comparison,
in the individual sporelings referred to the following facts are clear : —

(i) The stelar endodermis is an unbroken cylinder throughout, and no lateral

connection between cortex and pith has been observed.

(ii) The pith is of intrastelar origin, arising in the ontogeny by gradual increase

of wood-parenchyma at the core of the stele, with an accompanying
decrease of central tracheides. This intraxylic readjustment can be
explained most reasonably as due to change of destination of procambial
elements as the growing point advanced.

718

DR JOHN M‘LEAN THOMPSON ON NEW STELAR FACTS,

(iii) No marked increase has occurred in diameter of the xylem-cylinder or stelar
area during the transition from solid protostely to the medullated state.
Thus the progression to the well-marked parenchymatous pith is accom-
panied by reduction in transverse area of wood. The increase of stelar
parenchyma is in this instance greater than that of conductive tracheides,
the gain to the former constituting a loss to the latter. The inter-
dependence of medullation and xylem-increase, and the intrastelar origin of
pith in Schizsea malaccana, are thus indicated.

The progression from the juvenile to the adult stele has not been traced, but
the latter has been examined by Russow (.51), Prantl (50), Tansley and
Chick (58).

They have recorded a stelar structure similar to that shown in fig. 6, and were
united in considering the intraxylic parenchyma “ an intrastelar pith.” The differ-
entiation of the vascular system from the growing apex of the stem has been studied
(58). The recorded facts are in harmony with the intrastelar theory for the pith,
and in no way support the cortical hypothesis. Within the pith Tansley and Chick
have found in plants with large steles, strands of tracheides the arrangement of which
is “ quite capricious.” Thus the pith with internal tracheides described for the
sporeling has been shown to recur in the adult.

To other features which have been observed in adult stems of this species,
and particularly the presence of endodermal pockets decurrent from the axils of leaf-
traces through xylic gaps into the pith, further reference will be made., For the
moment it may suffice to note that, though the ontogenetic development of these
pockets has not been observed, they are absent from the medullated sporeling. It
may then be held that in the ontogeny of Schizsea malaccana stelar pocketing is
antedated by medullation.

Schizsea dichotoma.

A number of incomplete adult stems of this species have been examined.* In
general they revealed a stelar structure similar to that already described by Boodle
(3), (5). Throughout the greater length of stem examined the bulky sclerotic pith
has a selvage of thin-walled parenchyma. For the most part the xylem-cylinder is
solenoxylic, locally it is dictyoxylic, and xylic gaps independent of leaf-traces occur.
In text-fig. 1 the general distribution of stelar tissues is shown in longitudinal section
with leaf- and root-traces reduced to one plane. Endodermis is indicated by the line
(En.), thin-walled-parenchyma is white, phloem cross-hatched, xylem solid black, and
sclerenchyma dotted. Leaf-traces are lettered in ascending order (a, b, c, etc.), and
levels specially mentioned are numbered. At I there is a xylic gap with shallow
endodermal pocket, but no leaf-trace. In the axils of leaf-traces a, b, e, f, and g are
* These have been kindly provided by Professor Margaret Benson, D.Sc., and Rev. W. W. Watts, F.L.S.

AND THEIR BEARING ON STELAR THEORIES FOR THE FERNS.

719

shallow endodermal pockets. From the axils of
leaf-traces c, d, and.h, pockets of varying length
are decurrent through xylic gaps into the par-
enchyma between the xylem-cylinder and central
sclerenchyma. These pockets are of varying
depth ; they are basally closed and contain tissue
similar in structure to the inner cortex. In one
instance (c) the base of the pocket is forked. At
III a spindle of endodermis enclosing paren-
chyma is immersed in the pith. As this spindle
dies out it is followed upwards by a growing
mass of tracheides, which in like manner dwindles
to a vanishing point, leaving the stelar core com-
pletely sclerotic. The general distribution of
stelar tissues thus noted if similar to that already
recorded by Boodle (5) for this species. Certain
points of detail, however, merit special atten-
tion. Fig. 7 represents a transverse section of
the stele at I, and shows the shallow endodermal
pocket already referred to (Pk.). The stelar core
is a continuous mass of sclerenchyma separated
for the greater part of its circumference from
the xylem-cylinder by a selvage of thin-walled
parenchyma. At several points this selvage is
narrow, and at S the sclerenchyma abuts upon
the tracheides. At T and U tracheides are
isolated within the selvage, and may thus be in
contact with the sclerenchyma. The impres-
sion gained from many such sections is that the
general distribution of these stelar tissues is
fairly defined, but that intrastelar adjustments
may occur.

Fig. 8 represents a transverse section of the
stele at II. From it, it will be again apparent
that, by local increase or decrease of xylem,
parenchyma, and sclerenchyma, an histological
balance is struck within the stele. That the
consequent variations from point to point in the
proportions and distribution of the component
tissues are purely intrastelar matters cannot be
doubted. At W a group of endodermal cells is lodged in

parenchyma.

720

DR JOHN M'LEAN THOMPSON ON NEW STELAR FACTS,

This is the closed base of the endodermal pocket decurrent from the axil of
leaf-trace h, while the base of the trace itself is indicated. , by the protruding
arc of xylem and phloem (Y). Fig. 9 represents a transverse section of the
stele at a slightly higher level (III). Here the endodermal pocket is tubular, and
contains thin-walled parenchyma like that of the inner cortex. The stelar core
is no longer purely sclerotic, but comprises thin-walled cells (Z), among which
endodermal cells occur (En.). These are the basal cells of the endodermal spindle.
In fig. 10 the stelar structure at level IV is shown. The mouth of the pocket
has been reached, but the leaf-trace is not yet free from the stele. An increased
proportion of the intraxylic tissue is thin-walled parenchyma, here continuous
from the stelar core to the pericycle through the xylic gap (G-). At this level
the isolated endodermal spindle is a wide tube (En.) containing thin-walled cells
like those around it. At V the leaf-trace has been freed from the stele, and the
endodermal spindle in the pith has reached a vanishing point (En., fig. ll). Within
the central parenchyma is a group of tracheides (T). The 'succeeding sections show
no trace of the inner endodermis, but the inner tracheides are increased in
number (T, fig. 12).

Reduction of the xylem spindle marks the succeeding sections. Its vanishing
point is reached in fig. 13 (T), and the pith is once more mainly sclerotic.

All the structural facts accord with the view that the intrastelar changes in
balance and distribution of wood, thin-walled parenchyma, sclerenchyma, and
endodermis, thus traced from point to point, have been accomplished in the ontogeny
by change of procambial destination as the growing point advanced, and that the
tissues involved are of purely intrastelar nature.

Isolated endodermal spindles within the pith have also been recorded, by
Boodle (5), who, for lack of a ready explanation of their function as independent
structures, suggested that they may be relics of foliar pockets, which, having figured
in the ancestry, have lost their mouths during descent, and thus are no longer
linked to the outer endodermis at stelar gaps. Tansley and Chick (58) have
recorded not only deep foliar pockets but also an endodermal spindle within the
stele of Schizasa malaccana. The latter structure is decurrent from the axil of a
leaf- trace, but, though devoid of an opening into the cortex, is linked to the outer
endodermis through the xylic gap by a chain of endodermal cells. It is undoubtedly
a potential pocket, but its actual state and origin may be variously interpreted.
It may indicate a phyletic degeneration of foliar pockets, or may merely illustrate
how procambial destination determines mature structure. On the latter view an
endodermal spindle initiated within the stele may in one instance be linked to the
outer endodermis at a xylic gap, thus contributing to one unbroken endodermis ; in
another such an endodermal union may be partial or absent. The central position
of the endodermal spindle referred to in the adult stele of Schizsea dichotoma and its
independence of adjacent leaf- traces give point to the latter suggestion. It is indeed

AND THEIR BEARING ON STELAR THEORIES FOR THE FERNS.

721

obvious that a deep pocket is associated with the nearest leaf-trace (text-fig. l).
On this view endodermal cells within the stele may be regarded as new creations,
which in some cases form foliar pockets, in others remain isolated1 structures. But
in all instances they provide an increased endodermal area for the stele, and may
enhance the physiological control.

The ancestry of the Schizasaceas is unknown, and accordingly phyletic degenera-
tion of foliar pockets for the stock is matter of mere speculation.

That medullation and pocketing are distinct phenomena in Schizaect dichotoma
as in Schizasa malaccana is evidenced by the sporeling structure described by
Boodle (3). He has traced the transition from solid protostely to medullation by
acropetal increase of wood-parenchyma. The departure of the first leaf-traces is
protostelic and, though pocketing is initiated, medullation is independent of it. An
isolated endodermal spindle has also been observed in the sporeling pith. When to
these facts is added the structural evidence already recorded from sporelings of
Schizasa jpusilla (13), Schizasa rujpestris (ll), and Anemia phyllitidis (3), (ll), and
from adult plants of Schizasa digitata, and Schizasa jistulosa (3), belief in the
intrastelar origin of pith in these organisms is strengthened.

But on the theory of the cortical origin of pith the medulla of these plants “ must
be regarded as a derivative of the cortex which has become more or less completely
sequestered within the stele” (37). This view, as expounded by Jeffrey (30), (33),
finds support in histological similarity between cortex and pith. On purely histo-
logical grounds exception may be taken to it, for the pith and cortex in both Schizasa
digitata and Schizasa dichotoma are dissimilar. The sclerotic pith-cells of these
species have dense mucilagenous contents, and their highly lamellated walls have
numerous fine pits (fig. 15). Their chief duty is apparently water-storage. The
cortical cells are mainly sclerosed : where the walls are thick, the pits are wide ; where
the walls are thin, pitting is obscure. The contents of these cells are not apparently
mucilagenous, and starch-storage is an evident function (fig. 14). A reasonable view
of such facts may be that structural similarity between cortex and pith indicates
for them like functions, while histological differences indicate functional differences.
Thus, on histological grounds alone a cortical origin for pith cannot be deduced.

The main prop of the theory of cortical origin of pith is, however, cortical
intrusion. No ontogenetic evidence of actual intrusion has been advanced for the
Schizasaceas, and a phyletic flow of cortex into the stele can only be inferred. But
that pocketing is a result of change of procambial destination rather than a phase of
cortical intrusion appears to be the natural interpretation of the structural facts.
This view may be further supported from the adult stele of Schizasa dichotoma.
Text-fig. 2 shows the general distribution of stelar tissues in longitudinal section in
part of one of the stems examined. Certain xylic gaps have been omitted. Leaf-
traces and endodermal pockets are reduced to one plane. As in text-fig. 1, pockets
of diverse depth and form are shown, in some cases decurrent from the axils of

722

DR JOHN M‘LEAN THOMPSON ON NEW STELAR FACTS,

leaf-traces, in others through xylic gaps with which leaf-traces are not associated.
But the features to which special attention would he drawn occur between levels I

and II. Between these lies an isolated endo-
dermal spindle, while within the base of a pocket
is a chain of tracheides — here represented as a
black line (T). On the theory of cortical origin
of pith, tracheides would not be expected to figure
among the contents of a closed endodermal
pocket. In fig. 16 the stelar structure at the
base of the endodermal spindle is shown. The
xylem is dictyoxylic, and to the inside of the
narrower xylic gap is an endodermal cell (En.),
which is the base of the spindle. Fig. 17 repre-
sents the structure at a slightly higher level.
The xylic gaps have closed, and the endodermal
spindle is here tubular (En.). In fig. 18 the apex
of the spindle is reached (En.). At En.' is an
almost basal section of the pocket from the axil of
II leaf- trace a, at En" is a section towards the base
of the pocket associated with leaf-trace b. It
contains a mature tracheid. The details of this
pocket are shown on a larger scale in fig. 19.
The endodermis is a complete ring, and contains
a parenchymatous cell (P) and a tracheid (T).
Fig. 20 shows the structure of the pocket at a
slightly lower level. The endodermis here sur-
rounds two parenchymatous cells, and a tracheid
forms part of the endodermal ring. The actual
base of the pocket is shown in fig. 21. It com-
prises six endodermal cells, and partly involves
a tracheid. Tracheides of the xylem-cylinder are
close at hand (T'). Followed down into a lower
section (fig. 22), the tracheid in fig. 21 (T) consorts
with other tracheides which lie in the parenchy-
matous selvage. An inclusion of tracheides in
endodermal cells has been observed in another
stem. In this case the endodermis was an. iso-
lated spindle situated in the sclerotic pith. Fig. 23 shows a section through the
base of this spindle. The endodermal cells surround a single tracheid (T). Fig. 24
is from a section of the spindle at a higher level ; the endodermal ring (En.) is entirely
filled with tracheids. From these facts it may be concluded that by static change

AND THEIR BEARING ON STELA R THEORIES FOR THE FEB NS.

723

of destination of procambial cells from point to point mature elements of different
category may be variously grouped within the stele. To explain them, neither
phyletic nor dynamic flow of tissues is necessary, nor indeed is there evidence of
the latter.

On the evidence already dealt with the following conclusions are reached : —

(i) The species considered have in the adult state medullated protosteles.

(ii) In each the medulla is of purely intrastelar origin.

(iii) All the evidence shows that endodermal pockets through xylic gaps are

intrastelar structures, and arise by change of procambial destination
within the stele.

(iv) Endodermal spindles partially or completely isolated within the stele may

be regarded in some cases as potential and not vestigial pockets, in others
* as independent though sporadic formations of inner endodermis which
has not been continued permanently in the ontogeny.

(v) By pockets and endodermal spindles the endodermal area for the stele as a
whole is enlarged, and physiological control may be enhanced.

(vi) That the contents of endodermal pockets and the adjacent cortex are
commonly of the same histological character may be expected, since they
are continuous through the mouths of the pockets. Thus united, they
behave as one tissue which need not be expected to resemble the pith
either in structure or function.

Gleicheniacese.

Of Gleicheniaceous plants Gleichenia jlabellata, Gleichenia pectinata , and
Platyzoma microphyllum bear directly on the subject in hand.

The internode of Gleichenia Jlabellata shows a cylindrical stele with a solid
parenchymatous xylem-core. At the nodes shallow endodermal pockets associated
with the leaf-traces are decurrent within the stele. The ontogenetic development of
these pockets has not been traced, but by them the area of stelar endodermis is
enlarged. On the view of static procambial change within the stele, the pockets
would be considered intrastelar in origin and independent of any possible medullation.
On the theory of cortical origin of pith, their contents would be viewed as cortical
tissue by confluence of which within the stele any future medullation would be
attained.

The ontogenetic evidence for Platyzoma microphyllum is still wanting, but the
adult stele has been examined (4), (49), (61), (62). It is commonly a medullated
cylinder with both outer and inner endodermis, and phloem only on the outer face of
the xylem. It is devoid of leap-gaps and perforations. Thus the pith is completely
isolated from the cortex, and the inner endodermis is independent of its outer cor-
relative. Jeffrey (30) has affirmed that the medulla of .Platyzoma is “in reality

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 28). 112

724

DR JOHN M‘LEAN THOMPSON ON NEW STELA R FACTS,

extrastelar, but no longer communicates with the peripheral cortex.” For the
establishment of this condition he assumed that pocketing had occurred in the
ancestry of Platyzomco, but that the confluent pockets by which the medulla arose
have been “ completely cut off from the outside cortex” during descent. It has now
been shown that the adult stem of Platyzoma is at points protostelic (62) ; from this
state medullation arises by purely intrastelar changes. In like manner, an inner
endodermis is developed around the pith. In default of any evidence of loss of stelar
gaps and of cortical origin for the pith and inner endodermis, a reasonable interpreta-
tion of the adult stelar structure may be that the medulla and inner endodermis are
intrastelar in origin as in location. As in Schizaea malaccana and Schizaea dichotoma
so also in Platyzoma, the possession of inner endodermis locally or throughout the
adult medullated protostele increases the endodermal area, and may enhance the
physiological control.

The ontogeny of Gleichenia pectinata * has been studied, and the full facts of
progression from the protostele of the sporeling to the solenostele of the adult are now
known. At the base of the juvenile plant is a long protostelic tract bearing a number
of leaf-traces (a, b, c, d, e, text-fig. 3). A transverse section of the stele at the point
of departure of the fourth leaf-trace is shown in fig. 25. There is a solid parenchy-
matous xylem-core (X), around which the phloem forms an unbroken ring (Ph.).
There is no endodermal pocket in the succeeding sections, and the leaf-trace departs
in the protostelic manner. In some of the smaller plants the upper portion of the'
protostele showed a mixed pith (fig. 26), with large tracheides (T) scattered through
thin-walled parenchyma. In some instances the stele was in this condition im-
mediately below the growing point ; in others a definite pith existed. In all such
cases examined the ontogenetic progression to the mixed pith or pure medulla was
marked by segregation of tracheides and parenchyma, as in Schizaea malaccana, and
the stellar endodermis was maintained unbroken. Thus the pith is an intraxylic
tissue. The protostelic tract of the juvenile axis is followed in the ontogeny by a
long medullated stage (text-fig. 3, levels I to II), the central region of tracheides
being entirely replaced by parenchyma. At first there are no xylic gaps at the
departure of the leaf-traces or elsewhere. Fig. 27 shows the transverse section of the
stele during the protostelic departure of a leaf- trace (f) and a root- trace (g) in this
region. Later the xylem is interrupted above the leaf-traces, and opposite the xylic
gaps thus formed there may be slight pocketing of the endodermis (h, text-fig. 3).
The ontogenetic stelar changes already traced for Schizaea malaccana are thus in
general repeated. As the stele is followed forward, phloem appears in the pith as an in-
definite and incomplete ring of sieve-tubes lining the xylem internally (III, text-fig. 3).
These are shown at Ph.' (fig. 28), which represents a transverse section of the
stele during departure of leaf-trace (k). A solenostelic structure soon follows,

* The materials for this study were kindly provided by Mr Vm, Harris, F.L.S., Director, Hope Gardens,
Jamaica.

AND THEIR BEARING ON STELAR THEORIES FOR THE FERNS.

7 25

— is

n

and results from the following steps. Inner endodermal cells appear in the pith
(En.', fig. 28). These mark the base of a tube of inner endodermis closed below and
widening upwards in the pith (En., text-fig. 3). It is
continuous on the one hand with the endodermis of the "

next departing leaf-trace (k), and thus provides the
increase of endodermal area which an axillary pocket
would secure. It is continued as the inner endodermis
lining the internode immediately above (text-fig. 3 ; and
En/, fig. 29). In this figure is shown the stelar structure
slightly above the point of departure of leaf-trace k.

That the tissue which this inner endodermis involves
(M, fig. 29) is histologically similar to the inner cortex is
matter for no surprise, for they are continuous through a
stelar gap and act as one tissue. The stelar structure
thus attained is in fact solenostelic. This will be further
apparent from fig. 30, which represents a transverse section
of the stele at IV (text-fig. 3) after the first leaf-gap is
closed. The medulla (M), inner endodermis (En/), inner
phloem (Ph/), xylem-cy finder (X), outer phloem (Ph.), and ^
outer endodermis (En.) are all shown. Followed upwards,
the inner endodermis is continuous with the endodermis
surrounding the next succeeding leaf-trace (L, text-fig. 3).

In this way the solenostelic structure, once secured, is
maintained.

Thus it has been shown that, as in Schizsea malaccana,

Schizsea dichotoma, and others, so also in Gleichenia
pectinata, medullation and pocketing are distinct pheno-
mena. That medullation antedates pocketing in the
ontogeny has indeed been demonstrated. There is no
structural evidence of intrusion or flow of cortical tissue

by which the pith of the solenostelic tract might arise, I

and study of the growing point itself supports the view
that by change of procambial destination the stelar changes
here described are effected. By such a static change alone
the continuity of endodermis across the pith could be
secured while solenostely is maintained. This formation
of inner endodermis follows upon the interruption of the
stelar xylem above a leaf- trace (text-fig. 3), which is thus
the first associated with a true foliar gap. The immediate
result of these changes is the physiological delimitation of the stelar pith of the
protostele below the inner endodermis from the later formed pith of the solenostelic

f

J

I

Text-fig. 3.

726

DR JOHN M'LEAN THOMPSON ON NEW STELAR FACTS,

tract. But that both these tracts of pith are referable in origin to the pro-
cambial mass of the growing point seems the only reasonable interpretation of the
ontogenetic facts. They are thus viewed as one continuous medullary field across
which a barrier of inner endodermis has been constructed by static changes of
procambial destination during the ontogeny. Here, as in the other organisms
considered, the inner endodermis provides an increase of endodermal area by
which the physiological control of the conductive cylinder may be enhanced.
Of these organisms Gleichenia pectinata is considered the most advanced in stelar
structure, the collective attainments of the others in regard to stelar amplification
being repeated in its ontogeny. To these are added solenostely and foliar gaps.
On the view of static ontogenetic change of stelar structure here adopted, the foliar
gaps of Gleichenia pectinata are of similar origin to endodermal pockets ; but while
the latter are closed at their base and do not permit of histological continuity
of pith and cortex, the former are open channels through which pith and cortex,
originally distinct, unite to form one continuous tissue.

The general position adopted on the facts described is that stelar tissues of any
category may be formed where and when required within the confines of the stele.
Protostely with medulla, endodermal pockets, and sporadic inner endodermal spindles
is the high-water mark attained by intrastelar changes for the species of Schizsea
dealt with : Gleichenia fidbellata has acquired endodermal pockets alone, and Platy-
zoma has not advanced to the solenostely shown by Gleichenia pectinata , being
still devoid of stelar gaps and inner phloem.

Gleichenia pectinata is a solenostelic member of the “ Superficiales,” thus
resembling such organisms as Metaxya, Lophosoria, Matonia, and Dipteris, which
possess solenosteles and superficial sori. The ontogenetic facts for the latter genera
are unknown, but grounds for further comparison may be found in the stelar
ontogeny of examples of the “ Marginales,” which will now be considered.

Lindsaya adiantifolia.

Pacts of adult structure for a number of the Lindsayas have been provided by
Tansley and Lulham (59), Gwynne-Vaughan, and others. The internode in the
Lindsaya type shows a dorsiventral stele with parenchymatous xylem-core in which
a continuous internal strand of phloem and parenchyma runs close to its dorsal
surface, and is linked to the outer phloem of both leaf-trace and stele at a xylic gap.
These gaps are in the axils of leaf-traces, and through them endodermal pockets are
decurrent into the internal phloem. In the adult stele of a Lindsaya the stelar
endodermis is thus unbroken. The general arrangement of stelar tissues as seen
in longitudinal section is represented in text-fig. 4 ; xylem is solid black, parenchyma
is white, phloem cross-hatched, and endodermis an unbroken line. The ontogeny
of Lindsaya adiantifolia opens with a solid protostele. Its xylem-core becomes
parenchymatous as it expands (fig. 31). Phloem forms an narrow unbroken band

AND THEIR BEARING ON STELAR THEORIES FOR THE FERNS.

727

in close contact with the wood (Ph.), and the pericycle is broad (Pr.). Text-fig. 5
shows the general arrangement of tissues in median plane of a sporeling. Leaf-
traces are lettered, root-traces are on the left, and levels referred to are numbered.
Fig. 31 is from level I. At a slightly higher level a mass of parenchymatous cells
is seen in the xylem. At the core of this parenchymatous pith sieve-tubes appear
in the immediately succeeding sections. This inner phloem (Ph/, fig. 32) soon
bulks largely in the pith, so that the parenchyma around it is reduced to a narrow
selvage (S). This is the stelar state at II, a leaf-trace (a, text-
fig. 5 ; L, fig. 32) is departing in the protostelic manner, and to-
wards the upper side of the stele the xylem is thin and commonly
interrupted. The number of leaf-traces which depart in the proto-
stelic manner varies from plant to plant, but frequently a xylic
gap follows the formation of the first or second trace after inner
phloem is established. Through these xylic gaps inner and outer
phloem may or may not be united. In text-fig. 5 their union is
shown through the first xylic gap. There is no structural evidence
of intrusion of outer phloem and parenchyma into the xylem-
cylinder, and in these early stages there are no endodermal
pockets. The medulla of thin-walled parenchyma and the sieve-
tubes are then of intraxylic origin. As the stele is followed forward
the arrangement of tissues just described is maintained. This is
indicated in the upper portion of text-fig. 5, and inner and outer
phloem are linked through the xylic gaps above leaf-traces b, c,
and*d. The stelar structure during departure of trace c is shown
in fig. 33. The outer phloem (Ph.) is a narrow band, the inner
phloem is bulky (Ph/). They are united through the xylic gap
before the arc of leaf-trace xylem and phloem (X) is free. As
the stele expands by conical enlargement, the xylic gaps are suc-
cessively longer. Through them pockets of increasing depth are formed within the
stele. Thus the adult Lindsaya state is reached.

It is thus apparent that in the ontogenetic progression from solid protostely to
the adult Lindsaya structure, the formation of medulla and inner phloem involves
neither cortical intrusion nor tissue -flow, but is the result of static intrastelar change
by which also the pockets in the adult stele arise. The position taken on the
structural facts is that in Lindsaya adiantifolia medulla and inner phloem are
initiated in close succession at an early stage in the ontogeny. The phloem is from
the first a bulky solid core around which the parenchymatous pith is a narrow
selvage. The maintenance of this balance and distribution of stelar tissues with the
addition of xylic gaps, endodermal pockets, and phloem-continuity through the gaps,
gives the so-called Lindsaya type of adult stele. It is matter for no surprise that
the Lindsaya state has been recorded for diverse genera, and figures in the early

Text -fig. 5.

728

DR JOHN M‘LEAN THOMPSON ON NEW STELA R FACTS,

ontogeny of a number of higher ferns, in which medulla and inner phloem are early
secured ; for the Lindsaya structure is a stage in intrastelar amplification from an
original protostele which may persist into the adult axis, or be passed through to a
higher stelar state. The Schizaeoid-Dicksonioid affinity of the Lindsay as has been
indicated by Bower (ll) and others. It will accordingly be of interest to know
the ontogenetic facts for a plant of like general affinity but more advanced adult
stelar structure. The facts for Loxsoma Cunninghamii will now be given.

Loxsoma Cunninghamii.

The adult solenostele of this species has been fully described by Gwynne-
Vatjghan (24). The ontogeny opens with a basal region of protostely of varying

length, and bearing both leaf- and root-
traces. The bases of text-figs. 6 and 7 show
the general arrangement of tissues in this
region ; xylem is solid black, parenchyma is
white, phloem is hatched, and endodermis is
an unbroken line ; leaf-traces are lettered,
and levels referred to are numbered. Fig. 34
shows the stelar structure at I, text-fig. 6,
and shows the solid protostely. As leaf-
trace a is departing a medulla appears.
At its core is a strand of sieve-tubes. Fig.
35 shows the stelar structure at this level
(II) ; there is a xylic gap, but the inner
(Ph/) and outer phloem (Ph.) are not united
through it. Fig. 36 shows the structure at
III ; the xylic gap is closed, and the slender
core of phloem persists. For a consider-
able distance leaf-traces are liberated in the
protostelic manner (b, c, d, and e). The
pith is well developed, and the central strand
of phloem is locally indefinite, IV and V.
At VI it is tubular, and involves parenchy-
matous cells. Fig. 37 shows the stelar structure during departure of leaf-trace
d, and illustrates the Lindsaya condition with protostelic departure of a trace.
At VII there is a second xylic gap independent of a leaf-trace. At this level the
inner phloem is indefinite, but no indication of a phloic connection through the
gap is seen. As leaf-trace f is departing, the inner phloem becomes definitely
tubular. Above f is a xylic gap through which inner and outer phloem are
linked for the first time. Fig. 38 shows the stelar structure at VIII ; the inner
phloem (Ph/) is a tube within the parenchymatous pith. The stele is in this

n

AND THEIR BEARING ON STELAR THEORIES FOR THE FERNS.

729

condition as it merges into the undifferentiated mass of
the growing apex. From these facts it may he concluded .
that by early initiation of inner phloem within a purely
intraxylic pith a Lindsaya structure arises. By change
of procambial destination the inner phloem may vary in
amount and position, and the final structure so far evolved
for the specimen considered is a medullated protostele with
inner phloem and xylic gaps, but devoid of endodermal
pockets.

In text-fig. 7 the general arrangement of stelar tissues
in another specimen is shown. In its essential features it
repeats the stelar changes from solid protostely to medullated
protostely with tubular inner phloem, and xylic gaps recorded II
from the first plant. The full resemblance to a Lindsaya
type, as illustrated by Lindsaya adiantifolia, lies in the fact
that the outer and inner phloem may be connected through
xylic gaps (I and II). Towards the apex of the stele definite
endodermal pockets of variable depth are associated with the
leaf-traces. Text-fig. 8 shows the stelar state in part of an
adult plant. At the base the inner phloem is tubular, and
a leaf-trace is departing in the protostelic manner (a).
Associated with leaf-traces c, d, e, and f are endodermal
pockets of variable depth. The pith is bulky, and the
inner phloem is a wide tube more or less coaxial with the
xylem-cylinder. Fig. 39 shows the stelar structure at I,
and illustrates the stelar expansion which has been effected.
Passing upwards, endodermal cells appear in the pith (En/,
fig. 40). These form the base of a pocket (P, fig. 41 ; p, text-
fig. 8), which opens into the first foliar gap (g ; andG, fig. 42),
and extends onwards into a succession of them (g1, g2, and
g3). Thus, as in Gleichenia pectinata , a continuous inner
endodermis is established, and solenostely is attained. Fig. 43
shows the full solenostelic structure at II. The pith (M) is
here a continuous column of more or less sclerotic tissue like
the cortex without, to which it is directly linked through the
succession of foliar gaps. t

No structural evidence has been found of sliding growth
or cortical intrusion to which the establishment of soleno- <
stely could be referred, but the study of the growing point
has led to the conclusion that here, as in other organisms
already considered, the successive changes of stelar structure

Text-fig. 8.

730

DR JOHN M'LEAN THOMPSON ON NEW STELAR FACTS,

from protostely to solenostely are due to a static change of quality of procambial
cells as the growing point advanced.

Acrostichum aureum *

In Loxsoma the internodes are long, and solenostely is delayed until many
leaves have been formed. But in many ferns with upright stocks the internodes
are short, and certain ontogenetic steps are apt to be abbre-
viated or even omitted. This may be illustrated by Acro-
stichum aureum , a plant of Pterid affinity. Text-fig. 9 shows
the arrangement of stelar tissues in the juvenile axis of a
sturdy plant. The stage of solid protostely is brief (I), only
one leaf-trace departing in the protostelic manner before
medullation occurs (a). As the stele expands upwards, a
medullated stage with a xylic gap is reached (II). Higher up
a core of phloem appears in the pith (Ph.). At first it is solid,
but later is tubular and encloses parenchymatous pith (M).
In the latter an inner endodermis is initiated, as in Gleichenia
pectinata, Loxsoma Cunninghamii, and others considered (e).
It is held to be potentially the base of the pocket which opens
into the first leaf-gap, but is not maintained and dies out as
a spindle when followed upwards. Higher up an inner endo-
dermis is permanently established, beginning as a pocket
opening on the one hand into the first leaf-gap (g), and on the
other being continued as the inner endodermis of a typical
solenostele (III). Thus in the ontogeny of Acrostichum aureum
the Lindsaya stage is brief, and enclodermal pocketing is
omitted from the early solenoxylic stages. That its omission
goes hand in hand with the early attainment of solenostely
is the view here adopted. In some of the specimens examined
solenostely arises even earlier in the ontogeny than in the
example described. In them isolated spindles of inner endo-
dermis do not occur, but solenostely is secured with the
first formation of inner endodermis within the pith. The
ontogenetic condensation thus shown for Acrostichum aureum
appears also in Cheilanthes, and in Pteridium aquilinum (33),
(48), and Psesia podophylia (ll), and may be expected to figure generally for
plants of Pterid affinity.

There is already a body of ontogenetic facts which are known for a number of
genera,, The early stages of the sporeling of Anemia phyllitides have been
described by Boodle (3), but the actual transition to solenostely was not traced.

* The materials on which this study was made were kindly provided hy Professor I. B. Balfour, F.R.S.

AND THEIR BEARING ON STELAR THEORIES FOR THE FERNS.

731

Osmunda regalis has been studied in the sporeling stages by Leclerc du Sablon (48),
Jeffrey (33), G-wynne -Vaughan (27), and Sinnott (55), while Faull has provided
the facts of sporeling structure for Osmunda cinnamomea (23). Sew are and Fore
have performed a like service for Todea (54).

Brebner (12), Charles (17), and Farmer and Hill (21) have described the
sporeling structure of members of the Marattiacese ; while Lang has studied in
detail certain of the Ophioglossacese (43), (44), (45), (46), (47). The facts for
Alsophila have been described by Gwynne-Vaughan (25), for Doodia aspera by
Chaneler (16), and for Pteridium aquilinum by Leclerc eu Sablon (48) and Jeffrey
(33). All these, with the records of adult structure for a wide range of Ferns
(see Bibliography), may easily be freed from the theoretical statements which cling
around them. When this is accomplished, they are easily brought into line with
the stelar facts recorded in these pages, and are most reasonably interpreted as
illustrative of a process of static change of procambial elements by which stelar
changes are affected. On this view it is held that the sum of the .ontogenetic
evidence so far obtained for the Ferns shows that medullation and the solenostelic and
dictyostelic states arise in the ontogeny entirely by intrastelar readjustment in
which change of procambial destination is mainly operative.

The object of the present memoir is to provide structural facts, but the meaning
of stelar change will be dealt with latter. A list of the publications which appear
to bear most directly on the subject of stelar development is appended.

Summary.

1. The object of this memoir is to provide the ontogenetic facts of stelar
development for a number of Ferns.

2. The sporeling structure of Schizsea malaccana is first dealt with. From
serial sections it is concluded that the pith is of intrastelar origin.

3. A number of incomplete adult stems of Schizsea dichotoma have been
examined in serial sections, and the stelar structure reconstructed. From these it
is concluded that this species has a medullated protostele, the medulla being of purely
intrastelar origin. Endodermal pockets in this species are held to be intrastelar
structures which arise by change of procambial destination within the stele.

4. The ontogentic progression from protostely to solenostely is described for
Gleichenia pectinata, Loxsoma Cunninghamii, and Acrostichum aureum , and the
evolution of the Lindsay a type of stele is traced in Lindsay a adiantifolia. It is
held that in these, medulla and inner phloem are of purely intrastelar origin, and
that the inner endodermis and solenostelic structure of the adult stele of the former
species arise by purely static change of quality of procambial elements as the apex
of the stem advances. There is no evidence of intrusion of cortical tissues into
the stele during medullation.

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 28).

113

732

DR JOHN M'LEAN THOMPSON ON NEW STELAR FACTS,

5. It is held that the sum of the ontogenetic evidence so far available for the
Ferns shows that medullation, and the solenostelic and dictyostelic states arise in
the ontogeny by intrastelar readjustment in which change of procambial destination
is mainly operative.

DESCRIPTION OF FIGURES IN TEXT.

In these the general arrangement of stelar tissues is given in median longitudinal section, with leaf-
and root-traces reduced to one plane. Leaf-traces and stelar gaps are lettered, and levels referred to
are numbered.

Fig. 1. Schizsea dichotoma, plan of stelar construction in an adult stem. Outer endodermis, En. ; phloem,
cross-hatched ; xylem, black ; sclerenchyma, dotted ; parenchyma, white. There are endodermal pockets
of varying depth, and isolated spindles of inner endodermis and wood in the sclerenchyma.

Fig. 2. Schizsea dichotoma , plan of stelar construction in an adult stem. There is an isolated inner
endodermal spindle in the thin-walled portion of the pith, and tracheides (T) are lodged in the bottom of
an endodermal pocket.

Fig. 3. Gleichenia pectinata, plan of stelar construction in a juvenile plant.

Fig. 4. Lindsaya adiantifolia, plan of stelar construction at an adult node.

Fig. 5. Lindsaya adiantifolia , plan of stelar construction in a juvenile plant.

Figs. 6, 7. Loxsoma Gunninghamii , plan of stelar construction in two juvenile plants.

Fig. 8. Loxsoma Cunninghamii, plan of stelar construction in the upper portion of a juvenile plant which
has attained a solenostelic structure towards its. apex.

Fig. 9. Acrostichum aureum, plan of stelar construction in a juvenile plant.

DESCRIPTION OF FIGURES IN PLATES.

Plates I-IY.

. Figs. 1-6. Schizsea malaccana. Transverse sections of the stele of Schizsea malaccana, at successive
levels in a juvenile plant. 1 is from the protostelic base with solid xylem-core ; 6 is from a level transitional
to the adult stem, and shows the well-developed pith. A, xylem parenchyma ; B, endodermal bridge ;
En., outer endodermis ; L, leaf-trace ; M, pith ; 0, parenchymatous gap in xylem ; Ph., phloem ; Pr., pericycle ;
T, tracheides in the pith ; X, xylem. (All x 40.)

Figs. 7-23. Schizsea dichotoma.

Figs. 7-13. Transverse sections of the stele of adult specimen, reconstructed in text-fig. 1. The sections
are from successive ascending levels referred to in the text ; En., inner endodermis ; G, xylic gap ; Pk., endo-
dermal pocket ; S, sclerenchyma abutting on tracheides ; T, U, isolated tracheides ; W, base' of endodermal
pocket ; Y, base of leaf-trace ; Z, parenchymatous part of pith. (All x 50.)

Fig. 14. Transverse section of inner cortex of an adult plant. ( x 80.)

Fig. 15. Transverse section of sclerenchyma of pith. ( x 80.)

Figs. 16-18. Transverse sections of the stele of an adult plant, reconstructed in text-fig. 2. The sections
are from successive descending levels referred to in the text ; En., inner endodermis : En/, En.", bases of
endodermal pockets. (All x 50.)

Figs. 19-22. Transverse sections of pocket En.", text-fig. 2, at successive descending levels until the
pocket is closed. P, enclosed parenchyma; T, enclosed tracheides which form a spindle situated partly in
the endodermal pocket and partly in the thin-walled selvage and continuous through the endodermis from
the one to the other; T', adjacent tracheides in the selvage. (All x 80.)

AND THEIR BEARING!- ON STELAR THEORIES FOR THE FERNS.

733

Figs. 23-24. Transverse sections of an isolated spindle of inner endodermis in the sclerotic pith of an
adult stem ; En., the endodermis ; T, enclosed tracheides. (All x 80.)

Figs. 25-30. Gleichenia pectinata. Successive transverse sections from the protostelic base to a solenostelic
tract of a juvenile plant reconstructed in text-fig. 3. En.', inner endodermis ; En., outer endodermis ; M, pith ;

Ph.', inner phloem ; Ph., outer phloem ; T, tracheides in a “mixed-pith” ; X, xylem. (All x 45.)

Figs. 31-33. Lindsaya adiantifolia. Successive transverse sections from the stele of a juvenile plant
reconstructed in text-fig. 5. Fig. 31 is from the base; fig. 33 is from a level showing the full “Lindsaya”
structure. L, leaf-trace ; Ph/, inner phloem; Ph., outer phloem; Pr., pericycle ; S, parenchymatous portion
of pith; X, base of leaf-trace. (All x 50.)

Figs. 34-43. Loxsoma Gunninghamii. Transverse sections of the stele of juvenile plants at successive
ascending levels referred to in reconstructions (text-figs. 6, 7, and 8). They show the transition from the
protostelic base (fig. 34) to the full solenostelic structure of the adult (fig. 43) ; En.', inner endodermis ;
G, leaf-gap; M, pith; P, pocket; Ph., outer phloem ; Ph/, inner phloem. (All x 50.)

1 gratefully acknowledge my indebtedness to the Executive Committee of the Carnegie Trust for a.
grant to defray the expense of the plates illustrating this memoir.

BIBLIOGRAPHY.

In the following list of titles no attempt has been made at completeness, but works on stelar anatomy
bearing directly on the subject of this memoir alone are cited.

(1) Bertrand, Paul, “L’etude anatomique des Fougeres anciennes et les problemes qu’elle souleve,”

Progressus Rei Botanicse, Jena, 1913.

(2) Boodle, L. A., “On the Anatomy of the Hymen ophyllaceae,” Ann. of Bot., vol. xiv, London, 1900.

(3) Boodle, L. A., “On the Anatomy of the Schizaeaceas,” Ann. of Bot., vol. xv, London, 1901.

(4) Boodle, L. A., “On the Anatomy of the Gieicheniaceae,” Ann. of Bot., vol. xv, London, 1901.

(5) Boodle, L. A., “Further Observations on Schizeea,” Ann. of Bot., vol. xvii, London, 1903.

(6) Boodle, L. A., and Hiley, W. E., “ On the Vascular Structure of some Species of Gleichenia,”

Ann. of Bot., vol. xxiii, London, 1909.

(7) Bower, F. O., The Origin of a Land-Flora, London, 1908.

(8) Bower, F. O., “ Plagiogyria,” Ann. of Bot., vol. xxiv, London, 1910.

(9) Bower, F. O., “ On the Primary Xylem, and the Origin of Medullation in the Ophioglossacese,”

Ann. of Bot., vol. xxv, London, 1911.

(10) Bower, F. O., “Medullation in the Pteridophyta,” Ann. of Bot., vol. xxv, London, 1911.

(11) Bower, F. O., “The Pteroideas,” Ann. of Bot., vol. xxxii, London, 1918.

(12) Brebner, G., “On the Anatomy of Dans&a and other Marattiaceae,” Ann. of Bot., vol. xvi,

London, 1902.

(13) Britton, N. L., and Taylor, N., “Life History of Schizeea pusillaf Bull. Torrey Bot. Club,

vol. xxviii, 1901.

(14) Campbell, D. H., “The Development of the Ostrich Fern,” Mem. Boston Soc. Nat. Hist., 1887.

(15) Campbell, D. H., The Eusporangiatse, Washington, 1911.

(16) Chandler, S. E., “On the Arrangement of the Vascular Strands in the ‘Seedlings’ of Certain

Leptosporangiate Ferns,” Ann. of Bot., vol. xix, London, 1905.

(17) Charles, G. M., “ On the Anatomy of the Sporeling of Marattia alata,” Botanical Gazette,

vol. xlii, Chicago, 1911.

(18) Conard, H. S., “The Structure and Life-History of the Hay-Scented Fern,” Carnegie Institution

Publication, Washington, 1908.

734

DR JOHN M‘LEAN THOMPSON ON NEW STELAR FACTS,

(19) Coulter, J. M., “Vascular Anatomy and the Reproductive Structures,” American Naturalist,

vol. xliii, 1909.

(20) Farmer, J. B., and Freeman, W. G., “ On the Structure and Affinity of Helminthostachys Zeylanica'f

Ann. of Bot., vol. xiii, London, 1899.

(21) Farmer, J. B., and Hill, T. G., “ On the Arrangement and Structure of the Vascular Strands in

Angioptevis evecta and some other Marattiaceae,” Ann. of Bot., vol. xvi, London, 1902,

(22) Faull, J. H., “The Anatomy of the Osmundacese,” Botanical Gazette, vol. xxxii, Chicago, 1901 ;

and University of Toronto Studies, Biological Series, No. 2, 1902.

(23) Faull, J. H., “The Stele of Osmunda cinnamomea,” Trans. Canadian Institute, vol. viii, 1909.

(24) Gwynne-Vaughan, D. T., “ Observations on the Anatomy of Solenostelic Ferns : I, Loxsoma ,”

Ann. of Bot., vol. xv, London, 1901.

(25) Gwynne-Vaughan, D. T., do., II, Ann. of Bot., vol. xvii, London, 1903.

(26) Gwynne-Vaughan, D. T., “On the Possible Existence of a Fern-Stem having the Form of a

Lattice- work Tube,” New Phytologist, vol. iv, 1905.

(27) Gwynne-Vaughan, D. T., “ Some Remarks on the Anatomy of the Osmundacese,” Ann. of Bot.,

vol. xxv, London, 1911.

(28) Gwynne-Vaughan, D. T., “On a ‘Mixed-Pith ’ in an Anomalous Stem of Osmunda regalis,” Ann.

of Bot., vol. xxviii, London, 1914.

(29) Iossa, M„ Le Developpement de V Appareil Conducteur dans les Rhizomes des Osmundacees et

Gleiche'niace'es, Geneva, 1914.

(30) Jeffrey, E. C., “The Morphology of the Central Cylinder in Vascular Plants,” Trans., Section K,

Report of British Association for the Advancement of Science, Toronto, 1897, London, 1898.

(31) Jeffrey, E. C., “ Gametophyte of Botrychium virginianum,” Trans. Canadian Institute, vol. iv, 1898.

(32) Jeffrey, E. C., “The Morphology of the Central Cylinder in the Angiosperms,” Trans. Canadian

Institute, vol. vi, 1900.

(33) Jeffrey, E. C., “The Structure and Development of the Stem in the Pteridophyta and

Gymnosperms,” Phil. Trans. Roy. Soc., B, vol. cxcv, London, 1902.

(34) Jeb'frey, E. C., “ Morphology and Phylogeny,” Science, N.S., vol. xxiii, 1906.

(35) Jeffrey, E. C., “ Are there Foliar Gaps in the Lycopsida? ” Botanical Gazette, vol. xxxix, 1908.

(36) Jeffrey, E. C., “ The Pteropsida,” Botanical Gazette, vol. xli, 1910.

(37) Jeffrey, E. C., The Anatomy of Woody-Plants, Chicago Press, 1917.

(38) Kidston, R., and Gwynne-Vaughan, D. T., “ On the Fossil Osmundacese,” pt. 1, Trans. Roy. Soc.

Edin., vol. xlv, 1907.

(39) Kidston, R., and Gwynne-Vaughan, D. T., do , pt. 2, Trans. Roy. Soc. Edin., vol. xlvi, 1908.

(40) Kidston, R., and Gwynne-Vaughan, D. T., do., pt. 3, Trans. Roy. Soc. Edin., vol. xlvi, 1909.

(41) Kidston, R., and Gwynne-Vaughan, D. T., do., pt. 4, Trans. Roy. Soc. Edin., vol. xlvii, 1910.

(42) Kidston, R., and Gwynne-Vaughan, D. T., do., pt. 5, Trans. Roy. Soc. Edin., vol. 1, 1914.

(43) Lang, W. H., “On the Prothalli of Ophioglossum pendulum and Helminthostachys Zeylanica,” Ann.

of Bot., vol. xvi, London, 1902.

(44) Lang, W. H., “ On the Interpretation of the Vascular Anatomy of the Ophioglossacese,” Mem. and

Proc. Manchester Literary and Philosophical Soc., vol. lvi, pt. 2, 1912.

(45) Lang, W. H., “ Studies in the Morphology and Anatomy of the Ophioglossacese,” pt. 1, Ann. of

Bot., vol. xxvii, London, 1913.

(46) Lang, W. H., do., pt. 2, Ann. of Bot., vol. xxviii, London, 1914.

(47) Lang, W. H., do., pt. 3, Ann. of Bot., vol. xxix, London, 1915.

(48) Leclerc du Sablon, “ Recherches anatomiques sur la formation de la tige de Fougeres,” Ann. des

Sci. Nat. Bot., ser. vii, vol. xi, Paris, 1890.

(49) Poirault, G., “Recherches anatomiques sur le Cryptogames Vasculaires,” Ann. des Sci. Nat. Bot.,

s4r. vii, vol. xvii, Paris, 1893.

(50) Prantl, Unters. z. Morph, d. Gefasskryptogamen : II, “ Schizseaceen,” 1881.

(51) Russow, “ Vergleich. Unters. d. Leitbiindel-Kryptogamen,” Mem., de V Acad, des Sci. de St-Petersbourg ,

xix, 1872.

AND THEIR BEARING ON STELAR THEORIES FOR THE FERNS

735

(52) Schoute, J. C., Die Steldr Theorie, Jena, 1903.

(53) Seward, A. C., “On the Structure and Affinity of Matonia pectin ata,” Phil. Trans. Roy. Soc.,

B, vol. cxci, London, 1899.

(54) Seward, A. C., and Ford, S. 0., “The Anatomy of Todea, with Notes on the Geological History

and Affinities of the Osmundacese,” Trans. Linn. Soc. Bot., ser. ii, vol. vi, London, 1903.

(55) Sinnott, E. W., “Foliar Gaps in the Osmundacese,” Ann. of Bot., vol. xxiv, London, 1910.

(56) Strasburger, E., “Bau. u. Yerrichtungen d. Leitungsbahuen,” Hist. Beitr., iii, 1891.

(57) Tansley, A. G., “ Lectures on the Evolution of the Filicinean Vascular System,” New Phytologist,

Reprint, No. 2, London, 1908.

(58) Tansley, A. G., and Chick, E., “ On the Structure of Schiz&a malaccana,” Ann. of Bot., vol. xvii,

London, 1903.

(59) Tansley, A. G., and Lulham, R. B., “ On a New Type of Fern Stele, and its Probable Phylogenetic

Relations,” Ann. of Bot., vol. xvi, London, 1902.

(60) Thomas, E. N., “ Some Points in the Anatomy of Acrostichum aureum,” New Phytologist, vol. iv,

London, 1905.

(61) Thompson, J. M., “ The Anatomy and Affinity of Platyzoma microphyllum,” Trans. Roy. Soc. Edin.,

vol. li, 1916.

(62) Thompson, J. M., “ The Morphology of the Stele of Platyzoma microphyllum,” Trans. Roy. Soc.

Edin., vol. Iii, 1918.

(63) Zenetti, P., “ Das Leitungssystem im Stamm von Osmunda regalis und dessen Uebergang an den

Blattstiel,” Botanische Zeitung, Jahrg. 53, Leipzig, 1895.

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 28).

114

Trans. Roy. Soc. Edinr-

Plate I. Vol. LII.

Dr J. M. Thompson on New Stelar Facts, and their Bearing
on Stelar Theories for the Ferns.

Zineo-Collolype Co.. Edinburgh,

| J. M. Thompson, del.

Trans. Roy. Soc. Edin1

Plate II. Vol.LII,

Dr J. M. Thompson on New Stelar Facts, and their Bearing
on Stelar Theories for the Ferns.

J- M. Thompson, del.

24..

25.

Zinco-Collotype Co., Edinburgh.

Trans. Roy. Soc. Ed i nr Plate III. Vol. LIT

Dr J. M. Thompson on New Stelar Facts, and their Bearing
on Stelar Theories for the Ferns.

J* M. Thompson, del.

35.

Zir.co-Collotype Co., Edinburgh.

Plate IV. Vol. LI I ,

Trans, Roy. Soc. Edinr-

Dr J. M. Thompson on New Stelar Facts, and their Bearing
on Stelar Theories for the Ferns.

J* M. Thompson, del.

Zinco-Collotype Co., Edinburgh.-

( 737 )

XXIX. —ISLE OF WIGHT DISEASE IN HIVE BEES.

(1) The Etiology of the Disease. By John Rennie, D.Se. ; Philip Bruce White,
B.Sc. ; and Elsie J. Harvey. (With One Plate.)

(Read November 1, 1920. MS. received December 4, 1920. Issued separately March 25, 1921.)

Introductory.

Isle of Wight Bee Disease has been known in this country certainly since 1904,
when it was first recognised in the island from which it derives its popular name.
According to Imms it was probably present in Derbyshire in 1902, and was also known
in Cornwall and other districts in 1904. Prior to these dates periodic losses of bees of
a serious character are on record, dating as far back as the middle of the eighteenth
century. Bullamore and Malden (1912) haye summarised fully these outbreaks in.
historical series in their report in Journal of Board of Agriculture, Supplement 8,
xix. From a study of the records which they have brought together and from
personal inquiries which we have made at various bee-keepers of wide experience,
it would appear that none of these earlier outbreaks attained the general distribu-
tion throughout the country which we know in Isle of Wight Disease at the present
date, nor did any of them remain established over such an extensive period of years
as that which has continued without interruption from 1902 until the present time.

Another striking characteristic, and one which has an important bearing in any
investigation which seeks to trace the source of this malady, is the fact that no
disease of such a permanent and extensive nature has so far been recognised unmis-
takably outside the British Isles. All the facts we are at present aware of suggest
its definitely localised and insular character.

Since 1907 investigations into the cause of this disease have been carried out by
a series of workers who have from time to time reported upon the subject : Imms
(1907), Malden (1909), Graham Smith, Fantham, and others (1912 and 1913) in
England ; and Anderson (1916), Anderson and Rennie (1916), Rennie and Harvey,
No. 1 (1919) and No. 2 (1919), in Scotland.

The net result of these investigations has been that the English workers (Graham
Smith et alii, 1912 and 1913) put forward as the causal organism the protozoan,
Nosema apis. It is due to Anderson amongst the Scottish workers to state that
he was the first to call this conclusion in question, and we think that the later work
referred to above (Anderson and Rennie, and Rennie and Harvey, 1919) has succeeded
in establishing (l) that Isle of Wight disease and Nosema infection are not
coincident, and (2) that there exists a distinct 'disease due to Nosema apis, which

TRANS. ROY. SOC. EDIN., VOL. LII,. PART IV (NO. 29). 115

738 DR JOHN RENNIE, MR PHILIP BRUCE WHITE, AND MISS ELSIE J. HARVEY

exhibits, however, totally different external symptoms and a distinct pathology in
the individual bee. Collateral work by G. F. White (1918) in America supports
this latter conclusion. The problem of the cause of Isle of Wight disease until now
has thus been left unsolved.

Characteristics of the Disease as Hitherto Observed in the
Colony as a Whole.

The diagnosis of Isle of Wight disease from “ symptoms ” has always been a more
or less unsatisfactory procedure. Hitherto the presence of the disease in a colony
has not been recognised until infection has been well advanced in a high proportion
of the bees. At this stage of disability, the most usual features recognisable by the
bee-keeper are inability to fly, accompanied sometimes with imperfect folding of the
wings. In fine weather a proportion of the affected bees may leave the hive and
crawl around, climbing grasses, etc. Later, in the cooler part of the day, they
commonly collect in small clusters. Such bees are lost to the colony, since they do
not return to the hive, and in any case are useless as workers at this stage. Some-
times large numbers come out and loiter on the alighting board in the sun, returning
to the hive when the sun has gone. Associated with the incapacity for flight there
is- usually a congested condition of the colon. In certain circumstances dysentery
may be present as a complication. Most of these symptoms may be present in
other disorders of a more temporary kind, and we have been accustomed to regard
as true Isle of Wight disease only those cases where such visible conditions, once
commenced, continued in the stock, affecting succeeding broods of bees. There is a
continuous mortality from the disease. Bijllamore and Malden regard no single
symptom as characteristic, and state that “ the only essential feature is the death of
large numbers of bees.”

The association of the causative organism now to be considered will henceforth
afford an exact means of diagnosing the disease, which we suggest should now be
designated Acarine disease.

Discovery of the Causal Agent of Isle of Wight Disease.

The present and following papers announce the discovery of a parasitic organism
invading the respiratory system of the adult bee, which after exhaustive investiga-
tion we now bring forward as the causal agent in this disease. This parasite is a
hitherto undescribed mite, identified by one of us (J. R.) as belonging to the genus
Tarsonemus. It was first observed by one of us (E. H.) in December 1919, when
a single example was found in a portion of trachea present in a preparation, per-
manently preserved, of the thoracic glands (fig. l).# It was significant of the fuller
knowledge of the disease, soon to be attained, that the bee in which it occurred was

* This find was followed up at the time by a systematic search for mites in hives, upon frames, bees, etc., which
resulted in the finding of no fewer than five different species in definite association with bees, dead and alive. (J. R.)

ON ISLE OF WIGHT DISEASE IN HIVE BEES — ETIOLOGY.

739

“ healthy” in the. sense that it belonged to a colony which had no history of disease
and was regarded as free from such. In the following May, Mr White made the
further and independent discovery that mites in all stages of development occurred
in certain of the major thoracic tracheae of “ crawling ” bees. In reporting this
discovery he stated that he had found this condition in at least 150 sick bees, repre-
sentative of several diseased stocks, and also that he had failed to find mites in
95 per cent, of apparently healthy bees.* On this occasion he expressed the view
that the parasites seen by him bore a definite etiological relationship to the disease.

That this discovery was one of very great, significance was obvious, and the
senior author immediately proceeded to its further verification. The first stock of
bees examined was one in a highly prosperous condition. The bees were occupying
twenty frames in the latter half of May ; they were working splendidly and a source
of great satisfaction to the owner. Twelve flying bees were captured as they entered
the hive. They were taken to the laboratory and the first bee opened was found to
contain the parasite in limited numbers. All the twelve were examined, and of the
remaining eleven, one other was found also to be harbouring the mite in question.

These facts presented, as in a nutshell, the problem confronting us. It was
evident that the distribution was not limited to those bees or stocks hitherto
regarded as “ sick,” and as the result of’ the extensive investigations which followed
and which are now recorded, we are able to announce that notwithstanding
variations in the course of the disease in different stocks, we regard it as established
■that there is an invariable association of the parasite with diseased stocks, and that
there is a definite pathology in relation to infection in the individual bee.

Incidence of Infection within the Colony.

Brood. — Except in one doubtful case we have not found infection with Tarso-
nemus in brood of any stage, nor has it been found amongst the very youngest
of adult bees.

Workers. — Amongst workers the infection is most marked in the older bee, and
in any case it has not been found except in incipient stages amongst nurse bees
whose adult life has been brief.

Drones. — We have found that drones of affected stocks suffer equally along with
the workers.

Queens. — The important and interesting question of “the relation of the queen in
this connection has been investigated by one of us (J. R.) in a limited number of cases.
Of fifteen queens examined of stocks known to be affected with this disease, ten were
found to be free, whilst the remaining five harboured this organism. That queens
may undergo infection is specially interesting, in view of the familiar fact that the

* The systematic examination which followed showed the number of apparently healthy stocks harbouring the
parasite to be as high as 36 per cent. (J. R.)

740 DR JOHN RENNIE, MR PHILIP BRUCE WHITE, AND MISS ELSIE J. HARVEY

queen of an infected Isle of Wight diseased stock survives usually' until the colony is
extinct. That her survival is in a measure due to the fact that she remains within
the hive, is supported by the knowledge we already possess that workers affected
with the disease may live for months after they are incapable of flight, and are thus
useless to the colony.

Region of Infection within the Bee.

The mite, Tarsonemus, occupies a very restricted region in that part of the
tracheal system which has its origin at the anterior thoracic spiracle. In a well-
established case of infection it will be found that, extending inward from this
spiracle on either side indifferently, parasites in all stages of development may be
present in any part of this portion of the respiratory system, whilst the ill effects of
their presence may be seen not only in the region of occupation, but in the muscular
tissue to which these extend. It is not an infrequent occurrence in advanced
cases of the disease for these wider tracheae to be occupied with mites in closely
packed formation. All stages of development occur ; e.g. ova, larvae, nymphs, and
adults may be found together (figs. 2 and 5). In the smaller branches frequently
these are occupied as far as their diameter will permit, when a single individual
may be found practically blocking the tube, and sometimes a linear succession of
individuals may be seen in such a position.

The facts which have led us to the conclusion that the occurrence of this
organism in the position indicated is to be regarded as causally related to this
disease, are to be found not alone in the presence of Tarsonemus in the respiratory
system of the bee. There is the universal coincidence of its occurrence in diseased
bees. Further, we have been able to trace the development of the disease within
bee colonies from the earliest stages of infection to its complete manifestation in
crawding and other definite symptoms. We have observed that the total effects
resulting from its development, feeding upon the bee Jnd life generally within it,
renders it useless as a working unit, disorganises the social system and eventually
shortens the bee’s life. Further, these vital effects are accompanied by visible
pathological conditions in the tissues. The most obvious of these is a browning or
blackening and thickening of the tracheal wall (figs. 6 and 7). The thickened
tracheae become progressively hardened and brittle in texture, and certain muscle
fibres become atrophied. This latter aspect of the problem is the subject of
separate detailed treatment, in the paper which follows by Mr White.

These pathological appearances in an infected bee may be present on both sides
of the anterior tracheal system. What we have described is the condition in a well-
established instance where breeding has been in progress for some time, but as has
been mentioned early stages of infection have been frequently witnessed in which
the number of parasites present have been observed to be as few as a single mite and
no abnormal condition apparent.

ON ISLE OF WIGHT DISEASE IN HIVE BEES — ETIOLOGY.

741

Cumulative Evidence that Tarsonemus is Causally Belated
to this Disease.

In the course of our investigation we have searched over three thousand individual
bees representing 250 separate stocks scattered throughout Great Britain. These
examinations covered over 110 stocks reported to us by reliable bee-keepers or certi-
fied by ourselves as suffering from Isle of Wight disease. The parasite was present
in every one of those stocks. A striking result of this part of the inquiry, which
involved the examination individually of 700 bees at least, was the discovery that in
every case showing the familiar symptoms of Isle of Wight disease the parasite was
present. No exception has been found. There is apparently an invariable and clear
association of this organism with all bees suffering from Isle of Wight disease.

These examinations applied not only to bees obtained during 1920, but included
samples representative of all seasons of the year, and dating back as far as September
1916. These observations relating to the earlier dated bees were made upon diseased
bees which had been preserved by Bennie and Harvey on the dates mentioned
(see p. 752).

Keputed Healthy Stocks.

Amongst the 250 stocks above mentioned there were about 50 which were
reported to us as healthy and in which we found the parasite Tarsonemus to be
present. That is to say, of 140 stocks believed by the owners to be healthy, 50, or
nearly 36 per cent., harboured this parasite. Concurrent with such discoveries
we ascertained by direct examination ourselves of flying bees (l) which were
members of colonies in which the disease was definitely established and (2) which
were taken from colonies believed to be healthy and showing no indications otherwise,
that amongst these were to be found considerable numbers harbouring the parasite.
This was further complicated by the fact that in those infected flying bees certain of
those pathological conditions — eg. the blackening and hardening of the tracheal
tubes — were very marked. As an example it may be quoted that this condition was
found in bees entering the hive carrying pollen or nectar, both belonging to stocks in
which crawling and other symptoms were well established, and also to those reputed
healthy stocks.

A Particular Case.

As an illustration of this aspect of the disease we may quote the following : —

At the door of the hive of a sick stock showing habitual crawling in fine weather
and steadily declining from the disease, we captured as they alighted 27 foraging
bees in the course of a single afternoon. Tarsonemus was found in every one of these
bees, all stages of development being represented. In a number of the cases, soiling
and destruction of the tracheal tubes was very marked, quite as bad as anything we
have observed in bees crawling from the disease.

742 DR JOHN RENNIE, MR PHILIP BRUCE WHITE, AND MISS ELSIE J. HARVEY

In several other stocks showing the disease in an advanced stage, every bee
taken over a period extending to weeks, including drones, flying and crawling
workers, was found on examination to be infected. The hying workers were
frequently more heavily parasitised than were the bees of the same stock which
were unable to hy.

These facts have shown us that “ crawling ” is only one of the phases of the disease
and that it cannot be dependent exclusively upon the intensity of the infection, as
shown by numbers of parasites ; it may be incidental, in part at any rate, to a critical
position of certain of the mites, so that oxygen starvation of groups of fibres of the
muscles of flight results. Evidence in support of this is brought forward in Mr
White’s paper which follows, where also other possible factors are considered.

The following Table summarises the typical results of our examinations of bees
for the presence of Tarsonemus. H. indicates that the stock from which the bees
were obtained was showing no indications of disease, and was described by the
owners as Healthy.

S. stands for Sick stock, and invariably indicates that the ordinary symptoms of
Isle of Wight disease known to bee-keepers were present. The figures quoted
under the heading Tarsonemus indicate the number of bees in which Tarsonemus
was found.

Table I.

Ref.

No.

Date.

Locality.

Condi-
tion of
Stock.

Numbers

Examined.

Tarso-

nemus.

Remarks.

1

22 May

Stonehaven

H.

12

2

Developed the disease subsequently.

6

26 „

Edinburgh

H.

20

o

7

Rubislaw

H.

2

1

No. 44 in record.

10

14

Aberdeen

S. .

13

11

Requeened. Reduced to 1 frame, 30th
October.

18

22

Kintore

S.

12

12

Crawling marked.

22

27 ’’

N orthumberland

s.

5

5

Stock died out.

37

4 June

Copford

H.

31

0

Infected artificially later. Became
queenless. Died out subsequently.

48

6 ,,

Shipton

H.

3

0

57

8 „

Cults

H.

12

2

Died out from disease and queenlessness.

55

9 ,,

Banchory

S.

3

3

62

14 „

Kirriemuir

H.

9 .

0

70

17 „

Huntly

H.

5

5

Early stage of infection. Disease very
prevalent in immediate neighbour-
hood and robbing noticeable.

74

17 „

Glasgow

H.

34

3

No disease signs so far (30/12/20).

75

18 „

Drumlithie

s.

6

6

A stock dwindling from the disease.
Minor tracheae were blocked.

78

20 „

Cawdor

H.

30

0

(4 stocks.)

82

20 „

Aboyne

H.

6

6

Last stock of 6, — 5 of which have
already died of disease.

86

22 „

Laurencekirk

S.

20

17

An ordinary case.

92

21 „

Dingwall

H.

6

6

Stock much reduced and in bad condi-
tion, but no crawling had been seen.

94

21 „

Aberdour, Fife

H.

20

0

No record of disease in this apiary.

ON ISLE OF WIGHT DISEASE IN HIVE BEES — ETIOLOGY.

743

Table I. — continued.

Eef.

No.

Date.

Locality.

Condi-
tion of
Stock.

Numbers

Examined.

Tarso-

nemus.

Remarks.

97a

26 June

Bristol

S.

6

6

Heavily infected from a cast which up
till a few days previously appear
healthy. Crawling sudden and ex-
tensive.

99

28 „

Banchory

s.

6

6

An ordinary advanced case.

100

26 „

Rubislaw

H.

30

0

See record, No. 60.

102

29 „

Edinburgh

S.

13

10

Has shown none of usual signs of Isle of
Wight disease, but has dwindled since
spring. Queen found free of infection.

103

29 „

Northumberland

s.

12

12

105

30 „

Turriff

S. 1 ■

7

5

Intermittent signs of disease shown.
These bees not known to have the
disease.

108

2 July

Ellon

s.

8

6

Reported 17th September: — “Now
very healthy and closely packed down
on ten standard frames. Gave one and
a half crates shallow frames surplus.”

111

2

Ellon

H.

13

0

Same apiary as No. 108.

112

2

Ellon

H.

1

Strong stock all along, swarmed twice.
Gave two and a half crates shallow
frames surplus. Never showed any
signs of disease.

115

3 „

Port Elphinstone

H.

30

0

See record, Re No. 1.

116

3 „

WarriloWj Sussex

H.

7

0

118

5 „

Cluny

S.

6

6

Intermittent disease history from pre-
vious September. Died out about
this date.

119

5 „

Nairn

H.

5

0

120

5 „

Cults

S.

13

13

An 'ordinary advanced case.

124

3 „

Inverurie

s.

6

6

Do.

126

8 „

Glasgow

H.

30

3

There have been no signs of the disease.

128

8 „

Keig

S.

6

6

299

8 „

Witney, Oxon.

s.

6

6

Slight crawling only, at this stock.

129

8 „

Inverness

s.

6

6

Crawling observed for months pre-
viously. Infection very heavy.

134

8 „

Rubislaw

H.

30

5

Died subsequently of the disease.

140

9 „

Rubislaw

s.

19

19

From 4th June onward this stock by
repeated examination (4 times in one
month) showed practically every bee
infected. It appeared to be doing
moderately well. Mass crawling was
never observed, but the stock became
queenless and died out before the
end of July.

141

10 „

Inverurie

s.

10

10

146

11 „

Boat of Garten

H.

3

3

Stock believed healthy; from apiary
where other four stocks had died
from the disease.

148

13 „

Park

s.

4

4

Stock crawling. Drones were included
in this examination.

149

14 „

Boat of Garten

H.

14

10

Not known to be sick, but some bees
seem to have dislocated wings.

165

15 1

Coull

S.

5

5

Yielded a , good top swarm, and crawl-
ing appeared subsequently. Tracheae
were closely packed with mites in
all cases.

744 DR JOHN RENNIE, MR PHILIP BRUCE WHITE, AND MISS ELSIE J. HARVEY

Table I. — continued.

Ref.

No.

Date.

Locality.

Condi-
tion of
Stock.

Numbers

Examined.

Tarso-

nemus.

Remarks.

166

15 July

Rubislaw

H.

20

3

Infection slight. This stock died out
in the course of the summer.

167

16

”

Rubislaw

H.

20

0

This stock later developed an infection.
It is still in existence.

168

16

,,

Dingwall

H.

8

8

169

16

Italy

H.

22

0

Two samples. Attendants accompany-
ing queens. Both lots infected with
Nosema.

170

16

,,

Bandon (Ireland)

H.

26

3

153

11

”

Kinaldie |

7

7

Roof stock. All these bees were
crawlers.

154

11

,,

Kinaldie j

30

8

Same stock. All these bees were fliers.

175

18

Nairn

H.

11

0

180

19

’

Rubislaw

H.

30

0

Standing in an apiary beside diseased
stocks. It has since become in-
fected, but there are no external
indications.

190

21

”

Dorridge

S.

6

4

A strong stock of Italian hybrids with
a 1919 queen.

191

21

, ,

Dor ridge

H.

28

0

An Italian nucleus with a 1920 queen.

195

29

Fyvie

S.

7

7

Stock reported crawling for weeks pre-
viously.

199

6

Augt.

Norfolk

H.

12

3

Stock remains apparently healthy.

201

14

Norfolk

H.

3

2

A light infection.

202

14

,,

Old Meldrum

S.

6

6

Died out from the disease.

216

18

„

Dingwall

s.

10

10

A bad case.

217

18

Birkenhead

H.

10

0

219

20

”

Birkenhead

s.

10

10

From 2 stocks, examined separately ;
both infected.

222

20

J9

Lucerne

H.

29

0

From 3 stocks.

230

26

Lugano

H.

44

0

From 2 stocks.

235

28

”

Wick

S.

10

10

A suspected stock. A definite case of
infection in an area only recently
affected by the disease.

239

30

,,

Washington, U.S.A.

H.

37

0

243

1 Sept.

Beauly

H.

9

0

263

22

j.

Bucksburn

S.

10

5

264

23

Washington, U.S.A.

H.

23

0

268

27

Eddieston

H.

21

0

Maintained in an area free from disease.

269

25

”

W eston-s.-Mare

S.

11

11

Two stocks ; examined separately, both
infected.

273

28

Austria

H.

20

0

276

28

Stoke-u nder-Ham

H.

4

0

277

28

»

Stoke-under-Ham

S.

4

4

Not known by owner to be sick, but
crawling bees found.

283

17

Oct.

New Machar

H.

9

0

287

17

New Machar

S.

10

10

Trachea; badly blackened.

290

28

»

Rubislaw

H.

25

4

In infected apiary ; no external signs
of disease.

302

30

”

Bungay

H.

30

27

Some crawling about middle of month.
Put up for winter.

303

30

”

Bungay

H.

28

23

No sign of crawling. Put up for
winter.

ON ISLE OF WIGHT DISEASE IN HIVE BEES — ETIOLOGY.

745

Table II.

The following illustrates the progress of infection as observed in the periodic
examination of 4 stocks of the same apiary : —

Date.

Locality.

Condition
of Stock.

Number

Examined.

Tarsonemus.'

20 May
1 July

Fintry No. 1

H.

H.

6

3

Flying bees.

3 Aug.

,,

S.

7

7 Crawling bees.

»

s.

5

4 Fliers.

20 May

Fintry No. 2

H.

5

0 Fliers.

8 June

H.

8

2 Fliers.

3 Aug.

>»

S.

4

4 Crawlers.

20 May

Fintry No. 3

H.

24

0 Fliers.

8 June

H.

16

1 Fliers.

1 July

S.

10

3 (8 fliers, 2 crawlers).

3 Aug.

..

s.

22

22 Crawlers.

20 May

Fintry No. 4

H.

15

0 Fliers.

3 Aug.

S.

6

6 Crawlers.

The data set forth in the foregoing Tables are thoroughly representative of our
results as a whole. Examination of the figures quoted will show : —

1. That in every case of a “ Sick” stock, Tarsonemus was present in the stock,

and in a high proportion of cases it was found in every bee examined.

2. That in the majority of stocks marked “ Healthy,” Tarsonemus was not

found.

3. That in a proportion (36 per cent.) of supposed “ healthy ” stocks, Tar-

sonemus was found. Of these within the period of observation (five
months at most) a proportion eventually developed the usual symptoms,
and died out from the disease. A proportion died out without having
shown the symptom of “ mass crawling,” and a few remain apparently
healthy.

Regarding the admittedly diseased stocks in which Tarsonemus has been found,
we deem further illustration unnecessary. Concerning those other cases, some of
which may appear to present difficulties, we consider it important that details should
be submitted, and of these we now quote typical examples.

For ease of comparison the main facts regarding each are summarised at the end
of the series.

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 29).

116

746 DR JOHN RENNIE, MR PHILIP BRUCE WHITE, AND MISS ELSIE J. HARVEY

Stock Records.

R. No. 1. — This stock at the end of May was covering fully twenty frames and
was in very good condition. On the 22nd of this month twelve bees taken entering
the hive were examined for the presence of Tarsonemus. Two of the twelve were
found affected at the initial stage. The parasites were few and the tracheae were
perfectly clean. As stated, there were no signs of disease. About a month after-
wards the stock, which meantime had worked well and shown no signs of disease, was
again examined. On this occasion fifty-two bees were searched, and of these forty-
three contained the parasite. In most cases the tracheae were heavily infected but
the tubes were comparatively clean. A further sample was obtained upon the 6th
July, and at this time twenty -five bees out of a total of twenty-eight taken were
infected. It should be stated that these bees were taken at random by shaking off
a frame into a box placed below. A number of these showed a bronzing of the
tubes, especially at the forks. By this time the bees were showing some listlessness
and not working so well. The owner made an artificial swarm, removing the old
queen and supplying the main stock with a virgin Italian queen. Twenty-three
pounds of drained honey were obtained at this time. The two stocks were
subsequently placed side by side. On 21st July, in a sample of thirty-five bees,
twepty-eight contained the parasite. About the third week of August, after a period
of cold weather, crawling became evident in both stock and swarm. About the end
of August a sample of twenty-eight bees was supplied from the parent stock, and of
these twenty-six were badly parasitised. Both stocks continued to crawl in large
numbers, and as robbing by other bees was going on, the owner destroyed them
about the end of September.

No. 44. — This stock was obtained upon 11th April from an apiary which has
been in existence for many years, and in which Isle of Wight disease has never
been known. It was placed on the date mentioned in a new hive upon its own
frames in an experimental apiary in which there were stocks suffering from Isle of
Wight disease. The stock progressed normally throughout the summer and by the
middle of June the bees were working in a super. On the 26th May an examination
made on two bees taken from the stock showed one to be infected with Tarsonemus.
Two days later one out of eight was found similarly affected. Subsequently,
periodic examinations as follows were made on the dates mentioned.

Date.

No.

Examined.

Result.

Infected.

Not

Infected.

12 June

31

3

17

22 July

10

5

5

31 Aug.

33

32

1

ON ISLE OF WIGHT DISEASE IN HIVE BEES — ETIOLOGY.

74 7

Although during the wh'ole of this period crawling symptoms never were in
evidence and the bees appeared to be working normally, the numbers did not
increase, nor were stores accumulated. In the later part of July and August the
decline in numbers was rapid, and crawling developed towards the end. The stock
was robbed actively and became extinct towards the middle of September. The
apiary from which the stock was obtained remains clear of the disease.

No. 61. — On the 20th May a stock of Italian bees infected with Nosema apis
was obtained from Glassel. It built up rapidly in the course of the following weeks
and by the middle of July the bees were covering fifteen frames.

On the 30th May four bees were taken entering the doorway ; these were active
and inclined to sting ; two of them were found infected with Tarsonemus and two
were clear. On 7th July thirty bees were taken and of these five were infected.
Again, on 14th July twenty bees were taken ; ten were infected with Tarsonemus
and ten were free. Up to this date there were no external signs of disease in the
stock. The season being poor there were no surplus stores, but, as already stated, the
stock was strong in bees.

On 30th August thirty-four bees were examined and of these twenty-seven were
found infected, and about this date crawling amongst the bees was observed for the
first time. By this date the stock was reduced to about seven frames of bees, and
robbing by other bees was being persistently attempted. It eventually died out at
the end of September.

No. A. Ch.— In the month of May a stock was obtained from an apiary in
Dyce, where there had been no disease for many years. It was a swarm of the
previous year, and after transference was isolated from other stocks. On 17th
July six bees taken from the stock, which appeared perfectly healthy, were all
found infected, but to a slight degree. The infection appeared to be recent. A
fortnight later crawling became evident in the stock, and six crawling bees
supplied were found to be all infected and more extensively than in the previous
sample. By the 1st September the stock had declined to about four frames of
bees. There were no stores. Nine flying bees were taken ; these were all found
infected and' having their tracheal tubes much blackened. The owner at this date
destroyed the bees.

No. Glasgow , I. — On the 8th July a sample of thirty bees was taken from a
stock of Dutch bees obtained from West of Scotland College of Agriculture on
22nd March. The stock at the time the sample was taken appeared perfectly
healthy and was doing well. Of the thirty bees, three were found infected with
Tarsonemus. The stock swarmed, and the swarm for a time appeared strong and
healthy. At the beginning of August the parent stock covered eight frames, with
stores and brood on six. No loiterers or crawlers have been observed at either the
parent or swarm stocks. Thirty -five bees of the parent stock were examined at this
date, and of these three were found infected with Tarsonemus.

748 DR JOHN RENNIE, MR PHILIP BRUCE WHITE, AND MISS ELSIE J. HARVEY

It remained so at the end of October, although at this date it showed a propor-
tion of infected bees of about 27 per cent. (33 bees, 9 infected).

At the present time (30th October) this stock is strong in numbers, without
visible signs of disease. The owner has united it with one half of the swarm.

No. Glasgow, P. — This stock at the beginning of June was strong and working
well. The bees covered ten frames, six of which were very well filled with brood.
Brood was also present upon the other four and there were plenty of stores. The
stock swarmed at this date but the swarm, secured with difficulty and with a loss
of bees, was returned to the parent stock. On 17th June a sample of thirty-four
bees was examined, and of these three were found infected with Tarsonemus.
There were no external signs of disease and matters appeared normal with the
stock. During the next three months the owner paid little attention to it, and
in September the bees were reduced to four frames with brood and were without
food. None of the usual signs of Isle of Wight disease had ever been seen about
the stock, which was now being fed. A sample of twenty-nine bees was taken on
7th September, and of these two only were infected. This shows a slight decrease
as judged by the samples. A later sample supplied at the end of October, how-
ever, showed an increase in proportion of infected bees. The stock as a whole does
not appear affected by the presence of the parasite, but it is not particularly strong
in numbers.

No. 62. — Early in May of this year a nucleus of three frames of bees with queen
and brood was obtained.

On 27th May fifteen bees, and again on 3rd June four bees, were examined for
Tarsonemus, with a negative result. The bees were standing in an infected apiary
and at this time were working well and rapidly increasing in numbers.

On 14th June twenty -five bees were examined, and of these twenty-two were
found clear of the parasite ; of the remaining three, two contained several
parasites and one a single adult female. The bees multiplied rapidly and swarmed
twice in the course of the summer.

On 15 th September thirty -three bees were examined, and of these nineteen
proved infected.

At this date there were no signs of disease as far as behaviour of the bees
was concerned. The numbers were well maintained and the stores sufficient.

At the end of October, forty bees were taken and all except three were found
infected. The pathological features were not marked. The stock is apparently in
a strong condition as regards numbers at this date.

No. 60. — On 20th May a small lot of bees covering three frames was obtained,
which on examination was found to be harbouring the parasite Nosema apis.
Apart from this there were no external signs of disease about the stock, and it
built up moderately well. By 2nd August the bees covered over nine frames, with
brood upon seven. The season was poor and stores were short.

ON ISLE OF WIGHT DISEASE IN HIVE BEES — ETIOLOGY.

749

At the middle of September the stock appeared well, apart from a shortage of
stores. Further at this date Nosema was still present.

In the course of the summer, bees of this stock were periodically examined
between 17th June and 27th September for the presence of Tarsonemus. In all
one hundred and fifty-six bees were tested and on only two occasions, namely upon
20th August and 27th September, was Tarsonemus found. In each case only one
bee was found infected. This and a preceding stock (No. 61 ) were obtained from
the same apiary and have stood together, but a little way apart from the other
stocks, during the period of observation. Several of these other stocks in the
same apiary were at this time suffering from Isle of Wight disease. At the end
of October a sample of thirty-five bees was taken off the frames, and all were found
free from Tarsonemus.

W. No. 2. — This stock, on 4th August, headed by a young queen, appeared
normal and in good condition. There were no visible grounds for suspecting
infection. Of twelve bees taken at this date, three were found harbouring
Tarsonemus.

On 6th September the bees were covering most of the frames, and there was a
good amount of sealed brood and eggs in the inner frames. There were no signs of
disease. At 20th September one crate of sections honey was obtained and about
20 lbs. of stores were left in the hive. At this date fourteen bees were taken at
random, and of these twelve showed infection with Tarsonemus and two were clear.

At the end of October, of twenty-eight bees supplied, twenty-three were found
infected.

No. 24. — This was a nucleus of five frames of Italian bees obtained upon 11th June.

Twenty -two bees were examined at this date and these were found to be free
from Tarsonemus infection. A fortnight later thirty-two bees were examined and
one bee only was found containing this parasite ; everything appeared normal with
the stock. Throughout the summer the stock prospered only moderately well. In
the first week of September twenty-three bees were taken at random, and of these
one only contained Tarsonemus. The stock has yielded no surplus stores and there
has been no indications of disease. Nosema ajpis is not present in the stock and the
apparent weakness cannot be attributed to it or to Tarsonemus.

Twenty-two bees were examined, and all found negative, on 29th October 1920.
It would, therefore, appear that although Tarsonemus was present in the stock as
early as the 25th June, there was no apparent increase in the incidence of infection
as late as the end of October.

Re. No. 1. — This stock was brought to Aberdeenshire in the month of June
from Caithness-shire. The bees, headed by a 1919 queen, were bred in this district
in an apiary which had existed for many years and has had no experience of the
disease. Thirty bees were examined on 3rd July and all were found free from
Tarsonemus infection. The bees were placed in an area which has not been free

750 DR JOHN RENNIE, MR PHILIP BRUCE WHITE, AND MISS ELSIE J. HARVEY

from Isle of Wight disease for a long time. On the 9th September the owner
reported “ so far the bees have done well and the stock is strong, but I have made
no attempt to take honey ; indeed, I have been feeding a little recently just to keep
the queen breeding so as to supply young bees for winter.”

At this date a second sample, consisting of twenty -two bees, was supplied, and of
these, one bee was found harbouring Tarsonemus. The infection was localised just
inside the spiracle of one side, and was limited to one adult and a few ova.

It is practically certain that infection of the stock in this case was effected
within three months and probably not much earlier. In other words the stock
stood in a highly infectious area for over two months without contracting the disease.

At the end of October, the owner reported the stock as “ specially strong.” A
sample of six bees was received, and of these, one was found harbouring Tarsonemus,
the other five being free.

Re. No. 2. — A second stock of similar origin, and with queen of same age, was
obtained and placed alongside No. 1, just described, upon the 9th July. Exactly
two months afterwards a sample of fifteen bees was taken and found free from
infection. The stock is strong and is receiving similar treatment to the other.

Of a sample of twenty-one bees of this stock examined at the end of October,
twenty were free from infection and one showed an initial infection, consisting of
a few mites near the spiracle on one side. The owner reported it as “ lively, and
taking in pollen. There have been no signs of crawling about this hive. There is
plenty disease in the neighbourhood.”

R. No. 2. — Early in August a presumed healthy stock of bees was placed along-
side two stocks both at the crawling stage of Isle of Wight disease. At the end
of August thirteen bees were taken from amongst the foragers as they entered the
hive. Ten were clear of parasites and the remaining three were affected, all of
them slightly. Two of the infected cases showed only one or two adults and a few
ova just within the spiracle of one side. Infection had evidently taken place during
the period the stock was upon this site and not before. Upon advice given, the
stock was removed at the end of August some distance from the others referred
to above. At the end of October the owner reported : “It has filled up fairly well
on the heather, is very lively and seems all right. To-day they are gathering in
pollen and I send you a sample from those that were flying out and in.”

The sample contained thirty bees, and of these twenty-nine were infected with
Tarsonemus. There was a fair amount of bronzing of the tracheae, all stages of
development were present, and in a number of cases the mites were densely packed
in the outer tubes.

W. No. 3. — This stock, requeened in the middle of July, was normal and in
good condition on 14th August, when, out of a small sample of three bees, two were
found to be infected. On the 6th September the stock was examined and found to
have bees covering nearly ten frames, with plenty brood and eggs. This stock has

ON ISLE OF WIGHT DISEASE IN HIVE BEES — ETIOLOGY.

751

yielded four crates of sections and had fifteen pounds of stores left in the hive. The
owner writes, “ I am pleased to state there are no signs of any trouble.” Seven
bees were supplied from the stock, and of these five were harbouring Tarsonemus
at the same date.

About the middle of October a slight amount of crawling was observed in this
stock, but weather conditions have prevented further observations. At the end
of October a sample of bees supplied was found heavily infected. The stock
continues under observation.

Notes on Stock Records.

R. No. 1.— In about six weeks after the infection was first discovered, but not
until the incidence of infection had risen to over 89 per cent., did visible signs of the
presence of the disease appear.

No. 44. — This stock certainly developed an infection of Tarsonemus within the
period of 11th April to 26th May, i.e. about six weeks. The examinations showed
a rapid spread of Tarsonemus within the colony, so that in a little over four months
from the arrival of the stock the incidence of infection was 97 per cent. And yet
crawling was never in evidence* until near the end.

No. 61. — This stock shows a striking parallel to the previous. Within the three
months from 30th May to 30th August, the infection rose to about 80 per cent.,
and only now did crawling symptoms appear, although meantime the stock had
visibly declined in numbers.

No. A. Ch. — Examination of bees from the original apiary in September, which
were showing suspicious signs, showed that they too were infected with Tarsonemus.
It appears probable that this stock was infected before leaving the original apiary,
and from the fact that on 17th July the percentage of infection was so very high
it would appear that the distribution of Tarsonemus was well established, though
probably of recent origin.

Glasgow, I. — This stock was known to have a definite infection on the 8th July,
which rose at the end of October to 27 per cent., without disease signs appearing.
This is a case in which the spread of the disease within the stock is progressing
with relative slowness.

Glasgow, P.— This is a similar case to the foregoing in that the spread of infection
has been slow, and in which the ordinary disease symptoms have never appeared.

No. 62. — Nearly six weeks in an infected apiary elapsed before this clean stock
was found to be harbouring Tarsonemus. After four and a half months from the
time the presence of Tarsonemus was first discovered the incidence had increased
to over 92 per cent., and no disease symptoms have ever been seen and the stock
appears in a prosperous condition.

No. 60. — This is a stock which, although found to have Tarsonemus present in
August, appears to have lost the infection.

752 DR JOHN RENNIE, MR PHILIP BRUCE WHITE, AND MISS ELSIE J. HARVEY

W. No. 2. — Tarsonemus has been known to be present in this stock since the
beginning of August. It is now wintering, and no signs of disease have ever been
observed.

No. 24. — Infection from near the end of June to the beginning of September and
not found later. Stock has not prospered.

Re. No. 1. — Slow progress of infection. No signs of disease ; stock wintering.

Re. No. 2. — Infection took place within three months, and the progress very slow.

R. No. 2. — Proximity to two heavily infected stocks is a feature of this case.
Infection probable within a month, rose in two months about 97 per cent., and the
stock has dwindled to very small dimensions.

W. No. 3. — Known to be infected for two months before any suspicious signs
appeared, and these only slight in character. The stock has done well, but its
possible survival till spring is doubtful.

Table III.

Date.

Locality.

Condition,
of Stock.

Numbers

Examined.

Tarso-

nemus.

Remarks.

24 Sept. 1916

Aberdeen

S.

5

4

Crawling bees.

24 Apr. 1917

Rubislaw

S.

1

1

Development stages of parasite.

26 „ 1917

5 July 1917

Glassel

S.

9

8

Stock crawling.

Aberdeen

S.

1

1

All stages of development of parasite.

1 Eeb. 1918

Stoneywood

S.

3

3

„ „

18 July 1918

Glassel

S.

1

1

,, „

11 Jan. 1919

Aberdeen

S.

7

7

>> >>

8 Dec. 1919

Aberdeen

s.

3

3

Stock crawling.

The bees referred to in the foregoing table were all obtained from stocks
recognised as suffering from the disease. The bees had been preserved on the dates
mentioned. The results confirm the presence of Tarsonemus in diseased bees over
the four years 1916—20, and also make clear the important fact borne out by the 1920
examinations that the breeding of Tarsonemus goes on throughout the whole year.

General Considerations.

For a sound appreciation of the foregoing records, and particularly of their
diversity in detail, it is necessary that the various factors likely to be present
affecting the course and culmination of the epizootic within the bee colony
should be clearly set forth.

A colony of bees consists normally of a population which particularly in summer
is undergoing a continuous and relatively rapid change, both as regards its con-
stituent members and also as regards the total numbers. Daily during the working
season there is both a steady mortality and a steady increase. This latter depends

ON ISLE OF WIGHT DISEASE IN HIYE BEES— ETIOLOGY.

753

upon the age and fertility of the queen. Also, for normal prosperity, there must be
maintained a definite proportion of nurse bees and foragers.

In a colony affected with disease of any kind, which significantly affects the
normal mortality rate, the age incidence of the mortality amongst the workers,
and the productivity of the queen, are characters of the highest importance as
affecting the maintenance of the colony as an effective and prosperous unit. These
two opposing factors struggle with each other — losses from idleness and crawling, and
mortality due to disease, added to the normal wastage on the one hand, and gains
from the production of new bees on the other. A young and prolific queen by sheer
production of new bees may so keep down the proportion of infected and hence more
or less ineffective members as to render such a colony to some degree profitable.

Disease may be maintained within a colony in two important ways. It may
be instituted by the infection of a few members of the colony by contact with a
single bee carrying mites, which has mingled in the cluster. This may be a stray
bee from another colony or a member of the stock which has been robbing a diseased
colony, and such infection may constitute the only one from the outside. In this
case we may expect that progress will be slow, if indeed the infection does not die
out. The attacked bees may be old and die away from the hive before transmitting
Tarsonemud to other members of the colony, or the infection may be so swamped by
normal increase as to be practically ineffective. Whether a stock once infected is
doomed sooner or later in every instance we have not sufficient evidence as yet to
say. Some of the cases quoted, if the samples of bees taken may be regarded as
representative, appear to have lost the infection. And we do have some evidence
that extinction maybe delayed for a long time. Rennie and Harvey, No. 1 (1919),
have already directed attention to cases where the source and time of infection of a
stock was known in autumn and the usual symptoms did not become evident until
the following year.

A second and highly important factor, however, which we are satisfied is very
frequently in operation, is repeated or multiple infections continued from the outside
over a considerable period of time. We then have the disease spreading from many
foci. The drifting of bees into strange hives is common. Once the disease has
gained some ground, the social instinct of the colony is weakened, both by the
disturbance of the normal balance of worker types and by the illness of a high
proportion of bees. Robbing may now take place, and amongst the robbers there
may be infected bees which will intensify the trouble. This robbing, at first resisted,
is eventually allowed to become rife, and when this is established we have noticed
that extinction is practically inevitable.

Other factors which may tell against the stock are the presence of an indifferent
queen whose production may be poor, and from whose low racial vigour shorter-
lived bees result.

The varying character of the factors shows that a uniform course of spreading of

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 29). 117

754

ISLE OF WIGHT DISEASE IN HIVE BEES— ETIOLOGY.

infection within a stock is not likely to occur. Yet apart from any special combina-
tion of adverse factors, it appears to be common for the disease to steadily gain
ground within a stock once it is established. It is, however, clear that no sound
argument against the view that Tarsonemus is the cause can be built up from cases
where, in the presence of Tarsonemus, disease did not spread within the stock and
destroy it as a whole, so long as it can be shown that there is an associated pathology
which in due course renders the infected bees ineffective members of the colony, and
all the time causes loss of bees by crawling or directly by death.

LITERATURE CITED.

Anderson, John (1916), “The Connection of Nosema apis and Isle of Wight Disease in Hive Bees,”
Proc. Roy. Phys. Soc. Edin., xx, p. 16.

Anderson and Rennie (1916), “Observation and Experiments bearing on Isle of Wight Disease in Hive
Bees,” Proc. Roy. Phys. Soc. Edin., xx, p. 23.

Graham Smith, Fantham, and others (1912), “Report on the Isle of Wight Bee Disease,” Supplement 8,
Journ. Board of Agric., xix.

(1913), “Further Report on the Isle of Wight Disease,” Supplement 10, Journ. Board of Agric., xx.

Imms, A. D. (1907), “Report on a Disease in Bees in the Isle of Wight,” Journ. Board of Agric., xiv, p. 129.
Malden, W. (1909), “Disease of Bees in the Isle of Wight,” Journ. Board of Agric., xv, p. 809.

Rennie and Harvey (1919), (1) “Isle of Wight Disease in Hive Bees,” Journ. Scott. Board of Agric., ii,
p. 176.

(1919), (2) “ Nosema apis in Hive Bees,” Journ. Scott. Board of Agric., ii, p. 511.

White, G. F. (1919), “Nosema Disease,” TJ.S. Dept. Agric. Bureau of Entomology Bull. No. 780.

t

EXPLANATION OF FIGURES.

All the figures are photographed with Watson Service Microscope.

Fig. 1. Tarsonemus in trachea of hive bee. The first specimen observed. A. Tarsonemus. B. trachea.
C. thoracic glands. Tv in. oil imm. obj., ocular No. 2.

Fig. 2. Teased preparation of infected trachea showing various stages of Tarsonemus. § in. obj.,
ocular No. 2.

Figs. 3, 4, and 5. Tracheae containing Tarsonemus.

Fig. 3 shows the blackening of the tubes.

Fig. 5 shows blocking of wider tube with larval stages. £ in. obj., ocular No. 4.

Fig. 5, ocular No. 4.

Fig. 6. Section of infested trachea wall. A. not badly affected. Mites seen in section, -jt in. obj.,
ocular No. 4.

Fig. 7. Do., showing A. thickening of wall of tube. B. Two ovigerous females in situ. ^ in. obj.,
ocular No. 2.

Trans: Roy: Soc: Edinf

Vol: LI 1.

Rennie, White, and Harvey: Etiology of Isle of Wight Disease.

Zinco Collotype Co., Edinburgh.

( 755 )

XXIX.

(2) The Pathology of Isle of Wight Disease in Hive Bees. By P. Bruce White,

B.Sc., Bacteriologist to the Bee Disease Investigation, University of Aberdeen
and N. of Scotland College of Agriculture. Communicated by Dr John
Rennie. (With One Plate.)

( From the Department of Pathology , Marischal College, Aberdeen.)

(Bead November 1, 1920. MS. received December 7, 1920. Issued separately March 25, 1921.)

Isle of Wight disease is, as we have seen, primarily a disease of the respiratory
system, in which the organism remains localised throughout the entire course of
the attack.

The effects are, however, .far-reaching, and are registered in the disordered
functioning of several organs, and in visible pathological changes in some of them.

The parasitic invasion has two aspects.

We have, in the first place, to consider the active injury wrought upon the host
by a parasite developing and living at the expense of its body fluids. With this
aspect of the question may be coupled the possibility of a definite toxic action on
the part of the parasite.

In the second place, we have to consider the passive. role of the mites in hindering
or inhibiting the normal functions of the infected organs.

Before proceeding to consider the various pathological conditions, a few words on
the distribution of the parasite within the host is called for.

In infected bees the mites are consistently present in the tracheal system of the
thorax, and in a certain number of bees are also to be found in the air-vessels of the
head. No mites have, up to the present, been discovered in the abdominal system.

The primary parasitic invasion takes place through one or both of the first pair
of spiracular orifices, and apparently through these alone. The whole anterior
thoracic system of major tracheae and air-sacs is liable to infection. The infection
may be unilateral or bilateral, and in some cases where the mites have entered on
one side -only, they may migrate into the vessels of the other side, setting up
a secondary bilateral infection. The passage is effected through the roots of the
anterior air-sacs, which form a commissure between the large paired vessels
supplying the head.

A single mite may enter the bee, or several may enter together or at intervals.
Sometimes the pregnant female may advance as far as the secondary tracheae before
depositing eggs, but the primary trachea becomes involved sooner or later in the
vast majority of cases.

It is usually only in the later stages of the attack that the mites attain the
smaller tracheae, the thoracic air-sacs, and the vessels of the head. In such cases,

756

MR P. BRUCE WHITE ON

however, the young mites frequently advance as far as the .calibre of the vessels
permits.

The various pathological conditions which may be encountered in infected bees
will now be considered, the various systems being treated of in sequence.

Tracheal System.

Macroscopical appearances. — The first change visible to the naked eye is an
increased opacity of the infected vessels due to the aggregation of ova and the
younger forms of the parasite within the lumina. As the disease advances the
trachea assumes a brown tint, which gradually deepens and becomes flecked with
black. Finally considerable portions of the infected tracheae may become dead black.
This change in colour is accompanied by an increasing hardness and brittleness of
the parts, which become rigid. This brittleness results in a phenomenon which is
of some use in the field diagnosis of the disease. It is frequently found that upon
exerting moderate pressure upon the upper surface of the thorax of bees crawling
from the disease, that a drop of fluid — blood — will exude from the first spiracle of
one or both sides, the rupture of the trachea at its insertion having thrown the
hsemocoele open to the exterior.

Microscopic appearances. — During the earlier stages of the attack, the oval and
almost colourless ova and embryos may be seen lying within the lumina of the
tracheae. The parent mites, too, may often be found in the vicinity. The tracheal
wall may show here and there a few fragments of brownish matter, the faeces of the
invading adults.

This condition is maintained till, with the appearance of the later developmental
stages of the parasite and the young adults, the wall becomes encrusted with granules
of faecal matter. These granules, irregular in size and discoid or spherical in shape,
become arranged in the interspaces between the tracheal thickenings, forming an
irregular series of transverse bands upon the tracheal wall. They are of a brownish
or yellowish colour, and when densely aggregated appear black. The colour of the
deposit upon the wall therefore varies with the thickness of the crust and the
amount of pigment it contains. The pigment may become extracted, leaving
the pallid granules behind.

A similar deposit may collect in the lacunas between the parasites themselves.

A typically affected tracheal tube is shown in fig. 1, while a fragment of the
encrusted wall is shown further enlarged in fig. 2.

The faecal matter may, further, be inhaled and, though the bulk appears to be
trapped in the air-sacs and larger vessels, may attain the finer ramifications of the
system, sometimes forming small emboli in the tracheoles. This is particularly
frequent when the parasites are present in the air-sacs.

Careful study of the tracheal wall for perforations reveals little. In two cases

ISLE OF WIGHT DISEASE TN HI YE BEES— PATHOLOGY.

757

only has it been possible to observe the long piercing apparatus of the mite actually
passing through punctures in the wall. The material damage done in this way is
seemingly small.

Muscular System.

Visible pathological changes of the muscle fibres occur, but these are apparently
restricted to the thoracic muscles of flight. Though a considerable number of
fibres, in highly infected crawling bees, may show signs of the atrophic change
to be described, the number showing definite degenerative changes is usually small.
While no such changes have been noted in non-infected bees of whatever age, they
may occur in infected bees which show no outward symptoms of the malady. On
the other hand, a percentage of infected crawling bees show no marked muscle
changes.

Macroscopic appearances. — Upon teasing out in saline the thoracic muscle mass
of a bee crawling from the disease it is usually found that certain fibres — averaging
2-6 in number — contrast markedly with the flaccid, greyish-yellow normal fibres by
their opaque white colour, slenderness, brittleness, and rigidity.

Microscopic appearances. — Under the low power of the microscope these white
fibres are conspicuous by their slenderness, density, and granular appearance. The
ends show an irregular fracture quite unlike the frayed-out ends of the normal
fibres. A number of micrometer measurements on these and on healthy fibres of
the muscles of flight gave the following values : —

Average width of healthy fibres of muscles of flight = '24 mm.

,, ,, atrophied ,, ,, ,, ='12 mm.

In fresh preparations examined with the 1/6" objective it was possible to make
out the nature of the change which had taken place.

In the normal muscle of the bee the bulk of the fibre is composed of the fibrillse,
upon and between which lie the large flattened sarcosomes or myochondria. These
granules mask the transverse, but not the longitudinal, striation of the fibre. When
the fibre is teased out the fibrillse fray out, allowing the diaphanous sarcosomes
to escape.

In the case of the atrophied fibres of infected bees the appearances are different.
Microscopic examination may show little or nothing of the original fibrillar struc-
ture. It is often found that the bulk of the fibre is composed of densely arranged
longitudinal columns of closely packed and very coherent sarcosomes which do
not escape and float away when the fibre is teased out. Between these granular
columns, it is found, upon closer examination, that remnants of the fibrillse persist,
though many may be reduced to thread-like vestiges of their original form.

A drawing of the low-power appearances of normal and atrophied fibres is shown
in fig. 3, and a piece of atrophied fibre is drawn under the high power in fig. 4,

758

MR P. BRUCE WHITE ON

while in fig. 5 the substance of a highly atrophied fibre (A) is contrasted with two
normal fibrillse and normal sarcosomes (B).

The sarcosomes of the atrophied fibre are much denser and more cubical in shape
than the normal granules, and may be of relatively enormous size. In highly
degenerate fibres they may form conglomerate masses of considerable size.

When degenerate muscle is treated with dilute acetic acid the fibrillar vestiges
swell, forcing the granular columns apart, and a picture closely resembling normal
muscle when treated in the same way is achieved.

When freshly obtained normal and atrophied fibres are placed in a drop of dilute
eosin or methylene blue, it is found that while the former are only slowly and
superficially stained, the latter become rapidly and deeply stained through their
entire substance. This would seem to indicate that the degenerate fibres are dead.

Though various fixation and staining methods have been employed, they have
added little to the facts derived from the study of fresh material.

These muscle changes may be summarised as —

A general wastage of the fibrillar substance and loss of sarcous fluid, with the
condensation of the frequently enlarged sarcosomes in densely arranged longitu-
dinal columns, the process resulting in a shrinkage of the fibre with loss of function.

All stages in this process are, of course, to be encountered.

A further feature of some atrophied fibres, and occasionally of those which do not
show the typical signs of wastage, is the development of pigmented spots in their
substance. Such spots are represented in fig. 5.

These spots vary in colour from yellow to a deep brown or black, and often
appear to bear a definite relationship to the distribution of the tracheoles supplying
the muscles.

Careful scrutiny has not entirely elucidated their origin. It seems possible that
they may be caused by staining of the muscle by the fsecal dye of the parasite, which
has percolated into the final ramifications of the tracheal system.

Other explanations which have been considered are that the discoloration is due
to a degenerative process in the muscle or to an accumulation of waste products.
This point may perhaps be cleared up by further work.

In some diseased stocks these spots are found in almost every crawling bee ; in
others, apparently at the same stage of the disease, they are absent.

The Blood.

The blood of the crawling bee is often scanty, though when such bees are
warmed and fed they recover their normal complement of body fluid.

No qualitative cytological difference has been noted between the blood of healthy
and crawling bees, though the number of cells per unit volume may be increased in
the latter. This increase is probably entirely due to loss of plasma and cannot be
regarded as a leucocytosis.

ISLE OF WIGHT DISEASE IN HIVE BEES — PATHOLOGY.

759

The Alimentary System.

As regards the alimentary system, the investigation has but little to add to'’ the
observations of former workers. The disordered condition of the alimentary tract
has attracted much attention in the search for a clue to the causation of Isle of
Wight disease.

The facts may be briefly outlined.

In the majority of crawling bees the hind gut and small intestine are distended
to the limit of their capacity with accumulated faeces, and the contents of the lower
region of the chyle stomach may contain a large admixture of faecal matter. The
chyle stomach itself may present a rather wasted appearance, and its contents may
be of an unusually deep purple colour. These changes are in all probability merely
due to a reduction in the fluid contents of the organ.

In the vast majority of crawling bees no lesion is to be found in the alimentary
wall either macroscopically or in stained sections. On two occasions only have signs
of penetration of the wall by organisms (other than Nosema) been noted. In one
case there was an infiltration of the wall near the insertion of the Malpighian tubules
by a large filamentous bacillus ; in the other case fungal hyphrn had invaded the
epithelium of the lower portion of the chyle stomach. Such phenomena are to be
regarded merely as terminal infections.

The flora of the alimentary tract of the normal bee has been carefully investigated
and compared with that of bees crawling from the disease. Very little qualitative
difference- has been found between the two.

In Isle of Wight disease there is a colonisation of the chyle stomach by the
intestinal organisms, and certain organisms, such as coliform bacilli and yeasts, are
more frequent and abundant than in healthy bees. Certain streptococci, to be
described elsewhere, have also a predilection for the alimentary tract of Isle of
Wight bees.

Malpighian Tubules.

In a proportion of crawling bees certain of the Malpighian tubules, when mounted
in saline, may possess a bright yellow colour due to the presence of large amounts of
the excretory pigment. In these coloured areas the excretory granules within the
epithelium may be abnormally large and spicular. In fresh preparations the cells
appear to be filled with large bacilli. A similar condition may arise in bees after a
period of confinement.

Nervous System.

The examination of the nervous system for pathological changes is as yet very
incomplete. The observations of the writer have been mainly restricted to the
thoracic ganglia. No changes have been noted in the posterior thoracic ganglionic

760

MR P. BRUCE WHITE ON

mass, and the few sections which have been examined of the anterior thoracic ganglia
of sick and healthy bees show no alterations which cannot be accounted for as
physiological variations due to senility.

Discussion.

With these facts before us an attempt may be made to discuss the correlation
between the action of the parasite, the pathological changes, and the symptomatology
of the disease.

We have alluded to the two aspects of the primary effect of the parasite upon the
host : the active injury produced by a parasite living upon the host fluids, with the
added probability of a toxic action, and the passive obstruction of the- respiratory
system of the head and anterior thorax.

The pregnant parasites producing many, relatively large ova, the developing
brood and the young adults must make considerable demands upon the host. It has
been pointed out that the blood of crawling bees is often scanty, but it is improbable
that this is in any significant degree directly due to the mites, but arises from the
fact that fluid lost by excretion and transpiration is not replaced owing to the
inability of the stricken bee to take or to obtain food. As many heavily infected
bees continue to forage, though their tracheae are bronzed and blackened by long
sojourn of the mites, it would seem probable that nutritive sapping does not per se
render the bee effete.

The same uncertainty surrounds the question of a toxic action. One member at
least ( T . intectus) of the genus to which the parasite belongs is known to be
venomous, but the exact importance of this factor in the disease must, like the
foregoing, remain for the present a matter of surmise.

The passive action of the parasites and their products in partially or completely
blocking the infected tracheae is a factor of which the importance is much more
readily estimated.

It is obvious that any obstruction of the tracheal lumen must reduce the
efficiency of the respiratory exchange of the organs supplied. In the vast majority
of crawling bees the effective lumina of certain of the major tracheae are obviously
very much reduced, and in some all but obliterated. The organs supplied by such
tracheae must be reduced to an acute degree of oxygen starvation, and among the
organs of which the respiratory supply is thus endangered are those of the head and
the thoracic muscles of flight.

It is clear that the effects must vary from case to case : —

(a) With the degree of the obstruction.

( b ) With the position of the . obstruction.

(c) According as to whether the obstruction is bilateral or unilateral.

The actual number of parasites distributed through the respiratory system is from

ISLE OF WIGHT DISEASE IN HIYE BEES — PATHOLOGY.

761

this point of view of secondary importance, a fact which may explain the apparent
vigour of many heavily infected bees.

In order to obtain some idea of the effects actually arising from mechanical
obstruction of the spiracles, a series of experiments were undertaken upon healthy
bees. In these experiments melted paraffin wax was applied to the first spiracle of
one or both sides of each bee in such a way as to give, on solidification of the
wax, complete closure of the spiracular orifice without impairing the free play of
the wings.

Bees treated in this way were maintained in boxes and were examined at in-
tervals. In each experiment ten to twenty experimental bees were employed, and
parallel controls were kept under the same conditions.

Upon closure of one spiracle the experimental bees at once lost the power of
flight, but remained otherwise active in their movements, running quickly over the
bench and beating the air with their wings. Upon the second and third days it was
sometimes found that a proportion of the bees were capable of flight — which was,
however, usually of very short duration. In these it is probable that the wax had
become partially dislodged. The majority of the bees continued to crawl. After
the lapse of several days these crawling bees became more sluggish in their move-
ments, sometimes showing a tendency to drag their hind legs, and about the
sixth to seventh day, bees were noted which showed a dislocation of the wings
similar to that so common among bees crawling from Isle of Wight disease.
About this time, too, some of the bees began to die : many were, however, maintained
up to the beginning of the third week. During this period also a few of the control
bees died, but the remainder retained the power of flight throughout.

At intervals experimental and control bees were killed for examination. Both
in the “ artificial crawlers ” and in those control bees which had not been given
opportunity to void their faeces on the wing, the hind gut was found distended with
faecal matter. At the end of the first week of experiment it was found that the
thoracic musculature of the experimental bees showed, in many cases, atrophy of
exactly the same type as had been found in infected bees. The degree of this
atrophy and the number of fibres affected varied with the duration of the experi-
ment. No such changes were noted in the control bees.

In those experiments in which the first spiracles of each side were closed with
wax, the phenomena were different. As before, the power of flight was at once lost,
but after twenty-four to forty-eight hours the bees had developed a reeling gait and
appeared to be continually falling over their own heads. It was seldom that any
survived the third day. No muscle atrophy was to be discovered, death having
supervened too rapidly for the accomplishment of this change.

From these experiments it may be stated that : —

Through closure of the first spiracle of one side, a condition of crawling is induced
which bears a close resemblance in its symptoms to Isle of Wight disease, and that,

TRANS. ROY. SOC. EDIN., YOL. LII, PART IV (NO. 29). 118

762

MR P. BRUCE WHITE ON

further, the procedure may occasion atrophic muscle changes which are only known
to occur in that disease. When the first pair of spiracles is closed, a state of com-
plete incapacitation results, ending rapidly in death.

Though too close a parallel must not be drawn with the natural disease, these
experiments are illuminating in that they give a basis to the view that the r61e of
parasites in partially preventing thoracic respiration is of prime importance in the
disease — possibly in itself capable of occasioning all the symptoms by which we are
wont to diagnose the disease and the muscle atrophy so often associated with it.

The pathological syndrome of Isle of Wight disease is undoubtedly complex.
Apart from the sapping of the host fluids and the probable injection of a venom,
the mites may impair the mechanism of the bee either by destroying the respiratory
supply of the individual organs or by cutting off that of the nerve centres which
control and co-ordinate their activities. It is possible that the indirect effect
through the nervous system, possessed as this is of a dual respiratory supply, is
particularly acute when there is considerable bilateral obstruction of the tracheal
system.

Through the combined influence of these factors the power of flight is lost, and
a series of secondary conditions arise.

The faeces normally voided on the wing accumulate, thus increasing the difficulty
of locomotion and compressing the abdominal air-sacs — another blow at the respira-
tory function. Intestinal pressure must hinder the excretory activities of the
Malpighian tubules, and this excretory stasis, together with the absorption of toxins
from the stagnant gut, must be reflected back upon the body of the insect.

As soon as the power of flight is lost death of the bee becomes imminent, for
once it leaves the warmth and stores of the hive, unable to return, it perishes of cold
and starvation. Should it elect to remain within the hive it is faced with a prospect
of functional stagnation which cannot be indefinitely maintained. It would seem too
that in the colder months sick stocks often perish en masse through inability to
maintain the hive temperature.

It seems that in rare cases individual bees may recover from the attack upon
being abandoned by the parasitic brood. Such cases are recognised by their bronzed
and blackened tracheae, which, however, contain no living mites. Bees in this con-
dition have been found foraging for infected stocks.

Concluding Remarks.

It is somewhat remarkable that the macroscopic changes of the thoracic tracheae
and muscle have so long escaped observation in spite of the detailed examinations of
several independent workers.

Imms (l) held that “ the disease is eminently one of the digestive system, and
might be described as a condition of enlargement of the hind intestine,” while

ISLE OF WIGHT DISEASE IN HIVE BEES — PATHOLOGY.

763

Malden (2) was of opinion that “ the disease must be regarded as an infections one
which primarily affects the chyle stomach.”

The latter states that in his investigations “ no changes were discovered in the
salivary glands, brain, fat body, tracheae, air-sacs, Malpighian bodies, or honey
stomach,” but that “ the chyle stomach in many cases showed marked changes in
section.” Of these changes in the chyle stomach the present investigation has seen
little or nothing. Two exceptional cases have been noted where the epithelial lining
had been definitely invaded.

Malden, as a result of his bacteriological work, suggested a “plague-like”
bacillus, called by him B. pestiformis apis, as the cause of the malady. He, however,
made the suggestion with some reserve, and later, when his work had been over-
shadowed bv the “ Nosema theory,” considered that toxins produced by various
species of bacteria played an important secondary role in the disease.

There is, however, an underlying truth in his summing up : “ The actual cause of
death is uncertain, but it is probably brought about by malnutrition, possibly
combined with the absorption of a specific poison and of the products of decomposi-
tion in the colon, and probably aided to some extent by the imperfect oxygenation
of the tissues, owing to the pressure exerted by the distended colon on the abdominal
air-sacs.”

Imperfect oxygenation, and possibly malnutrition and a toxic condition, are the
main factors in the disease as we see them to-day.

In this paper it has been sought to outline those pathological facts which have so
far come to light, and to relate them as reasonably as possible to the action of the
parasite on the one hand and the symptoms of the disease on the other. Certain
points have been merely touched on and others left in doubt, but it is hoped that
further details will be soon forthcoming.

There are many to whom my thanks are due. To Dr J. Rennie, who, as director
of the research, has been an inspiring leader throughout, I tender my warmest
thanks for help and advice and much personal kindness. My sincere thanks are
due to Professor T. Shennan and the staff of the Pathology Department, Marischal
College, Aberdeen, who have given me every facility for carrying out this work. It
was in the Pathology Department that the parasite was independently discovered
on the 11th May 1920, and the theory of its significance in the disease formulated.
I wish also to express my gratitude to Professor J. Arthur Thomson, Mr A. H. E.
Wood, and the members of the Joint Committee of the University of Aberdeen and
N. of Scotland College of Agriculture for all their interest and support.

REFERENCES TO LITERATURE.

(1) Imms, J. Board of Agric., vol. xiv, No. 3, June 1907.

(2) Malden, Ibid., vol. xv, No. 11, February 1909.

764

ISLE OF WIGHT DISEASE IN HIVE BEES — PATHOLOGY.

DESCRIPTION OF FIGURES.

Fig. 1. Infected .trachea showing typical changes. The black faecal deposit is arranged in transverse
lines, (x 70.)

Fig. 2. Fragment of wall of infected trachea, shown much enlarged. Granular deposits of faecal matter
lie between the tracheal thickenings.

Fig. 3. Normal ( n ) and atrophied (a) fibres from the thoracic muscle of a crawling bee. Note the
density, slenderness, and fractured ends of the atrophied fibres. ( x 50.)

Fig. 4. Fragment of an atrophied muscle fibre showing dense longitudinal columns of sarcosomes
from between which a few fibrillar remnants (/) project. ( x 500.)

Fig. 5. A teased-out fragment of an atrophied muscle fibre (A) is contrasted with normal fibrillae and
sarcosomes (B). ( x 600.)

Fig. 6. Portion of atrophied muscle fibre showing blackened spots in its substance which follow the
distribution of the tracheoles

Trans: Roy: Soc: EdinT

Vol: LI 1

White: Pathology of Isle of Wight Disease.

Zinco Collotype Co., Edinburgh.

( 765 )

XXIX.

(3) Isle of Wight Disease in Hive Bees — Experiments on Infection with Tarsonemus
woodi, n. sp. By Elsie J. Harvey. Communicated by Dr John Rennie.

(Read November 1, 1920. MS. received November 27, 1920. Issued separately March 25, 1921.)

Introductory.

The following experiments and observations have been undertaken with a view
to discovering the means by which Tarsonemus is transmitted from one bee to
another. It is obvious that a stage of the parasite exists outside the bee, and that
there are also several possibilities (which may occur). One is the passage from bee
to bee within the hive either directly or through the medium of frames or combs.
Another, also within the hive, where the mites may in wandering upon the frames
enter the cells and invade the body of the developing larvae or pupae, and in this way
be present in the bee when it hatches. A third possibility is that whereby foraging
infected bees may leave the mite upon flowers, vegetation generally, drinking-
grounds, or other situations, to be picked up later by other bees chancing to
visit these. Crawling or dead bees may in a similar manner prove to be a source
of infection through the contamination of the ground about the hive or of the
actual hive itself.

Exhaustive investigation of this last possibility is a somewhat difficult matter,
for which no opportunity has yet been found, and owing to the short time at my
disposal it has been set aside in favour of the more promising one of infection by
direct contact between bee and bee.

Review of Former Experiments.

The conclusions arrived at from former experiments published before the organism
was discovered, pointed to the probability of the disease being of an infectious
character. This was shown when, e.g., a frame of infected bees, say of a black colour,
were placed in a healthy stock of Italians, and in due course the disease became
evident in the yellow bees. It is admitted, of course, that only a probability is
indicated in such an experiment.

What evidence we have from experiments with brood, in which frames of sealed
brood of bees of one colour from an infected hive have been placed in the hive of a
healthy stock of a different colour, points to the disease being an affection of adult
bees only. Many experiments of this nature have been tried, and the results have
been on the whole uniformly in favour of the view that brood hatched out under
such circumstances was free from disease.

766

MISS ELSIE J. HARVEY ON

Experiments.

1. To Discover whether Bees become infected before Emergence from Cells.

Early in the spring of this year a few frames of brood from a badly affected stock
were placed in an incubator ; 155 of the bees which hatched out were examined, with
the result that only the tube on one side in one bee was found infected with the
parasite. This evidence, as far as it goes, therefore, does not exclude the possibility
that mites may enter the cells and invade a bee’s body before it emerges from the
cells, but its rarity, as shown here, would seem to suggest that at best it is only
an incidental occurrence, and is not one of the regular ways in which infection
is conveyed.

2. To Discover the Stage or Stages which occur normally outside the Bee.

{a) By examination of the individuals in the tubes, in cases where the infection
had only newly commenced, it is often found that an ovigerous female with a few
eggs in different stages of development are the only parasites present. This would
suggest that the migratory stage of the parasite is the fertilised female. In such a
case migration of the male does not seem to be necessary, although, as is seen from
the results of the experiments recorded, they leave the body of the dead bee.

(b) Equal numbers of living and of dead bees were placed in separate petri
dishes, and these were kept as nearly as possible at the temperature of the hive.
The dishes were examined microscopically at intervals of from two to forty-eight
hours. Fifteen such experiments were carried out, with the result that only 10
mites, all of which were females, were recovered from the living bees : one of these
was alive and active. As many as 75 (62 female and 13 male) were found in the
dishes containing the dead bees. Of the 62 female mites, 8 were alive and active,
while all the 13 males appeared dead. None of the females seen outside the bee
were carrying fully developed ova. It would appear from these experiments that
migration of both sexes takes place from the dead bee.

3. An Endeavour to produce Infection artificially.

(a) By contact with living sick bees.

These experiments were carried out in small queen cages. The bees were fed
with soft candy and kept as far as possible at the temperature of the hive. The
infective bees were in five cases crawlers picked up from in front of the hive ; in two
cases the bees were caught on entering the hive of a stock which was known to have
a high per cent, of infection, and in the remaining three, the bees were taken from
the frames of a sick stock. Virgin queens headed six of the experiments.

In these experiments the healthy bees were maintained in contact with sick bees
for periods extending from four to seventeen days. In only one case was a positive

ISLE OF WIGHT DISEASE IN HI7E BEES — EXPERIMENTS ON INFECTION. 767

result obtained. On the fourteenth day of the experiment one bee was found to
contain one ovigerous female at the entrance of the tube.

(b) By contact with dead bees.

Before commencing these experiments, observations were made to discover how
long the parasites lived after the death of the host. It was found that a few female
mites were still capable of feeble movement on the fifth and sixth days.

Seven of these experiments were put up in the same way as the foregoing, sub-
stituting newly killed bees for the living sick bees. The result in this group was
negative.

(c) By placing tracheal tubes containing the parasite on the thorax of healthy
bees.

Preliminary observations were made on the behaviour of the mites when the
tubes had been dissected out of the bee. These were placed in welled slides, and
both sexes of the mites were seen to emerge within an hour. These wandered about
actively, and were occasionally seen to re-enter the tubes. In most cases the mites
became inert and passive 'within twenty -four hours of leaving the tubes. It cannot
be said with certainty that in any of these experiments the mites are really dead.

Bees were now taken from a stock known to be free from infection, and tubes
containing the parasite were placed on the thorax near the first spiracle. Twenty-
four bees were treated in this manner, and were examined after twenty-four hours.
It was found that no infection had taken place.

Examination of the stocks from which the experimental bees were taken, for the
presence or absence of the parasite, were being regularly carried out.

It is to be clearly understood that the whole of the foregoing experiments are
provisional in character, and all of them, as well as others, are being repeated upon a
larger scale.

From the results obtained, as far as they go, however, it appears that experi-
mental infection with the parasite Tarsonemus is difficult to effect ; it must be taken
into consideration that any deviation from the normal habits of the bee host such as
is involved in these experiments may have a corresponding effect upon those of
the parasite.

I wish to express my sincere thanks to Dr Bennie, both for his help and advice
in the preparation of this paper, and for his kindness personally whilst I have
worked under him.

( 768 )

XXIX.

(4) Isle of Wight Disease in Hive Bees — Acarine Disease : The Organism asso-
ciated with the Disease — Tarsonemus woodi, n. sp. By John Rennie,
D.Sc. (With One Plate and Two Figures in the Text.)

(Read November 1, 1920. MS. received November 27, 1920. Issued separately March 25, 1921.)

The organism which has been found living in the anterior tracheal system of
hive bees, and whose presence is associated with Isle of Wight disease, I have
identified as a hitherto undescribed species of the genus Tarsonemus. This
genus was founded in 1876 by Canestrini and Fanzago, and since then a moderate
number only of species has been established. The true systematic position of
these Acarines has been much in doubt, and their position in the order has
from time to time been revised. Canestrini (1888) constituted the Tarsonemes
the type of a special family, the Tarsonemini ; they have been associated with
the Oribatidse by Berlese, and with the Cheyletidse by Trouessart (1892).
Banks (1904) regarded them as showing resemblances to the Tyroglyphidse, and
placed them in a super-family Sarcoptoidea. An important character of the
Tarsonemes is the existence of a tracheal system in the adult female, which is
not found in the male nor in any pre-adult stage of either sex. This feature was
adopted by Berlese (1897) as the basis of his sub-order, Heterostigmata, and
by Oudemans (1906) in his division Trachelostigmata. This super-family includes
two families — Tarsonemidse and Scutacaridse * (Oudemans, 1916). This last is
the Disparipedidse of Berlese.

The Tarsonemidse are a small family of soft-bodied mites, the females of
which are tracheate, and which usually exhibit prominent hairs upon . the tarsi
of the last pair of legs. The body is more or less clearly segmented dorsally.
The mandibles are needle-like, the palps slender and minute. The females
possess in most instances, between the first and second pair of legs, a pair of
delicate rounded or club-shaped organs which have been designated pseudo-
stigmata by Oudemans. The legs are short, with six or fewer joints. They
are bedecked with a limited number of stout hairs, and terminate in claws. The
tarsi of the first pair possesses a single claw, the second and third, two. The
fourth tarsus varies in the different genera. Suckers are frequent. There may
be distinct sex dimorphism, especially in the genus Tarsonemus.

The Genus Tarsonemus.

Canestrini’s original description defining the genus is as follows : —

“ Rostro normale e libero. Zampe del quarto pajo nella femmina poco sviluppate,
sfornite di uncini e di ventosa e terminate da duo setole ; zampe del primo pajo,

* I desire here to gratefully acknowledge the courtesy of Dr Oudemans in guiding me to the literature of the
Tarsonemidee and Acarina generally.

ISLE OF WIGHT DISEASE IN' HIVE BEES — ACARlNE DISEASE.

F69

pure nelle femmine, normali, conformate come quelle del secondo e terzo pajo,
colla differenza che hanno un’ ungliia sola. Zampe del quarto pajo nel maschio
robuste, constituent! insieme unachela, terminate da un’ unghia robusta. Epimeri
del terzo e quarto pajo nel maschio assai lunghi e forti e convergent insieme
verso l’avanti. Scudo dorsale diviso in segment. Animali viventi su piante.”

The adult female of the species to be described conforms well to the generic
characters given above, and cannot, in my opinion, be separated from the genus
Tarsonemus. The only morphological character upon which such a separation
could be . based is the conformation of the fourth pair of legs, and possibly the
absence of pseudostigmata (not included in the original generic description). The
male is undoubtedly more specialised in the fourth leg characters, but to con-
stitute a new genus upon this fact, or upon the parasitic habit with which this
specialised feature appears associated, seems to me inadmissible, at any rate in
the present state of our knowledge.

A distinctive feature of the genus Tarsonemus is the fourth pair of legs,
which in the female are slender, terminate in two hairs, and are devoid of
claws. In the male, in the gall-inhabiting and free-living species, the last pair
of legs is robust and terminates in a claw-like segment, usually incurved and
frequently strongly developed. In some of the species recorded as endoparasitic,
these characters in the male appear less well marked, and in the main show a
reduction* in size of this pair of appendages. In the species to be described the
hind legs in both sexes present, especially in the male, features which I regard
as related to the parasitic mode of life and restricted habitat of the insect’s
tracheal system.

Tarsonemus woodi, n. sp.

I propose to designate this species, which is parasitic in the anterior thoracic
tracheal system of the hive bee, Apis mellijica , and which does not appear to
have been described before, by the name of Tarsonemus woodi, n. sp. The adult
ovigerous female measures from ’14 to T9 mm. in length, the male about ’ll to
T5 mm. (fig. 1). Viewed with reflected light, these mites are more or less bean-
shaped in form, greyish in colour, and scantily bedecked with hairs. When removed
from the tracheae of the bee they progress slowly upon glass, but when seen within
the tube, although continued observation has not revealed much progression, a good
deal of active and vigorous leg movement may be observed.

Ovigerotjs Female. — Seen from above, the body presents a somewhat oval
form, broadest in the neighbourhood of the second pair of legs. The following are
typical dimensions for a fully grown adult : —

Total length from tip of gnathosoma to hinder end of body, T9 mm.

Total length from tip of gnathosoma to tip of longest hair of fourth pair
of legs, ’25 mm.

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 29).

1 19

770

HH JOHN RENNIE ON

Width at broadest part of body, ‘10 mm.

Length of gnathosoma, ‘03 mm.

Length of first, second, or third legs, ‘05 mm.

Dorsally the body shows definite segmentation.

On the ventral side there is a distinct transverse furrow behind the second
pair of legs. The gnathosoma is bluntly triangular and freely movable, and in

Text-fig. 1. — Dorsal view of body of immature female,
showing distribution of hairs. The anterior por-
tion of the two main tracheae are shown.

Text-fig. 2.

a, Fourth leg of male ; b, fourth leg of female ; c, second leg
of male ; d, first leg of male.

the middle line in front the mandibles are frequently to be seen protruded as a
pair of slender curved needles. The two anterior pairs of legs are directed forward
or outwards, and some distance apart from the third and fourth pairs.

Body Hairs. — On the dorsal surface there are eight pairs of hairs fairly
evenly distributed along the body (text-fig. l). There is a short pair anteriorly,
directed forward close to and upon each side of the gnathosoma, arising on the
medial side of the stigmata. A second pair arises opposite the second pair of legs,

ISLE OF WIGHT DISEASE IN HIVE BEES— ACARTNE DISEASE.

77 1

about one-third of the body width from the margin, and slightly in front of the
dorsal furrow. Behind these are two pairs at about the middle of the body,
one pair near the margin and the other in a nearly direct line behind the second
pair. A pair arises nearly opposite the origin of the third pair of legs ; behind
these are two pairs constituting an almost straight transverse row, closely followed
by a final pair close to the hinder end of the body, and separated by about a
third of the width.

Tracheal System. — A pair of stigmata opening dorsally occurs anteriorly close
to the base of the gnathosoma. From these the more dorsal and anterior portions
of the tracheal system pass backward as a pair of curving, slender tubes, which
cross each other at the level of the anterior limbs. These pass ventralwards
curving inwards toward each other, and again outwards towards the bases of
the posterior limbs.

Legs (text-fig. 2). — The anterior pair is jointed, and terminates in a single hook,
with circular sucker in addition. On the penultimate segment upon the dorsal side
there is a sensory spine-like organ — “riech-haar” of Oudemans. The second pair
also carries a similar spine and ends in a double claw and sucker. The third pair
of legs resembles the second, but a “riech-haar” is absent. These limbs arise near
the margin and are directed outward and backward, and usually lie clear of the body
for the greater part of their length. The fourth pair of legs, which arises closely
behind the third pair, exhibits distinctive features. In contrast to what is usual
in other species, instead of being long and slender, this limb is moderately stout at
its basal segment and slender beyond. But the whole limb consists of only two
segments and is much reduced in length. Upon it are four hairs, two of which are
terminal and of considerable length. One of the remaining two arises nearer the
base, but is usually seen along with the terminal hairs, forming a group of three
projecting beyond the body (text-fig. 2, h).

Immature Female. — This differs in size and general shape, being more trun-.
cated behind (fig. 9).

Length of body from '13 to '15 mm.

Width ,, ,, '08 to ‘09 mm.

Adult Male. — The adult male is smaller than the female and more truncated in
form behind.

Total body length 'll to T5 mm.

,, width '06 to '08 mm.

Dorsally, it appears to consist of three segments, besides the gnathosoma. Only
five pairs of hairs are present, all placed slightly inward from the lateral margin,
and nearly equidistant along the side. On the ventral side the following features
are observable. The transverse furrow behind the second pair of legs is well marked.
There are no longitudinal epimeral grooves such as are prominent in most free-living

772

DR JOHN RENNIE ON

species. The external genitalia consist of two rounded lobes, projecting posteriorly
on each side of a tapering triangular penial sheath.

Legs. — Special features regarding the limbs are : On the penultimate segment
of the second pair of legs there is a prominent sensory organ. The fourth pair of
legs is relatively slender, and presents the most distinctive characteristic of the
species in this sex. There are three joints, all of which are comparatively weak ;
the terminal one bears the very long, stout, and finely curved hair distinctive of
the genus. Just within this, and almost at the tip of the last segment, is a small
straight spine, which occupies the position of the incurved terminal claw, charac-
teristic of free-living species. Close examination has suggested that this structure
is of the same character as that on the second pair of legs of the males, and on the
female limbs also, and that it is probably sensory in function (fig. 7).

Larva. — The larva is large, being about ’2 mm. in length by about '08 mm. in
width at its broadest part. The mouth parts resemble those of the later stages.
There are three pairs of short legs ; the first terminates in a double claw (fig. 4).

Ovum.-!— The egg is large, being about T4 mm. long by '06 mm. broad, and
slightly concave along one side (fig. 2).

All the Heterostigmata appear to be parasitic, some on plants, others upon
insects, and doubtfully on warm-blooded vertebrates. The majority of the described
species of Tarsonemus are found upon plants ; the others are from the bodies of
vertebrate animals, in which a number have been found associated with malignant
growths. T. woodi, n. sp., appears to be the only species known to occur in
insects. Its discovery as an endoparasite within the hive bee therefore constitutes
a significant advance in our knowledge of these mites, and of the general importance
of the genus, apart from the far-reaching suggestiveness of its causal relationship
to a disease in hive bees which has baffied inquiry for the last sixteen years.

As giving a more detailed indication of the habits of the genus, the following
brief references to the best-known species are submitted. In view of the importance
of habit and habitat in the case of T. woodi, I have preferred to refer to these,
not in the order of their original discovery or description, but to group them
from this point of view.

Hall-forming Species.

T. jloricolus On. and F., 1876. — This species is described as occurring on the
flowers of Verbascum, forming galls in foliage of Vitis vinifera, Coryllus avellana,
Salix alba, etc., in putrefying stuff, and in frass of bacon beetle. T. ( Cheylurus )
socialis, according to Berlese, is of the same species. It is described from the skin
and base of the feathers of birds of very diverse species, both terrestrial and aquatic.

T. buxi Canestr. and B., 1884. — Occurs as an inquiline in Phytoptus galls and in
Diplosis galls, See Canestrini (1886), pp. 320-1,

ISLE OF WIGHT DISEASE IN HIVE BEES — ACARINE DISEASE.

773

T. canestrini Massal., 1897. — Forms small rounded galls on several European
grasses.

T. phragmitidis Schlechtendal, 1897. — A species resembling T. canestrini, which
occurs as an inquiline in phragmitid galls.

T. contubernalis Reuter, 1906. — An inquiline in galls upon Galium verum.

T. latus Banks, 1904. — Causes galls on the main shoots of the mango plant.

T. intectus Karpelles, 1885.— In barley; producing severe irritation on hands of
workers in the Danube region in Hungary and Russia.

T. spirifex Marchal, 1902. — -On grasses; causes elongated swellings on oat.
Occurs in colonies.

Described as doing Damage to Plants but not Apparently Associated

with Galls.

T. oryzce Targ., 1878. — Infests the ears of the rice plant.

T. culmicolus Reuter, 1900. — From spikes of meadow grass ; produces “ silver
top” in grasses in Finland, where it is found in the leaf sheaf above the upper-
most node.

T. anasce Tryon. — Described as causing injury to pine apples in Australia.

T. frag aria Zimmerm., 1904. — Has been found on strawberries.

T. graminis Kramer, 1886. — So named by Kramer because it occurs in abnormally
rolled-up grass leaves.

T. bancrofti Michael. — Has been ' described as causing damage to sugar-cane in
Queensland and in Barbados.

Described from Animals.

T. floricolus.- — Already quoted above as occurring in the bases of birds’ feathers.

T. soricola Oudms., 1903. — Found on Sorex vulgaris.

Regarding the placing of these species here, it may be mentioned that Oudemans
is of opinion that when forms occur on animals these are probably no more than
transporting agents.

T. hominis Dahl, 1910. — From human ovary in carcinoma and fibroma, and
from bladder in cystitis (Blanc and Rollet, 1910).

T. sauli Dahl, 1910 (T. equi ; T. muris ; T. canis ). — From tumours in mammals.

T. woodi, n. sp., Rennie, 1920. — In thoracic tracheae of Apis mellifica.

In this connection reference should also be made to a form described by Myake
and Scriba (1893) from the urine of the human subject as Nephropliages sanguinarius.
According to Oudemans this form is a Tarsonemus. It is tracheate, blind, has
needle-formed manibles, and the pear-shaped pseudostigmatic organ has been mis-
taken for eyes. According to him the mouth parts are not described or figured
correctly.

The group of species which have been obtained from mammalian tissues presents

774

DR JOHN RENNIE ON

some features of interest in relation to T. ivoodi, and it seems worth while con-
sidering these in some detail. Dahl has described a species termed by him
T. hominis, which was obtained by E. Saul, the female from a fibroma of the human
ovary, and the male from a carcinoma of the same organ. Later in the same year,
Saul published micro-photographs of T. hominis and of others obtained from a
cancer of the mouse, a papilloma of a horse, and a sarcoma of a dog. Following
the publication of Saul’s photographs, Blanc and Rollet (1910) published a state-
ment that they had in their possession an acarid obtained in 1909 from the urine
of a patient suffering from a refractory cystitis. They describe the specimen in
detail and recognise it as a male of T. hominis.

T. hominis is distinguished, according to Dahl, from all previously described
species by the following. In the female the fourth pair of legs is more shortened
than in other species. Except for the end bristles it does not reach to the hinder
end of the body. The third pair has a longer, thinner, two-segmented end part sharply
marked off from the basal segment by the greater width of the latter. The two
bristles at the hind end of the body are wider apart than is the case in other species.
The male is distinguished from all other known males by the size and thickness of
the long bristle at the end of the last pair of legs, and by the presence of a thick,
club-shaped appendage on the second pair of legs (riech-haar of Oudemans).
Both sexes are further differentiated by the course of the epimeral grooves on the
posterior ventral surface.

Dahl groups all the forms from mouse, horse, and dog tumours as T. sauli.
Amongst these there are two males, distinguished from T. hominis in that, of the
five longitudinal furrows, the three innermost are united by a well-developed trans-
verse furrow, and the sensory organ on the second pair of legs is not more developed
than in the first pair. In the females constant distinguishing characters could not
be made out. He states that the same difficulty applies to the females from gall-
inhabiting species.

Mr Stanley Hirst has kindly directed my attention to the fact that the con-
clusions of Dahl have been severely criticised by Reuter (1910) both as regards
the probable accidental introduction of the mites in question to the preserved tissues
from which the preparations were made, and as to the identity of the species. It
appears to me that Dahl has not shown sufficient care in differentiating the forms
found from species already described.

Affinities of Tarsonemus woodi, n. sp.

Dahl (1910) regards the genus Tarsonemus as representing a transitional stage
between the gall-forming mites, Eriophyidae or Phytoptidae, and other mite families.
He bases his conclusions largely upon the characters of the fourth pair of legs. In
the species which are not endoparasitic in animals, but lead a life in relatively free
space and where mating may be effected in the open, the fourth pair of legs in the

ISLE OF WIGHT DISEASE IN HIVE BEES— ACARINE DISEASE.

775

male is relatively large and of robust build, and terminates in a stout curved spine.
These features are regarded as of value in mating. In the male of T. hominis this
limb appears definitely smaller in size and general build in proportion to the other
parts as compared with other species. Dahl interprets this as related to an endo-
parasitic life. In view, however, of our very slight knowledge of this species, and
especially on account of the doubt which exists as to its normal habitat, the
conclusion must be received with reserve.

If we apply such a comparison to T. muris and T. equi (T. sauli Dahl), these
occupy an intermediate position between such a species, e.g., as T. jloricolus and
T. ■ liominis , and so far affords some support to Dahl’s view.

Including T. woodi in this comparison, we regard its place as undoubtedly at the
end of the series. T. woodi agrees with T. hominis in general appearance in both
sexes, but in detail more closely in the male than in the female. In the male they
agree in possessing on the second pair of legs a sensory organ (riech-haar of
Oudemans) of relatively large size as compared to the one on the first pair. The
reduction in size of the last pair of legs is also a common feature, although in
T. woodi the whole limb is markedly slighter in build than in T. hominis. With
respect to the terminal claw also the comparison is interesting. In T. hominis , though
showing the inward curvature characteristic of the genus, this is smaller than is the
case in all the hitherto known species. In T. woodi the limb appears to terminate
in a straight, sharp spine. The mite has been seen carrying this spine directed
inwards at right angles to the limb. Under a high power it exhibits an appearance
similar to that of the sensory organ upon the anterior limbs, and there is doubt
as to its homology with the terminal claw present in other species. Whatever view
we take regarding the reality of the endoparasitism in T. hominis , etc., I incline
strongly to the view that these special features in T. woodi are to be interpreted
in relation to the fact that the habitat of the male is probably limited to the
tracheae of the host, and also that mating takes place in this confined space. These
conditions largely obviate the necessity for specialised clasping limbs. At the
same time a sensory organ on the limb would obviously be of value.

Another feature worthy of notice is found in the nature of the ventral surface,
which is devoid of the five longitudinal epimeral grooves which are prominent in
most species, including some of those which are regarded as endoparasitic in
mammalia, e.g. T. hominis.

In the female of T. woodi fewer points of comparison can be laid hold of. The
most noteworthy are to be observed in the two hinder pairs of legs. These in
T. hominis and in T. sauli are comparatively weak and slender. In both cases the
fourth pair terminates in the usual two long bristles. In T. ivoodi the fourth pair
is hot slender, but is reduced to three segments, is somewhat stumpy, and its two
bristles are long and sweeping. Pseudostigmata have not been described in any
of the so-called endoparasitic species, and they have not been observed in T. ivoodi.

776

DR JOHN RENNIE ON

Reviewing the main features 'of T. woodi, it appears that there are good grounds
for regarding this as a species of specialised structure in relation to the particular
habitat in which it lives.

Biological Considerations.

I now propose briefly to consider the biological problem presented by T. woodi
in relation to Isle of Wight disease.

For the final acceptance of the thesis that T. woodi is exclusively responsible
for the condition known as Isle of Wight disease, careful consideration must be
paid to the biological aspect of the problem.

I. Although the numbers of bees examined from outside Great Britain in relation
to those from within have been comparatively few, yet considerable numbers have
been tested. Through the assistance of the Ministry of Agriculture, bees arriving
in this country accompanying queens from Italy have been obtained in a number of
cases for examination.

In all, several hundreds of bees were obtained from this source. These, along
with others obtained direct from Italy, were searched for the presence of Tarsonemus.
The result of these examinations was that the bees were found entirely free from
the parasite. The evidence is so far satisfactory that it may be accepted that
Tarsonemus is not being introduced to this country in Italian bees. Smaller
numbers of Dutch bees so imported have also yielded on examination a similar
result. Bees in limited numbers have also been obtained from Switzerland and from
North America, all of which were also free from this parasite. The evidence is not
complete by any means, but, as far as it goes, it is of one kind. Since this disease
has never been clearly demonstrated to exist outside the British Isles, nor certainly
any epizootic approaching in any way the dimensions of Isle of Wight disease in the
British Isles, and further, since all such evidence as we possess points to a causal
relation between Tarsonemus and Isle of Wight disease in bees, this coincidence in
distribution is noteworthy. If a geographical distribution limited to Britain should
be established in the hive bee — and to do this is a mere matter of time and favour-
able opportunity — in my opinion it would point to a relatively recent invasion of the
bee, although the opposite finding would not be against such a view.

It may be noted that Zander (1911), who has paid particular attention to
the recording of pests found in hives and upon hive bees in Germany, makes no
reference to Acarids of any kind. In the course of our investigations we have found
in hives or upon combs, dead or live bees, at least five different species, including
one other species of Tarsonemus* These mites will be dealt with in a subsequent
publication.

* In G. R. Acad. Paris , t. 62, 1866, M. Emil Duchemin records the occurrence of a microscopic Acarus on
diseased hive bees. He gives no description nor figure. This is clearly not an endoparasite, since M. Duchemin
found that it bred upon sunflowers protected from the bees.

ISLE OF WIGHT DISEASE IN HIVE BEES — ACARINE DISEASE.

777

II. Morphology, development, and habits all point to the fact that this is a
parasitic organism which must have been related to the tracheal system of some
host for an indefinite period. The habit is not new. If T. woodi has been a
parasite of bees for ages, it seems improbable that the disease phenomena which
accompany its presence, and such as we are now familiar with, could have escaped
notice. On the other hand, it may be that, although the parasitic relationship is
not new, the pathology is. But such is not very probable.

It is true that, as far as bee records go, there have been in the past periods of
epizootic disease^ in bees from time to time, but there is no evidence that a con-
tinuous epizootic extending from sixteen to eighteen years has taken place.

III. It may be suggested that earlier methods of bee-keeping, whereby destruction
of bees was annually resorted to, kept down this parasite. This would certainly
have been the case, if the parasite were present, and the method should be applied
to all existing diseased stocks before winter. But surely the disease would then, as
now, have manifested itself constantly in the working season to a degree sufficient
to attract attention. And it must be remembered that modern methods of bee-
keeping are not confined to Great Britain and Ireland.

IV. May it not be that Tarsonemus, owing to some unknown change in the
normal balance of inter-relations, is at present undergoing one of those periods of
undue increase such as occurs from time to time in various animal forms. We must
recognise that it may be a parasite of bees which normally does not attain such an
incidence as to attract special notice, and that in recent times there has been some
change in the “balance of nature” which has led to its excessive increase. Bee-
keeping has increased in Britain within the last twenty-five years ; can it be said
that, apart from the ravages of this disease, our Islands are overstocked ? This
again is unlikely. .

V. It has been suggested that British bees of the present time are of a deterior-
ated breed, and have lost resisting power, so that Tarsonemus , a relatively non-
pathogenic parasite ordinarily, is able to breed excessively. My provisional answer
is that other racial forms are similarly affected. For example, Egyptian, Dutch,
Punic, and Italian bees can be readily infected, and in these Tarsonemus multiplies
with disastrous results, as in British bees. But the question of the ability of a stock
to survive a prolonged period of Tarsonemus infection is not a simple one ; amongst
other factors, it involves the question of relative fertility of particular queens, as
well as that of individual tolerance of the parasite (p. 753).

VI. Tarsonemus may be relatively new to hive bees and normal to some other
insect.

There remains the possibility that Tarsonemus exists normally in some wild
insect — possibly a hymenopteron — related to the hive bee, and that invasion of the
bee is recently established. In such a case, the unknown normal host will remain a
potential reservoir of the parasite.

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 29).

120

778

DR JOHN RENNIE ON

It is noteworthy that many species occur oh plants, but, as has been shown,
these, as far as known, do not possess the specialised characters of T. woodi, and for
this reason it seems improbable that bees have recently become infected from flower-
haunting forms. It is more probable that this took place through contact with
other insects, possibly robbing visitors to hives.

It may be remarked that along with my colleagues I have examined numbers of
wasps, humble bees, earwigs, wax moths, Braula caeca, and although other mites
were readily found upon the exterior, particularly upon the earwigs, the thoracic
tracheae in all cases were found clear.

The importance of finding answers to all of these questions raised is fully recog-
nised by the Investigation I have the honour to direct, and the work is continuing
with unabated vigour. I desire to take this opportunity of thanking my colleagues
for the skill and ability with which they have aided me in the work so far accom-
plished, and particularly Miss Elsie Harvey, my personal assistant, whose loyalty
and diligence have been noteworthy.

The foregoing researches have been carried out under the J oint-Committee upon
Animal Nutrition of the University of Aberdeen and the North of Scotland College
of Agriculture, with the aid of grants from the Development Commission, together
with the generous financial help of A. H. E. Wood, Esq., of Glassel. To all of these,
for their valuable assistance, and to the Local Advisory Committee, under Professor
John Arthur Thomson, whose encouraging advice has been unfailing throughout,
I desire to express the cordial thanks of my colleagues and myself. I also desire to
thank the Carnegie Trust for help in connection with this research.

Parasitology Laboratory,

University of Aberdeen.

REFERENCES.

Banks, Nathan (1904), “A Treatise on the Acarina or Mites,” Proc. N. S. Nat. Mus., Washington, xxxiii,
No. 1382.

Berlese, A. (1897), Atari, Myriopoda, et Scorpiones huiusque in Italia reperta.

Blanc et Rollet (1910), “De la presence chez l’Homme de Tarsonemus Kominis Dahl,” Compt. Rend, de la
Soc. de Biol., t. lxix, p. 233.

Canestrini, G. (1888), Prospetto dele Acarofauna Italiana, Padova.

Dahl, Fr. (1910), “Milbenals Erzenger von Zellwucherungen,” Gentralb. f. Baht. u. Parasitenk, Bd. liii,
Orig. p. 524.

Myake and Scriba (1893), “Vorl. Mitth. iiber ein. neuen Paras, d. Mensch,” Berl. klin. Wochenschr., xvi,
p. 374.

Oudemans (1906), Entomologische Berichten, v. 4, Nos. 91 and 92, p. 316 and p. 331.

Reuter, E. (1910), Gentralb. f. Bakt. u. Parasitenk, Bd. lvi, Orig. p. 339.

Saul, E. (1913), Gentralb. f. Bakt. u. Parasitenk, Bd. lxxi, Orig. p. 59.

Trouessart, E. (1892), “Considerations generates sur la classification des Acariens,” Revue Sci. Nat. de
VOuest, Paris, p. 20.

Zander, E. (1911), Krankheiten und Schddlinge der erwachsenen Bienen, Stuttgart.

ISLE OF WIGHT DISEASE IN HIVE BEES — ACARINE DISEASE.

779

EXPLANATION OF FIGURES.

Tarsonemus woodi, n. sp. Microphotographs.

(All tire preparations are photographed with a Watson Service Microscope.)

Fig. 1. Adult female, ventral view. Obj. in. Ocular No. 4.

Fig. 2. A, young tracheate female : focussed on ventral side to show tracheae. B, ovum. C, a male.

Obj. f in. Ocular No. 4.

Fig. 3. Adult male. Obj. -§ in. Ocular No. 4.

Fig. 4. Larva. Focussed to show double claw on first pair of legs. Obj. in. Ocular No. 4.

Fig. 5. Larva containing nymph. Focussed to show the latter : a female. Out of focus : A, posterior
end of larva. B, first pair of legs of larva. C, gnathosoma of larva. Obj. ^ in. Ocular No. 4.

Fig. 6. Male. Focussed to show sensory organ, A, on second pair of legs. Shows also spine on fourth
pair of legs. Obj. ^ in. Ocular No. 2.

Fig. 7. Posterior end of male, showing last pair of legs and external genitalia. A, spine. B, genital lobes.
Obj. I in. Ocular No. 4.

Fig. 8. Teased trachea showing : A, adult female. B, immature female (tracheate), C, two males.
Obj. § in. Ocular No. 4.

Fig. 9. Female, showing tracheae. Obj. ^--in. Ocular No. 4.

Vol: LI 1 .

Trans: Roy: Soc: EdinT

Rennie: Organism associated with Isle of Wight Disease

m

( 781 )

XXX. — Shackleton Antarctic Expedition, 1914-1917 : Depths and Deposits of
the Weddell Sea. By J. M. Wordie, M.A., F.G.S. Communicated by
Professor J. W. Gregory, F.R.S.

(MS. received December 6, 1920. Read February 7, 1921. Issued separately May 27, 1921.)

Introductory.

Previous to the voyage of the Endurance in 1914-1915, the depth and extent of
the Weddell Sea were either based on or surmised from the deep-sea soundings made
on the Scotia by Dr Bruce in 1903 and 1904; on Dr Otto Nordenskjold’s ship
Antarctic in 1901-1902 ; and on the German Expedition ship Deutschland under
the leadership of Lieut. FiLCHNER.in 1912-1913. The first set are confined to the
eastern and north-eastern portion, and may be said to run diagonally across the
mouth of the sea ; the second set were made in the extreme north-western area ;
whilst those of the Deutschland are disposed right down the centre of the sea, and
consist mainly of a south-to-north series approximately along the meridian of 43° W.

In the early part of her voyage the Endurance was on ground partly covered
by the Scotia , and she was thus able to supplement very considerably Dr Bruce’s
work off Coats Land; in January and February 1915 her track crossed that of
the Deutschland at about a right angle ; thereafter it lay on entirely new ground
to the west of all previous routes. She drifted right across that unknown tract
which some map-makers have called New South Greenland, and the majority of
people Morrell Land.* Apart, therefore, from the importance which attaches to the
soundings by reason of their being on ground previously unexplored, they are also
of value inasmuch as they definitely prove that “Morrell’s land” is no part of the
Antarctic continent, and probably does not exist even as an island.

Equipment.

The instruments used for sounding were three in number : a Lucas machine
graduated up to 5000 fathoms ; a smaller machine of the same type for depths
of 1000 fathoms ; and a Kelvin machine registering up to 300 fathoms. By splicing
on additional wire, however, it later on became possible to use the Kelvin for depths
of 450 fathoms. This, of course, meant that the dial, which was arranged for a 300-
fathom drum, did not register quite accurately for 450 fathoms of wire. It must
have slightly under-registered, but there was no opportunity of actually checking
the error; in any case, it can hardly have been as much as.l per cent., and can
well be neglected.

* Morrell himself called it “ New South Greenland,” and stated that it had been discovered by Captain Johnson.
To refer to it as “ Morrell’s land,” rather than “ Morrell Land,” would be more correct.

TRANS. ROY. 'SOC. EDIN., VOL. LII, PART IV (NO. 30).

121

782

MR J. M. WORDIE ON

The type of attachment employed for collecting bottom samples varied with the
machine. In the case of the Lucas, there was used either a snapper or a group of
four short tubes, each about 4 inches in length and |-inch diameter ; both were fitted
with a detachable weight of 50 lbs. In the smaller Lucas machine, the 28-lb. weight
was not detachable ; this machine, however, was very seldom required, and was finally
stowed away as unnecessary. In the case of the Kelvin a 14-lb. sounding lead was
at first in use, but later on a snapper with fixed weight (belonging to the small
Lucas machine) was found to give better samples. When the deposits ultimately
became so monotonous that they were no longer desirable as specimens, a return was
made to using the simple 14-lb. lead. Experience showed that in Antarctic waters
the snapper is the best all-round form of attachment.

This equipment was found to be quite adequate to its purpose.* Sufficient
detachable weights were carried to take one hundred deep-sea soundings. As the
great majority of the casts, however, were made in shallow waters where a machine
with recoverable sinker could be employed, not quite half of these weights had been
used up to the time when the ship was crushed and had to be abandoned. It is
hardly necessary here to give details or hints on the methods.! It may be mentioned,
however, that winding-in was done by hand ; in the case of the deeper soundings
(1500-2000 fathoms) this took almost an hour.

Position of the Soundings.

While the ship was a free agent and fighting her way southwards through the
ice, soundings were made whenever she was held up by close pack. The ship’s
course was never interrupted to take a sounding, for the number of times both in the
pack and off Coats Land when she was stationary was quite sufficient for the purpose.
After she was beset (January 1915), casts were made more frequently, and, when the
depths became moderate, sounding became a routine task practically every forenoon.
Later (August 1915), when deep water was again met with, an attempt was made to
sound twice in every degree of latitude. This at least was what was aimed at ; the
programme only lasted for a short time, however, as at the end of October the
ship was crushed and abandoned, and sounding gear was naturally not among the
essentials salved from the wreck.

In the Table of Soundings the positions up to and including January 19, 1915
(when the drift commenced), are those at which the sounding was made, calculated by
dead reckoning from an observed noon position. Onwards from January 20, however,
the positions entered are the daily noon positions observed by the ship’s officers ; their

* This was the only branch of oceanography which was well provided for — a result, of course, of planning the
scientific equipment of the Expedition almost entirely for a base on land. In view of the opportunities for oceano-
graphical work, whether the ship is beset or not, it is desirable that this science, as it was on the Scotia voyage, should
he well in the forefront of the programme of any future expedition.

f Dr Pirie gives very full and useful information about methods, etc., in his report.

DEPTHS AND DEPOSITS OF THE WEDDELL SEA.

783

accuracy, by reason of the occultation observations obtained during the winter, is of
the order of one nautical mile.* During the latter period the ship was drifting in
the ice at an average rate of four sea miles per day. As the soundings were
practically all made within three hours of noon, none of them are likely to be distant
from the latter position by more than one mile — that is to say, they differ by no more
than the possible error of the ship’s noon position.

Deposits.

The actual samples were all abandoned when the ship was crushed. At the time
they were collected, however, brief notes were always made, and these are quite
sufficient to provide a short geological description.

Everywhere in the Weddell Sea the deposit was found to be terrigenous in origin ;
in keeping with Philippi and Pirie, the term “ Glacial Mud (and Clay) ” is therefore
employed. Pirie rather emphasises the distinction between glacial mud and glacial
clay ; on board the Endurance it was not thought necessary to do so, or indeed
practicable without a detailed examination of the deposit. In the table the term
“ Glacial Mud ” covers both muds and clays. Where the deposit was sandy, or where
pebbles were commoner than usual, attention is drawn to these features. The origin,
of course, of a deposit such as this glacial mud — namely, the agency of drifting ice —
makes the presence of rocks from the size of a pebble up to boulders weighing several
hundredweights everywhere possible. Traces of sand should also have the same
wide distribution ; yet a deposit which could be truthfully called sand rather than
mud was something unusual.

In former days this characteristic deposit of Antarctic waters was classed as
“Blue Mud.” Everything terrigenous in origin found below 100 fathoms (if not
separated off into the small subdivisions such as Volcanic and Coral Muds) was in fact
liable to be put down as such. This cannot have been Sir John Murray’s original
intention, but after the publication of the Challenger volume the term certainly
had a tendency to become almost synonymous with a deep-sea terrigenous deposit.
For these Antarctic muds,’ however, the term was by no means a fortunate one.
To begin with, the deposit in question characterises the continental shelf as well
as oceanic depths. The colour, moreover (owing probably to deficiency in organic
matter), is practically never blue, but rather dark grey or brown to brownish-grey.
As a rule, when dried it becomes lighter in colour. Blue Mud, for these and other
reasons detailed by Philippi, can no longer be considered a fit name. A subdivision
of the same standing as Volcanic or Coral Mud is necessary. Philippi and Pirie
have accordingly adopted Glacial Mud as being the most suitable term. To sum up :
Glacial Mud differs from most terrigenous deposits in being found not only in
shallow water ( i.e . on the shelf which in the Antarctic is 200, not 100, fathoms

* The occultation observations have since been checked by Mr A. O. D. Crommelin of Greenwich Observatory,
and the positions adjusted accordingly by Mr James, physicist to the Expedition.

784

MR J. M. WORDIE ON

in depth), but also in depths of as much as 2000 fathoms ; it differs from Deep-Sea
Terrigenous Deposits in there being no arrangement of the constituents according to
size, the transporting agency being a solid, namely, ice ; and it differs from Blue
Mud both in colour and in the fact that there is practically no carbonate of lime
present in the deposit and very few organic constituents.*

As regards the Weddell Sea deposits, Pirie seems to imply that the distribution
of clay lies farther out from the continent than the mud. On the Endurance ,
however, it was noticed that deposits on the shelf were fully as worthy of being
called clays as those from deep water, at least as far as macroscopic evidence went.
The consistency, indeed, of the bottom deposit might vary very rapidly in a very
small area. One day the lead would plunge in to the depth of a foot into soft mud,
and the next, but a few miles away, sink only about half an inch into tough clay.

Sand or grit was frequently present, but never in large quantity except off Coats
Land, where some of the deposits were labelled as sand rather than mud ; and in
70° S. lat., 40° W. long., where for the space of a week in the middle of April 1915
the deposits were all very sandy. This unusual feature makes it quite possible that
at the latter place the Endurance was nearer land than at any other time after
leaving Coats Land ; absolutely no sign, however, of what might be land was seen
from the crow’s-nest.

The proportion of rock fragments, mainly small pebbles, was highest along the
Coats Land coast. Pebbles and boulders, however, were likely to occur at any time,
and particularly when a bottom sample was procured by means of the dredge. These
rocks were always very carefully examined (as hand specimens, and not under the
microscope) in view of the complete lack of knowledge of Coats Land geology.
Considering the ice-bound nature of that land, it seems very unlikely that exposures
of rock will ever be found there in any quantity, so that any data, however scanty,
obtainable from bottom samples such as these must be of value.

Geology of the Rock-Fragments.

The rock specimens obtained directly off Coats Land are best treated separately
from those got farther west in the Weddell Sea, as their source is not so much a
matter of doubt. On January 12, 1915, in 74° 07' S., 23° 02' W., a dredging was
made in 103 fathoms and brought up a sandy deposit full of sponge fragments and
small pieces of rock. Among the latter, granite with white felspar was the commonest
igneous rock ; basalt and dolerite were not quite so abundant ; a grey grit was much
the commonest sedimentary rock, and a purple sandstone was found, but not in any
quantity. At a later date, in 76° 34' S., 31° 27' W., the dredge yielded, in addition
to granite with white felspar, another variety with hornblende and red felspar ;

* Murray and Renard’s classification as given in The Depths of the Ocean , p. 161, requires a slight alteration'
accordingly. “ Glacial Mud (and Clay) ” should be inserted after “ Coral Mud,” grouped on the one side with “ Deep-
Sea'Deposits” and on the other with “ Terrigenous Deposits,”

DEPTHS AND DEPOSITS OF THE WEDDELL SEA.

785

basalt also occurred, and a rock which was provisionally classed as a peridotite. On
other occasions grit (both red and green), shale, and pebbles of hard mud were
obtained. The shales were generally marked with glacial striae. On all these
occasions the glacial mud which formed the bulk of the sample was extremely sandy.
Among the material collected, two rock types only could be called common, namely,
grey granite and grits, grey to green in colour. Some of the sediments recalled the
Beacon Sandstone of Victoria Land in general appearance, but they were never
sufficiently distinctive types to warrant one in saying definitely that Coats Land
geologically resembles the lands east of the Ross Sea.

The remainder of the dredgings were made farther out from land; but there is
every reason for saying that here, as well as nearer Coats Land, the source of th‘e
material was to the east. This conclusion is founded on the clockwise motion of the
Weddell Sea ice, proved by the Deutschland and Endurance drifts. Among these
later dredgings shale, quartzite, and green grit -were the commonest rocks ; grit
particularly was very common. On March 26,. 1915, in 76° 27' S., 38° 43' W., the.
haul was remarkable for including a red grit boulder weighing over 70 lbs., and a
block of limestone (with fossils) about half that weight. Another block of similar
limestone was brought up not ten miles away a few days later. The specimens
have, of course, been lost, but it is quite possible that two of them represent the
Archseocyathus Lst. of Cambrian age, previously known from Victoria Land and
from a Scotia dredging in 62° 10' S., 41° 20' W. Other sediments identified were
oolitic limestone, white and dark-grey quartzite, arkose (noted as being fairly
common), banded shale, spotted shale, and chert. Of igneous rocks there was a
considerable variety. Granite, although a hornblende variety with red felspar is
noted more than once, was never so abundant as farther east. Dolerite, however,
was common. Peridotite appears again, also pitchstone, porphyrite, and quartz
felsite ; diorite and basalt sometimes occurred, but not often. On one occasion a
fragment possibly of tuff was obtained, but its identification was apparently a matter
of doubt. Metamorphic rocks were represented by gneiss, garnetiferous gneiss, and
mica-schist. None of the igneous or metamorphic rocks, however, occurred in
quantity at all comparable to grey grit and white quartzite.

On the whole the data are insufficient for determining whether Coats Land is
geologically similar to Victoria Land or to Graham Land. The balance of probability
has always been in favour of the former alternative. The Endurance observations,
however, do not bring the problem any nearer solution, but they have certainly
produced nothing to render it improbable that Coats Land, like Victoria Land,
belongs to the plateau type of Antarctica.

Bathymetrical Results.

Previous bathymetrical maps of this area have in two instances put down
hypothetical contours over the region explored by the Endurance. Bruce on the

786

MR J. M. WORDIE ON

one hand has crowded his contours in a somewhat unnatural way to ensure “ Morrell’s
land ” being part of the Antarctic continent. Brennecke of the Deutschland , on
the other hand, in prolonging his 1000-, 2000-, and 3000-metre depth-contours
has made them sweep across to Graham Land in natural curves, and to the south-
west of them has inserted the term “ Flachsee.” The result of the Endurance
soundings definitely removes “ Morrell’s land” from the realm of probability, and
substantiates practically all that Brennecke had inferred as likely to occur in the
west. His “Flachsee” has proved to be a remarkably shrewd guess, though it still
remains doubtful what grounds he had for inferring such a broad continental shelf.

The main results of the Endurance soundings are these : —

(1) The 1000-, 1500-, and 2000-fathom lines apparently sweep across without
disturbance from Coats Land to Graham Land, passing south and west of the position
where Morrell claimed land and where Boss charted “strong appearance of land”
(but without, however, deeming it sufficiently important or well founded to mention
it in his book of travels).* The latter “land” was first called in question when
Nordenskjold made a sounding of 2050 fathoms within thirty miles of Ross’s
“ appearance of land.” The behaviour of these contours, therefore, makes “ Ross’s
land ” and “ Morrell’s land ” highly improbable. It is true that there is plenty of room
for an island here ; but it would have to rise from oceanic depths, and the rule for
other Antarctic islands is that they are generally linked to the continent by a rise or
something similar. Such cannot be the case here. Moreover, there are very good
grounds for concluding that Morrell’s chronometer was far from correct, making
(if his “north-east cape of New South Greenland” is Joinville Island) his positions
fully eight degrees too far east ; if his statejnents are to be believed, the land off
whose coast he claims to have been sealing was simply the east coast of Graham
Land. Map-makers of the future, therefore, are hardly likely to insert land in 68° S.,
48° W., unless it is vouched for by someone else besides this somewhat discredited
American sealer.

(2) The 500-fathom line follows a much more irregular course, and at one point
is deflected south as far as the face of the Wilhelm Barrier. This suggests the
existence just here of a deep channel, with a somewhat sinuous course, running
south approximately between the 36th and 37th meridians.

(3) East of the above-mentioned deep channel the bottom shoals gently towards
Leopold and Caird Coasts. The continental shelf off Coats Land is in fact compara-
tively narrow. Deep water, therefore, is found off such features as the protruding
Stancomb Wills Promontory, a barrier remnant of a once larger area ; under the
ice-cliffs 676 fathoms, no bottom, and ten miles off a depth of 1355 fathoms were
recorded. The shelf is narrow but by no means regular ; the soundings of the three
expeditions which have visited Coats Land make it pretty certain that although that

* Ross in his narrative (vol. i, pp. 177-178) tells how his inexperienced officers were often deceived by appear
ance of land, and would not be persuaded that it was otherwise until they had actually sailed over the place.

DEPTHS AND DEPOSITS OF THE WEDDELL SEA.

787

country is now completely ice-covered, except for some nunataks in the extreme
south in 78° S. lat., the solid rock below is nevertheless sculptured into hills and
valleys. This is indicated not only by the irregularity of the soundings, but also by
the uneven, mountainous character of the snow-covered land itself.

(4) West of the deep channel down the centre of the sea there is a somewhat
sharper rise than to the east ; and there are also more uniform conditions on the
shelf itself. The Endurance drifted on to the shelf in the last week of March 1915.
Up till then the total number of casts which had found bottom was thirty-five ;
besides these, there were some doubtful and some incomplete soundings, not without
importance. Once on the shelf, soundings could be made without the sinking weight
having to be sacrificed each time ; a daily cast therefore became the rule. From
March 31 till July 31 (both inclusive) 103 soundings were made. All of these
were in depths under 275 fathoms.

In a way this series of soundings on the continental shelf was unique, and it
would have been a matter of surprise if some important result had not been the
outcome of it all. The majority of the soundings clustered round 180-190 or
250-260 fathoms. Over a large area the depth was about the same, day after day,
and then it would suddenly change to another level. The back and forward drift
of the ship now had its uses, for the numerous soundings are so distributed as to

N.W. s 5 S-E-

The stepped terraces on 'the continental shelf. Vertical scale much exaggerated. The most north-westerly terrace
is eighty miles broad, and at least seventy in length (i.e. from south-west to north-east), and so on.

prove the existence of a series of stepped terraces with boundaries running N.E.-
S.W. ; these run at right angles, apparently, to the presumed coast line farther to
the south-west, but are parallel to the known coast line of the Leopold and Caird
Coasts of Coats Land. The stepped terraces extend from 76° 18' S., 38° 23' W., to
72° 37r S., 47° 47' W., a distance of about 270 sea miles. The line so measured (i.e.
across the terraces) represents very nearly the mean course of the ship’s drift from
S.E. to N.W. ; at the same time, however, she was drifting sometimes more to the
north-east, at other times more to the south-west ; and in neither direction did the
water show the least sign of shoaling. The extent of this shallow area therefore
came to be known as at least forty miles across in the south-east, and over seventy

788

MR J. M. WORDIF, ON

in the north-west. These terraces are shown diagrammatically on the accompanying
figure. The slope from one terrace to another is of course much distorted ; if drawn
to scale it would be almost imperceptible, as it works out on the average at only
1 in 200.

On the east coast of Graham Land the Charcot Expeditions have shown the shelf
there to be at least one hundred miles broad and of similar depth to what it is in
the Weddell Sea. In the past indeed the Antarctic continental shelf has aroused
considerable interest by reason of its depth being double the average depth of the
shelf round the other continents — 200-250 as against 100 fathoms. Philippi,
commenting on this, thought it might be due to planing down by ice during the
once greater extent of the Antarctic ice-cap. Nordenskjold, on the other hand,
thinks there may be some connection between the existence of the great Antarctic
ice-cap and a resulting land submergence. The terraced structure, which the
Endurance observations show, at once disproves Philippi’s' idea. These features,
however, can be explained by faulting ; * and it may therefore be the case that the
deep Antarctic shelf here and elsewhere is the result of earth movement of some sort
rather than of erosion.

(5) The slope of the continental edge as deduced from the Endurance figures is
1 in 62. This result is certainly misleading ; for unfortunately the edge was crossed
in a three-day blizzard at the beginning of August 1915, when the ship was in
considerable danger and sounding therefore of secondary importance. There is an
interval of forty miles or more in which to place the edge ; the 185-fathom sounding
is just over sixty sea-miles from a depth of 1146 fathoms. The sounding of 370
fathoms, no bottom, is of practically no value, as it is only ten miles, from 1146
fathoms. That the edge is much steeper than 1 in 62 is more than likely, judged by the
paired soundings of other expeditions. The steepest slope recorded by any Antarctic
expedition is 1 in 17, and the average 1 in 26 ; the real slopes, moreover, will
probably be a trifle steeper, as the chance that a ship crossed the edge at right angles
is small. The Endurance figure should therefore be put aside.

Sounding
on Shelf.

Sounding in
Deep Water.

Distance
Apart in
Sea-miles.

Slope.

Belgica

279

1476

25

1 in 29

Gauss ....

209

1267

18

1 „ 17

Scotia . . . .

159

1950

45

1 „ 25

Deutschland .

305

820

18

1 „ 35

Endurance .

185

1146

60

1 „ 62

(6) In the forefront of all bathymetrical work in the American sector of Antarctica
is the question of the one-time relationship to each other of the various island groups
which border the Weddell Sea to the north. On this important point, however, the

* Nordenskjold states that Bransfield Strait is due to faulting.

DEPTHS AND DEPOSITS OF THE WEDDELL SEA.

789

Endurance results have nothing fresh to bring forward. Much that is new has,
indeed, been found in recent years and been published, mainly in foreign periodicals.
Such are, for instance, the finding by the Swedish Expedition of a rise between
Joinville Island and the Powell Group, and the extremely important line of soundings
made by the Deutschland between the South Orkneys and the South Sandwich
Group. The latter show not a simple rise, but a series of alternate rises and deeps
trending E. — W. and perhaps linking these two island groups. Reference may also
be . made to a rock reported by a whaler in 1916 as lying in latitude 58° 31' S.,
longitude 41° 48 ' W. Much more evidence, however, is certainly required before it
can be decided whether the necessary link between the geologically similar regions
of Graham Land and Patagonia connects all the island groups forming the so-called
Southern Antilles, or passes somewhat more to the west, as Professor Gregory would
have it.

Table of Soundings.

Abbreviations: — K = Kelvin machine ; L = Lucas ; S.L. =Small Lucas; l = lead; sn = snapper;
t= tubes ; Gl.m. = Glacial mud or clay.

Position.

Depth

Depth

Attach-

ment.

No.

Date. 1

Lati-
tude S.

Longi-
tude W.

Fathoms.

in

Metres.

Machine.

Nature of Bottom. Remarks.

1914

1

Dec. 18

62° 42

18° 14

(2810)

(5139)

L.

Very doubtful sounding.

„ 26
1915

65 43

17 36

2819

5155

L.

sn.

Gl.m., brownish-grey in colour. The consistency
is like that of clay. Under the microscope frag-
ments of quartz and of some coloured mineral
were noticed.

3

4

Jan. 6
„ 10

70 44
72 02

21 25
16 07

2400

200

4389

366

L.

K.

Gl.m., brownish-grey. Quartz fragments very
common ; besides other unidentified minerals.

5

,, 10

72 43

18 47

216

334

K

Barrier edge only about 100 yards distant.

6

„ 11

73 20

20 55

155

283

S.L.

sn.

Pebbles and sand grains ; the former include
dolerite and basalt (1 with olivine). At this
point barrier cliff turns sharply to S.E.

7

,, 11

73 29

21 50

190

317

S.L.

1.

sn. 1
1. /
sn.

Stony bottom. (A sounding half an hour later
gave 210 fathoms, no bottom.)

8

,, 12

74 06

22 51

95

174

/S.L.
\ K.

K.

Pebbles (of grit), sand, and mud.

9

„ 12

74 10

22 58

128

234

Greenish-brown sand, pebbles, and mud ; sponge
spicules, and foraminifera.

10

,, 12

74 07

23 02

•

103

188

K.

sn;

Sand and pebbles.

A dredging made in this position brought up a
mass of sponges. The bottom deposit was
sandy, with many pebbles (the largest 2 inches
across) of : — Grey grit, common ; purple sand-
stone ; granite with white felspar, common ;
basalt ; dolerite ; syenite (possibly).

11

„ 13

74 2

26 12

676

1236

...

Stancomb Wills Promontory bore E. 2J miles.

' 12

,, 14

74 10

27 21

1355

2478

L.

sn.

Gl.m., greyish-brown. No pebbles.

13

„ 15

j 75 2

25 25

268

490

. L.

sn.

Gl.m., somewhat sandy, with a few pebbles.

14

,, 15

75 20

26 41

120

219

K.

1

Some sand on the arming.

15

16

: 76 26

28 40

134

245

K.

1.

Sand grains and sponge spicules.

16

,, 16

76 22

28 31

136

249

K.

sn.

Pebble of hornblende-granite, coated with bryozoa.

17

,, 19

1 76 34

31 27

312

570

L.

sn.

Gl.m., bluish-grey, sand, and small pebbles.

A dredging made later in the day brought up
sandy mud with pebbles of: — Peridotite (pos-
sibly) ; granite with white felspar ; granite
with red felspar ; arkose ; basalt ; pebbles of
indurated mud.

TRANS. ROY. SOC. EDTN., VOL. LI1, PART IV (NO. 30). 122

790

MR J. M. WORDIE ON

Table of Soundings — continued.

Position.

Depth

Depth

Attach-

ment.

No.

Date.

Lati-
tude S.

Longi-
tude W.

Fathoms.

Metres.

Machine.

Nature of Bottom. Remarks.

18

Jan. 20

76° 39

32° 08

275

451

Very doubtful sounding.

19

,, 21

76 44

32 45

350

640

L.‘

sn.

Gl.m. and sand ; about half a dozen small pebbles.

20

,, 22

76 49

33 22

337

616

L.

sn.

21

,, 25

76 48

33 36

382

699

L.

sn.

• Gl.m. and sand, the former grey and bluish-grey
in colour ; probably grey the surface colour.

22

„ 26

76 50

33 57

360

658

K.

sn.

Mud, dark grey, sand, and pebbles.

A fish trap which was in use came up filled with
waterlogged mud, sand, and a few pebbles, the
latter include shale and hornblende-granite with
i red felspar.

27

76 50

34 02

The fish trap was in use at a depth , of about
355 fathoms. It filled with sand, mud, and
pebbles: — Hornblende syenite ; red grit; green
grit ; shale (some pieces showing glacial strife).

23

,, 28

76 46

34 05

449

821

L.

sn.

Gl.m. , dark grey ; very little sand and a few very
small pebbles.

24

„ 29

76 47

34 12

449

821

L.

sn.

Gl.m., dark grey ; very little sand and a few
pebbles.

Gl.m., dark grey, a trifle sandy; a few small
pebbles.

25

Feb. 1

76 49

34 31

510

933

L.

sn.

26

„ 3

76 50

34 47

520

951

L.

sn.

Gl.m., dark grey, sandy.

27

,. 4

76 53

34 38

520

951

L.

sn.

28

6

76 53

34 48

530

969

L.

sn.

Gl.m., dark grey.

29

,, 8

76 55

35 01

529

967

L.

sn.

Gl.m., dark grey and sandy.

30.

,, 13

76 50

35 18

529

967

L.

sn.

Gl.m., dark grey, sand, and small rock fragments ;
dark grey mud when dried becomes light dusty
grey in colour.

31

„ 19

Mar. 5

76 55

34 54

522

955

L.

sn.

Gl.m., dark grey.

32

76 53

35 54

561

1026

L.

sn.

33

,, 18

76 53

37 17

606

1108

L.

sn.

Gl.m.

34

„ 22

76 37

38 15

1 442

808

L.

sn.

Gl.m.

35

,, 23

76 3;

38 22

443

810

L.

. sn.

Gl.m.

Dredge brought up much Gl.m., together with
pebbles of black shale and grit.

36

,, 24

76 36

38 22

419

766

L.

sn.

Gl.m.

Dredge yielded pebbles of shale and grit, one of
the latter being false-bedded.

37

,, 25

76 32

38 37

406

742

L.

t.

Gl.m.

3S

,» 26

76 27

38 43

380

695

L.

t.

Gl.m. , light grey, at least 2 inches deep.

Dredge was down twice. The first haul yielded a
porphyrite pebble 3 inches across ; the second
brought up numerous pebbles of grit and shale
besides a block of red grit 70 lbs. in weight, and
one of fossiliferous limestone 7 inches across.

39

„ 27

76 22

38 50

I 358

1

1

655

L. .

t.

Gl.m.

Dredge contained pebbles of : — Granite with red
felspar ; grit ; shale ; banded shale, brown and
dark grey.

40

,, 30

38 28

338

618

L.

Gl. m. A Rutherford tube was used for the first
and only time.

Dredge brought up a block of fossiliferous lime-
stone 6 inches across.

41

,, 31

76 18

38 23

256

468

K.

1.

Gl.ni.

42

Apr. 2

76 17

38 24

262

479

K.

1.

Gl.m.

Dredge besides abundant mud contained two
small pebbles, one a mica schist.

43

,, 3

76 17

38 34

264

483

K.

1, »

44

,, 4

76 9

38 30

250

457

K.

1.

45

,, 5

76 9

38 43

245

448

K.

1.

Gl. m.

Dredge included a few small pebbles.

46

,, 6

76 12

39 04

244

446

K.

1.

Gl.m.

47

,, 7

76 18

39 48

242

443

K.

1.

Gl.m., slightly gritty.

48

,, 9

76 29

40 14

273

j

499

K.

1.

Gl.m., gritty.

Dredge brought up sandy mud. Two pebbles,
one a limestone, the other an oolite.

DEPTHS AND DEPOSITS OF THE WEDDELL SEA,

791

Table of Soundings — continued.

Position

Depth

Depth

Attach-

No.

Date.

in

in

Machine.

Nature of Bottom. Remarks.

Lati-

Longi-

Fathoms.

Metres.

ment.

tude S.

tude W.

49

Apr.

10

76° 27

40

09

253

463

K.

1.

Gl.m., gritty.

Dredge contained a block of white grit about the

size of one’s hand.

50

”

11

76 24

40

03

255

466

K.

1.

Gl.m., gritty.

Dredge contained pebbles of : — Grey shale ; quartz-
ite ; grit ; diorite (?).

51

,,

12

76 14

39

14

238

435

K.

1.

52

12

232

424

K.

1.

Sounding made about six hours later than No 51.

53

13

76 04

39

20

212

388

K.

1.

Gl.m. and sand.

54

14

75 59

39

13

197

360

K.

1.

55

15

75 54

39

16

192

351

K.

sn.

Gl. sand and mud.

56

16

75 55

39

31

192

351

K.

sn.

Sand and pebbles.

57

17

75 57

39

41

196

358

K.

sn.

Moist sand and mud.

58

18

75 57

39

49

193

353

K.

sn.

Gl.m.

59

19

75 59

10

11

181

331

K.

sn.

Firm dark grey sand and mud.

60

19

177

324

K.

1.

Sounding about six hours later than No. 59.

61

20

76 00

40

48

178

325

K.

sn.

62

20

178

325

K.

sn.

Sounding about six hours later than No. 61.

63

21

76 03

41

48

181

331

K.

sn.

Sand, mud, and pebble.

64

22

76 01

42

01

180

329

K.

Gl.m. and sand.

Dredge yielded over 100 pebbles up to 3 inches
across. For the most part they were rounded
and enclosed in tough sandy gl.m. Some ap-
peared to be facetted : — Grey grits, common ;
quartzite, white ; pink grit ; arkose, fairly
common ; dolerite, common ; diorite ; granite
with pink felspar ; granite-gneiss ; garneti-
ferous gneiss ; peridotite (possibly) ; chert
breccia ; spotted shale : shales.

Gl.m.

65

,,

23

76 02

41

48

179

327

K.

sn.

66

,,

25

75 56

41

40

178

325

K.

sn.

Gl.m.

67

26

75 49

41

42

179

327

K.

sn.

Gl.m.

68

,,

27

75 45

41

42

177

324

K.

sn.

Gl.m., stiff.

69

,

28

75 38

41

33

174

318

K.

sn.

Gl.m.

70

,

29

75 29

41

25

169

309

K.

1.

Gl.m.

71

May

30

75 25

41

43

172

314

K.

1.

Gl.m.

72

1

75 27

42

00

175

320

K.

1.

Gl.m.

73

3

75 22

42

18

175

320

K.

1.

Gl.m.

74

f 9

4

75 23

42

26

173

316

K.

1.

Gl.m.

Fish trap was lowered and filled with bluish-grey

75

5

75 19

42

26

170

311

K.

sn.

gl.m.

Gl.m. and sand.

76

,,

6

75 10

42

20

161

294

K.

1.

Gl.m. with a little grit.

"

7

75 07

41

52

159

291

K.

1.

Gl.m. and a little grit.

Dredge filled with 2 feet of tenacious mud. About
a dozen pebbles : — Arkose ; grits, white and
pink ; green igneous rock.

78

,,

8

75 03

42

17

162

296

K.

1.

Gl.m., slightly gritty.

79

10

74 59

42

06

152

278

K.

1.

Gl.m., light grey, gritty to the feel.

80

11

75 00

41

58

157

287

K.

1.

Gl.m.

81

,,

12

75 08

41

54

157

287

K.

1.

Gl.m., gritty.

82

13

75 19

42

22

170

311

K.

1.

Gl.m., very little grit.

83

14

75 23

42

52

163

298

K.

1 I-

Gl.m.

84

15

75 27

43

09

161

294

K.

1 1-

Gl.m.

85

,,

16

75 26

43

37

153

280

K.

1.

Gl.m., very slightly gritty.

86

,,

17

75 24

43

57

157

287

K.

1.

Gl.m., slightly gritty.

87

18

75 23

44

01

156

285

K.

sn.

Gl.m., slightly sandy.

88

20

75 27

44

26

155

283

K.

1.

Gl.m., with a little sand.

89

,

21

75 24

44

45

155

283

K.

1.

Gl.m., with a little grit.

90

,,

22

75 23

44

59

157

287

K.

1.

Gl.m., light grey ; practically no sand,

91

24

75 22

45

45

165

302

K.

1.

Gl.m., light grey.

92

”

26

75 14

44

54

172

314

K.

1.

Gl.m.

Dredge contained following pebbles : — Quartzite,

white ; chert ; grits ; dolerite (1) ; shale, black
and grey.

93

”

27 j

75 04

44

51

187

342

K.

1.

Gl.m.

792

MR J. M. WORDIE ON

Table of Soundings — continued.

Position.

Depth

Depth

Attach-

No.

Date.

in

in

Machine.

Nature of Bottom. Remarks.

Lati-

Longi-

Fathoms.

Metres.

ment.

tude S.

tude W.

94

May 28

74° 59

44° 56

197

360

K.

1.

Gl. m., dark grey.

95

,, 28

205

375

K.

1.

Gl.m., grey (brownish); no sand. Sounding
about twelve hours later than No. 94.

96

.. 29

74 55

44 41

204

373

K.

1. '

Gl.m.

Dredge brought up a few pebbles Quartzite,

dark grey ; slate.

97

„ 30

74 46

44 57

224

410

K.

1.

Gl.m.

98

,, 31

June 2

74 49

45 30

253

463

K.

1.

Gl.m., grey.

99

74 47

45 12

254

464

K.

1.

Gl.m.

100

,, 3

74 45

45 05

256

468

K. .

1.

Gl.m. , slightly gritty.

101

,, 4

74 44

44 57

256

468

K.

1.

Gl.m. ; colour had a bluish tinge.

102

, 5

74 43

44 55

254

464

K.

1.

Gl.m,, grey.

103

, 7

74 32

45 03

256

468

K.

1.

Gl.m., grey ; no grit.

104

„ 8

74 27

45 13

258

472

K.

1.

Gl.m.

105

,, 9

74 26

45 23

259

474

K.

1.

Gl.m., slightly sandy.

106

,, 10

74 25

45 28

259

474

K.

1.

Gl.m.

107

11

74 27

44 45

254

464

K.

1.

Gl.m. Minute ice crystals present in the mud ;

make determination of grittiness by feel quite
impossible.

108

„ 13

74 29

46 21

259

474

K.

1.

Gl.m.

109

„ 14

74 30

46 21

258

472

K.

sn.

Gl.m., brownish-grey.

110

„ 15

74 33

46 23

258

472

K.

1.

Gl.m.

111

,, 17

74 39

46 39

252

461

K.

1.

Gl.m.

112

,, 18

74 37

46 48

255

466

K.

1.

Gl.m.

Dredge brought up mud with very few pebbles.

One of these is a white quartzite, 3 inches across.

113

„ 21

74 30

47 39

260

475

K.

1.

Gl.m.

114

,, 22

74 21

47 39

262

479

K.

1.

Gl.m.

115

,, 24

75 02

47 23

249

455

K.

1.

116

,, 25

73 57

47 18

239

437

K.

1.

Gl.m., grey.

Dredge contained numerous subangular and an-
gular pebbles ; — Hornblende-granite with red
felspar; arkose ; quartz felsite ; quartzite;
grits, grey and green, common ; pitchstone ;
indurated mud pebbles.

117

„ 26

73 58

47 17

238

435

K.

1.

118

,, 28

73 59

47 20

255

466

K.

1.

Gl.m., grey.

119

,, 29

74 03

47 29

260

475

K.

1.

Gl.m.

120

July 1

74 08

47 54

255

466

K.

1.

Gl.m.

Dredge brought up mud ; no pebbles.

121

,, 3

74 11

48 39

226

413

K.

1.

Gl.m.

122

„ 4

74 09

48 50

203

371

K.

1.

Gl.m.

123

„ 4

194

355

K.

1.

Gl.m., brownish. Sounding made twelve hours

later than No. 122.

124

5

74 06

49 08

184

336

K.

1.

Gl.m., brownish.

Dredge yielded nearly 60 pebbles : — Grits, quart-

zites .(white), very common; shales; tuff (?);
gneiss .; garnetiferous gneiss ; aiorite or gabbro ;
felsite ; basalt ; arkose.

125

,, 6

74 06

49 11

185

338

K.

1.

Gl.m., with very small rock pieces.

Instead of plunging the usual distance of about 1

foot into the mud the lead only went in J inch.

126

u 7

74 07

49 14

192

351

K.

1.

Gl.m., distinctly sandy.

127

,, 8

74 08

48 53

192

351

K.

L

Gl.m.

128

,, 9

74 06

49 20

192

351

K.

1.

Lead absolutely clean.

129

>, 11

74 06

49 17

189

346

K.

1.

Little mud in cup ; otherwise lead clean.

130

,, 13

74 12

48 57

190

347

K

1.

Mud only in cup.

131

,, 16

73 36

48 27

202

369

K.

1.

Gl.m., brownish-grey ; lead nearly clean.

132

„ 17

73 38

48 28

196

358

K.

1.

Gl.m., covering lead to depth of i foot.

133

,, 20

73 26

48 05

190

347

K.

1.

Gl.m., 6 inches deep;

134

,, 21

73 26

48 07

190

347

K.

1.

Gl.m., brownish -grey.

Dredge yielded: — Grits, quartzites, 40-50 pebbles

4-3 inches across ; granite with red felspar.

135

i 22

73 19

47 50

185

338

K.

1.

Gl.m.

136

24

73 07

48 12

186

340

K.

1.

Gl.m. ; lead sank very deep in mud.

137

,, 25

72 59

48 04

187

342

K.

1.

Gl.m , 8-10 inches.

DEPTHS AND DEPOSITS OF THE WEDDELL SEA.

793

Table of Soundings — continued.

No.

Date.

Posi

Lati-
tude S.

don.

Longi-
tude W.

Depth

in

Fathoms.

Depth

in

Metres.

Machine.

Attach-

ment.

Nature of Bottom, Remarks.

138

July 26

72° 51

47° 34'

189

346

K.

1.

Gl.m., 6 inches.

139

27

72 47

47 36

190

347

K.

1.

Gl.m., 6 inches.

140

’ }

28

72 44

47 34

188

344

K.

1.

Gl.m., 6 inches.

141

29

72 40

47 34

189

346

K.

1.

Gl.m., 10 inches.

142

’3

30

72 39

47 34

187

342

K.

1.

Gl.m., dried to light grey colour.

Pebble of white quartzite.

143

31

72 37

47 47

185

338

K.

1.

Gl.m.

144

Aug.

4

370

677

K.

Sounding made three hours earlier than No. 145.

Ship drifting rapidly.

145

4

71 50

48 41

452

827

L.

1.

Doubtful sounding, owing to lightness of sinker.

146

9 9

5

71 42

49 16

1146

2096

L.

t.

Gl.m., light grey, 6 inches deep.

99

10

Extremely doubtful sounding on this date.

147

12

71 03

49 40

1550

2835

L

t.

148

17

70 38

49 54

1676

3065

L.

t.

Gl.m., light grey.

149

Sept.

25

70 09

50 08

1900

3475

L.

t.

Gl.m., grey.

150

6

69 54

50 22

1850

3383

L.

Gl.m., light grey.

151

, ,

20

69 38

50 35

1856

3394

L.

t.

Gl.m. , light grey.

152

”

28

69 32

51 12

1876

3431

L.

t.

Gl.m., grey.

BIBLIOGRAPHY.

Arctowski, H., Geographical Journal, xiv (1899), p. 77 : “Bathymetrical Conditions of the Antarctic
Regions.”

Brennecke, W., Ann. Hydrographie, xli (1913), pp. 134-144: “ Ozeanographische Arbeiten der Deutschen
Antarktischen Expedition : Die Eisfahrt.”

Bruce, W. S., Scottish Geographical Magazine, 1905 : “Bathymetrical Survey of the South Atlantic Ocean
and Weddell Sea.”

Morrell, B., A Narrative of Four Voyages to the South Sea, etc., 1822-1831 (1832).

Murray, John, and E. Philippi, Deutsche Tief see- Expedition, x (1908) : “ Die Grundproben der Deutschen
Tiefsee-Expedition.”

and A. F. Renard, Scientific Results of the Voyage of H.M.S. “ Challenger ” : Report on Deep-Sea

Deposits (1891).

Nordenskjold, O., Wissen. Ergeb. d. Schwedischen Siidpolar-Expedition, 1901-3, Bd. i, Lief, ii (1917): “Die
ozeanographischen Ergebnisse.”

Philippi, E., Deutsche Sudpolar- Expedition, Bd. ii, Heft 6: “Die Grundproben der Deutschen Siidpolar-
Expedition.”

Zeit. f. Gletscherk. ii (1907): “Ueber die Landeisbeobachtungen der letzen fiinf Siidpolar-

Expeditionen.”

Pirie, J. H. Harvey, Trans. Roy. Soc. Edin., xlix, pt. iii, No. 10 (1913): “Scottish National Antarctic
Expedition, 1902-4 ; Deep-Sea Deposits.”

Pryzbyllok, E., Zeit. Ges. Erd. Berlin, 1913, pp. 1-17 : “Deutsche Antarktische Expedition : Bericht fiber
die Tatigkeit nach Verlassen von Sudgeorgien.”

Ross, J. C., A Voyage of Discovery and. Research in the Southern and Antarctic Regions during the Years
1839-43 (1847).

Rouch, J., Deuxieme Expedition Antarctique Franyais : Oceanographie Physique, pp. 1-11.

Wordie, J. M., Geographical Journal, li (1918), p. 216 : “The Drift of the Endurance.”

TRANS. ROY. SOC. EDIN., VOL. LII, PART IY (NO. 30).

123

( 795 )

XXXI. — Shackleton Antarctic Expedition, 1914-1917: The Natural History of
Pack-Ice as observed in the Weddell Sea. By J. M. Wordie, M.A., F.G-.S.
Communicated by Professor J. W. Gregory, F.R.S. (With Nine Text-Figures
and Four Plates.)

(MS. received December 6, 1920. Read February 7, 1921. Issued separately June 21, 1921.)

CONTENTS.

I. Introduction .

II. Early Stages .

III. The Ice in Motion .

PAGE

795

800

804

IV. Changes in the Ice
V. Decay .

VI. Bibliography

PAGE

817

825

82

I. — Introduction.

The opportunities for observation were afforded by the voyage and subsequent drift
of the S.Y. Endurance. During December 1914 and January 1915 for a period of
six weeks she successfully bored her way through pack-ice of every description —
drift-ice, open-pack, and very frequently even close-pack. Continually fighting, she
penetrated from 59° to 72° S. lat., and finally reached the land water off Coats Land on
the latter parallel. As the crow flies, therefore, she was navigated through ice for
nearly 800 geographical miles On this voyage ; her actual course among the ice-fields
and floes was computed to exceed 2000 miles, an achievement without parallel in the
Antarctic. The principle adopted was to keep to the east, where presumably there is
less pack than in the west ; if the Endurance experience is a normal one, however,
the meridian of 20° W. long., which was followed, is certainly not far enough east.

The exploration of Coats Land and the discovery of Caird Coast followed, until
the ship was finally beset on January 19, 1915. From that date until October 1915
she drifted northwards and westwards round the Weddell Sea at an average rate of
four sea-miles per day, till finally crushed and abandoned on October 27. There-
after the crew tried to sledge to land ; or, camped on the ice, drifted from the position
where the ship was wrecked (69° S. lat.) through a further seven degrees of latitude.
So passed a second summer in the ice ; and it was not till April 1916 that boats could
be launched and escape effected, through the fringe of the pack, to Elephant Island.

These seventeen months of close association in some form or other with pack-ice
have led to an opportunity of studying its origin and decay, and some of the laws
governing its behaviour, which has never before been afforded to a British expedition
in the Antarctic. The conclusions thus obtained, particularly as regards the westerly
drift, will, it is hoped, be of use to all future Antarctic navigators.

From comparative study it appears as if the sea-ice cycle round the North Pole
is much the same as in the South — spread, however, over a longer period and

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 31). 124

796 MR J. M. WORDIE ON THE NATURAL HISTORY OF PACK-ICE

influenced by much higher summer temperatures. The Arctic ice, however, has
never been subjected to quite the same methods of observation, and it is therefore
better to regard this paper as the natural history, not of pack-ice in general, but of
Weddell Sea pack-ice.*

Historical. — -Previous to 1820 the idea seems to have prevailed that no ice.
formed at sea. Daines Barrington and Higgins, for instance, tried to reconcile
the reports (now known to be erroneous) of whalers, that all ice found at sea was fit
to drink, with the fact that sea-water frozen in the laboratory was fairly salt ; they
maintained, accordingly, that “ no considerable congelation ever takes place in the
sea,” and that the floating ice observed in high latitudes is land-derived. In 1820,
however, Scoresby, with the experience of many whaling voyages behind him,
published what appears to be the earliest scientific account of the Arctic pack-ice by
the actual observer himself. He demonstrated for the first time that ice could form on
the sea far from land, and also showed how this ice might sometimes be fresh enough
for drinking purposes. The mistake made in the next four decades was to imagine
that freshness was the rule — a misconception which persisted till Walker, who
accompanied M‘Olintock as surgeon on the Fox , published his results in 1860.
References to ice in the Franklin Search Expeditions are of course numerous, but
very seldom reliable ; and practically none of the writers except Rae take account
of the physical and chemical side of the subject. Rae in later life read a short paper
on some of his observations to the Physical Society, and made Guthrie sufficiently
interested to make further experiments.

When in the seventies interest became once more directed to the Arctic, a good
deal more attention than formerly was paid to the natural history of sea-ice. Both
Payer and his' associate Weyprecht published clear and direct accounts of what
they saw. On the Hares Expedition, however, ideas on sea-ice became again some-
what confused ; the term “ palseocrystic,” for instance, was introduced to include
“ floebergs ” and heavy floes. Nares himself thought floebergs were due to direct
freezing of sea-water ; but Moss, and later Greely, realised that they were due to
accumulated snowfall ; and many years later Peary ' actually found them forming
from glaciers on the north coast of Greenland. The ‘‘ palseocrystic floes,” on the
other hand, are now regarded as simply hummocky-pack and heavy floes ; later
travellers over the Polar Sea, however, have done little to advance knowledge in
this respect.

In more recent years Pettersson, Drygalski, and Hamberg have made consider-
able additions on the physical side. Pettersson’s results especially call for
mention, for, if accepted, they are of extreme importance ; Buchanan and others,
however, have taken exception to his work. It now remains for someone to test in
nature what Pettersson found in the laboratory.

* In the forthcoming Ice Memoir of the Scott Antarctic Expedition, 1910-1913, Mr Priestley proposes to give
an account of the Ross Sea fast-ice and pack-ice.

AS OBSERVED IN THE WEDDELL SEA.

797

Slowly, piece by piece, the Arctic pack has come to be known, but a great deal
still requires to be done. A useful summary of present-day views is given in
Krummel’s Handbuch der Ozeanographie. The Antarctic pack, however, has
never had the same attention paid to it. Fricker made a short compilation in
1893 ; but more important, as it was the record of actual observation, was
Arctowski’s memoir in the Belgica Results. Scott, Ferrar, Gourdon, Priestley,
David, Mawson, and others have all in turn contributed additional knowledge.
Not through them, but in spite of their accounts, an impression seems, however, to
prevail that Antarctic pack is of very different nature from Arctic ; to this the strongest
exception must be taken. Wrong ideas on sea-ice have never been short-lived; it
is to be hoped, therefore, that the case of Arctic and Antarctic pack will not be
prejudged till more is known about them both.

Terminology. — No new terms are introduced, beyond those suggested for the
various types of cracks originating in the ice. The old terminology, used by the
Arctic whalers, had a natural and practical origin, and is therefore followed as
closely as possible. As there has been a slight tendency, however, for some recent
Antarctic voyagers to alter the original use of a word, a restatement of some of
the terms has become necessary ; and a short glossary is therefore conveniently
inserted here.

The best of all previous glossaries is, of course, that of Scoresby, who was whaler
before becoming scientist. A much fuller, and at the same time the most recent,
list of terms is that of Markham and Mill in the Antarctic Manual , 1901 ; most
of their definitions cannot be bettered, but in others some slight modification seems
desirable, since, after all, it is the navigators and sailors themselves who must have
the last say. Where possible, howevef, the actual phrases of Scoresby or of Mill
are used here.

Slush or Sludge. — The initial stages in the freezing of sea-water, when its
consistency is gluey or soupy. The term is also occasionally used for brash-ice still
further broken down.

Pancake-ice. — Small floes of new ice, approximately circular, and with raised rims.

Young-ice. — All unhummocked ice, no matter of what age or thickness, which
has platy structure and fibrous appearance when broken. Ice of this nature was
formerly known in the Arctic as “bay-ice,” but the term, unfortunately, has also
been largely used in the Antarctic for “ fast-ice,” and for exceptionally heavy
hummocky floes. With two such opposite meanings, “ bay-ice ” is therefore no
longer of use for descriptive purposes.

Fast-ice. — Sea-ice which remains fast in the position of growth throughout the
winter, and sometimes even during the ensuing summer. It may therefore attain
a thickness considerably above the average. Other names for this type of ice are
“land-ice” (Payer and Mill), “Schelfeis” (Drygalski), “shore-ice” (Nansen),
“ bay-ice ” (Shackleton and David), and “ coast-ice.” If it is thought necessary to

798

MR J. M. WORDIE ON THE NATURAL HISTORY OF PACK-ICE

employ a special name for fast-ice when it breaks adrift, “ land-floes ” is the most
suitable ; but generally it can be simply spoken of as heavy floes.

Floe. — An area of ice, other than fast-ice, whose limits are within sight. The
surface of a floe may be level or hummocked, and in size it may vary from “ pan-
cakes” on the one hand to “ fields” on the other. “Light floes” are. between 1 and
2 feet in thickness ; floes thicker than this are termed “ heavy.” The latter, however,
often owe their thickness to hummocking, and in the Antarctic at any rate are usually
covered with fairly deep snow.

Field. — An area of ice of such extent that its limits cannot be seen from the
masthead.

Crack. — Any fracture or rift in sea-ice.

Lead or Lane. — A navigable passage through pack-ice. Leads may form either
by the widening of a crack or by a general loosening of the floes. (On the Endurance
voyage it was customary to speak of the former as leads even when covered with
young-ice.)

Pool. — Any enclosed water-area in the pack other than a lead or lane. Pools
may be of any size: those called “polynia” by Admiral Wrangel were so large
as to give rise to the belief in an open polar sea.

Frost-smoke . — The fog-like clouds which appear over newly formed leads and
pools, owing to the contact of the colder air with the relatively warm sea-water.

Water-sky. — The dark streak on the sky due to the reflection of leads 'or pools
or the open sea.

Ice-blink. — The white or yellowish-white glare on the sky produced by the
reflection of large areas of sea-ice. The antithesis of water-sky.

Hummocking. — The processes of pressure formation whereby level young-ice
becomes broken and built up into hummocky-pack. “ Tenting,” “ rafting,” and
“ raftering ” are terms in use to describe different phases of the process.

Hummocky-jloes. — Floes composed wholly or partly of recemented pressure-ice.
They have also been described as “ old pack,” “ screwed pack ” (David), “ Scholleneis”
(German writers), and sometimes simply “pack-ice.” In contrast to young-ice, the
structure is no longer invariably platy or fibrous, but is generally spotted and
granular. There is less salt present, and the ice may appear almost translucent.

The Pack. — Term used in a wide sense to include any area of sea-ice, other than
fast-ice, no matter what form it takes or how disposed. The French term is
“ banquise de derive.”

Close-pack-ice. — Pack composed of floes mostly in contact.*

Open-pack-ice. — The floes for the most part do not touch.*

Drift-ice. — Loose, very open pack, where water preponderates over ice.* The

* Drift-ice is so open that ships can go full speed through it, and hardly ever need to change direction. In open-
pack, on the contrary, the speed is slow and changes of course continually necessary. In close-pack a sailing-ship’s
course is generally completely checked, while steamers can only progress by repeatedly charging the ice.

AS OBSERVED IN THE WEDDELL SEA.

799

floes are usually smaller than in close- or open-pack, being, in fact, the result of the
first stage in the breaking down of the ice.

Brash. — Small fragments and rounded nodules : the wreck of other kinds
of ice.

Bergy-bits. — Medium-sized pieces of glacier ice or of hummocky-pack washed
clear of snow. (Typical bergy-bits have been described as being “ about the size
of a cottage.”)

Growlers. — Smaller pieces of ice than the above, appearing greenish in colour
because barely showing above water-level.

Rotten-ice. — Floes which have become much honeycombed in the course of
melting.

The above list is by no means exhaustive. There are many other terms of less
importance, some of them being quite local, and many now obsolete. “ Sea-bar,”
“ sailing-ice,” “ tongue,” and “ calf ” seem to have gone out of use almost altogether.
Drift-ice may collect into “ streams” and “patches.” The pack edge may protrude
as. a “point,” or recede to form a “bight” or “ bay,” etc., etc. Local terms are
particularly numerous in Newfoundland, but are seldom found in print. Certain
terms peculiar to navigation in pack-ice should also be mentioned : “ sallying,”
for instance, describes how a crew rolls a ship by dashing across from side to side in
unison; a ship gets “nipped” or “beset” when open-pack closes up round her and
stops all progress ; and “ boring ” and “ slewing ” describe different ways of working
through close-pack.

It will be gathered from the above definitions that sea-ice in the first instance
is divided into (i) fast-ice and (ii) pack-ice. The latter is further subdivided,
according to the arrangement and size of the floes, into (a) field-ice, ( b ) close-pack,
(c) open-pack, and ( d ) drift-ice. The floes themselves in all four subdivisions may
be of young-ice or of hummocky-ice, and light or heavy according to thickness.
A chart of ice-conditions should first of all distinguish fast-ice and the above four
subdivisions, and then, if necessary, specify the nature of the floes.*

(i) Fast-ice.

Sea-ice

(ii) Pack-ice

(a) Field-ice

( b ) Close-pack

(c) Open-pack

( d ) Drift-ice

according to
1 arrangement
and size.

f Young-ice ) according to
\ Hummocky-ice / surface.

/ Light Hoes ) according to
\ Heavy floes / thickness.

* The Danish Meteorological Institute has for many years been publishing an annual chart of this nature for the
ice in the Arctic. The naming, however, is slightly different from that adopted here. Six types of ice are dis-
tinguished, namely, “unbroken polar ice (i.e. fast-ice) ; land-floe ; great ice-fields ; tight pack ; open ice ; bay-ice and
brash.” Certain of these names are no longer suitable in English. It will be noticed that four of the types depend
on the arrangement of the floes, but two (land-floe and bay-ice), whose usefulness on the chart is open to doubt, on the
nature and thickness of the floe itself.

800

MR J. M. WORDIE ON THE NATURAL HISTORY OF PACK-ICE

II.— Early Stages.

Formation. — In the Weddell Sea, formation of new sea-ice took place both in
rough water and in smooth ; and the resulting structure differed accordingly. One
had to distinguish, therefore, between young-ice formed in calm water, and ice
formed on a ruffled sea. The former condition was found to be much the commoner ;
the latter was only noticed as a consequence of the very strongest blizzards, when
the pack was completely broken up and rearranged. At no time during the voyage
of the Endurance was the open sea seen to freeze over except in a dead calm ;
though in higher latitudes there should be no reason against a rough open sea
freezing. The general rule, however, seemed to be for young-ice to form in the
ever-opening pools and leads among older ice, as the latter offered protection by
damping down any swell.

(l) Growth of young -ice in still water was studied very frequently in the leads,
lanes, and pools which were continually forming at all times of the year. In 1915,
when in lat. 77° S., young-ice was noticed for the first time on February 6, and a
week later all leads began to freeze almost as soon as- formed. This went on until
October, when for new leads to freeze became unusual ; the ship’s position was then
in lat. 69° SO7 S. Water-skies were noted as being numerous on October 9 ; and
on the 11th young-ice freshly formed on a pool had melted. In 1916, owing
to the N.N.W. drift of the ice-floes on which various camps were situated, observations
were made in much lower latitudes ; young-ice did not start forming that year
until the first week of March, and then only in a cold snap, the latitude at the
time being 64° S.

The ideal opportunity for observing the first stages of ice-formation was such as
was found on May 2, 1915. On that date a crack, formed in heavy hummocky
pack, opened to a lead, and from a distance was seen to be giving off abundant
frost-smoke. By reason of the crack having formed in heavy pack of composite
origin, the bounding walls were seldom perpendicular ; occasional tongues of ice
jutted out into the water at various depths, and by reflecting up the light showed
the intervening water filled with freely floating small platy crystals of ice about the
size of a finger-nail. They had not yet arranged themselves or coalesced in any
way, but seemed to fill the water for a depth of some feet. The water above such
a jutting ice-tongue was probably chilled to a much greater depth than farther out
in the lead, being bounded both below and on one side by ice, and above by the cold
air ; and this might account for the number of shimmering crystals in the water.
As the crystals became more definite, they rose to the surface, and one could almost
see them arranging themselves on to the film growing out from the edge of the lead.
Such a fringe of young-ice was generally referred to as “black-ice” ; the blackness,
however, was largely due to contrast with the surrounding snow-covered floes. As
a general rule it had a smooth but slightly damp upper surface formed of platy ice-

AS OBSERVED IN THE WEDDELL SEA.

801

crystals (as already described) set horizontally. This is a feature first noted by
Ferrar, and applies only to the uppermost \ \ inch ; the vertical arrangement of

the plates and grouping into bundles below the upper layer and throughout the
rest of the ice have been noticed by all observers, and described and figured in
detail, particularly by Drygalski. This is the structure often referred to as platy
or fibrous.

To the impression, however, given by Ferrar, that the upper layer is always
formed of horizontally arranged plates, exception must be taken. A note of
April 19 says “that from the edge of the old-ice thin narrow wedges (finger-like)
extend out into the black-ice ; these wedges have their apex against the old-ice and
broaden as they go outwards, often reaching a length of 6 inches. They appear

even darker than the surrounding black-ice, and the reason seems to be this : that
they are made up almost entirely of plates set vertically.” In the spaces between
the wedges, however, the plates at the surface were set horizontally, reflected up
the light a little, and did not appear so dark, therefore, as the wedges. Similar
structures were frequently observed, and are apparently always to be expected
where the water is bounded by a wall of old-ice at much lower temperature, from
whose edge the young sheet can grow out. On May 2 a sketch was made of such
a thin sheet in active process of growth (figs. 1 and 2).

Three bands, characterised by separate structures, were distinguishable. First
of all, an irregular network 5 to 6 inches broad ran along the immediate edge of the
old floe ; then came a strip about 1 foot broad of radiating black wedges separated
ftom each other by lighter interspaces ; and still farther out was a third strip made
up of an irregular mass of plates not yet systematically arranged on to the wedges.
The second figure shows how irregular the wedges are, both in shape and in relative
arrangement, and how the term “ wedge” is simply one of convenience.

In the wedges the ice-plates are arranged vertically, or in an almost vertical
position ; in the interspaces horizontally arranged plates shade downwards into a

802

MR J. M. WORDIE ON THE NATURAL HISTORY OF PACK-ICE

vertical series, and this latter position becomes the rule for further increase in
thickness.

Ice-flowers were a common feature whenever newly formed cracks and leads
froze over in winter ; they were at their best when the temperature was below
zero. They never remained perfect very long, however, for the clusters were
generally rimed over in about twenty-four hours ; or, since very low temperatures
were rarely of long duration in 1915, a certain amount of remelting soon
took place.

When ice-flowers were not present, young newly formed ice was always more or
less soft and damp on the upper surface, owing to the temperature being generally
higher than was conducive to their formation. On such a surface brine-bubbles
were numerous, and were apparently potential nuclei for ice-flowers. Observations
made on. these bubbles in the end of May showed that, from originally being small
and somewhat elliptical, they grew to be nearly an inch in length. In the
“wedges” the axis of the bubble ran parallel with the length of the wedge; but
in the interspaces, and where there was no guiding structure such as the wedge,
the direction of the bubbles was quite fortuitous. Those which reached nearly an
inch in length had been under observation for a week, when a fall of snow effectively
prevented further investigation. Perhaps the earliest stage of the bubbles are the
minute white specks such as were noticed on May 2 in black wedges of ice not
twenty-four hours old. Elsewhere, and at different times, both the small rudi-
mentary white specks and the undoubted brine-bubbles were often seen. It seemed
pretty certain, indeed, that it only required a sudden drop in temperature for an
ice-flower to form round such a bubble as nucleus.

The ice-flowers were only salt at the base, the distal crystal points being simply
rime. Their irregular distribution, however, over a sheet of young-ice requires some
explanation. A possible one is afforded by the way in which young-ice on a lead
crystallises outward from the sides, and so takes different lengths of time to form,
the resulting salt-content therefore being a variable one.

What has been described above either as “black-ice” a day or two old, or
more generally as “ yrtung-ice,” is simply what the Arctic whalers called “ bay-ice.”
As ice formed in bays was always level, they also came to apply the term to
undisturbed ice whether formed in bays . or on the open sea. Scoresby says :
“Bay-ice is that which is newly formed on the sea and consists of two kinds,
common bay -ice and pancake-ice ; the former occurring in smooth extensive sheets,
and the latter in small circular pieces with raised edges.” “ Bay-ice may be said to
extend from the first pellicle of ice up to a foot in thickness.” Payer, Brtjce, and
J. K. Davis use the term in its original and proper sense. In most of the Antarctic
expeditions of the last twenty years, however, the name has unfortunately been used
literally for ice formed in bays; in the Ross Sea, ice of this nature is sometimes
almost a miniature tabular berg. The latter, so-called “ bay-ice,” will be referred

AS OBSERVED IN THE WEDDELL SEA.

803

to here as fast-ice, or land-floes where adrift ; and, to prevent confusion, “ young-ice ”
will be always used to denote the old whalers’ “ bay-ice.”

(2) Opportunities for seeing young-ice forming on a rough sea were of rare
occurrence. One such, however, took place on July 15, when a blizzard opened
a lead near the ship 300 yards in breadth. Crystallisation started almost immedi-
ately, and was materially helped on by the amount of blizzard - driven snow
which had already formed a slush on the water surface. “It was like a rough
frozen sea, for the slush had been rippled as it froze. In one place the slush had
crystallised in circular patches — a type of pancake-ice.” The wind and the motion
of the water rounded otf these minute pancakes in the usual way ; “ and in one case
they had been rafted on to one another so as to appear apposed when viewed in
section.”

Without going further into detail, it may be stated that “pancake-ice” is the
result of the sea freezing while rippled by the wind or disturbed by swell ; on the
contrary, “black-ice” (with its frequent accompaniment of “ ice-flowers”) is confined
to still water. The term “ pancake-ice ” has a very definite and precise meaning ;
of all the terms used to describe ice, it is the one least liable to misconstruction.
It would be well, nevertheless, to emphasise that, though outwardly the same,
ice-pancakes may originate in two ways. Generally they form during the building
up of young-ice, when wind and swell roughen the sea and prevent any widespread
sheet being preserved. As such they range in size from 2 to 3 inches in diameter
to floes 2 to 3 yards across, capable of supporting a sleeping seal. The other and
the rarer type occurs during the decay of sea-ice, for the floes may then be broken
into small areas and assume the form of pancakes ; frequently, too, snow and
slush, by the wearing down of the floes in summer or from the capsize of an
iceberg, collect in patches and by a to-and-fro motion behave like young-ice and
form similar small pancakes.

Horizontal Banding. — A feature of young-ice, if it grew to any thickness, was
the frequent presence of horizontal banding. Though previously noticed by David
and Priestley, this banded structure has never been closely examined and
thoroughly explained. A block of young-ice of this nature was investigated in
detail during the winter, and gave the following results : —

Cl per cent . Spec. Grav.

cm. white opaque ice ...... ‘274 '894

1J cm. blue translucent ice . . . . . ‘238 '918

3 cm. white opaque ice ....... *212 '893

Thin blue translucent band (followed below by alternating *207 '918

narrow bands).

The Cl figures show that the difference between the two kinds of ice was not one
of salinity ; it is improbable, therefore, that it depends on the rate of freezing.
The actual difference was one of density, the white bands being the lighter.
TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 31). 125

804

MR J. M. WORDIE ON THE NATURAL HISTORY OF PACK-ICE

David and Priestley saj the cause is similar to what occurs in lake-ice, where
during pauses in crystallisation large quantities of gas rise from decaying organic
matter on the lake-bottom to the under surface of the ice. Some further
explanation, however, appears, necessary.

Plasticity. — The most striking physical character of young-ice is its plasticity.
Pancake-ice, however, and ice formed from slush were never so plastic as black-ice,
and this presumably was the result of their not being so salt. Pettersson contends
that plasticity is a peculiarity connected with the expanding properties of sea-ice
down to about - 20° C. All varieties of ice are more or less plastic at their
melting-point, but in the case of sea-ice, he says, the wide range of the eutectic
points for the different salts in it virtually means that it has already partially
begun to melt at —20° C., and is therefore plastic from there up to the melting-
point of the ice proper (— 1'8° C.).

Chlorine Percentage. — The saltness of young-ice must depend partly on the rate
at which it freezes ; and possibly at this stage, when it is only 2 to 3 inches thick,
the chlorine percentage may be an index to the total amount of included salt. The
methods on the Endurance , however, allowed of the amount of chlorine only
being determined. In young-ice about 2| inches thick, it was on different occasions
found to be present in the proportion of 7'23, 7'6 9, and 8'5 grms. per thousand.
On the one occasion on which sea-water, over which ice was forming, was tested,
the former contained 19 '46 per thousand Cl, as against presumably 7 to 8 per
thousand in the young-ice. #

III. — The Ice in Motion.

Along the Antarctic coasts the rule seems to be for the sea-ice to be at all times
liable to break up and drift away to the north and west. To this, however, there
are two exceptions : firstly, where the configuration of the land is such as to shelter
the ice from the prevailing winds, thus preventing disruption in winter, and
sometimes even in summer ; and secondly, where there are obstructions of the
nature, for instance, of stranded bergs, which hold up the ice both in summer and
in winter. The Weddell Sea fully bore out the general rule. At two places only
is the ice at present known to be stationary ; one of these is off Leopold Coast,
where a chain of stranded bergs a little distance out from the land effectively
anchors a long strip of ice and prevents it being continually subject to the restless
motion of the pack ; the other is on the coast of Graham Land between Eoss Island
and Joinville Peninsula, where, to judge by Eoss’s narrative and Nordenskjold’s
descriptions sixty years later, the ice very seldom breaks out. The former appears

* These figures were determined by Mr James, physicist to the Expedition, using what under the circumstances
is the most convenient method, namely titration with silver nitrate. Unfortunately, no experiments were made to
determine the specific gravity of ice as young as the above. The figures for specific gravity discussed later on were
all got from ice over 2 feet thick, where the chlorine content only averaged about 2 to 3 per thousand, and was
sometimes even less.

AS OBSERVED IN THE WEDDELL SEA.

805

to be a case of fast-ice, and may possibly be a type of what Drygalski calls
“ Schelfeis.” Whatever the naming of this ice may be (it is best called “ fast-ice ”
when anchored, “land-floes” or simply “heavy pack” when adrift), emphasis must
be laid on its occurrence being merely a temporary interruption of the normal
sea-ice cycle. Should such an interruption, however, persist so as to verge on
permanency, then an important new feature is produced, namely a barrier (or, as
Nordenskjold and others now call it, “ shelf-ice ”). Direct freezing from sea-water
ceases, but the ice-mass continues to grow by the addition above of successive
layers of snow, which in due course become neve. When the Weddell Sea becomes
fully known, several barriers will probably appear on the map. Here reference
need only be made to the one found by Nordenskjold in the north-east portion
of the sea, namely along the Oscar Coast of Graham Land. In his own descriptions
it is referred to as a “ low ice-terrace ” ; but its height (30 metres) and the origin
which he adopts for it make it quite comparable with the Great Boss Barrier ; and
it has therefore been proposed to call it in future the Nordenskjold Barrier.

Where the ice is not in motion, therefore, fast-ice is developed, and, in extreme
cases, barriers or shelf-ice. These are the exceptions ; the rule is for sea-ice to be
continually drifting. Bound up with the drift are the two related phenomena of
cracks and pressure.

Formation of Cracks. — Cracks were partly the effect of the ice being in motion,
but they were much more often the immediate factor allowing movement. There
was always a tendency (which will be explained later) for an ice-field, whether
of young-ice or of hummocky-pack, to have cracks formed across it. The smaller
areas so formed then shifted their relative positions under the driving influence of the
wind. Then either wide leads and lanes were formed, or, on the other hand, ice-
rafting or even heavy pressure resulted. So it comes about that, although the
formation of cracks gives rise to leads and lanes and therefore makes the pack
navigable, yet these very openings may themselves enable pressure and hummocking
to begin, with consequent danger and difficulty to a ship attempting the pack.

All cracks were a relief from strain, but the ultimate causes of the strain were
very different, and in some cases unknown. In the Weddell Sea it was convenient
to group the stresses as : —

(i) Sudden differences of temperature — contraction cracks.

(ii) Unequal loading — stress or strain cracks.

(iii) Pressure.

The treatment of cracks of the last category will be deferred till pressure comes
to be considered. Elsewhere in the Antarctic and in the Arctic similar types of
crack appear to be the rule.

(i) Owing to the heavy autumnal snowfalls experienced by the Endurance in
1915, the underlying ice was almost everywhere protected from the effects of

806

MR J. M. WORDIE ON THE NATURAL HISTORY OF PACK-ICE

sudden changes of temperature. Though some cracks, therefore, were possibly
temperature cracks, they were never proved as such ; and in the Weddell Sea,
at any rate, they are not considered important.

Elsewhere in the Antarctic they apparently do occur, and in the Arctic, accord-
ing to Weyprecht, their occurrence is extremely common. Since their forma-
tion, however, except at very low temperatures, has been called in question by
Pettersson, a detailed case must be mentioned. Priestley and David describe
how in the Ross Sea a crack formed with a fall of temperature from — 16° C. to
— 29° C., and then opened and closed according as the temperature fell or rose.
With the final rise of temperature to - 18° C. the young ice which had formed
in the open water of the crack was overthrust a distance of 5 feet 6 inches. The
importance of knowing the actual temperature comes in, in view of the remark-
able conclusions arrived at experimentally by Pettersson ; for he illustrates by
means of curves how. instead of being contracted by a fall of temperature, the
volume of sea-ice increases down to —20° C. Between — 4’4° C. and — 6'4° C.,
for instance, the ice he experimented with (comparable to frozen water of the
open ocean) expanded its volume 0'002957. Figures such as these, he says, are
only comparable to the changes of volume effected in a gaseous body. According
to Pettersson, therefore, sudden falls of temperature down to — 20° C. will not
contract but will bend the ice upward and rupture it in that way. Many con-
traction cracks, however, such as that of Priestley and David, are claimed at
higher temperatures than -20° C., and field experience, therefore, is so much
at variance with laboratory experiment that the question must be left an open
one.* In any case, detailed descriptions of other temperature cracks are desir-
able, and Pettersson’s experimental methods certainly want confirming.

(ii) For the second and by far the most important and commonest type of
crack it has been hard to find an entirely suitable name, and, though the term
“ strain crack ” has finally been adopted, it is realised that it is not an exact term,
as all cracks are a relief from strain.

That the ice and its overlying load of snow were seldom in a state of equilibrium
was a commonplace during all observations. The effect, for instance, of the irregular
distribution of the snow-covering was always very noticeable, for there were even
some places where the ice-surface was entirely below water. Consequently it is
next to impossible to determine the specific gravity of sea-ice from the proportion
above and below water-level ; even in places bare of snow there must still be
stresses of some sort. This seems obvious, and yet the specific gravity of sea-ice,
as given in Krummel’s Handbuch der Ozeanographie, is based on just such unsuit-
able measurements as these made by Makaroff.

* This discrepancy was pointed out to Mr Priestley in May 1919, who till then did not know of Pettersson’s
work. Although without access to his notes, he was most emphatic in saying that contraction cracks were possible
with a fall of temperature not reaching so low as — 20° G. ( — 4° F.).

AS OBSERVED IN THE WEDDELL SEA.

807

Weyprecht gives a good instance to show how a patchwork field of pack-ice
is always in a state of strain. He postulates the case of a fioe 2 metres thick
being divided by a crack, where young - ice immediately starts forming. It
will soon be 1 metre thick, and by then the proportion above water should be
0‘2 metre (taking 4:1 as the average proportion of submerged ice) ; but the
thicker, earlier-formed ice bordering the crack has only increased perhaps one-
tenth of this amount, and consequently in its case only a further 0'02 metre
should emerge. So at the junction either the younger ice is held down below
its proper level., or the older ice is buoyed up. In any case, there is a stress of
some sort along the line of contact.

Sea-ice when newly formed is extremely plastic and can conform to such stresses
as these ; but as it ages it becomes much harder and less bendable, so that finally
it must reach that state of strain at which it only requires some slight impulse
to break it. In most cases this impulse will be the wind, though occasions must
occur when the ice cracks simply because it has passed the breaking strain.

Proof of tension in a floe before it cracked was forthcoming in many ways : the
ice-surface on one side of a freshly formed crack might immediately take a different
level from that on the other side; the two sides might be displaced laterally; or,
again, the opposite sides of a crack might open out at once to a breadth of some
inches, to become stationary at that distance for a considerable time.

A special case of strain cracks, and one easily realised, occurs when a swell
from the open sea runs under a wide ice-field. Over the trough of the wave*
a thick ice-floe must be entirely unsupported ; a crack parallel to the wave-front
may result ; when more than one such crack occurs, they will of course be
parallel to each other. A case in point occurred on March 30, 1916. Whether
when more than two in number they will be at equal distances apart is open to
doubt ; should it be so, the interval from crack to crack may possibly be half
the wave-length. During the Endurance drift, never more than a couple of
parallel cracks, originated beyond all doubt by a swell, were seen. This was
probably due to the fact that, by the time the floes drifted into the region of
swells from the open sea, they were already much reduced in size by the
formation of ordinary strain cracks.

There were cases during the winter, however, when series of cracks formed
whose relationship was so nearly parallel as to make one inclined to invoke the
agency of a swell running under the pack from open water. The best of these cases
was a series of four cracks (one of which ultimately developed into a lead) which
opened on March 17, 1915 (fig. 3). Seven miles to the south-east there was open
water on Maxell 11, and beyond that open water as far as could be seen, perhaps
even as far as Coats Land, thirty miles farther on. A swell was therefore well
within the bounds of possibility. The cracks were not parallel, however, but seemed
to form a fan-shaped group. This suggests another explanation, namely torsion.

808

MR J. M. WORDIE ON THE NATURAL HISTORY OF PACK-ICE

If so, they belong to the third class of cracks, those due to heavy pressure. The
detailed treatment of the latter, however, is best deferred until the phenomena of
pressure have been discussed.

As a yet further case of specialised strain cracks, tide-cracks should be at least
mentioned. The drift of the Endurance was never near enough the coast to enable
them to be studied, but from published descriptions tide-cracks appear to be a
perfectly normal result of strain.

Pressure. — When a moving ice-field became cut up by cracks running in
different directions and many of them of considerable width, each floe as it was
formed tended to travel at a different rate from that of the parent ice-field. A stretch
of sea-ice was never a simple formation, but rather a great conglomeration of sheets

va

of young-ice and of floes which had survived the previous summer ; consequently,
each of the new areas detached from it had its own peculiarities of thickness, depth
of snow, extent, and so on. As was shown above, it was these very peculiarities
which produced the strains ; and these in turn caused the cracks. They were also
partly the cause of the different rates at which individual floes, formed • from the
break-up of an ice-field, travelled under the never-ceasing impulse of wind and
current.

The main cause, however, of the different rates of travel was not so much the
size and depth of the floes as the nature of their surface. Every hummock, in fact,
acted as a sail, and the rate of movement therefore depended to a certain extent
on the amount of hummocking in proportion to the area and weight of the floe. A
legacy of previous pressure, these hummocks in turn became the cause of still
further pressures. When two floes were moving at different rates, either the
distance between them increased and a lead or lane passable for ships was formed,
or the interval between floes decreased, one so to speak overtaking the other, and
then the result, if there was sufficient momentum, was pressure in some form or other.

AS OBSERVED IN THE WEDDELL SEA.

809

Very often there was first opening and then closing of the space between two
floes. Take an imaginary case, namely an area which has been almost stationary :
when movement starts the bigger floes will take some time to reach their proper
speed ; but, on the other hand, they will go on travelling after smaller floes have
stopped moving. In the early stages, therefore, the big heavy floe will be attacked
by smaller floes overtaking it ; in the latter stages it itself will be the attacker of
other smaller floes ahead ; by reason of its size and weight it is pretty sure to take
the upper hand whatever happens. Light pressure only may result, but, even so,
it will only still further complicate the floe surface and create possibilities of yet
further differences in speed.

Besides the forward motion imparted by the wind, there was also a swinging or
turning tendency in almost every floe ; this was either an effort to trim them-
selves to the wind, or the result of pressure from another Hoe ; but, as the floes
were continually hindered by each other, and the wind was not necessarily constant
in direction, this swinging habit seemed to have no end. The process was' of course
more easily seen and realised when the pack was loose. For instance, on April 6,
1916, the floe on which the Endurance party was camped swung as much as 180°
in one night ; the ice at the time was described as being fairly open and travelling
fast ; but it was still open-pack, not drift-ice. When the ice was closer, the actual
swinging was not so obvious, but the effects were greater ; and one saw the type
of pack produced which is so dangerous to sail through, and which the
navigator calls “ screwing pack.”

The screwing or shearing habit resulted in pressure being located mainly at the
jutting corners of floes. By one or other of the methods to be described, a hummock
was formed of loose ice-blocks mixed up with a certain amount of snow ; when
movement ceased it settled downwards and became covered with falling and drifting
snow, and in the end it would simply appear as a trifling inequality on the ice-
surface, no bigger than a haycock.

Pressure worked in three ways : —

(а) Bending ;

(б) Tenting ;

(c) Rafting.

These are the old terms given by early whalers and Arctic explorers, and are practi-
cally self-explanatory. Collectively, or when not particularised, the term generally
used is “ hummocking.” Taken as a whole, hummocking resembled fairly closely
the experiments which geologists make to illustrate mountain-building.

(а) Bending was characteristic of thin and very plastic ice. The most impressive
result of this nature was the formation, in March 1915, of an arch 3 feet high and
8 feet span, formed of young-ice about 8 inches thick.

(б) Tenting was, on the other hand, confined to heavy floes, which being thicker

810

MR J. M. WORDIE ON THE NATURAL HISTORY OF PACK-ICE

and less plastic were on that account less bendable; the ice swelled up, a crack
formed perpendicular to the direction of the. pressure, and a tent-like structure
resulted. Other radiating cracks were usually developed ; and, if movement
^continued, the blocks so formed were soon piled up into a pressure ridge.

(c) Rafting was the commonest of all the processes, and followed automatically
when either of the other two methods already mentioned was carried to an extreme.
Moreover, when rafting was going on, there was often a good deal of bending and
tenting also taking place. In very thin ice, for instance, a process of rafting was
developed in conjunction with bending, and gave a somewhat peculiar appearance
to young newly-frozen leads, namely a series of lines at or nearly at right angles
to the edge of the lead. This was due to the young-ice closing up when not more
than 1 to 2 inches thick ; the two sheets dovetailed into one another, like moving the
fingers of one hand over those of the other, in such a way that the little finger of
the right hand is above the little finger of the left, the third finger below, the middle
above, the first below, and so on. The vertical grain of young-ice allowed of the
process going on to a considerable distance, and in this way two sheets of young-ice
might dovetail to a depth of 5 to 10 yards.

This was rafting at its very simplest. A slightly more advanced stage occurred
where the screwing tendency brought jutting corner against corner, and forced one
floe to override the other. Most impressive of all, however, was the formation
of a great pressure ridge. Whether the work of a few hours, or a matter of days, it
was the most forceful and exciting incident of life on the pack. One case is worth
describing in detail : namely the occurrences in July 1915, which lasted for nearly
a fortnight and exhibited in that time all possible types of “ hummocking.”

About 300 yards forward of the ship’s bows a lead had formed in February 1915,
and when frozen over with level young-ice had been found on that account the
most convenient place on which to train the dog teams. By the latter half of
March, however, a number of cracks had already formed across this lead (fig. 3),
and their closing had led to a certain amount of light pressure. It was not, however,
till July 14 and 15, as a result of a strong S.W. blizzard, that the topography
ahead changed to any great extent. On these dates the floes broke up and re-
arranged themselves ; and in one case a small berg moved half a mile from its
former position and swung round through 90 degrees. The result was that the
ice-topography, with the exception of the 300 yards immediately ahead, was com-
pletely changed. It now consisted of islands of old pack-ice set among young-ice
of two to three days’ growth, the latter with a rough, uneven surface due to the
overlapping of small pancakes. Pressure at this time, it should be noted, did not
involve the older floes. For the next few days the ice was very restless, and the
small berg already mentioned moved back very nearly to its former position relative
to the ship, and at the same time swung a further 90 degrees.

The ice immediately surrounding the ship was in its component parts at least

AS OBSERVED IN THE WEDDELL SEA.

811

over a year old ; as floes and brash it had been the cause of the ship’s being beset in
January 1915, and it was cemented at that time into what may be termed an “ice-
conglomerate.” Its age and structure made it less liable to crack than young-ice,
there being no vertical lines of weakness, so that a ship enclosed in such a position
had all the appearance of security. Across this old ice to the frozen lead ahead a
track ran from the ship for a distance of 300 yards, and was marked by cairns of ice-
blocks, or pylons, as they were called.

On the 22nd of J uly heavy pressure was taking place beyond the farthest of these
pylons, and it led to the destruction of a couple of places where the ice had been till
then a subject of weekly investigation. The impact of the pressure caused a crack
the same afternoon in the old-ice round the ship — a “ shock crack,” which ran
parallel to and roughly 40 yards distant from the pylon road. On the 23rd pressure
still went on, and a ridge was formed built up of the youngest ice. The old-ice,
however, and the 3 feet thick young-ice formed since February had not begun to raft,
though some more shock cracks were developed in both. On the 24th the pressure
ridge had advanced another 10 yards over, but not involving, the 3 feet thick ice of
February.

It was not, however, till the next day that the thicker ice, both that formed in
February and that surviving from 1914, became involved. The already mentioned
pressure ridge dating from the 22nd began an irresistible advance over the old-ice,
weighing it down and breaking it up by means of “weight cracks.” Though it was
not really necessary, one almost felt impelled to walk fast or even to run in order
to avoid the advancing pressure. On the 26th there was a respite, and any further
movement took the form of shearing. Matters, in fact, remained quiet until the end
of the month.

On the 1st of August, however, there being a strong S.S.W. blizzard blowing, new
and unexpected developments took place. Not only did a crack form athwartships,
running out at right angles to the ship on either beam, but a shock crack due to
renewal of the July pressure started only 10 yards away on the starboard side.
Working immediately set in along the latter crack, and blocks of ice 4 and 5 feet
thick were soon involved in the pressure. The ship broke out of her ice-berth, and
so rapidly did the shearing develop to starboard that it gave her the appearance
of sailing over the top of the ice-floes. When in the afternoon things became
quieter, the ship, instead of being safely bedded in an old conglomerate floe and
300 yards from the February lead ahead, found herself right in the middle of broken-
up and unstable ice, where trouble might always occur.

This, however, was the end of the disturbances in the immediate neighbourhood,
and the ship was once more frozen in, this time among a maze of hummocks and
pressure ridges.

The tendency in the next two months was for the pack to become looser, and
unfrozen pools and lanes were not infrequent in October. On October 15, by

TRANS. ROY SOC. EDIN., VOL. LII, PART IV (NO. 31). 126

812

MR J. M. WORDIE ON THE NATURAL HISTORY OF PACK-ICE

means of one of these lanes, the ship moved to a new position 400 yards away, and
was secured in a narrow lead not much more than the width of her beam. All
would have been well had she not been warped round into a transverse crack four
days later, for this move left her at the meeting-point of three floes. When pressure
began again, instead of rising with it, she was twisted by the working of the three
floes and developed a bad leak at the stern-post on October 24. No longer
buoyant, she was now unable to weather such pressures as formerly would have
given little trouble, and finally on October 27 she was abandoned, the bottom torn
out of her and the water flush with the upper deck. She remained thus for a month,
and it was not till November 21 that' she finally sank under water.

A word may be said about height of pressure ridges , since there has always been
a tendency to exaggerate this feature. During these fifteen months in the Weddell
Sea it was quite exceptional for any ridges to be as much as 15 to 20 feet high ; those
of July and August 1915 were only 12 to 14 feet in height. On one occasion about
three miles from the ship a hummock was found 25 to 30 feet high, built up entirely
of blocks of sea-ice ; it was not, however, part of a pressure ridge, but was simply an
isolated hummock due probably to the encounter of two jutting corners. Again in
March 1916, a thick slab of sea-ice was tilted up so that the upper edge was 25 feet
above water-level ; in this case at any rate it certainly was a jutting floe corner
which had had to give way during screwing movement.

Accounts from the North have given the impression of extremely high pressure
ridges in the Arctic seas. All these accounts, where reliable, are found to be based
on phenomena seen north of Baffin Bay and Kobeson Channel, where the land
offers considerable obstruction to the drifting ice, and possibly causes higher ridges
than on the open sea. Nansen, on the contrary, definitely states that on the drift
of the Fram across the North Polar Basin itself the highest hummock was 30 feet,
and comparatively few exceeded 20 feet.

Should pressure ice, therefore, be met with in the South whose average is above
20 feet or even above 15 feet, it must point to the existence of land obstructing the
free drift of the ice. A case in point was the belt of heavy hummocky -ice found by
the Endurance on her southward voyage stretching from 70° S. 15° W. to 71° S.
22° W. ; the floes were so close and heavy that it took six days to find a passage
through. This very heavy ice was regarded as pointing to a deep embayment some-
where east of Coats Land, or at least to a very large ice-tongue protruding from the
coast and so checking the westerly drift.

Pressure ridges were highest and most prominent when newly formed. They
almost immediately sank to a position of equilibrium, as required by the specific
gravity of the ice, the ratio of ice below to that above water being either 4 to 1 or
5 to 1. Possibly, as Weyprecht suggests, the supporting floes and pieces, which at
least in the upper two-thirds were dry before the hummocking, when driven below
became waterlogged and so lost some of their buoyancy. Whatever the cause, the

AS OBSERVED IN THE WEDDELL SEA.

813

ridges certainly were less prominent in a few days, and soon lost their rawness and
angularity by getting drifted over with snow.

Formation of Cracks by Pressure. — Two kinds were definitely established : —

(a) Weight or hinge cracks.

( b ) Shock or concussion cracks.

Possibly there was a third, namely

(c) Torsion cracks.

(a) Ice of advanced age was no longer plastic and did not bend, therefore, under
the weight of a heavy pressure ridge piled on top of it, as it might once have done.
The result, if the breaking strain was reached, was a longitudinal crack in front of
the pressure, the ice being pressed down in much the same way that a door opens on
its hinge (fig. 4). At the same time a number of radial cracks were developed ; and
the ice in front of the pressure was thus broken up beforehand into blocks of suit-
able size for further hummocking. Between a pressure ridge and a weight crack of

Hinqe Crack

/_

. ~ Water Level.

Fig. 4.

this nature the surface of the ice was often below the level of the sea ; the pool
formed, however, was apparently concentrated brine rather than sea-water.

(b) The shock or concussion crack, -like the smaller radial cracks mentioned above,
was also transverse to the advancing pressure ridge. It is sufficient to state that
cracks of this nature were seen to form on many occasions (p. 811), and that the
proximate cause was the impinging of one moving floe on to another relatively inert.
Probably the development of the crack was made easier by the passive floe being
already in a state of tension.

(c) All the above types of crack, whether due to outside causes {i.e. pressure) or
to causes in the floe itself ( e.g . stresses), were actually observed and are beyond all
question. Torsion cracks are, however, more hypothetical. They have been discussed
by Arctowski and Hobbs in an attempt to explain leads taking the form in ground-
plan of a chain of pools. Citing Datjbree’s well-known experiment on the production
of two sets of cracks at or nearly at right angles, Arctowski makes the claim that
the same thing may occur in an ice-mass relatively passive. On opening, a zigzag
lead would result (fig. 5) ; and as it opens farther there would in the ideal case be a
chain of diamond-shaped areas ; but, as the effect in nature can hardly be as perfect
as this, it resolves itself (so it is claimed) into a chain of pools. On one occasion a
sketch was made of some opening cracks in the neighbourhood of the Endurance
which seemed to be a case in point (fig. 6). The whole matter, however, requires

814

MR J. M. WORDIE ON THE NATURAL HISTORY OF PACK ICE

further observation, for other explanations are possible. Chains of pools, in fact, can
be very easily accounted for by shearing and screwing.

Effect of Wind on the Drift. — During 1915 a drift indicator was set up on the
floe about 50 yards from the ship ; it was in no sense a current meter, but was an

approximation to a drift meter, as it was ultimately possible to deduce the speed of
the drift from the rapidity or otherwise with which the indicator reacted. It con-
sisted of iron piping about 9 feet in length set inside a slightly shorter length of ship’s
rail. The latter was firmly fixed in the floe and was filled with a mixture of salt and'

petroleum, so that there was very little chance
of the inner piping becoming fast frozen. A
vane had been fixed to the lower end of the
piping, and an arrow-head at the upper end, in
the same vertical plane as the vane. The arrow
accordingly always pointed in the direction in
which the floe was travelling.* Commander
Worsley made several observations daily, aqd
at each observation rotated the arrow and the
vane to 90°; he then waited until they swung
back, and the time which this took gave him
a measure of the rate of drift. The result of these observations was to shoAV that
the floes always travelled with the wind, but slightly to the left. The velocity,
however, did not depend entirely on the strength of the wind ; it was influenced
in great measure by the presence or absence of open water in the direction to which
the ice was being taken, no matter how far off this open water might be.

M'Clintock and later Nansen have discussed the right-hand deflection in the
Arctic. One case may be quoted to show the extent of the left- hand component in

* Of. E. H. Shackleton, Heart cf the Antarctic, vol. ii, p. 413.

AS OBSERVED IN THE WEDDELL SEA.

815

the Antarctic (fig. 7). From January 16 till January 22, 1916, a fierce blizzard blew
for six days from S. W. by. S. Despite this the floe on which the camp was situated
drifted N. by E., i.e. two points more to the left than the wind. This was no doubt
an extreme case ; but it is worth noting that the trend of the Graham Land coast
immediately to the west did not make things any easier for ice to move to the left.

The track of the ice-drift was influenced by three factors in all : — (a) the wind ;
(b) a left-hand component due to the earth’s rotation ; and (c) a true current. The
unsolved problem is the amount and direction of the last factor in the Weddell Sea.
Nansen’s method might, however, be used, namely : — To lay off on squared paper
lines representing the wind’s direction and force at
each observation (i.e. every four hours), the length
of each line being proportional to the strength of
the wind. When this is done an irregular track
should result (not unlike the drift of the ice) which
may cross itself at certain points. At these points,
accordingly, the wind resultant for the period in
question should be nil ; the left-hand component
should also be neutralised ; and the distance,
therefore, which the ship has travelled in the
interval, as determined by observations, should
be entirely due to the true current. The third
factor should therefore be kept in mind when
examining the drift chart of the Endurance. At
first sight it looks as if wind were the only agent
responsible, but it should not be forgotten that a
true current, not determined as yet and perhaps
only of limited extent, is quite within the bounds of possibility.

So far as drift control by wind went, it was possibly the result simply of
’long-shore winds. When one studies in this connection the Deutschland and
Endurance tracks wherever they are near known coasts, it will be seen that the
ships’ tracks and the coast adjoining are approximately parallel. The abrupt change
in the general direction of the two courses between 72° and 74° S. lat. may mean,
therefore, that the coast of the unknown land to the south-west also shows a similar
change in direction. It also makes it improbable that there is any strait separating
Graham Land from the rest of Antarctica.

Westerly Antarctic Drift. — The most striking feature of the Deutschland drift,
apart from the kink mentioned above, is the equally abrupt easterly deflection in
65° S., where the ice looks, indeed, as if it had come under the influence of the
westerly winds. This would be an unusually high latitude in which to meet the
westerlies, and it is probable that there is some other reason for the behaviour here ;
for after a time the general direction of the! Deutschland drift again becomes

816

MR J. M. WORDIE ON THE NATURAL HISTORY OE PACK-ICE

northerly. Taken as a whole, the winds met with off the Antarctic coasts are south-
easterly, with the result that the ice-pack moves towards the west. It has taken
years for this fact to be properly appreciated ; early circumnavigators, such as
Biscoe for instance, believed that the westerlies still prevailed far south, and sailed
eastwards, not knowing how much they would have gained by adopting the exactly
opposite course in high latitudes. The westerly drift round the Antarctic is well
known to-day, and it is obvious that the course of future explorations will be very
markedly influenced by this knowledge. It means, for instance, that the eastern
borders of such seas as the Weddell Sea are free from ice, at least towards the end
of the season, while on the west side the reverse is more likely. This makes it
improbable, therefore, that the east coast of Graham Land will ever be charted from
the sea, for it is unlikely that any ship will venture down the west side of the
Weddell Sea, or having crept down Coats Land will risk penetrating farther north-
west along the Wilhelm Barrier, with the ever-threatening possibility of being
jammed against the land by pack from the east.

Graham Land obstructs the westerly drift ; that and the low average wind-
velocity in the Weddell Sea compared with the Ross Sea are the causes of the
unusual congestion of ice which generally prevails in the former area. The
Endurance tried to make its southing between 15° and 20° W. long., but from
60° onwards there was a continual struggle with the pack. There must, however,
be an eastern limit to this ice congestion ; possibly a ship sailing south along the
Greenwich meridian may meet with no ice there even in January. One cannot say
probably, for there is also the likelihood of a considerable indentation (and nourish-
ing ground, therefore, for ice) into the continent between 15° E. and 15° W. long.
The reason for supposing the existence of such a sea has already been referred to,
namely the finding by the Endurance in 70° to 71° S. lat., 15° to 20° W. long., of
a line of impenetrable hummocky pack heavier than anything encountered elsewhere,
even off Graham Land. One felt compelled to think that it could only be due to
the westerly drift piling up the ice either against a great ice-tongue (e.g. Termina-
tion Ice-Tongue in 100° E. long.) or along a N.— S. coast obstructing the normal drift.
Bellingshausen’s discovery of land in 69° S. lat., 16° E. long., marks the eastern
limit of this possible indentation. A ship working westwards from Enderby Land
would be taking the most suitable course for entering this presumed sea.

Position of the Pack-Ice Edge. — The course of future exploration is bound to be
influenced to a great extent by the results of the voyage of the Endurance, for
much that was formerly mere supposition in regard to the ice is now accomplished
fact. It would also be most useful, therefore, if the average edge of the pack in the
Weddell Sea could be determined from the records of all who have previously visited
this region, and a chart drawn like that published annually for the Arctic by the
Danish Meteorological Institute. An attempt was made to do so, but achieved only
moderate success. The time does not seem suitable for it yet. In the first place,

AS OBSERVED IN THE WEDDELL SEA.

817

there are hardly enough observations available, and Such as there are, are for the.
most part valueless, owing to the observer seldom specifying in detail what type
of ice he saw. For the present not much can be done to amplify the various pack-
edges set down on the chart accompanying Mill’s Siege of the South Pole beyond
adding the Filchner and Shackleton conditions. Such a map to be useful
should, of course, have the day and month as well as the year inserted, as the edge
is continually moving northwards. The only data as to the rate of this movement
are supplied by the Deutschland, which drifted 6 sea-miles a day, and the Endurance,
nearer the coast, which averaged 4 sea-miles a day.

IY. — Changes in the Ice.

From its very beginning the history of the pack-ice was one of continual,
unceasing change. The field itself was never at rest. Driven by wind and current,
it was cracked the one day, and compressed the next. Outwardly, therefore, it
finally appeared very different from the original sheet of young-ice, which consisted
either of black-ice covered with ice-flowers or of cemented pancakes. Its surface
became a maze of hummocks and pressure ridges, for the most part drifted over with
snow. These were the outward signs of change due to its being in motion ; but
inwardly the ice had also become very different, independently of whether it moved
or not. It had altered in chemical constitution, in density, structure, hardness,
elasticity, and so on. These changes, it should be noted, varied greatly according as
the portion of ice examined was still floating in the water, or had been hummocked
so as to lie above sea-level.

Growth in Thickness. — The rate at which undisturbed ice grew depended largely
on whether there was a heavy snow-covering or not. Hence it could come about,
and did, that the earliest-formed ice in the winter of 1915 was not necessarily the
thickest at the close of the year. Observations on the thickness were made at several
different points, but ultimately these became destroyed one by one, so that by October
there was only one small area of 1915 ice whose entire history was definitely known.
The final measurements for thickness were made here, and gave the result that
by October 13 ice which started forming on February 6 was now 145 cm.
( = 4 ft. 9 in.) thick. This ice started forming in and about 77° S. lat. ; and on
October 13 the position was 69° 20' S. lat. Owing to the abandonment of the
ship a fortnight later, this was the last measurement possible at this locality. It is
more than probable, however, that the ice continued to increase in thickness until
possibly the end of December, but certainly no later than this date. On plotting
out the different measurements and continuing the diagrammatic curve so got in
keeping with the composite curve of past observers, an additional 3 inches can be
assumed. The maximum thickness of one-year ice in the Weddell Sea (except fast-
ice in high latitudes) is therefore taken as 5 feet. Anything thicker than that is
regarded with certainty either as having been rafted at some time or other, or as

818

MR J. M. WORDIE ON THE NATURAL HISTORY OF PACK ICE

being the result of two winters’ freezing. Both these winters may have been spent
as drifting pack-ice, or for the first of them the ice may have been held fast in some
sheltered bay.

Summary Table of Rate of Growth.

Locality.

Date.

Thickness.

1915.

inches.

Centimetres.

D

14th February

0

0

16th February

6

15-25

3rd March

12

30-5

13 th March

15

38-1

6th April

21

53-4

1st June

375

95-25

6th July

40-35

102-5

F

6th February
9th August

0

Small amount

7th September

1 36 | °fsn0W'

13 th October

57

1451 C0VeriDK

L

5th October

1301 Same floe as

10th October

131 1 F, but deeper

21st October

14 1 J snow-covering.

Certain floes, however, whose thickness was due entirely to freezing, exceeded
5 feet. Some of these were measured while still afloat, others immediately after being
hummocked. A selection is given below of those of most interest, and from some of
which important conclusions were drawn.

Table of exceptionally Thick Floes due to Freezing.

Date of Observation.

Thickness.

Remarks.

8th August 1915

145 cm.

Pressure ridge beside ship. The ice in which
the ship was beset in January-February
1915. (80 cm. formed in 1914; 65 cm.

formed in 1915, as proved by diatom layer.)

5th and 12th September
1915

7i ft.

Part of a hummock. Diatoms were present
only in the lowest part ; presumed, therefore,
to have formed almost entirely in 1914, and
to have grown a few inches only in 1915.

8th December 1915

9-10 ft.

A slab in a hummock, not very far distant
from the last locality.

31st December 1915

7 ft.

“ Ocean Camp ” floe.

7 th February 1916

6 ft.

“ Patience Camp ” floe.

19 th March 1916

6 ft. (approx.)

Portion of hummock nearly 25 ft. in height.
Upper part made up of 2-3 ft. of “ spotted
ice” with diatoms at base, lower of 1-3 ft.
of “ fibrous ice ” with current summer’s
diatoms at base.

30th March 1916

8 ft. 4 in.

“ Patience Camp ” floe. Less snow presumably
than where measured on 7th February.
Vertical lines pronounced from the very top
downwards, i.e. all fibrous ice.

AS OBSERVED IN' THE WEDDELL SEA.

819

Diatoms. — In the summer countless diatoms made their appearance on the
surface of the water, on the edges of the ice, and along its under surface ; they
inserted themselves between the plates and lived actually in the body of the ice
itself. This may have been simply an effort to get to the light. At the end of
summer, accordingly, a good number were entangled and became frozen into the ice,
and so gave it a yellow colour. This discoloured band, therefore, marked the
interval between two winter periods of freezing, and could be used as an index to the
age of the floe. The piece of ice, for instance, noted on March 19 had two diatom
layers ; the second and lower was formed in summer 1915-16 ; the upper presumably
in 1914-15 ; accordingly the uppermost ( i.e . oldest) portion of the ice must have
commenced freezing in winter 1914 at latest.

Change in Structure. — It had been well known to whalers for a long period that
old hummocked ice could be used for drinking purposes ; but this was knowledge
which had apparently been lost, for the Discovery in 1901 melted snow for this
purpose. It is an important point, however, in a ship’s economy to realise the
freshness of hummocked ice, as snow requires nearly double the amount of heat that
ice does to produce a given quantity of water. The ice which was suitable for this
purpose was very different in appearance from freshly formed young-ice with fibrous
structure.* It had become clear, almost transparent, had a conchoidal fracture,
might almost be called blue in colour, and had a slightly spotted look. Though
quite suitable for drinking, it was not entirely fresh ; tested with silver nitrate, the
solution always became cloudy, showing that some chlorine at least was still present.
All the water used for chemical experiments on board the Endurance had to be
distilled on this account.

This type of ice was apparently that examined by Drygalski near the 'Gauss
winter quarters in 1903, and by him called “ Blaueis.” He attributes the changes to
the influence of wind and evaporation (i.e. ablation), basing his contention on the
smoothed and rounded nature of the ice. That the latter appearance is due to
ablation is certainly true, but that the change in structure is due to the same cause
is unproved. The long imprisonment of the Endurance gave the opportunity for
showing that the structural change was due to the high summer temperatures ; it
might partly take place in ice still afloat (i.e. not hummocked), but it reached its
fullest development when the ice had been subjected to a summer’s temperatures as
a hummock above the water level.

About half the ice in the Weddell Sea was of this “ spotted” nature (Drygalski’s
“ Blaueis ”). That the change was not due to wind follows from the fact that ablation,
however important it may have been round the Gauss winter quarters, is nowhere
known to be at all effective in the Weddell Sea. That it may play a part nearer
the coast than the position of the Endurance is quite possible, but, so far as the

* Called “ fibrous ” for convenience ; young-ice is not really fibrous but platy in structure ; the edges of tbe plates,
however, give it a fibrous appearance.

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 31).

127

820

MR J. M. WORDIE ON THE NATURAL HISTORY OF PACK-ICE

measurements of ablation went, the results round the ship were nil. These
measurements took the form of a marked stake embedded in the ice of some
prominent hummock. One such was set up on May 21, 1915, and when examined
for the last time on September 4, as on previous visits, it was found still crusted,
like the ice around, by rime 1 inch thick. This was the rule everywhere through-
out the winter.

In the summer, conditions were not quite the same, and changes did occur, but
they were directly the outcome of the sun’s heat. This was the very season during
which quantitative measurements would have been of most value ; but such were
of course quite impossible after the ship was crushed in October 1915. The

summer changes are therefore the result
of comparative observation, and quanti-
tative tests are still required.

That the change from fibrous ice ( i.e .
striated ice showing vertical lines) to
spotted ice can take place in hummocks is
not new ; the details of what takes place
have been set down very clearly by Ham-
berg. They amount to this : that the
ice, when subjected to a temperature at
and about the melting-point, loses some
of the liquid salt inclusions, their place
being taken apparently by air-bubbles ;
the ice then ceases to show its pro-
nouncedly striated or fibrous appearance
and becomes bubbly and almost granular.
Traces, however, of the vertical lines can still be seen on very close examination
and are a convincing proof of what has taken place. What had not been settled
by Hamberg was the length of time necessary to effect this change. The length
of life of the Weddell Sea floes being known, it can be definitely asserted now
that the change can be effected in one summer.

One can also go further than this and say that the change may also commence
while the floe is still in the water. In this connection the thick floes examined
on August 8, 1915, and March 19, 1916, were of considerable value for what
they told.

That examined on August 8, 1915, which is roughly sketched in fig. 8, was
raised up into a pressure hummock on August 1 ; previously it formed part
of an old floe which had lain ahead of the ship since very early in February 1915,
and which was formed at that time by the cementing together of brash and ice
pieces (dating from 1914) during the north-east blizzards of the latter part of
January 1915. The upper portion, just over 80 cm. in thickness, was spotted

AS OBSERVED IN THE WEDDELL SEA.

821

ice, and of this amount the bottom half was discoloured by diatoms. It was
separated by a sharp line from 65 cm. of fibrous ice. From this specimen the
following conclusions were drawn : —

(1) The sharp line indicated a break in the freezing process during the summer

1914-15, and separated ice formed in 1915 from that formed the previous
year.

(2) The presence of diatoms showed that the ice was afloat in summer 1914-15.

(3) It also showed that the ice must have been very loose to allow of these

minute plants penetrating 40 cm. into the ice.

(4) This in turn pointed to a temperature at or about the freezing point —

just what was required to produce the structural change from fibrous
to spotted.

The block examined on March 19, 1916, had been tilted up into a hummock
a couple of days earlier. Its upper 2-3 feet consisted of spotted ice, in its lower
portion discoloured by diatoms. The variation in thickness may have been partly
due to summer melting in 1914-15 ; it resulted, at any rate, in an uneven boundary
between the upper portion and the lower 1-3 feet of ice ; the latter was also
slightly spotted, but with the fibrous structure not yet entirely obliterated ; in turn
its lower layers also were strongly discoloured by diatoms. This was an important
find, for it showed the effects by March of the summer heat of 1915-16 ; the change
had already begun, but was not so far advanced as to obliterate entirely the fibrous
structure. The arrival of diatoms seems to be the sign that the change is
beginning.

The earliest date on which the structural change was definitely known to
have commenced was on December 30, 1915. On that date a visit was made
to one of the numerous areas of brownish-looking flat ice (the brown appearance
was connected with a very thin snow-covering) and an attempt made to bore
through the floe. Merely by working with the point of an ice-axe, it was soon
possible, despite the water oozing up in the first few inches, to reach down 2 feet
6 inches. At one time it would have been impossible to have reached even a depth
of 6 inches without excavating with a pick and shovel. The ice now was soft
right through and through ; and the whole ice-thickness must, therefore, have
been practically at the melting-point.

On the same date, however, it should be noted that a floe with a 3-foot snow-
covering was not so far advanced ; for in this case it was only possible to work down
a few inches with the point of the ice-axe. The snow-covering was now retarding
rise of temperature, just as once it retarded freezing ; this was evidence, therefore,
of how the almost universal thick snow-covering on Weddell Sea floes worked
towards prolonging their life.

Physical Changes.— The fracture and hardness of spotted ice are very different

822

MR J. M. WORDIE ON THE NATURAL HISTORY OF PACK-ICE

from the same physical properties in fibrous ice. This was to be expected, and does
not call for detailed mention.

Experiments were made from time to time to determine the specific gravity of
the various types of sea-ice, and to discover especially if there was any relation
between specific gravity and the amount of chlorine present. On the whole the
results were not as satisfactory as they might have been, and, under the circum-
stances, were of course incomplete. There is good ground, however, for supposing
that something may eventually be gained from them. The method used was that
recommended by Mr James, who himself made a good many of the determinations.
The sample of ice was first of all weighed in air, in a temperature which varied from
— 4° to — 12° C. (and which of course was not quite the same as that to which the
ice had been exposed in situ) ; it was then weighed in paraffin whose specific gravity
and coefficient of expansion were known ; and the result calculated accordingly.

In a previous paper it was stated that the specific gravities of fibrous ice ( i.e . ice
showing vertical striae) clustered round 0'92, whilst those of comparatively pure
spotted ice were nearer 0'91. This was a somewhat hasty conclusion to draw from
the results. At the time Dr Otto Pettersson’s paper “ On the Properties of Water
and Ice ” had not come to notice. The very numerous references to contraction
cracks in Polar literature made it appear that sea-ice below its freezing-point behaved
like pure ice, and contracted on cooling — the exact opposite to Pettersson’s con-
clusions. Inasmuch, therefore, as the determinations of specific gravity were made
with the ice varying in temperature from —4° to —12° C., the possible expansion
according to Pettersson would be nearly 2 per cent., and all figures have, therefore,
had to be re-examined. What has been done is to plot the specific gravity against
the temperature of the ice at the time it was weighed. The result for fibrous ice
shows a grouping of the values along lines which are perhaps comparable to
Pettersson’s curve. In the case of spotted ice, however, there are many serious
discrepancies ; but as this type contained considerably less salt than fibrous ice,
it is quite natural to expect that its coefficient of expansion approximates rather to
that of pure ice than to that of young-ice.

Leaving aside spotted ice, and considering only the platy or fibrous type, our
measurements showed that its specific gravity was 0‘92 at — 4° C. and 0'915 at — 15° C.
The amount of chlorine present was about 3 per cent. These figures indicate the
same order of things as Pettersson puts forward, but of a very different magnitude.
Between — 4'4° and — 6'4° C. he found an expansion of 0'002957 (i.e. more than three
times the Endurance result) for ice of a kind which seems to be the same as the
ice spoken of here as fibrous. The method used in the Endurance , however, was
not sufficiently exact to merit more weight being attached to the results got by it
than to Dr Pettersson’s made in the laboratory. Even if they could be, any dis-
crepancy could be easily explained by the fact that practically no two pieces of ice
are identical.

AS OBSERVED IN THE WEDDELL SEA.

823

Spotted ice .has probably a lower specific gravity than fibrous ice, but the
Endurance figures at any rate do not prove an invariable difference. They do
prove, however, that Krummel’s statement that sea-ice has a specific gravity of 0'92
in the Arctic and 0‘95 in the Antarctic is quite untrue.

Chemical Changes. — The amount of chlorine present in the various ice-types was
directly determined by titration with silver nitrate, Mr James being responsible for
the actual measurement. Results are given not as total salinities but in grammes
Cl per 1000 c.c. of melted ice. This is made necessary by the fact that the amount
of salt present relative to chlorine cannot be deduced by formula, as it can in the
case of sea-water, but must vary, to what extent being at present unknown. To
have determined the total salinity was quite out of the question on a non-oceano-
graphical expedition.

The results showed that : —

(1) The amount of chlorine initially enclosed (presumably in the form of brine)

depended probably on the rate of growth — in other words, on the

temperature at the time of formation.

(2) Ice in the position of growth ( i.e . unhummocked) slowly became fresher by

losing its chlorine.

(3) The chlorine was removed downwards and was not, as previously imagined,

mainly pressed out on the surface. Just how this took place was not

quite clear.

(4) The loss of chlorine in winter was probably slow compared with that which

took place in summer.

(5) For ice to become drinkable, it required to be hummocked and exposed to a

summer’s high temperatures.

The most important data were obtained from three vertical pits put down in ice
whose history was known and had been recorded. Those of September 7 and
October 13, 1915, were put down in ice which commenced freezing and had gone on
doing so uninterruptedly since February 6 ; samples of ice were taken every 20 cm.,
and examined for structure, salinity, and specific gravity. Those on September 7
showed invariably fibrous structure ; those on October 13 (in ice of exactly the same
history, the pit being only a couple of yards away), showed fibrous structure only in
the lower two-thirds, for in the upper third there was also a tendency towards being
spotted, which became very pronounced in the topmost sample. These series have
been plotted diagrammatically, chlorine content increasing towards the right, and
depth in the water downwards from the top of the page (fig. 9).

The third series, that of September 14, was collected from ice of a totally
different history ; the upper layers, formed in the winter of 1914, were spotted in
structure and discoloured by diatoms ; the middle portion was spotted, but vertical
lines appeared on a sample being melted ; and the bottom layers were pronouncedly

5 ■

10.

15

20

25

30

35

40

45-

50-

55-

60-

65-

70

75

60

85-

90-

95-

■ 100-

105-

110-

1 15

120-

125

130

135

140-

145-

150

Depth of Ice -Centimetres

MR J. M. WORDIE ON THE NATURAL HISTORY OF PACK-ICE

AS OBSERVED IN THE WEDDELL SEA.

825

fibrous and were in fact simply young-ice. The spotted ice formed in 1914 or earlier
is often referred to as old-ice ; unlike young-ice, it had experienced high summer
temperatures.

The two series in young-ice show how even in five weeks there has been a distinct
freshening of the ice ; in the case of the salter layer (25 cm. down) the loss in
chlorine amounts to 25 per cent. Layers less salt have not freshened in the same
proportion. The curve has become smoother, and it looks on that account as if
finally the amount of chlorine were going to be equally distributed through the ice,
except perhaps in the very bottom layers where the ice is but newly formed.

The series in old-ice shows a freshening of the same order, but much more
advanced. The very high chlorine percentage in the sample near the surface seems
out of place : it is explained, however, by the upper surface of the ice at this point
having been below water level, owing to its heavy load of snow ; for in such cases
salt water always seeped in laterally from the nearest crack and caused the upper
layers of old spotted ice like this to appear even salter than the young fibrous
variety. The fact remains, however, that spotted ice as a rule is fresher than
fibrous ice.

To summarise : — Young-ice had two possible lines of development open to it :
either it remained floating in water or was hummocked. In both cases its structure
became spotted instead of fibrous ; in the case of hummocked ice, it also became
translucent in appearance. The saltness was appreciably lessened in ice afloat, but,
if hummocked, the ice became fresh enough even for drinking purposes. That
a little chlorine, however, was still present even then was proved by titration with
silver nitrate.

V. — Decay.

To the rule that Arctic and Antarctic pack are much the same there appears
to be one noteworthy exception. In the matter of decay, sea-ice in the Antarctic,
by all accounts, behaves very differently from that in the Arctic. It was a very
rare thing, for instance, to see bare ice in the Weddell Sea, and when it did occur
it was owing to the ice having formed late in spring and never having had an
opportunity of getting covered by much snow. In no case was a pool of fresh
water due to natural causes seen on the ice ; salt pools certainly were seen, but
they were all due to seepage of water into depressions caused by the weight of
pressure ridges. Fresh-water pools, formed on camp sites owing to the amount
of soot lying around on the snow, may be set aside as unnatural. Under excep-
tional conditions, however, fresh pools do occur in the Antarctic ; Nordenskjold
mentions having found them on the ice near Snow Hill Island, but gives no
details ; and they also occur in the Ross Sea in places where there is a good
deal of dust scattered about over the ice.

In the Arctic affairs are very different, and melting of the snow to form
fresh-water pools and lakes is extremely common, particularly in the American

826

MR J. M. WORDIE ON THE NATURAL HISTORY OF PACK-ICE

Arctic Archipelago. Nansen and Weyprecht both mention them as occurring
in the neighbourhood of Franz Josef Land ; the latter cites melting of the ice on
the upper surface as the rule, and even maintains that this process may go on
while the ice is still increasing by freezing on the lower surface. The evidence
for this, however, was based on observations of the ice at the ship’s stern, in the
very place where abnormal conditions would be expected ; practically no weight,
therefore, can be given to Weyprecht’s statement that ice in northern latitudes
melts about 1 metre on the surface each year.

Experience in the Weddell Sea makes it certain that there, at any rate, surface
melting as a factor did not count. It must be regarded as abnormal. Except in
summer there was no ablation, and even then it only affected the snow covering
and not the ice beneath. The snow became granular and almost moist.

Other processes must be invoked to explain the decay, namely melting on the
under surface and mechanical attrition. Probably there is little or no ice melted
from below till the whole thickness reaches the melting-point. By all accounts,
however, this is what happens in the Arctic ; the diagrams obtained by
plotting time against ice -thickness, based on the figures of the different expedi-
tions, show at first a steep gradient until the thickness is about 8 inches ; then
a steady falling off in steepness, down to what appears to be the local average
thickness ; and then a line almost horizontal, until it suddenly steepens at the
melting period. The steepness of the diagram at the melting period is ever so
much greater than that at the initial freezing, indicating that the ice as a
whole first reaches the melting-point and then melts practically all at once. On
one occasion there was proof that this might happen in the. Weddell Sea — on
December 30, 1915, when it was possible to drive an ice-axe up to its head into the
soft ice; but no large floe was ever actually seen to melt. Melting of this nature
may be looked for, however, in sheltered bays, and is known to occur in the
Ross Sea; but in the Antarctic ice-fringe as a whole the pack seems hardly
ever to reach the melting stage until it has come under the influence of swells
from the open sea.

In distinguishing drift-ice from pack-ice, it was pointed out that the former is
opener and looser and moves faster accordingly ; it lacks the inertia belonging to
close-ice. This has a very important consequence, for it means that the outer edge
is continually scaling off, so to speak, and the swell getting access to more extensive
areas of ice than it would otherwise have operated on. The swell soon reduces the
size of the floes, and this in turn helps on the formation of more drift-ice ; and so
on. Meantime, the movement of the floe up and down in the water is mechanically
eroding the ice, apart from actual melting ; melting alone would produce honey-
combing, but this was a rare feature in the drift-ice fringing the pack ; it looked,
therefore, as if mechanical wear and tear was the most successful factor in producing
decay.

AS OBSERVED IN THE WEDDELL SEA.

827

In the case of the Weddell Sea ice, there was a local peculiarity, which must
accelerate the break-up. In its progress north, the ice was being continually pressed
against Graham Land and Joinville Island ; S.E. winds accordingly had little effect
in moving the ice, but S.W. ones drove it out towards the open sea ; the former
had a closing, the latter an opening effect. There was a point, however, where
the ice did get away to the north-west, namely when it reached the latitude of Join-
ville Island. In the middle of March, 1916, in 64° S. lat., a south-easterly blizzard had
practically no effect ; in the beginning of April, in lat. 62° 30' S., a lighter wind from
the same quarter drove the ice fast and far to the north-west. In this way separate
areas are probably sliced off the main pack and driven into Bransfield Strait. There
the noted currents of the Strait give it little peace, and it travels towards the South
Orkneys often as fast as 20 miles a day. The Endurance party’s experience of the
current gave some idea of its force. From the 9th to the 12th of April the winds
were easterly and east-north-easterly ; and in this period during the daylight the
boats were steered on a N.W. course ; at night camp was generally made on a
drifting floe. Observations on the 9th made the position 61° 56' S., 54° 05' W. ; but
on the 12th showed the boats to be in 62° 15' S., 53° 17' W. — a net loss of about
30 miles S.E. Dead reckoning, on the contrary, would have made the position at
least 25 miles N.W. of that on the 9th, to say nothing of steady winds from E.N.E.,
which should, have made it still farther west. These figures give an idea, though an
inexact one, of the strength of the easterly running currents in Bransfield Strait.

To summarise what probably happens during the decay of the Weddell Sea ice : —
The most important factor is the swell, which breaks down the floes, if not already
in that condition, into smaller-sized pieces. Mechanical erosion then takes place by
washing of sea-water against the ice. Finally, when the temperature allows of it, the
ice-blocks are destroyed by melting. By the time that stage is reached, the ice is
much comminuted, forming brash. The latter includes “ bergy bits ” (hummocked-
ice finally taking this appearance) or “ growlers,” which are merely ice-fragments
greenish in colour and so small that they hardly show above water. For several
hours one morning the boats of the Endurance passed through a maze of bergy bits.
At that time the ice was under the influence of the Bransfield Strait current ; and
this and the wind were carrying it, much battered and decayed, to its final melting-
ground towards and beyond the South Orkneys.

The author wishes to take this opportunity of acknowledging his indebtedness
to the Trustees of the Scott Polar Research Fund for the generous grants made by
them to meet the expenses of printing this and a previous paper on “ Depths and
Deposits of the Weddell Sea.”

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 31).

128

828

MR J. M. WORDIE ON THE NATURAL HISTORY OF PACK-ICE

VI.— BIBLIOGRAPHY.

Amundsen, Roald, in F. A. Cook, Through the First Antarctic Night, London, 1900 : App. v, “The Navigation
of the Antarctic Ice-pack,” pp. 448-452.

Anon., Annates de la Oficina Meteorologica Argentina, tomo xvii, “Clima de las Islas Orcadas del Sud,”
segunda parte, 1913, pp. 163-169.

• Nature, xi, March 11, 1875, p. 366: “Scientific Report of the Austro-Hungarian North Polar

Expedition of 1872-74.”

Arctowski, H., Resultats du voyage du S.Y. “ Belgica” : Oceanographic ; Glace de Mer et Banquises, 1908.
Barrington, Daines, The Possibility of Approaching the North Pole asserted : Observations on the Floating
Ice which is found in High Northern and Southern Latitudes, edit. Beaufoy, 1818 (1st edit. 1775).
Belcher, E., Last of the Arctic Voyages, 1855.

Borgen, C., Die Zweite Deutsche Nordpolarfahrt, Bd. ii, Wissen. Ergebnisse, pp. 685-701: “Uber die
Grosse der Eisbedeckung im Polarmeer.”

Brennecke, W., Ann. Hydrographie, xli (1913), pp. 134-144 : “ Ozeanographische Arbeiten des Deutschen
Antarktischen Expedition : Die Eisfahrt.”

Bruce, W. S., Polar Exploration (Home University Library), c. 1911, pp. 54—71.

Buchanan, J. Y., Proceedings, Royal Institution, May 8, 1908, p. 243 : “ Ice and its Natural History.”
David, T. W. E., and R. E. Priestley, British Antarctic Expedition, 1907-9: Reports of the Scientific
Investigations : Geology, London, 1914, vol. i, pp. 180-192.

Davis, J. K., With the “ Aurora ” in the Antarctic, 1911-1914, London, 1919.

Drygalski, E. von, Ostgronlandexpedition, Berlin, 1897, vol. i, pp. 419-429.

Sitz. d. Konig. Bayer. Akad. Wissen., Math.-physik. Kl., 1910, 9. Abhand. : “Das Schelfeis der

Antarktis an Gaussberg.”

Arch. Sciences phys. et not., Geneve, tome xxx, Oct. 1910, pp. 356-374 : “ La Glaciation des Mers.”

Sitz. d. Bayer. Alcad. Wissen., Math.-physik. Kl., 1919 : “Die Antarktis und ihre Vereisung.”

Ferrar, H. T., National Antarctic Expedition, 1901-1904- : Natural History, vol. i, Geology, pp. 55-62.
Fricker, Karl, Die Entstehung und Verbreitung des antarktischen Treibeises , Leipzig, 1893, pp. 160-168.
The Antarctic Regions, 1900, pp. 262-265.

Gourdon, Expedition Antarctique Frarigaise (1903-05): “ Geographie physique — Glaciologie,” pp. 121-139.
Greely, A. W., Three Years of Arctic Service, London, 1886, vol. ii, pp. 43-60, 351.

Guthrie, Fred, Proc. Phys. Soc. London, vol. i, 1874-6, p. 53: “On Salt Solutions and Attached Water.”
Hamberg, Axel, Bihang till Konigl. Svenska Vet.-Akad. Handl ., 1895, Bd. xxi, Afd. ii, No. 2 : “ Studien
fiber Meereis und Gletschereis.”

Harvey-Pirie, J. H., Trans. Royal Society of Edinburgh, xlix, iv, No. 15, 1913: “Scottish National
Antarctic Expedition: Glaciology of the South Orkneys,” p. 861.

Hobbs, W. H., Characteristics of Existing Glaciers, pp. >186-244.

Kane, E. K., U.S. Qrinnell Expedition, 1850-51 : 1853.

Krummel, Otto, Handbuch der Ozeanographie, 1907, Band i, pp. 498-526.

Markham, C. R., and H. R. Mill, in The Antarctic Manual, 1901, pp. 14-15: “Ice Nomenclature.”
Mawson, Douglas, in Shackleton, Heart of the Antarctic, 1909, vol. ii, Appendix, p. 334 et seq.
M‘Clintock, Leopold, The Voyage of the “Fox” in Arctic Seas, popular edit., 1908, pp. 293, 294.

Moss, E. L., Proc. Roy. Soc. London, xxvii (1878), p. 544 : “ Observations on Arctic Seawater and Ice.”
Nansen, F., Farthest North, London, 1897.

Nares, G. S., Narrative of a Voyage to the Polar Sea during 1875-6 in H.M. Ships “Alert” and
“ Discovery ,” 1878.

Nordenskjold, A. E., The Voyage of the “ Vega” round Asia and Europe, London, 1881.

Nordenskjold, Otto, Wissen. Ergebnisse d. Schwedischen Siidpolar-Expedition, 1901-08, Bd. i, Lief, i : “Die
Schwedische Siidpolar-Expedition und ihre geographische Tatigkeit,” pp. 117-119, 158-168,
174-178.

Payer, New Lands within the Arctic Circle, London, 1876, vol. i, pp. 3-61.

AS OBSERVED IN THE WEDDELL SEA.

829

Pettersson, Otto, “ Vega ” Expedition : Vetenslcapl. Iakttagelser II, Stockholm, 1883, pp. 247-323 : “ On
the Properties of Water and Ice.”

Przybyllok, E., Zeit. Ges. Erd. Berlin , 1913, pp. 1-17 : “Deutsche Antarktische Expedition: Bericht uber
die Tatigkeit nach Verlassen von Siidgeorgien.”

Rae, John, Proc. Phys. Soc. London , vol. i, 1874-6, p. 14: “On some Physical Properties of Ice, etc.”

Scoresby, W., An Account of the Arctic Regions, 2 vols., Edinburgh, 1820 ; vol. i, pp. 225—322.

Speerschneider, C. I. H., in Publikationer fra det Danske Meteor ologiske Institut, “The State of Ice in the
Arctic Seas.” [Appears annually ; summary for 21 years published in 1916 (1917).]

Sutherland, P. C., Journal of a Voyage in Baffin's Bay, etc., 1850-51 ( under the command of Mr William
Penny), London, 1852, pp. 477-486.

Walker, David, Journal Royal Dublin Society, ii (Jan. 1860), p. 371.

Weyprecht, Karl, Die Metamorphosen des Polareises, Wien, 1879.

Wordie, J. M., Geographical Journal, April 1918 : “The Drift of the Endurance ,” with map.

Wrangell, Ferd. von, Narrative of an Expedition to the Polar Sea in the years 1820-23, edit, by Lt.-Col.
Sabine, London, 1840.

Wright, C. S., in Scott’s Last Expedition, London, 1913, vol. ii, pp. 441-451.

Field-Ice. A field of hummocky pack reaching to the horizon.

Field-Ice. Mainly hummocky pack ; in the middle distance a frozen lead of young-ice.

Trans. Roy. Soc. JEdin.

Vol. LII.

Mr J. M. Wordie on “The Natural History of Pack-Ice as observed in the Weddell Sea.” — Plate I.

Trans. Roy. Soc. Edin.

Yol. LII.

Mr J. M. Wordib on “The Natural History of Pack-Ice as observed in the Weddell Sea.” — Plate II.

Close-Pack. Ship’s progress stopped. Only a few patches of water left between the floes.

Close-Pack. Ship steaming through large floe of very thin young-ice.

Trans. Roy. Soc. Edin.

Vol. LIT.

Mr J. M. Wordie on “The Natural History of Pack-Ice as observed in the Weddell Sea.” — Plate III.

Open-Pack. Floes hummocked, but sufficiently light to be broken by ship charging.

Drift-Ice. Floes moderately light.

Trans. Roy. Soc. Edin.

Yol. LII.

Mr J. M. Wordie on “ The Natural History of Pack-Ice as observed in the Weddell Sea.” — Plate IY.

Pressure Ridge. Hummocking of young-ice ; in foreground, still younger ice covered by ice-flowers.

Ice-Flowers. Detail of crystals.

( 831 )

XXXIL — On Old Red Sandstone Plants showing Structure, from the Rhynie Chert
Bed, Aberdeenshire. Part IV. Restorations of the Vascular Cryptogams,
and Discussion of their bearing on the General Morphology of the Pteridophyta
and the Origin of the Organisation of Land-Plants. By R. Kidston, LL.D.,
D.Sc., F.R.S., and W. H. Lang, D.Sc., F.R.S., Barker Professor of Cryptogamic
Botany in the University of Manchester. (With Five Plates.)

(Read May 2, 1921. MS. received May 2, 1921. Issued separately August 26, 1921.)

This part of the account of our examination of the plants preserved in the
silicified peat-bed of early Old Red Sandstone age found at Rhynie will be devoted
to the consideration of a number of general questions concerning the four Vascular
Cryptogams which it has yielded.

1. In the first place the morphological characters of Rhynia Gwynne-V aughani ,
R. major, Hornea, and Asteroxylon will be reviewed and the attempt made to
reconstruct the external appearance of these plants as they grew. In relation to this
a few additional features of the plants will be described.

2. This will lead naturally to a consideration of the general bearings upon plant-
morphology of the facts mentioned in this series of papers.

Reconstructions of the External Morphology of the Vascular
Plants of the Rhynie Deposit.

The structure of the various parts of the four Vascular Cryptogams found in the
Rhynie peat-bed has been described in the three preceding Memoirs of this series. *
In each case proofs or indications of the connection of the parts have been mentioned
and discussed. The chief features are summarised in the diagnoses of Rhynia
Gwynne-Vaughcmi, Rliynia major, and Hornea Lignieri in Part II, and of
Asteroxylon Mackiei in Part III ; these descriptions need not be repeated.

The reconstructions on Pis. I and II represent our conception of the external form
and relative sizes of the four plants. The absence of specimens preserved as
impressions necessitate such reconstructions being to a certain extent imaginative.
This especially applies to the more complex plant of Asteroxylon. Even for this,
however, and certainly in the cases of the simply constructed sporophytes of the
Rhyniacese, the reconstructions probably give a fairly correct idea of the habit of
the plants.

The brief notes that follow will serve to indicate both those features in the
reconstructions for which we possess clear and direct evidence, and those which have
to be supplied more or less conjecturally. The opportunity will be taken to add
* Trans. Roy. Soc. Eclin., vol. li, p. 761 ; vol. lii, pp. 603, 643.

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 32).

129

832 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

some new facts regarding the various plants which have come to light since the
descriptions were published.

Rhynia Gwynne-Vaughani (PL I, fig. l).

We have some direct evidence as to the mode of growth, height, and general habit
of this plant. The block represented in Part I, fig. 5, in which the plants were
preserved standing erect as they grew, shows the crowded, rush-like growth of the
erect, tapering stems. These could be traced for a length of 13 cm., but were
incomplete. In the reconstruction the plant is represented as about 17 cm. high,
which may fairly be taken as about the mean height. Since, however, stems have
been met with of 3 mm. diameter, while the stems in the specimen in Part I, fig. 5,
were of the more usual diameter of 2 mm., the plants of Rhynia Gwynne-Vaughani
probably attained a somewhat greater height than is shown in the reconstruction ; a
height of 20 cm. was suggested in the diagnosis.

The aerial stems tapered from a diameter of two (to three) millimetres at the
base to about one millimetre. Examples of dichotomous branching have been met
with sparingly. We have represented the sporangia as terminating some of the
tapering stems, and take the size of the sporangia, from the few specimens seen, as
about 3 mm. (or slightly more) in length by 1 to 1'5 mm. across. The rhizomes, of
which only a few good specimens have been seen, were stem-like and about the same
diameter as the aerial stems ; they bore rhizoids on the lower side.

Most specimens of Rhynia Gwynne-Vaughani have shown the structures which
we have spoken of as hemispherical projections. The size and distribution of these are
best shown in figs. 7 and 8 of Part I ; these represent pieces of stem, about 5 mm. in
length, and bear on the surface turned towards the observer five and three projections
respectively ; there are no projections on the piece of stem in fig. 6. The projections
are somewhat longer in the direction of the length of the stem, and have thus an oval
outline ; they are about ^ mm. long. These projections are simply indicated by
dots in the reconstruction. The larger or smaller adventitious branches, which
occupied the position of some of the hemispherical projections, are also represented
in this.

Some further particulars with regard to the hemispherical projections may
conveniently be given here. Their general structure was illustrated in Part I,
figs. 50 and 51. It was there shown that the projection is a localised pro-
tuberance of small cells referable to divisions in the epidermis and the layers
immediately below this, and bursting the thick cuticle. Both a vertical elonga-
tion of the cells and their septation by thin walls are concerned in the localised
growth. A similar longitudinal vertical section of another projection is given here
(PI. Ill, fig. 5).

To this we have now to add the fact that in a number of examples the projection
has been clearly seen to have formed more or less directly underneath a stoma.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 833

Whether this can be assumed to hold generally is, of course, not established, but
may be regarded as probable.

A good example is represented in figs. 6-8. As the general low-power view
(fig. 6) shows, the plane of section has passed tangentially to a stem of Rhynia
Gwynne-Vaugliani. A portion of the stem bears two hemispherical projections
cut in oblique tangential section and in part seen from the outer surface. Con-
tinuing the line of the stem, another hemispherical projection that was evidently
borne on the same surface has been shaved off by the section and is viewed directly
from above. The two portions are more highly magnified in figs. 7 and 8. A
stoma ( s .) is seen clearly in relation to the lower of the attached projections in fig. 7,
while the position of the stoma is indicated by a dark patch on the surface of the
other projection. The stoma ( s .) is 'Seen quite clearly on the isolated projection in
fig. 8. A number of examples similar to this have been observed, and in them also
the position of the stoma was near one end of the oval projection. Whether this
was the end directed towards the base or apex of the plant is uncertain. The
additional example in fig. 9 shows the stoma (,s.) at one end of a small projection,
apparently in an early stage of development.

In the light of these specimens it is clear that the hemispherical projections
cannot be regarded as part of the primary construction of the plant as at first
developed.

Many of the hemispherical projections did not undergo further development or
change. What appears to have been an exudation, in which fungi often grew,
sometimes formed to the outside of the projection as the stems lay in the matrix.
In other cases the projections were the seat of further developments. A single
specimen has been seen ill which a projection, prolonged vertically from the surface,
resembles an emergence (PI. Ill, fig. 10). Sometimes their superficial cells grew
out as rhizoids (Part I, figs. 52-54). At other times growth resulted in the
development of lateral, adventitious branches; these in some cases remained small,
and may have fallen off and served as a means of vegetative reproduction ; often,
however, they were larger and more strongly attached, though their vascular system
was separate from that of the stem bearing them (Part I, figs. 55-60).

Comparisons of the peculiar hemispherical projections on the stems of Rhynia
Gwynne-Vaugliani are possible both with normal and pathological developments
in recent plants.

They show certain resemblances to lenticels which, when formed in relation to
the original surface of a stem, usually develop below the stomata. Nothing in the
details of structure of the hemispherical projections of Rhynia indicates, however,
that they were related to normal gaseous exchange. This comparison is thus not a
very close one.

A more suggestive comparison is with certain intumescences that form on the
stems and leaves of various plants, either as a result of the excessive moisture of

834 DR R. K1DST0N AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

the atmosphere, or of the presence in this of injurious vapours. Numerous examples
of such intumescences are given in works on plant-pathology.* Perhaps the most
striking figures for comparison are the photographs of intumescences t formed
beneath the open stomata of leaves of cauliflower, etc., as a result of the entry of
dilute vapour of ammonia. The elongation of the affected cells at right angles
to the surface, and their subdivision by thin walls in the intumescences, present a
remarkable similarity to the changes which result in the hemispherical projections
of Rliynia.

There are no close comparisons to advance for the production of definite organs
as new growths from intumescences. The tissue composing the latter does not,
however, differ profoundly from callus, which, as is well known, is often the seat of
adventitious growths.

hi connection with the hemispherical projections and adventitious branches of
the stems of Rliynia Gwynne-Vaughani another feature which is clearly patho-
logical must be briefly described. A considerable number of the stems show dark
necrosed areas extending more or less deeply from the epidermis ; in other
specimens the dark necrosed mass is wanting or has more or less completely
disappeared leaving a larger or smaller cavity. Examples of these appearances are
shown in figs. 11-14. In some cases the necrosed area appears to be related to the
place of detachment of an adventitious branch (Part I, fig. 55), and sometimes it
corresponds to the position of a hemispherical projection. Often, however, no such
relation is to be traced. The necrosis cannot be proved to be related to the position
of a stoma, though this possibility is by no means excluded. There is no evidence
that the necrosis is caused by the presence of fungi ; fungal invasion may be
extensive without any signs of necrosis. Fungal filaments and other organisms
may, of course, be present in the resulting cavities.

The feature of special interest is the behaviour of the cells abutting on the
necrosed area or bounding the cavity. As shown in figs. 12-16, they had undergone
changes that must have taken place during the life of the tissues. The cells are
often elongated at right angles to the dead tissue of the cavity, and frequently show
evidence, by the presence of thin walls, of an active cell-division that can without
doubt be interpreted as a reaction to the wound. The arrangement of the cells and
the mode of growth of the tissue resembles what is found in the formation of the
hemispherical projections.

The further interest of these peculiarities of the remains of Rliynia Gwynne-
Vaughani will be considered in the section of Part V dealing with the conditions of
accumulation and preservation of the deposit.

* Of. Sorauer, PJlanzenkranldieiten, Aufl. 3, Bd. 1, p. 435 ff. Kuster, Pathologische Pflanzenanatomie,
Aufl. 2, p. 44 ff.

f Erwin F. Smith, “Mechanism of Tumor Growth in Crown Gall,” Jour. Agric. Res. ( Washington ), vol. viii,
pi. 60, etc.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 835

Rliynia major (PL I, fig. 2).

In the case of this plant there is no direct indication as to height or habit. All
parts of the plant, however, are known and are distinctly larger than the correspond-
ing parts of Rliynia Gwynne-Vaughani. The diameter of the cylindrical stems
ranged from 5 or 6 mm. to under 1‘5 mm. The stems are known to have branched
dichotomously, and doubtless tapered as in the case of the smaller species. In the
reconstruction we have represented the plant as about three times the height of
R. Gwynne-Vaughani. The sporangia are known to have terminated some of the
slender axes about 1'5 to 2 mm. in diameter. They attained the large size of
12 mm. x 4 mm. Sometimes two sporangia lay side by side in the peat as if they
had terminated two branches resulting from a dichotomy. Hemispherical pro-
jections and adventitious branches are not known to occur in this species. The
most satisfactory specimens of rhizomes in situ (Part I, fig. 13) show that they grew
horizontally and were cylindrical, stem-like, and dichotomously branched. They
bore rhizoids on the lower surface, sometimes on large bulges. The features thus
briefly touched upon will be found expressed in the reconstruction.

We have little to add to the description of this species, but the opportunity may
be taken to figure an obliquely longitudinal section of the apical region of a branch
(PI. IV, figs. 17 and 18). The actual apical meristem is contracted and collapsed,
but the progressive enlargement and elongation of the cells behind this can be traced.
The fact that the layer ofi cells with dark contents just below the outer cortex can
be traced near to, but not up to, the apex is of interest (fig. 17).

As pointed out above, hemispherical projections like those of Rliynia Gwynne-
Vaughani have not been found in R. major. Further, while the stems of the latter
plant may be very perfectly preserved or badly decayed, they do not so commonly
exhibit the wound-reactions around cavities due to necrosis described for R. Givynne-
Vaugliani. Examples of necrosed areas and of wound-reactions that must have
taken place during life have, however, been met with. This is shown in a very
striking fashion by the two stems represented in fig. 19 on PI. IV. In the case of
the stem on the left, the cortical tissue had disappeared from a sector of the stem as
seen in transverse section. Regenerative activity involving enlargement and division
of the cells thus exposed had taken place. This proceeded both from cells of the
cortex and of the phloem (fig. 20). In the other stem there was no cavity, but the
region of cortex below and around a dark necrosed patch had been the seat of
enlargement and division of cells. It contrasts with the normal cortex in the
arrangement of the groups of newly formed cells and the whole character of the
tissue. A portion of the disturbed growth in the neighbourhood of the stele of this
stem is shown more highly magnified in fig. 21.

The enlargement of the cells that here precedes their division can be compared
with another appearance, represented in fig. 22 for R. major. In the case of this

836 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

stem certain of the cortical cells at two or three regions of the transverse section
are markedly enlarged as compared with the others that are of normal size. A
similar appearance was recorded for a stem of R. Gwynne-Vaughani in Part 1
(p. 770, fig. 74).

This feature, like the wound-reactions, suggests comparison with a reaction of the
living plant to some prolonged stimulation.

Hornea Lignieri (PI. II, fig. 3).

For this plant also we have only the fragmentary remains of all the parts and no
specimens to show the general habit or height. It has been represented as some-
what smaller than Rhynia Gwynne-Vaughani since its stems did not, on the whole,
attain the diameter of those of the latter plant, and also give the impression of
having been less rigid.

The reconstruction shows the tuberous lobes of the rhizome from which the erect
stems arise. These were repeatedly branched dichotomously. Some of the finer
branches terminate in the sporangia, which were considerably smaller than those of
Rhynia Gwynne-Vaughani. The sporangia may stand singly or in pairs, or may
themselves be actually involved in the dichotomy of the branch-apex which has
produced them, and therefore appear lobed.

We have no further facts to add to the description of Hornea given in Part II.

Asteroxylon Mackiei (PI. II, fig. 4).

Since Asteroxylon was a much more complex plant than Rhynia or Hornea, the
reconstruction of its general external appearance from the fragmentary remains is
more difficult, although there is evidence for most of the parts represented in the
reconstruction. It has, however, been found necessary to infer the relative pro-
portions and connection of the parts in forming a representation of the whole plant.

Mention may first be made of features regarding which evidence will be found
in the description of Asteroxylon in Part III. The existence of a dichotomously
branched, leafless rhizome without absorbent hairs was fully established. Certain finer
root-like branches of this were found to penetrate fragments of plants in the peat.
A gradual passage from the leafless rhizome into the stem, which became clothed with
small simple leaves, was also demonstrated ; in the intermediate, transition region the
leaves make their appea'rance as small scales. The occasional, dichotomous branching
of large leafy shoots was met with, but this contrasted with the predominantly
lateral relation of the branches on most main axes. In the case of the lateral
branches and the finer ramification of the plant, dichotomous branching appeared to
be more probable. The leaves on the main stems and branches were closely crowded
and attained a length of over 5 mm. The shoot must have had much the same
appearance as that of various species of Lycopodium.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 837

In the impressions of Thursophyton ( Lycopodites ) Milleri, a plant of the Middle
Old Red Sandstone flora of Scotland, which also occurs in Norway, we have a similar
general habit to what is here assumed for the leafy shoots of Asteroxylon.

The repeatedly dichotomous leafless branches, which are believed to have borne
the dehiscent Sporangia, are represented in the reconstruction as the fertile region
of the plant. Since, however, the connection of these structures with the vegetative
organs of Asteroxylon is not established, a break is left in the restoration at this
point.

It is with regard to the general habit of the plant as a whole that direct evidence
is lacking. We have met with nothing pointing to the branched leafy shoots having
been horizontal or prostrate, and have therefore represented them as erect. The
transition from the leafless rhizome to the base of the leafy shoot also appears to
have stood vertically. In order to connect these parts naturally, we have assumed
the existence of a more or less horizontal branched rhizome, giving off, on the one
hand, downwardly growing root-like branches, and, on the other, erect leafy shoots.
This assumption seems natural and justified in the light of our knowledge of the
mode of growth of other plants with somewhat similar characters {c.g. Psilotum and
various species of Lycopodium ).

It is further supported by the large specimen of a rhizome of Asteroxylon repre-
sented in fig. 24 on PI. V. This specimen dips downwards and penetrates a large
stem of the plant lying in the peat. The basal region of the rhizome is of interest
since it has a well-marked epidermis and bears a number of scale leaves ; this region
has the structure intermediate between a leafy shoot and a rhizome. On tracing it
on to the more distal portion, the well-defined surface is lost and the structure is
that typical of the rhizomes of Asteroxylon. This specimen thus affords proof of
the direct passage from a leaf-bearing axis into a rhizome, while the transition from
a rhizome to a leafy shoot was followed in Part III (p. 647 ff). Fig. 23 on PI. IV is
a photograph of a portion of epidermis of a stem of Asteroxylon in surface view. It
shows the median line or ridge on the outer wall of each of the epidermal cells ; this
feature of Asteroxylon was recorded on p. 653 of Part III, but not figured. The
similarity in this detail of structure between Asteroxylon and Rhynia Gwynne-
Vaughani will be evident if fig. 23 is compared with fig. 31 of Part I.

The remaining figures on PI. V are from a series of transverse sections cut
through the terminal region of a small leafy shoot of Asteroxylon. The crowded
tips of the leaves were noticed at the surface of a block of the chert and a complete
series of nine sections prepared in the hope of obtaining the actual growing point.
Unfortunately the apical meristem itself was ground away in the process of cutting ;
it came between the second (fig. 25) and third (fig. 26) sections of the series.
Fig. 25 shows a complete and almost perfectly transverse section of the bud just
above the growing point. The leaves of the bud arranged in a complex spiral (much
as in a corresponding section of Lycopodium Selago ) diminish in size on passing

838 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

inwards towards the centre. The next section, represented in fig. 26 and more highly
magnified in fig. 27, passed through the tip of the stem below the growing point ; it
was so close to the latter that the widening of the stem leads to the superficial tissues
being cut somewhat tangentially. Epidermis (ep.), outer cortex (o.c.), and the outer
zone of the inner cortex ( i.c.o .) can be distinguished in order, although the cortical
tissues are not completely differentiated. The condition of the middle trabecular
zone of the inner cortex ( i.c.m .) is of interest. The trabecular structure is clearly
marked, but the radial depth of the zone is short. The inner zone of the inner
cortex ( i.c.i .) is of full width, and its rather imperfectly preserved tissue passes
without a sharp limit into the stele.

In the stele the xylem (cc.) forming the forked poles of what would become the
stellate .xylem is alone differentiated, there being no tracheides developed in the
central connecting portion. Leaf- traces (l.t.) departing from the poles of the stele are
seen around the latter. The soft tissues of the stele are preserved, though somewhat
imperfectly. It is clear from this section that the poles of the stelar xylem were
differentiated close behind the apical meristem.

The other sections of this series show that the central portion of the stelar xylem
did not become differentiated for some distance behind the apex. A stele from the
middle of the series is represented in fig. 28, the lowest section of the series is shown
complete in fig. 29 and its stele more highly magnified in fig. 30. This lowest
section shows that at the base of the series the differentiation of the stelar xylem
was hardly further advanced than in figs. 26 and 27, although a few elements of the
connecting metaxylem can be recognised. The outer cortex has attained the
structure found in mature stems ; the tissues of the inner cortex are too poorly
preserved and crushed to be instructive.

This complete series confirms and extends the information as to the arrangement
of the leaves in the bud shown in Part III, PI. V, fig. 36, and as to the structure of
the stele with its xylem imperfectly differentiated shown in Part III, PI. VIII,
figs. 59, 60. The series affords conclusive evidence that the general differentiation
of the stellate xylem of the shoots of Asteroxylon was centripetal ; the position of
the protoxylem in the ends of the rays was shown in Part III to be as a rule enclosed
or mesarch.

Some Bearings of the Rhynie Plants on the General Morphology of the

PTERIDOPHYTA AND THE ORIGIN OF THE ORGANISATION OF LAND-PLANTS.

The deposit at Rhynie gives us a remarkable glimpse of the vegetation of the
Early Devonian * period. The four higher plants, which are all Vascular Cryptogams,
have the double interest of being the most simply organised Pteridophy ta known
and of being the most ancient fully-known forms of this group. In relation to the

* The term Early Devonian is here used to include the period represented by the Lower and Middle Old Red
Sandstone, hut to exclude the Upper Old Red Sandstone.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 839

descriptions of the plants in the three preceding papers of this series, comparisons
have been made with existing and extinct plants, and some theoretical aspects of the
facts have been indicated. Although this course involves a little repetition, some
of the general bearings of these plants on the problem of the origin of the morpho-
logical construction of land-plants must now be considered.

As regards the origin of the type of alternation of generations and the balance of
the sexual and spore-bearing stages in the life-history, the Rhynie plants give no
information. The plants known to us were clearly the sporophyte generation. The
development of the spores in tetrads similar to those of other Vascular Cryptogams
points clearly to the existence of a sexual generation 'or gametophyte. Of the nature
of this we have no evidence. The absence of evidence is so far in favour of the
gametophyte having been much simpler and more delicate than the sporophyte, as
is the case in all the Vascular Cryptogams the life-history of which is fully known.

It is on the organisation of the sporophyte in the Vascular Cryptogams that most
light is thrown by the Rhynie plants. Apart from questions of physiological
morphology, with which we are not here concerned, this is a comparative and
historical problem. The historical data are necessarily imperfect, and their lack has
led to much speculation of doubtful value, although opposed speculations have served
to define aspects of the problem. It is in the additions they make to the ascertained
historical data that the great importance of the Rhynie plants consists. Without
entering fully into speculations on the origin of the Pteridophyta and their character-
istic organisation, it will be necessary to set the new facts in relation to them.

Existing Vascular Cryptogams may be broadly said to present us with a few main
types of organisation of the sporophyte. These are the Fern type, the Equisetum
type, the Lycopod type, and the type of the Psilotacese. The general characteristics
of these are familiar and need not be stated. It is only necessary to direct attention
to the facts that all these types agree in having shoots composed of stem and leaves,
and that in all, except the Psilotacese, the plant-body has definitely characterised
roots. In the Psilotacese roots are wanting, the aerial stems, with their smaller or
larger leaves, arising from a system of leafless rhizomes clothed with absorbent hairs.

The extinct Pteridophyta of earlier vegetations, back to and including the
Carboniferous and Upper Devonian periods, add, so far as we know, no essentially
different types of organisation. They make us acquainted with varied and interesting
forms of the Fern, Lycopod, and Equisetum types, and they add the wholly extinct
type of the Sphenophyllales. The vegetative body of all these plants is, however,
composed of root, stem, and leaf, and the comparison of the plants known from the
Upper Devonian to the existing period does not throw light on the origin of this
prevailing differentiation of the plant-body.

Even before the discovery of the material described in this series of papers
interest was naturally focussed on the plant-remains earlier than the Upper Devonian.
These were always incomplete, a serious drawback in making comparisons, but on

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 32). 130

840 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

the whole indicated plants of more primitive organisation than those known from
the Upper Devonian onwards. Without entering into an enumeration or survey of
these Early Devonian plants from Scotland, Norway, Canada, Bohemia, and elsewhere,
some of their features may be mentioned. In many of them the stems bore more or
less evident and usually relatively small leaves, while others resembled fern-fronds
without laminae. The appearance of most of the Middle Devonian plants was such
as to show that a distinction of stem and leaf had been established at that period.
On the other hand such plants as Pseudo sporochnus , and especially the Lower
Devonian Arthrostigma and Psilophyton, strongly suggested an even more primitive
organisation on the whole. Many of these remains, though now clearly shown to
have been vascular land-plants, have at various times been regarded as Algae. Other
anomalous plants of the same period still imperfectly known may not improbably
find their place in this latter group, or may have combined characters that are now
distinctive of Algae and Pteridophyta.

These indications of plants of simpler and more archaic organisation than any
Vascular Cryptogams of later periods, and the consequent interest of the Early
Devonian period for the study of early steps in the land-sporophyte, are not
inconsistent with more highly organised plants having already existed at that period.
The information on this point is insufficient, but large stems which bore regularly
arranged lateral appendages and other stems with suggestively complex internal
structure are known. It is the undoubted existence at that period of very simply
organised vascular plants that is the fact of critical importance.

Though Dawson’s restoration of Psilophyton was used by Lignier as an example
of such more primitive types of land-plants, it was impossible to overlook the lack of
critical evidence for some features in the reconstruction. The earlier data, though of
great interest, were indeed almost always rendered less valuable owing to uncertainty
regarding essential facts. The study of the Rhynie chert has fortunately added
definite knowledge of the organisation of certain simple and very completely
preserved plants of at least a near geological age.

The information afforded by the study of Asteroxylon, the most complex of the
Rhynie plants, would not by itself have added fundamentally to what was already
known or inferred, although it would have rendered this much more complete and
certain. In Asteroxylon we have a plant-body with shoots consisting of stems
bearing numerous relatively small leaves. The plant thus conforms, on the whole,
to the type of shoot known in Tliursophyton ( Lycopodites ) Milleri, Psilophyton
princeps , and at the present day in Lycopodium and the Psilotacese. In the absence
of definite roots and the possession of a simply constructed cylindrical rhizome
continued directly (with the appearance of leaves and complication of the internal
structure) into the aerial shoots, Asteroxylon shows a simpler organisation than any
Lycopod. This condition is, however, found in the existing Psilotacese, and it
presumably held also, as was stated by Dawson, for Psilophyton . While some

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 841

uncertainty must remain as to the reproductive region of Asteroxylon, the evidence
of association is in favour of the sporangia having been borne on leafless branch-
systems. In this respect also Asteroxylon would agree with Psilophyton.

The general organisation of the sporophyte in Asteroxylon and Psilophyton gives
us a distinct type of plant-body which contrasts with the Fern, Equisetum, and
Lycopod types, but may be more closely compared with the plant-body in the
existing Psilotacese. The Lycopod type is distinguished from the type of Asteroxylon
in its vegetative construction by possessing true roots ; the range of diversity in the
roots and root-bearing organs of the Lycopodiales must, however, be noted in this
connection.

The great interest of the organisation of Asteroxylon and Psilophyton , though
both these plants are in their respective ways more or less specialised, is the
suggestion it affords of a synthetic type, combining external and anatomical features
found in distinct groups of the Vascular Cryptogams. The morphological, and to a
less extent the anatomical, characters of the Psilotacese can be interpreted in the
light of what we know of Asteroxylon and Psilophyton. The external morphology of
the shoot, and especially the anatomy of Lycopodium , find parallels in Asteroxylon.
The comparison between the characters of Asteroxylon and those found in plants of
the Fern type is less apparent, but comes out especially when the Zygopteridese are
taken into consideration. The geological age of Asteroxylon and Psilophyton adds
importance to the synthetic type they exhibit. The definition of the various types
of Pteridophyta, that might be regarded as possibly derived from such a synthetic
type, is afforded partly by vegetative form and structure, but largely by the relative
positions occupied by the sporangia. The diversity attained in this respect by the
Fern and Lycopod types, for example, can be most readily conceived if a branch-
system bearing sporangia on some of its ultimate subdivisions is taken as the point
of departure. Such a structure appears to have held for the fertile regions of the
sporophytes of Asteroxylon and Psilophyton , while the vegetative shoots bore more
or less crowded small leaves.

The other three plants ( Rhynia Gwynne-Vaughani, Rliynia major , Hornea
Lignieri) found at Rhynie along with Asteroxylon exhibit still greater simplicity
of organisation in being not only rootless but leafless. They have been placed
together in a very natural family called the Rhyniacese which, along with the
Asteroxylacese, to which Asteroxylon and probably Psilophyton belong, represent
the distinct class of the Pteridophyta which has been named the Psilophytales after
its first clearly recognised member. The simple Rhyniacese afford a surprising
addition to our knowledge of the plant-forms realised in the Vascular Cryptogams.
With merely differences of detail the plant-body in Rhynia and Hornea consisted
of (a) a rhizomatous, subterranean portion, bearing absorbent hairs, without leaves,
and with nothing suggestive of the organisation of a root, ( b ) a dichotomously
branched, aerial system of cylindrical axes without leaves, and (c) large sporangia

842 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

terminating some of the finer branches. This type of plant-body is simpler than
anything that was previously known in plants with stomata, tracheidal vascular
system, and cuticularised spores. Its demonstration is the outstanding contribution
of fact which the study of the Rhynie deposit has made to the comparative
morphology of the Vascular Cryptogams, and through them of the land-vegetation
composed of sporophytes.

In the comparative study of the Vascular Cryptogams we must now include
plants with independently living, rootless, and leafless sporophytes ; these may,
without exaggeration, be described as thalloid branch-systems attached to the soil
by a basal region bearing rhizoids. As previously mentioned, the existence of a
gametophyte or prothallus, probably of small size, can be inferred from the pro-
duction of the spores in tetrads. There is nothing in these very simple sporophytes to
suggest their origin by reduction from more highly organised types. Their antiquity
and what we know of the peculiarities of the vegetation of the Early Devonian
period add to their interest and significance.

The type of the Rhyniacese is so simple that a gap remains between it and even
the simplest sporophytes with leafy shoots. The mode of origin of the latter type
of construction, which has long been a speculative morphological problem, is thus
not clearly demonstrated by the ancient leafless and leafy types that co-existed at
Rhynie. This problem will be dealt with further below.

On the other hand the simplicity of organisation of the sporophyte in the
Rhyniacese facilitates comparisons with the Bryophyta and the Algse. While
concentrating attention on the Rhynie plants as most fully known, some other Early
Devonian plants must be borne in mind in such comparisons. Prominent among
these plants is the incomplete and less satisfactorily preserved, stalked, columellate
sporangium or sporogonium which has been named Sporogonites by Halle. Another
plant that must be mentioned is the peculiar Parka, of uncertain systematic position,
in which masses of cuticularised spores have recently been shown to fill the areolae.

As regards Bryophyta, our practical ignorance of undoubted plants of this group
from any but the more recent geological formations must be remembered. It was
suggested by Halle that Sporogonites was the sporogonium of a Bryophyte ; in the
light of our knowledge of Hornea his alternative view that it might be the upper
portion of a plant on the line of descent of the Pteridophytes appears to be more
probable. All that can at present be said is that the simplicity of the independent
sporophytes of the Rhyniaceae makes comparison between the asexual generation in
Bryophyta and Pteridophyta easier, and to this extent narrows the gap between
these two great groups. Further than this it is not possible to go, so long as we
have no knowledge of simple sporophytes permanently borne on the gametophytes ;
this becomes more than ever the essential distinction between Bryophyta and
Pteridophyta.

The organisation of the Rhyniaceae also facilitates comparison with Algae. In

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 843

the case of some other Early Devonian plants, that are, however, very imperfectly
known, we may with fuller knowledge have to recognise such combinations of
characters as would break down any sharp distinction between the Algae and the
simplest Pteridophytes. This does not, however, appear to us to be the case for the
Rhyniaceae, and still less for the more complex Asteroxylaceae. The combination of
characters which the Rhynie plants possess (independent sporophytes, epidermis with
cuticle and stomata, vascular system with characteristic tracheides, thick-walled
sporangia with cuticularised spores) justify our recognising them as Vascular
Cryptogams or Pteridophyta. The differences between them and other Pteridophyta
are those of degree only, and are sufficiently expressed by placing them in a distinct
class (Psilophytales) of this group. It can, however, be said that the members of
this class are the Pteridophyta, which are most readily comparable with the Algae ;
this holds most strongly at present for Hornea, in which stomata have not been
demonstrated.

Terms implying actual relationship by descent have been deliberately avoided
in the comparisons which have been made above between the plants described in
this series of papers and Algae, Bryophyta, and other Pteridophyta. The suggested
comparisons are based on the observed facts of form and structure. In expressing
the comparisons in terms of possible phyletic relationship we enter into a more
speculative region. The characters in question are not such as to establish beyond
question the course of evolution either as regards the divergent origin of the more
specialised Pteridophyta from a common source, or, on the other hand, the direct
derivation of the simplest Pteridophyta from any particular Algal group.

It has been pointed out in the discussion in an earlier part of this series of papers
(Part III, p. 673) that the characters of Asteroxylon and Psilophyton , as representing
a synthetic type, are consistent with such a divergence of the great classes of
Pteridophyta from a common type as had been suggested by a number of investi-
gators. The Psilotacese would be of all existing plants the least altered from this ;
the Lycopodiales, while specialised in various directions, would have preserved
many features found in the archaic type ; the Fern type, if the large leaves are
regarded as specialised branch-systems, would have been more profoundly modified.
It is perhaps better to regard the point of divergence as represented by plants
somewhat simpler than the Asteroxylaceae in that they did not possess definite
small leaves ; this assumption is rendered a natural and legitimate one by the
fact that in the Rhyniaceae we have Vascular Cryptogams which actually exhibit
this simplicity.

The Rhyniaceae lend themselves naturally, on the other hand, to comparison with
Algae, as regards external form, the absence of definite roots, and especially in their
large sporangia having arisen more or less evidently by the transformation of the
tips of certain branches of the thalloid plant-body. These general resemblances, as
V^as pointed out (Part II, p. 622), “ are consistent with the Rhyniaceae finding their

844 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

place near the beginning of a current of change from an Alga-like type of plant to
the type of the simpler Vascular Cryptogams.”

It appears to us unwise to go further than this at present in view of the
absence of evidence justifying a connection of these Vascular Cryptogams with
particular Algal forms. We may anticipate sure, if slow, progress in our knowledge
on this question to come from the fuller study of other plant-remains of this early
geological period ; caution rather than boldness is therefore advisable in speculating
on the subject.

The facts described in these studies of the Rhynie plants, together with other
recent work on the Early Devonian flora, do not solve the question of the mode of
origin of the simplest Vascular Cryptogams. It is, therefore, only necessary to
touch briefly on their bearing upon various hypotheses which have been advanced
regarding this. It has been pointed out (Part II, p. 623) that the morphology
of the Rhyniaceae (while not inconsistent with the hypothesis that regards the
sporophyte as an interpolated phase in the life-history) supports the views which
hold the sporophyte of the Pteridophyta to have been derived by the modification
of a plant-body such as is seen in the asexual stage of a number of the higher
Algae. The former views have been especially worked out by Celakovsky and
Bower, while the conception of an origin of the Pteridophyta by transformation
of advanced Algal forms underlies the views of Potonie, Lignier, Schenck, Church,
and others. It is not necessary to enter here into the details of the various theories,
which require to be considered in the light of the state of our knowledge at the
times when they were respectively advanced.

When the attempt is made to relate in detail any of the particular hypotheses
to the features of the simple Pteridophyta recently discovered it becomes evident
how difficult it is to frame such speculations without assuming too much. The
further knowledge of the course of the life-history in such Algae as Laminaria,
Cutleria, Dictyota , and the Red Sea-weeds has revealed the existence of a definite
alternation of generations, with all proportional sizes of the two generations, within
the Algae. There is thus no reason, as in the earlier statements of the antithetic
theory, to start with the post-sexual complications in such Green Algae as GoleocJuete.
On the other hand the absence of leaves and of any specialised root-like branches
in the Rhyniaceae gives us within the Vascular Cryptogams a simpler starting-point
for the organisation of the sporophyte than is postulated by either Lignier or
Church, whose conceptions on the origin of the plant in the Vascular Cryptogams
by the transformation of an Algal plant-body are the most explicitly formulated.
The possibility of the type of plant in the Rhyniaceae having been reached by an
evolution on land, parallel to that of the marine algae, is not to be dismissed as out
of the question. In this connection some of the conceptions associated with the
antithetic theory may possibly be found to hold.

The facts so far known seem to favour the idea of the origin of the land

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 845

sporophyte by transformation of an Algal plant-body, but are equally consistent with
such ancestral Algae having possessed an antithetic alternation of generations.
They do not, however, support the detailed suggestions as to methods of advance
made by either of the opposed theories, e.g. on the one hand the relegation of the
differentiation of stem, leaves, and roots to the Algal ancestry, and on the other
intercalation of the sporophyte by progressive sterilisation. If such detailed
assumptions as these are put aside there is no longer the antagonism between the
antithetic and homologous origin of a sporophyte that at one time seemed to hold.
The simple sporophyte of Rhynia or Hornea was part of an antithetic life-history,
while its organisation can be properly treated as homologous with the plant-body as
realised in both the sexual and spore-bearing stages of many Algae.

The important change in the working position of comparative morphology that
has come about with the advance of our knowledge, both of Algae and of extinct
Vascular Cryptogams, is that a common consideration of the differentiation of the
plant-body in Pteridophyta, Bryophyta, and Algae is permissible and desirable. We
may leave questions of phylogeny on one side, with the recognition that the
characters of the Psilophy tales are such as on the one hand to bring these three
groups of plants closer together, and on the other to enable us to form a clearer
conception of a divergence towards the other classes of Pteridophyta and the land-
plants generally.

While the exact methods of advance, whether these implied actual genetic
relationship or were parallel developments, must remain open questions, progress
may be possible with such an extended comparative morphology. The bearing of
the plants under consideration on some problems of general comparative morphology
may, therefore, be referred to in conclusion. These remarks, which must necessarily
be brief, may be divided into those on the origin of roots and the morphology of the
protocorm, the origin of the leafy shoot and the relation between large-leaved and
small-leaved types of this, the vascular system, and the nature and original position
of the sporangia of Pteridophytes.

{a) The Origin of Roots ; Morphology of the Protocorm.

The absence of definite and characteristic roots is a common feature of all the
Rhynie plants ; as a presumably primitive condition this is only retained by the
Psilotacese among existing Vascular Cryptogams. In Asteroxylon , as in the
Psilotaceae, the extensive rhizome system suggests the possibility of the derivation
of roots by its further modification, and the occurrence of exogenous roots in some
'Lycopodiales further narrows the gap.

The subterranean region in the Rhyniacese was less extensive than in Asteroxylon
even when, as in Rhynia , it consisted of cylindrical branched axes resembling the
stems in general structure, but bearing rhizoids. It was especially interesting in the
case of Hornea on account of its tuberous lobed nature, the lobes being often rather

846 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

distinct, but in other specimens more or less confluent. A single, branched aerial
axis sprang from each lobe. The features of this type of rhizome, though on a larger
scale, have so much in common with the protocorm of the young embryonic plants
of some species of Lycopodium that in the description it has been spoken of as the
protocormous type of rhizome.

This resemblance is of great interest on account of Treub’s view that the proto-
corm of Lycopodium was the persistent relic of an organ, by which the sporophyte
became physiologically related to the soil before roots had evolved. This compari-
son has been briefly but sufficiently treated of in Part II (p. 620). It is only
necessary to repeat that the underground parts of Hornea appear to be consistent
with and to support Treub’s view. Just as in some species of Lycopodium a single
leaf or protophyll springs from each lobular portion of the protocorm, so in Hornea
a single aerial branch system occupies the corresponding position. This comparison
would involve a correspondence of the protophylls or leaves of Lycopodium with
the thalloid axes of Hornea.

The rhizome of Rliynia , and the more extensive rhizome-systems in Asteroxylon
and the existing Psilotacese, would then correspond morphologically to specialised
extensions of the protocorm. This, as Hornea shows, would have been the basal
region of the originally thalloid plant-body.

On the other hand the subterranean rhizome-system of these rootless plants may
be compared with the attaching region of certain Algse. The comparison is closer
with those Algse in which a rhizome-system ramifies in sandy soil and gives off erect
assimilating branch systems, than with the hold-fasts that attach many Algse to a
rocky substratum.*

Treub’s view of the protocorm involved the idea that the embryo of the early
Pteridophyta was dependent on the prothallus and required to become physiologically
independent. This is so in the ontogeny and may, as he assumed, have occurred in the
phylogeny, though this is not a necessary assumption. There is nothing inconsistent
with this in the morphology of Hornea. The repetition of the protocorm, though
primarily an embryonic organ, in individuals isolated in vegetative reproduction is
met with in Lycopodium, t If this view of the original nature of the protocorm and
protocormous rhizome is entertained it should be noted that it affects, though it does
not destroy, the comparison with the attaching bases of thalloid Algse. These
various comparisons cannot be unravelled so long as we are completely ignorant of
the history of the association between gametophyte and sporophyte in the ancestors
of the Pteridophyta.

From the simplicity of the rhizome-system in the Rhyniacese it would seem that
we are justified in assuming the appearance of specialised roots to have taken place

* Since we are dealing with general comparative morphology, and not with relationship, another more distant
comparison may be made with the leafless rhizome-system of the gametophyte of some Liverworts {e.g. Calobryacete).
f The repetition of the specialised tuberous protocorm of Phylloglossum may also he compared here.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 847

within the evolution of the vascular land-sporophyte and are not obliged, to account
for them by the transformation of organs already present in complex ancestral Algae.
We may, however, as above, trace comparisons with the corresponding region of the
Algal plant-body.

( b ) Nature of the Leafy Shoot ; Large-Leaved and Small-Leaved Shoots.

The simplicity of organisation of the Rhyniacese establishes the fact that there
were early Vascular Cryptogams in which the aerial parts consisted of a thalloid
branch-system. This is in favour of the view that the differentiation of a shoot,
with axis and leaves (though attained in the higher members of several groups of
Algae) was independently elaborated in the early land-plants. It is of course possible
that in some cases a more completely differentiated land-plant might have been
directly derived from an Algal form with stem and leaves, but the existence of these
simpler vascular plants renders it unnecessary to assume this without evidence.
They bring the origin of the distinction of stem and leaf in a shoot within the
evolution of the land sporophyte, although a corresponding differentiation is attained
in various Algse and in the gametophyte of Bryophyta. Without entering into a
full consideration of the problems of the relation of stem and leaf in the shoot, we
have to note the comparisons which may be made with the Rhynie plants and how
these bear on this much-discussed problem.

In the first place it may be noted again how closely the thalloid plant-body of the
leafless Rhyniacese can be compared with that of many Algse. The comparative
problem of the origin of shoots with small or large leaves subordinated to a stem
thus presents itself in both the Algae and the Vascular plants. Further insight as
regards either group will bear on the problem in the other, quite apart from genetic
relationship.

The question of the comparison of a shoot in which the stem is clothed with
small leaves with the leafless thalloid axis may be taken first, since it is presented
to us directly by the co-existence of Asteroxylon with the Rhyniaceae. In Asteroxylon
a gradual passage from the leafless rhizome into the shoot and the appearance at
first of small scale-leaves without vascular bundles can be followed ; this is essentially
similar to what is seen in Psilotum and Tmesipteris. The insertion, and in the case
of the living forms the development, of these leaves appear lateral. Certain features
of the shoots of the Psilotacese, especially of Tmesipteris, indicate, however, that this
is not inconsistent with an interpretation of such small leaves on the lines of their
being specialised branches subordinated to, and overtopped by, the further growth
of the shoot. There is no clear evidence of this, however, and all that can be said
in its favour at present is that the dichotomous branch system of the Rhyniacese
provides us with the sort of starting-point from which small leaf-like branches
originate in this way in certain Algse.

Between the leafy shoots of Asteroxylon and the leafless, dichotomous thalloid

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 32). 131

848 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

plant-body of the Rhyniacese there is, as has already been pointed out, a distinct gap.
We have kept our minds open to the possibility that thfe hemispherical projections
on many stems of Rhynia Gwynne- Vaughani might be regarded as indicating a step
towards the origin of small and truly lateral leaves ; this was indicated as one of
two alternatives in Part III (p. 674). There were always difficulties in such an
interpretation apparent to us. The late origin of these hemispherical projections
underneath the stomata that we have now been able to demonstrate (pp. 832-834)
seems to further weigh against entertaining it, although the specimen represented in
fig. 10 lends it some support. The only possible conclusion at present appears to
be that the Rhyniaceae afford no clear indication as to the first origin of leaves.

Another question raised by speculative morphology on which the bearing of
the Rhynie plants must be considered is as to whether a distinction should be drawn
between the small leaves (e.g. of Lycopodium ) and the large leaf or frond of such
plants as the Ferns. The clearest statement of such a view is that by Lignier, who
derived it partly from a consideration of the reconstruction of Psilophyton. He
distinguished the small leaves of Lycopods as phylloids from the specialised branch-
systems or cauloids which gave rise to the fern-frond.

On this question also the morphology of the Rhynie plants does not afford
direct evidence — though, as has been pointed out in Part III, it is not inconsistent
with such ah interpretation. The most significant fact ' provided by the newly
discovered plants is the demonstration in the Rhyniaceae of Vascular Cryptogams
with a thalloid plant-body without small leaves or phylloids. It is, therefore, not
necessary to assume that such phylloids have been lost from the branch-systems
forming fern-fronds, though this possibility is open to evidence as to whether in
particular examples it was or was not the case.

On the further question of whether a distinction is to be drawn between the
origin of relatively small leaves and of larger subordinated branch-systems the
Rhynie plants also do not give any clear evidence- On the whole their organisation
seems to weigh against drawing too sharp a distinction and to support Tansley’s
modification of Lignier’ s view.

It should be borne in mind that there is evidence in the Early Devonian period
of stems bearing definitely arranged, large, lateral appendages that can clearly be
regarded as specialised subordinated branch-systems or fronds. With fuller know-
ledge of such plants more definite evidence on the comparative problems indicated
above may be anticipated.

( c ) The Vascular System.

All the Rhynie plants had a well-developed, though simple, vascular system
with characteristically thickened tracheides forming the xylem, which was
surrounded by a tissue differing from the parenchymatous ground tissue and
corresponding to the phloem.

SHOWING STRUCTURE, FROM THE RHYJS’IE CHERT BED, ABERDEENSHIRE. 849

The structure of even the more simple Rhyniaceae in this respect differs
markedly from what has been observed in any known Algae, and no pertinent com-
parisons can be made with these plants. It is also unnecessary to do more than
mention without entering into detailed comparisons the simple strands of narrow
lignified cells found in the midrib of the thallus in a few genera of Hepaticae.

In the Rhyniaceae the structure of the simple vascular strand or stele is similar
throughout the plant. A feature of interest is the distinction of central and
peripheral xylem seen in the steles of Rhynia major and Hornea and in the stouter
strands of xylem of Rhynia Gwynne-Vaughani. Though the central xylem of
these plants is not clearly protoxylem, it is perhaps justifiable to compare the
arrangement with the centrarch construction of the leaf-traces of Asteroxylon and
many species of Lycopodium and the stem-steles of some existing and extinct
Vascular Cryptogams.

The complicated vascular system of the stem of Asteroxylon is not easy to
adequately compare with the simple vascular structure of the Rhyniaceae. This
difficulty is intimately connected with the gap between the leafless condition
of the latter and the leafy condition of the stem of the former plant which
has been emphasised in the preceding section. Another difficulty that may
be pointed out is that those regions of Asteroxylon in which the vascular
strand is simplest (rhizome, branch-traces) have a solid strand of xylem
with no indication of protoxylem. Protoxylem is clearly present in the
leaf-traces and appears in the arms of the stellate xylem of the stem-stele. The
attempt to interpret the complex xylem of the leafy stem of Asteroxylon at once
introduces the problem of the respective parts played in its construction by a cauline
strand and by decurrent leaf- traces. Plowever this may be, the resulting stele is
remarkably comparable with the stele of Lycopodium, although the type of thicken-
ing of its tracheides is quite peculiar to A steroxylon. The resemblance to the stellate
stem-steles of some Zygopteridese, though not so close, must also be mentioned. It
may be pointed out that there is also a possible comparison between the leaf-trace
“ receptors ” and the central mass of metaxylem of the Asteroxylon stele and the
receptors and the apolar connecting bar in the petiolar bundles of the Zygopteridese.
These comparisons are the more significant in the light of the various points of
resemblance between Asteroxylon and both Stauropteris and Lycopodium.

Such general questions concerning stelar anatomy are too wide to be entered
upon here, and may be left with these brief comparative remarks.

(d) Position and Nature of the Sporangia of Pteridophyta.

The Rhynie plants appear to afford some definite information on the nature
and original position of the sporangia in the Pteridophyta. In the absence of
evidence of actual origin in descent of one great group of plants from another,
we can only recognise the way in which the sporangia of Rhynia and Hornea

850 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

facilitate comparison on the one hand with the spore-bearing organs of the more
specialised Pteridophyta, and on the other hand with Bryophyta and especially
with Algae.

As regards the position of the sporangia, it is clear that in Rliynia and Hornea
they occupied the ends of some branches of the thalloid plant-body. There is
strong reason to hold that the position of the sporangia on certain branch-systems
of the more complex plant of Asteroxylon and Psilophyton was similar. If, as
appears to be the case, we are here dealing with generalised early plants representing
the precursors of the more specialised groups of Pteridophyta, it is reasonable to
expect that the position of the sporangia in the Psilophytales should be consistent
with and explain the diverse specialised positions they occupy in later derived groups.
This is as a matter of fact the case ; the position of the sporangia of the Rhyniaceae
at the tips of thalloid branches appears to fit naturally with conclusions arrived at
from other points of view by a number of investigators of the Pteridophyta.

Thus this conception applies readily to sporangiophores bearing one or more
sporangia such as are met with in what have been termed the sporangiophoric
Pteridophyta and in some others. The sporangiophores would appear to represent
the last persisting remains of the original leafless branch-systems of the Rhyniaceae.
It is unnecessary to follow this comparison into detail, but it may be pointed out
that it applies to such cases as the sporangiophores of Helminthostachys , of the
Equisetales and Sphenophyllales, in a somewhat different fashion to the fertile
appendages of the Psilotaceae, and with greater difficulty and obscurity to the
position of the sporangia of the Lycopodiales.

In the case of the Ferns there is no difficulty in the comparison of the terminal
position of sporangia on the ultimate ramifications of a branch-system with the
marginal position on a webbed cladode-system of pinnules forming a definite
flattened frond. This was suggested in detail by Lignier (taking Psilophyton as a
starting-point), and he further followed the change in position of sporangia to the
lower surface of the frond. Bower’s more recent conclusions envisage the marginal
position of the sporangia as the probable starting-point in the ease of Ferns and
support these comparisons from the side of the spore-producing organs of the higher
Ferns.

As regards the structure of the sporangia themselves, there are two very distinct
types or grades in the Rhynie plants. The fairly large sporangia referred to
Asteroxylon (which are, however, relatively small compared with those of the
Rhyniaceee) had a definite dehiscence depending on the structure of the sporangial
wall. They are thus comparable with the specialised sporangia found in most
existing and extinct Pteridophyta, and are very similar to those of Stauropteris.
The sporangia of Rhynia and Hornea , on the other hand, which in Rliynia major
were of relatively huge size, show no arrangements for dehiscence although the
epidermal layer of cells is specially thickened.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 851

A very important and striking feature of the sporangia of Rhynia and Hornea
is the indication they afford of the sporangium being the transformed and specialised
tip of a branch of the thalloid plant-body. This was shown especially in some less-
defined specimens of the sporangia of Rhynia Gwynne-Vaughani, which appeared
to have a group of spore-tetrads developed within a structure like the tip of a
vegetative branch (Part II, figs. 10, ll). It is also shown in the way*in which
the dichotomy of the tip of a branch in Hornea affects the sporangia developed
from it ; we meet with both simple and branched sporangia. While from one point
of view and as regards some specimens the sporangia of the- Rhyniaceae can be
regarded as specialised organs contrasting with the vegetative axes, from another
point of view they are clearly seen to be the modified tips of the latter.

These considerations are significant when the sporangia of the Rhyniaceae are
compared with the corresponding structures in the Bryophyta and the Algae.

As regards the Bryophyta, it is only possible to recognise that the types of
sporangia in Rhynia and Hornea facilitate comparison with certain sporogonia.
This applies to the enclosed position in which the spores are produced, protected by
a more or less massive wall, and to the presence in some cases of a sterile columella,
so that the spores lie between this and the wall. In the absence of any indication
as to whether the sporophytic generation in Bryophyta is to be regarded as reduced,
or as illustrating an independent line of progression, it seems inadvisable to pursue
the comparison in greater detail. The comparisons to be made with the Algae
may, however, apply to the Bryophyta as well as to the Rhyniaceae on which
they are based.

When the comparison is extended to the Algae it evidently bears on the primitive
nature and origin of the structures which in the Rhyniaceae and throughout the
Pteridophyta are known as sporangia. These have for long been regarded as organs
sui generis and not explained on the grounds of possible derivation wTith modification
from any part of the plant-body of the Algae. It has, however, been pointed out by
Schenck *■ that, when the comparison is extended to what we term sporangia in
various Algae, the strictly comparable structures are the spore-mother-cell and
resulting tetrad in the Vascular Cryptogams and the sporangia (tetra-sporangia, etc.)
in which reduction and spore-formation takes place in the Algae. Schenck therefore
suggested employing the name “sporotheca” for the organ usually known as the
sporangium in Pteridophyta. On this view the sporangium (or sporotheca) of the
Vascular Cryptogams (and the sporogonium of Bryophyta) would correspond to an
association of a large number of sporangia of the thallophyte type enclosed within a
special region of the plant-body.

The type of sporangium (sporotheca) in Rhynia and Hornea appears to
strengthen and support the view based on the above comparison. It has been
pointed out in Part II (p. 622) how their construction suggests comparison with the
* Engler’s Bot. Jahrb ., Bd. xlii (1908), p. 31.

8 52 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

specialised branches' of some Red Sea-weeds, in which the numerous tetrasporangia
are enclosed by a wall-like outer layer of the branch and may surround a central
axis that would correspond to the columella. The specialised tetraspore-bearing
branches are called stichidia. They not only occur as terminal organs of the thallus,
but may stand in definite relation to laterally placed vegetative branches.

Advancing knowledge of the spore-producing structures in the Early Devonian
plants may be expected to throw further light on such comparisons, which it is not
advisable to press too far at present. The facts appear, however, to fully warrant
the suggestion that the structure known as the sporangium in the Vascular Crypto-
gams may be regarded as corresponding to the tip of a branch of a thalloid sporophyte
enclosing the sunken tetrasporangia represented by the spore-mother-cells. On this
view the sporangium of the Pteridophyta would not be an organ sui generis, although
in considering its further modifications in the Pteridophyta it may be treated as if
it were a primary member of the plant.

In this section we have placed together our conclusions as to the main bearings
of the new facts obtained by the study of the Vascular plants of Early Devonian age
found at Rhynie on the comparative morphology of certain extinct and existing
plants. While it appears inadvisable to omit such a general discussion, we would
explicitly point out that we regard all such comparisons as to a certain extent
tentative. They are thus distinct from the addition to our definite knowledge of
the plants which it is the primary object of this series of papers to make known.

In conclusion a word may be said as to the bearing of the Vascular Cryptogams
in the Rhynie peat on the geological age of the deposit. That some plants of similar
morphological character exist in both the Lower and Middle Old Red Sandstone of
Scotland seems to be certain, although no clearly defined species has yet been
recorded as common to both. In the Middle Old Red Sandstone neither Psilopliyton
nor Artliro stigma, two characteristic Lower Old Red Sandstone plants, have yet
been recorded on satisfactory evidence. On the other hand, no such leafy plant as
Thursophyton Milleri, which is a most characteristic species of the Middle Old Red
Sandstone, has been discovered in the Lower Old Red. If the leafy shoots of
Asteroxylon Mackiei and those of Thursophyton Milleri are compared, taking the
probable sizes of the plants into consideration, the similarity is very striking. The
similarity is indeed so great as to suggest that they may be only different conditions
of preservation of one species. It seems at present desirable, however, to treat them
as distinct in the hope of additional light being thrown on their relation to each
other. In any case it is clear that Asteroxylon Mackiei in its leafy shoots exhibits
a close similarity to Thursophyton Milleri, and, even if not specifically identifiable
with it, represents a characteristic type of plant which so far as at present known is
absent from the Lower Old Red Sandstone. These facts taken together lead us to
believe with but little reservation that the Rhynie chert-band is almost certainly of
Middle Old Red Sandstone age.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE.

853

EXPLANATION OF PLATES.

(We are indebted to Mr J. Barrie Robinson for re-drawing our reconstructions of the vascular plants
reproduced on Plates I and II. All the figures on Plates III, IV, and V are from untouched photographs.)

Plate I.

Fig. 1. Restoration of Rliynia Gwynne-Vaughani.

Fig. 2. Restoration of Rliynia major.

Plate II.

Fig. 3. Restoration of Hornea Lignieri.

Fig. 4. Restoration of Asteroxylon Mackiei.

Plate III.

Fig. 5. Vertical section of a typical hemispherical projection from a longitudinal section of a stem of
Rliynia Gwynne-Vaughani. x 100. (No. 2514.)

Fig. 6. Tangential section of stem of Rhynia Gwynne-Vaughani bearing two hemispherical projections,
while another has been isolated by the section from the same stem and is seen on the extreme right. x 30.
(No. 2514.)

Fig. 7. Portion of specimen in fig. 6 showing the position of the stomata in relation to the two attached
projections, s., stoma. x 60. (No. 2514.)

Fig. 8. The detached projection in fig. 6 showing the stoma (s.) with its two guard-cells, x 60.
(No. 2514.)

Fig. 9. Small hemispherical projection forming beneath a stoma ( s .). From a tangential section of the
stem of Rhynia Gwynne-Vaughani. x 200. (No. 2423.)

Fig. 10. Hemispherical projection on a stem of Rhynia Gwynne-Vaughani , elongated to form a cylindrical
emergence, x 24. (No. 2515.)

Fig. 11. Longitudinal section of a stem of Rhynia Givynne-Vaughani showing a hemispherical projection
on the right, and on the left two necrosed areas, x 14. (No. 2516.)

Fig. 12. Necrosis cavity from a longitudinal section of the stem of Rhynia Gwynne-Vaughani. The
cavity is still roofed over by the cuticle and is bounded by cells that have undergone enlargement and cell-
division. x 50. (No. 2414.)

Fig. 13. Transverse section of stem of Rhynia Gwynne-Vaughani showing a necrosis cavity and the
.enlargement and division of the cells around this. x 38. (No. 2406.)

Fig, 14. Oblique section of stem of Rhynia Givynne-Vaughani showing a large necrosis cavity, the cells
around which are proliferating. x 36. (No. 2517.)

Plate IV.

Fig. 15. Portion of fig. 14 more highly magnified, showing the proliferating cells in the neighbourhood
of the xylem. x 100. (No. 2517.)

Fig. 16. Stem of Rhynia Gwynne-Vaughani with vertically elongated and dividing cells bordering on a
necrosed area, x 50. (No. 2427.)

Fig. 17. Longitudinal section of tip of stem of Rhynia major, x 28. (No. 2518.)

Fig. 18. Apex from fig. 17 more highly magnified showing the gradual enlargement of cells behind the
apical meristem. x 65. (No. 2518.)

Fig. 19. Two stems of Rhynia major in transverse section. The stem on the right shows a dark necrosed
area and in relation to this callus-like disturbance and proliferation of the tissues. The stem on the left
shows a large cavity and similar pathological changes in the tissues. x 15. (No. 2519.)

Fig. 20. Portion of the wound-tissue from the stem on the left in Fig. 19. x 100. (No. 2519.)

Fig. 21. Portion of the wound-tissue from the stem on the right in Fig. 19. x 100. (No. 2519.)

854 OLD RED SANDSTONE PLANTS FROM RHYNIE CHEBT BED, ABERDEENSHIRE.

Fig. 22. Stem of Rhi/nia Gwynne-Vaughani in transverse section showing enlargement of the cells in
certain regions as compared with others where the cells are of normal size, x 22. (No. 2520.)

Fig. 23. Asteroxylon Macldei. Portion of cuticle of a stem viewed obliquely from the surface. The
longitudinal dark lines on the epidermal cells are shown, x 100. (No. 2479.)

Plate Y.

Asteroxylon Macldei.

Fig. 24. Section of the chert showing a large rhizome bending downwards into the peat and penetrating
a large stem in the latter; the basal region of the rhizome hears small scale-leaves, x 3|. (No. 2557.)

Fig. 25. Transverse section of the bud of a leafy shoot, passing just above the apical cone and showing
the arrangement of the small leaves. x 14. (No. 2554.)

Fig. 26. Transverse section of the stem from the same series as the preceding figure passing immediately
below the apex. x 14. (No. 2553.)

Fig. 27. Portion of the section in fig. 26 more highly magnified, ep., epidermis; o.c., outer cortex ; i.c.o.,
outer zone of inner cortex; i.cm., trabecular middle zone of inner cortex ; i.c.i., inner zone of inner cortex;
l.t., departing leaf-traces ; x., xylem of the ends of the rays of the stellate xylem. x 35. (No. 2553.)

Fig. 28. Stele of a section lower in the series showing the poles of the stellate xylem still unconnected
by central xylem. x 50. (No. 2550.)

Fig. 29. Transverse section of the stem at the lowest level in the series., x 14. (No. 2547.)

Fig. 30! The stele of the section shown in Fig. 29. The central xylem is still undifferentiated except for
a few tracheides. x 50. (No. 2547.)

We again gratefully acknowledge our indebtedness to the Executive Committee of the Carnegie Trust
for a grant to defray the expense of the plates illustrating this memoir.

Trans. Roy. Soc. Edinr-

Kidston and Lang : — Plate I.

Vol. LII.

MTarlanc & Srokine, litli., Bdin.

Trans. Iioy. Soc. Edinr-

Kidston and Lang : — Plate II.

Vol. LI I.

Trans: Roy: Soc: EdinP

Vol: LI 1

Kidston and Lang — Plaie HI

Zinco Collotype Co., Eoinburch.

Kidston and Lang Photo :

Rhynia C wynne -Vaughani , K&]

Trans: Rov: Soc: EdinT

Vol: LI 1.

Kidston and Lang - Plait IV.

KlDSTON AND Lang Photo: , _ T Zinco Collotype Co.. Edinburgh.

Figs. 15, 16 - Khynia G wynne -Vaughani , K&L
Figs. 17-22 Rhynia Major, K&L.

Fig. 23. Asteroxylon Mackiei, K&L.

Trans: Roy: Soc: EdinT

Kidston and Lang:— Plate Y

Vol: LI 1.

30

24-

Kidston and Lang. Photo-.

Zinco Collotype Co., Edinburgh.

Asteroxylon Mackiei, K&L.

( 855 )

XXXIII. — On Old Red Sandstone Plants showing Structure, from the Rhynie
Chert Bed, Aberdeenshire. Part V. The Thallophyta occurring in the
Peat-Bed ; the Succession of the Plants throughout a Vertical Section of
the Bed, and the Conditions of Accumulation and Preservation of the
Deposit. By R. Kidston, LL.D., D.Sc., F.R.S., and W. H. Lang, D.Sc.,
F.R.S., Barker Professor of Cryptogamic Botany in the University of
Manchester. (With Ten Plates and One Figure in the Text.)

(Read May 2, 1921. MS. received June 28, 1921. Tssued separately November 30, 1921.)

In this concluding part of the series of papers # on the plants preserved in
the Rhynie chert-band, various remains of lower plants that occur in the peat
will be described, and some general questions concerning the accumulation and
preservation of the deposit considered. The paper is divided naturally into the
parts enumerated below.

1. An account will be given of a number of forms of Fungi which are met
with in the decayed remains of the vascular plants, and also in the peaty matrix.
The fungal remains are so generally distributed that they can be regarded as
forming an integral part of the peat. In connection with them, the question as
to whether there is evidence of a mycorrhizal relation between any of the Fungi
and the Vascular Cryptogams will be considered.

2. The occurrence of Schizophyta will then be considered and a few well-
characterised forms, that can be placed in this group, described.

3. A remarkable algal organism, to which the name Algites ( Palseonitella )
Cranii is given, will be described ; scattered remains of this have been found
at a few spots in the deposit, and we provisionally associate with these some
interesting but more doubtful remains.

4. Certain fragments that belong to an organism with the characteristic structure
of N ematophyton will be described and discussed. These have been met with
in a limited region of one block of the chert.

5. The succession of the plants, as determined by the study of a vertical series
of microscopic preparations, extending from the bottom to the top of a complete
section collected when the Rhynie deposit was exposed in situ, will be described.

6. In the light of these facts and of some others obtained from loose blocks of
the chert, the question of the conditions of accumulation and preservation of the
Rhynie deposit will be briefly considered.

More exhaustive study of the deposit will doubtless lead to the recognition of
further types of Thallophyta, and add to our knowledge of some of those described

* Part I, Trans. Boy. Soc. Edin., vol. li, p. 761 ; Part II, ibid., vol. lii, p. 603 ; Part III, ibid., vol. lii, p. 643 ;
Part IV, ibid., vol. lii, p. 831.

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 33).

132

856 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

below. It is probable, however, that the forms which have been met with in the
examination of the numerous sections prepared for the study of the vascular plants
and the composition of the deposit afford a not inadequate idea of the remains of
lower plants preserved in the Rhynie peat-bed. Their description will fittingly
complete the account of our investigation of the Botany of the deposit.

1. The Fungi of the Rhynie Deposit.

The most abundant remains of lower plants in the Rhynie deposit are clearly
Fungi, of which a considerable number of forms can be distinguished. Along
with Bacteria, which will be considered in the next section, they formed an integral
constituent of the peaty mass, as they do of recent peats.

Although the fungi are often as perfectly preserved as were the vascular plants,
they do not afford characters that are reliable as evidence of their classificatory
position. All the fungal remains to be described below consist of stout or fine
hyphse of the vegetative mycelium and various types of vesicles and resting-spores
(chlamydospores) borne on this. The mycelium is usually non-septate, although
one example of a clearly septate condition will be described.

Even when they appear to retain recognisable forms that recur in the deposit,
it is hardly possible to be sure of the specific distinctness of the various types of
mycelium with resting-spores found in the more or less decayed plant-remains and
the peaty matrix. No student of existing fungi would, we believe, feel justified
in distinguishing fungi found in a mixed growth in decayed organic matter when
he had only vegetative mycelium and chlamydospores to deal with. Distinctive
reproductive organs or constancy under culture would be required, even in the
case of Imperfect Fungi, and the investigator would be prepared for the same
species assuming a variety of forms as regards diameter of hyphse and size of
resting-spores.

It is necessary, however, to bring as much order as possible into the description
of the various fungi met with so abundantly in the fossil condition in the Rhynie
peat. The plan adopted will be to describe and illustrate the main form-types
depending on appearance, relative size, and also in part on the place of occurrence.
The possibility of distinguishing specific forms will then be considered, and some
well-chaTacterised forms will be named and diagnoses given for them.

We have deliberately ignored doubtful remains, and in all cases sought for
evidence of definite organisation, such as the attachment of vesicles or spores to
hyphse, or the structure of the spore-wall.

The fungi occur mainly in the portions of the vascular plants, but to some
extent in the matrix. In most cases they had evidently lived as saprophytes.
The question as to whether any of them may have been mycorrhizal during the
life of the vascular plants will require consideration in the general discussion which
will follow the particular descriptions.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 857

{a) Description of Form-types of Fungi recognised in the Rhyme Peat.

Fungus No. 1. (Figs. 1-4.)

The mycelium of the various types of fungi to be described below will be seen
to be, as a rule, non-septate. The specimen here described and figured agrees in
the general characters of the mycelium with some other fungi to be considered later,
but shows more or less numerous transverse walls in some, at least, of the hyphse.
It is further interesting in that the hyphse are associated in strands which, as is
not unusual, occupied the place of the phloem in a partially decayed stem of Aster-
oxylon ; the stem was cut longitudinally, and therefore shows a strand of hyphse
on each side of the xylem (figs. 1 and 2). The stout hyphse run on the whole
parallel to one another, and are branched without any relation to the position of
the septa. At places the branches form peculiar tangles or knots (fig. 2), but,
though anastomoses not improbably occurred, there are no indications of clamp-
connections. The transverse septa, the middle portion of which is often thicker
and of a darker colour than the outer walls of the hyphse, occur at irregular
intervals. They were frequent and distinct in the strand seen on one side of the
stele' (fig. l), while they were present, though less frequent, in the strand on the
other side (fig. 2). Some of the septate hyphse are shown more highly magnified
in fig. 3, and some of the non-septate hyphse in fig. 4. The same stem of Aster -
oxylon also contained the large resting-spores of the fungal type that will be
next described as Fungus No. 2. The mycelium here described as No. 1 not
improbably belongs to the fungus which bore the large resting-spores, though direct
proof of this is wanting. Similar strands of hyphse replacing the phloem in a stem
of Asteroxylon, associated with the larger resting-spores of Fungus No. 2 in the
cortex of the stem, are seen in the section represented in fig. 5.

Fungus No. 2.* (Figs. 5-11.)

This fungal type is probably the most widely distributed throughout the chert-
bed. It occurs in fragments of all the vascular plants and is associated with their
greater or less decay. It is most abundant and characteristically developed in the
stems of Rhynia major and of Asteroxylon ( cf. Part III, figs. 31, 42, 43, 96).

In its typical form it has stout hyphse, as well as finer branches. The hyphse
usually appear non-septate, but in relation to this the remarks under Fungus No. 1
should be borne in mind. They bear large oval, thick-walled resting-spores, the
wall of which consists of a number of layers and tends to exhibit a double contour.
The general appearance of the fungus as it occurs in the decayed cortex of a stem
of Asteroxylon is well shown in fig. 6, and more highly magnified in fig. 7. In
this specimen the hyphse were clearly shown, and bore smaller thin-walled vesicles
as well as the mature resting-spores ; the smaller structures were evidently stages
* Named Palxomyces Gordoni below (p. 868).

858 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

in the development of the latter. The description will be based on such a typical
growth of the fungus, in which practically everything that a wide comparison of
other specimens has . shown can be seen. *

The longitudinally running hyphse in the specimen represented in figs. 6 and 7
occurred in the inner cortex of a stem of Astevoxylon , the cellular structure of which
had practically disappeared. Though not as perfectly preserved as is often the case,
the hyphse can be clearly traced. They range in diameter from about 6 u (fig. 8)
to 16 or 24 u (fig. 9). A fine hypha sometimes widens out into one of the stouter
brown-walled hyphse (fig. 9). The hyphse, here and in other cases where they
have been seen both in the cortex and occupying the place of the phloem (cf. fig. 5)
appear non-septate.

In fig. 7 some of the hyphse widen out into oval or pear-shaped thin-walled
vesicles, collapsed specimens of which also occur. All stages intermediate between
these and the mature thick-walled resting-spores are found. The oval or spherical
resting-spores are of large size (about 240 m in major diameter, but with considerable
variation *), and usually show the double contour of the wall that is evident in
figs. 6 and 7, and to which reference, has already been made. In favourable cases
they can be seen to be borne on a hyphal stalk, often, as in fig. 9, arising as a lateral
branch from a longitudinally running hypha.

The double contour when best marked (fig. 10) is seen to depend on the presence
of a clear space separating an outer boundary layer of a brown colour from an inner
more or less thick layer. Within this, again, there may be remains of the contracted
and more or less indistinct contents. Sometimes an intermediate layer of the wall
can be distinguished in place of the clear space ; in other specimens this intermediate
region is occupied by a brownish material ; in yet other specimens the inner layer
appears to be connected with the outer layer by fine brown strands crossing the
space (fig. ll). Comparison of numerous specimens supports the impression given
by these facts that we have to do with the distinction of layers in an original wall
and the contraction of an inner layer.

The resemblance which the spores with double contour at first sight present
to an oogonium containing a single oospore is thus a superficial one. Nothing
to establish or justify such an interpretation has been found, though carefully
looked for.

Nothing could be ascertained with certainty regarding the contracted contents
sometimes present within the inner wall, and seen in the figures. The occurrence
of smaller intrusive fungi within these resting-spores will be dealt with separately
below (p. 864).

Fungus No. 3.t (Pigs. 12-14.)

It is necessary to distinguish as a separate form-type of fungus certain thick-
walled spherical resting-spores borne on stout hyphse and resembling in construction

* Thus the specimens in fig. 5 are over 350 /x. + Named Palazomyces Gordoni var. major below (p. 868).

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 859

those of the fungus described above under Fungus No. 2. These resting-spores,
specimens of which are represented in figs. 12 and 13, were about twice the size of
those of the preceding fungus, their diameter being about 500 m, though the
specimens in this case also exhibit considerable differences in size. The thick
hyphal stalk bearing the resting-spore is all that has been seen of the mycelium
of this type. Although when typically developed this is a very distinct form,
it must be kept in connection with Fungus No. 2, since specimens of resting-spores
intermediate in size have been met with (of. fig. 12 on the right). These large
resting-spores do not show the distinct double wall that is so often, though not
always, met with in the preceding form. The thick wall (fig. 14) is here best
described as distinctly stratified. There is first an outer layer, the surface of which
is brown ; then a wide, intermediate, more or less clear layer ; and lastly, an inner
brown layer. The intermediate layer is sometimes laminated or stratified (fig. 13).
These layers remain connected, but the contents, limited by a thin or thick brown
membrane, often contract more or less from the wall.

Fungus No. 4. (Figs. 15 and 16.)

The fungus represented in figs. 15 and 16 is placed here, since its resting-spores,
like those of the two preceding forms, have thick walls showing a differentiation into
layers. It occurred in a decaying portion of a stem of Asteroxylon. A moderately
fine mycelium was present in the tissue, and in relation to it were the thick-walled
oval or spherical resting-spores shown in the figures. These are evidently mature,
and, though exhibiting a rather wide range in size, the diameter of most of them
is in the neighbourhood of 120 u to 150 u. This is considerably smaller than the
spores of Fungus No. 2, from which these resting-spores also differ in the absence
of any indications of separation of the layers of the wall. A few of the resting-
spores in the group (which are not shown in the figure) are distinctly larger, and
thus narrow the interval between this form and Fungus No. 2 ; on the other hand,
some are distinctly smaller than the majority, and approach in this respect the
resting-spores to be described next.

Fungus No. 5. (Figs. 17 and 18.)

This form is only placed here on account of the relatively thick wall of the
small oval resting-spores, the major diameter of which was about 60 m. It does
not call for more detailed description in addition to its representation in figs.
17 and 18.

Fungus No. 6. (Fig. 19.)

We may place here as naturally as anywhere the rather ill-defined form of
fungus shown in fig. 19. This occurred in a decaying stem of Rhynia major, and
had the branched, non-septate mycelium, swelling into moderately small vesicles,
well preserved. Since adjoining stems of Rhynia major showed oval thick- walled

860 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

resting-spores like Fungus No. 2, it is not improbable that we have here a similar
fungus in an earlier or different condition of development. On the other hand, it
presents resemblances to the following type, in which the vesicles have never been
seen to pass into thick-walled resting-spores.

Fungus No. 7.* (Figs. 20-28.)

This fungus occurs characteristically between the cells of the inner cortex of
many of the rhizomes of Asleroxylon, and in the axes intermediate in structure
between rhizomes and stems. The region occupied is usually more or less decayed,
while the fungus seems to avoid the outer cortex and phloem. The general distri-
bution of the fungus was evident in figs. 16 and 22 of Part III.

Fig. 25 on Plate III shows the relation of the fungus to the inner and outer
cortex and its general appearance. It consists of stout and fine hyphse and vesicles.
The intercellular position of the hyphae is shown in fig. 26. Fig. 27 and fig. 20
show the hyphae and vesicles in more decayed tissue of the inner cortex. The
fungal hyphae are non-septate, only an occasional transverse septum having been
seen. The hyphae are abundantly branched (fig. 21), and sometimes connected by
anastomosing branches (fig. 22). The branches are sometimes recurved (figs. 20, 28).
The ends of many of the finer branches expand as oval or spherical thin-walled
vesicles which are usually not cut off by a septum from the cavity of the hypha
(figs. 23 and 24). The walls of the vesicles are not specially thickened, and do
not show any tendency to separate into layers or to give the appearance of a
double contour.

The fact that, as a rule, when the cell walls of the tissues are preserved the
fungus is intercellular has been emphasised above. Fig. 28 shows a superficial
cell of the outer cortex of the rhizome in fig. 25 which has been entered by the
fungus, that in this specimen was growing on the outer surface of the rhizome ;
a vesicle has formed on the fungus within the cell. Both the position and the
behaviour of the fungus in this case are, however, exceptional.

The figures show the range in diameter of the non-septate hyphse and also of the
vesicles. This makes it impossible to state a definite characteristic size. It may
be generally said that the hyphse range in diameter from about 4 n to 20 n, while
the vesicles attain a size of about 80 n.

Fungus No. 8.t (Figs. 29-35.)

This fungus is present in many of the protocorm-like rhizomes of Hornea ,
extending also into the region of the stem attached to the rhizome. It seems to
be almost always present in these positions in certain beds of the peat, while it is
completely absent from the same portions of Hornea in other beds.

The non-septate hyphse are stout (about 10-15 n) and of a dark colour (fig. 32).

* Named Palxomyces Asteroxyli below (p. 869). f Named Palxomyces Hornem below (p. 869).

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 861

They form a characteristic reticulum as they grow between the cells of the rhizome
(figs. 29, 3l). Sometimes an entering hypha can be traced giving rise to this
reticulum by branching (fig. 31). Felted masses or strands of hyphse or tangled
knot-like masses are present on the lower surface of some rhizomes among the
bases of the rhizoids (figs. 30, 33). The entering hyphse.seem to come from this
external mycelium.

In relation to the mycelium within the tissues, and sometimes outside the
rhizome, there are oval dark-walled resting-spores (figs. 29, 34). In the early
stage the spore has the form of a vesicle with its cavity continuous with that
of the hypha. The wall of the spore, though it acquires a brown colour, does
not become markedly thickened, and does not show a double contour. These
spores or vesicles measure about 100 u. Larger resting-spores of the same type,
and apparently derived from these, have been met with in the adjacent peaty
matrix (fig. 35).

Fungus No. 9* (Figs. 36-38.)

The general distribution of this fungus, which is of common occurrence in the
cortex of stems of Asteroxylon, and also enters the resting-spores of Fungus No. 2,
is well shown in fig. 36. The apparently non-septate hyphae occur abundantly
in the decaying tissue. The hyphae bear spherical resting-spores with a smooth
wall that is either thin or becomes slightly thickened. The wall of the resting-
spore, however, frequently ■ appears definitely thickened and rough, but critical
examination shows that this appearance is due to the spore having been closely
invested by a covering of the fine hyphae. These can be traced over the original
thin wall, and their tubular cavities are apparent where the invested wall is cut
in section (figs. 37 and 38). The diameter of the hyphae is about 3 u, and of the
resting-spores about 45 u.

Fungus No. lO.f (Figs. 39-41.)

This form of fungus is especially common in relation to Rhynia Gwynne-
Vaughani. It is found in the decaying stems (Part IV, fig. 13) and also in the
adjoining matrix, sometimes in a sort of exudation from the hemispherical projec-
tions. Fig. 39 shows a very characteristic mass of the fungus from the peaty
matrix derived from a broken-down stem, the remains of the xylem of which is seen
in the upper part of the figure. This figure and the portions of the same mass more
highly magnified in figs. 40 and 41 show that the fungus consists of a fine non-
septate mycelium, the hyphae of which form numerous dilatations that are some-
times intercalary, but more commonly terminal. These are cut off by septa and
enlarge into the thin-walled globular resting-spores that occur scattered or more
commonly in masses. What appears to be the same organism also forms longi-

* Named Palseomyces vestita, below (p. 869). j Named Palceomyces agglomerata below (p. 870).

862 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

tudinal rows of vesicles or spores in the cortex of stems or occupies a considerable
area of the cortex.

The hyphae are about 4 or 5 m in thickness, and the vesicles about 40 n in
diameter.

Fungus No. 11. (Figs. 42 and 43.)

The fungus which we place here resembles the preceding Fungus No. 10 in
miniature. It is not impossible that with further growth it may have come to
resemble it more closely. It is represented in figs. 42 and 43. This fungus has
been met with most commonly in decayed stems of Rhynia Gwynne-Vaughani, but
also occurs in other plants. Most frequently it appears as relatively small masses
of fine hyphae bearing small vesicles (fig. 42). As the latter figure shows, these
masses often appear as if they had been at first defined and enclosed by single cells
of the tissue and had burst out from them. It sometimes occurs in the region of
cells with dark contents immediately below the outer cortex of Rhynia (fig. 43), but
may be situated much more deeply, and even close to the stele. In some cases the
group of fine hyphae of this fungus appears to have originally occupied and grown
out from a resting-spore of Fungus No. 2.

Fungus No. 12. (Fig. 44.)

This form has much in common with that described as Fungus No. 10. It
occurred, as shown in fig. 44, in somewhat decayed stems of Rhynia Gwynne-
Vaughani, but at a slightly higher level of Bed A" 1. The moderately fine hyphse
enlarge into thin-walled vesicles, but the enlargement is usually a gradual one, so
that the vesicles are pear-shaped rather than spherical. Their diameter measured
across the widest part is rather greater than in the case of the type 10, measuring
about 60 .a.

Fungus No. 13.* (Figs. 45-47.)

The specimen represented rather inadequately in figs 45-47, owing to the thick-
ness and opacity of the section in which it occurred, we owe to the courtesy of
Mr John B. Simpson. The fungus was present in a decayed stem of Rhynia Gwynne-
Vaughani. It occurred in small oval patches, and appeared as if bursting through
the cuticle or epidermis of the stem ; the distribution might also be explained by
the patches occupying .the site of thin-walled hemispherical projections that had
decayed (fig. 45). The prominent structures of each patch of fungus are hyphae
radiating outwards and dilating into thin-walled vesicles. Even the larger vesicles
are not divided by a septum from the hypha. The vesicles may be 100 u in diameter,
but smaller and larger specimens occur in the same patch. Fig. 46 represents one
patch of fungus, while a few of the vesicles are shown more highly magnified in
fig. 47.

* Named Palseomyces Simpsoni below (p. 869).

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 863

Fungus No. 14. (Figs. 48—56.)

Fine mycelium with vesicles and resting-spores is not only met with in cells
of the parenchyma of the vascular plants, but even more commonly invading, and
developing within, the resting-spores of other fungal types. An example of this
in the case of a fungus that also occurs distributed through the decaying tissues has
already been seen in the case of Fungus No. 9, represented in fig. 36. A very
definite and recurring example of this mode of occurrence in the large resting-spores
of Fungus No. 3 is now to be described. Under the following type (No. 15) similar
intrusive fungi within the smaller resting-spores of Fungus No. 2 will be described.
The figured specimens will further afford additional examples of the resting-spores
of Fungi Nos. 2 and 3.

Fungus No. 14 was repeatedly met with in the peat of the lowest bed (A"l),
especially in relation to Rhynia Gwynne-Vaughani, and recurred among Hornea
in the upper part of the chert-bed (Bed N 20).

Specimens such as that shown in fig. 48, which were first noticed, appeared like
a distinct fungus of definite and peculiar shape. This had the form of a shortly
stalked, spherical body about ’5 mm. in diameter. The stalk was evidently formed
of numerous fine hyphse associated together. Similar fine hyphse formed a
peripheral layer of the spherical body and extended among the thin-walled spores
composing the central mass of this. The spores were spherical or oval, and could
sometimes be seen to be borne on the fine hyphae. Some of the spores are shown
more highly magnified in fig. 49.

The explanation of the definite shape of this mass of fine hyphse bearing thin-
walled spores was afforded by specimens such as those shown in figs. 50 and 51. In
the specimen in fig. 50 the group of thin- walled spores is seen to occupy the cavity
of a large resting-spore with the characteristic thick wall of Fungus No. 3. In this
case the intrusive fungus completely filled the cavity of the large resting-spore.
The spores of the intrusive fungus were less numerous and more loosely arranged
in the otherwise corresponding specimen in figs. 55 and 56.

The appearance of the stalk-like portion seen in fig. 48 is explained by the
specimen shown in fig. 51. In this the outline of the partially decayed thick wall
of the large resting-spore and of the stout hypha which bore it can be traced. The
cavity both of the hypha and the spherical spore is filled by a fine mycelium, the
spores on which are as yet undeveloped. In the light of this specimen it is easy
to understand how the specimen in fig. 48 originated by the complete decay and
disappearance of the hypha and the wall of the resting-spore that originally
determined the shape of the intrusive mass of fungus.

In other cases vesicles or spores of a similar fungus were developed on the
outside of the large resting-spore. A specimen of this nature cut through the
middle is shown from the cut surface in fig. 53. The thick wall and the contracted

TRANS. ROY. SOC. EDIN., VOL. HI, PART IV (NO. 33). 133

864 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

inner membrane can be traced. The vesicles belonging to the associated fungus are
seen covering portions of the thick wall. The same specimen seen from the other
side (that is, looking at the outer surface of the wall from which only the central
projecting portion is removed) is represented in fig. 52. The groups of vesicles are
here seen from above, and not in section. The cavity of the large spore was filled
with a fine mycelium bearing small, immature vesicles. Another specimen with the
hyphse and vesicles on the surface is shown in fig. 54.

Fungus No. 15., (Figs. 57-71.)

Many of the thick-walled resting:spores of the fungus described above as Fungus
No. 2, which, though large, are smaller than those of Fungus No. 3, are also occupied
by intrusive fungi, the mycelium of which had produced numerous spherical spores.
In so many cases these small spores can be seen to be borne on a mycelium that we
are justified in assuming that this was always the case. There is nothing to indicate
that they were in any case spores normally developed in, and belonging to, the large
thick-walled resting-spore of the original fungus. The enclosed spores may fill the
cavity more or less completely, and may also be present in the space between the
contracted inner layer and the outer layer of the wall of the large enclosing resting-
spore. They, may also burst through the wall of the latter or be formed on its
exterior. The spores of the intrusive fungus may themselves be occupied by an
intrusive mycelium bearing still smaller spores.

While the contained spores thus differ in size, it does not appear possible or
desirable to carry out a distinction of form-types on this ground, though it is
probable that we are concerned with a number of forms. An adequate idea of the
structures in question and the relative positions they occupy will be obtained by
referring to a number of examples that are represented in figs. 57-71. These
examples could have been considerably extended without, however, substantially
adding to our knowledge.

In fig. 57 three resting-spores of Fungus No. 2 are seen in decayed tissue of
Rhynia major. There is no intrusive fungus to be seen in the spore on the right,
but the other two are loosely filled with thin-walled spores. There were only slight
remains of the hyphse which bore these. The middle specimen is more highly
magnified in fig. 58. The specimen represented in fig. 59 is similar ; it further
shows clearly the presence of spores of the intrusive fungus between the layers of
the wall of the large resting-spore. The thick-walled spore in fig. 60 is large for
Fungus No. 2, and small for Fungus No. 3. It is filled with spores of the intrusive
fungus, between which the fine mycelium could be seen, though it is not brought out in
the figure. The specimen in fig. 61 similarly shows the spores of the intrusive fungus,
some of which are more highly magnified in fig. 62. In all these cases the thin-walled
spores approach in size those of Fungus No. 14. This is evident on comparing figs. 49
and 56 with figs. 60 and 62, the magnification in each case being 250 diameters.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 865

The resting-spores of Fungus No. 2 shown in fig. 63' contain an intrusive
fungus, the mycelium of which bears spherical spores of a smaller size. Two of the
specimens from the group in fig. 63 are more highly magnified in figs. 64 and 65.
These figures show that the small spores exhibit differences in size. They also show
clearly that the spores are borne on a fine mycelium which may be traceable for a
considerable length, or may only remain like a tail or appendage to the spore.
Similar remains of the mycelium attached to small spores from another specimen
are seen in fig. 66.

The large resting-spore represented in fig. 67, on the other hand, contains a
number of rather large thin-walled spores. Within these, again, are small spores that
are seen to have belonged to an intrusive fungus by the remains of hyphse attached
to some of them (fig. 68).

The intrusive fungus with it’s thin-walled vesicles or spores completely fills the
oval resting-spore in fig. 69. At one point hyphse ending in vesicles have burst
through the enclosing wall of the large resting-spore (fig. 70). The behaviour of
this specimen may be compared with that of the fungus described under No. 11.

Lastly, as in the case of the intrusive Fungus No. 14, related to the larger
resting-spores, the thin-walled spores may be formed on the exterior of these
resting-spores of type 2. An example of this condition is shown in fig. 71.

Some Other Figured Specimens.

Small spores similar to those described above as occurring within large resting-
spores are also met with distributed through the more decayed fragments of the
vascular plants. As a rule, the mycelium is not to be recognised. No advantage
is to be gained by detailed descriptions of such remains, which, in some cases,
probably belong to forms recognised above. It will be sufficient to illustrate one
or two examples.

The small resting-spores shown in figs. 72 and 73 were well defined and rather
thick-walled. They were scattered through a decayed stem of Rhynia Gwynne-
Vaughani. The still smaller dark spores shown in figs. 74 and 75 were similarly
distributed through amorphous decayed stems of Rhynia major. Among them
(fig. 74) were a few larger spherical thin-walled spores. Within these latter there
were often some four spore-like bodies without any indication of mycelium. An
example is shown in fig. 76.

Spore-like bodies from another decayed fragment of tissue are represented in
figs. 77 and 78. The appearance of these is rather different from the well-defined
fungal spores, and suggests comparison with small cells of Algse or Cyanophycese,
forming an irregular colonial mass. Such a possibility, though there is no ground
for a definite conclusion, must be borne in mind in considering a number of the
less-defined remains in relation to which no mycelium can be demonstrated.

This applies, for instance, to the spherical colonies of small dark cells which are

866 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

represented in fig. 79. In this case also it is only possible to point out the resem-
blance to some Protococcacese and Cyanophyceae, for the peculiar appearance might
result from partial decay of a group of fungal spores preceding the preservation.

Though placed for convenience with the fungi, these specimens must be regarded
as incertse sedis.

The peculiar spore-like body represented in figs. 80a, 80b, occurred in the decayed
cortex of a stem of Asteroxylon. The brown wall of the spore on surface view
(fig. 80a) showed a reticulate thickening. In section the wall appeared covered by
short thin-walled tubes standing at right angles to the surface (fig. 80b). The
bases of these tubes seen in optical section give the appearance of the reticulum
in the first figure.

The fungi so far described have, as a rule, been situated within the decaying
tissues of portions of the vascular plants. Some of them, especially resting-spores,
are also met with isolated in the matrix by decay of the surrounding tissues, while
type 10 appears to have grown and increased in amount in the amorphous matrix
adjoining decaying stems. Hyphee of fungal mycelium are also met with, branching
in the matrix in a fashion that indicates their growth through it rather than their
isolation by decay of portions of the vascular plants. No specially characteristic
forms of fungi have, however, been met with in the matrix only.

The very well-defined dark spores of medium size and distinctly borne on hyphse,
shown in fig. 81, occurred in the peaty matrix. Somewhat similar specimens have,
however, been noted within the decaying tissues.

( b ) Distinction and Description of Species of Fungi.

The examples described and illustrated above fairly represent the forms of fungi
which have come to our notice in studying sections prepared from the Rhynie chert.
A survey of them shows that they all consist of branched mycelium bearing terminal
or intercalary vesicles or resting-spores. The mycelium is usually and typically
non-septate. The hyphse show a considerable range in diameter, and the vesicles
or resting-spores differ widely in size and in the characters of their wall. It must
remain an open question to what extent the form-types distinguished above are
specifically distinct. The differences in thickness of. hyphse and in size of resting-
spore may have been affected by the conditions of life, and perhaps by the occurrence
of the fungi in parts of different plants. It does not appear advisable to multiply
species or specific names on the imperfect differential characters afforded by fungal
remains without any reproductive organs other than chlamydospores. On the other
hand, some of the forms are evidently distinct and maintain their characters
wherever met with in the deposit. Some are further of interest on account of
their regular association with the remains of particular plants or parts of plants.
We propose to adopt a middle course and, while not giving names to all the forms
recognised above, to distinguish some of the most characteristic of these as species

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 867

of the genus Palseomyces. This generic name is employed in the comprehensive
sense suggested by Seward.# Some of the other types figured may be grouped, on
grounds of general resemblance that does not imply specific identity, in relation to
those few to which we venture to give names.

The forms which we thus propose to name may be indicated by reference to the
numbers used in the preceding descriptions. The specific characters can then be
summarised in diagnoses.

The name Palseomyces Gordoni will be given to the well-marked and widespread
form described as Fungus No. 2 above. The specific name is given to mark the
association of Professor W. T. Gordon with the study of the Rhynie chert. As
recorded in Part III, we owe to him the unique block of chert containing what we
assume to be the reproductive organs of Asteroxylon. Under this name also, as
P. Gordoni var. major , the form with similar but larger resting-spores, described
as Fungus No. 3 above, may be placed. We may further associate with this species
the forms which we do not name, but have distinguished as Fungus No. 1 (hyphal
strands) and Fungi Nos. 4 and 5 (thick-walled resting-spores of smaller size).

The name Palseomyces Asteroxyli will be given to Fungus No. 7, which is so
commonly and typically present in the more or less decayed inner cortex of the
rhizomes of Asteroxylon. Some other forms with fairly stout hyphse enlarging into
vesicles ( e.g . Fungus No. 6) may perhaps be associated with this without placing
them under the same name.

On similar grounds, both of structural features and of regular position of occur-
rence in some rhizomes and stems of Hornea, the form described as Fungus No. 8
above may be named Palseomyces Hornese.

There is considerable variety in the forms with relatively small resting-spores
or vesicles borne on fine hyphse. We shall only distinguish and name three
of these.

Fungus No. 9 will be named Palseomyces vestita. With this we might associate
forms that are occasionally met with, and have their larger thin-walled spores
sometimes covered in the same fashion by fine hyphae.

The peculiar and insufficiently known form with hyphae bursting the surface
and terminating in associated pear-shaped vesicles (Fungus No. 13) will be named
Palseomyces Simpsoni, after Mr Simpson, who noticed it and kindly sent us his
preparation for study.

The name Palseomyces agglomerata is given to Fungus No. 10 as the most char-
acteristic and repeatedly occurring form with fine hyphae and thin-walled spherical
resting-spores. With this may be associated more or less closely, but without
bringing them under the same specific name, Fungus No. 11 and Fungus No. 12.
Further, the various types of fine mycelium described as Fungus No. 14 and Fungus
No. 15, and found as intrusive fungi in the large resting-spores of Palseomyces

* Fossil Plants, vol. i, p. 222.

868 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

Gordoni and P. Gordoni var. major , have much in common with the type of
Palseomyces agglomerata, and may be placed beside it without prejudice to their
probably including a number of distinct kinds.

Brief diagnoses of these named forms may now be given before entering upon
some general considerations on the Rhynie fungi.

The generic name Palaeomyces # is employed here as a “ useful and comprehensive
designation ” under which to place fossil fungi, of the precise systematic position
of which there is insufficient evidence. It is, therefore, not further defined, but
the fungi from the Rhynie deposit that are described below as species of this genus
agree in their usually non-septate mycelium bearing spherical or oval vesicles or
resting-spores.

In the case of each species a particular slide in the Kidston collection is cited
that can be regarded as the type specimen. Most of the species occur more or less
frequently in preparations of the Rhynie chert.

Palseomyces Gordoni , Kidston and Lang, n.sp. (PI. I, figs. 6-11. Slide No. 2522.)

Hyphse usually non-septate (in some cases probably septate) ; branched ; branch-
ing without relation to the septa, if present ; ranging in diameter from about 6 n
to 24 m.

Resting-spores developed from vesicular swellings on hyphse ; usually terminal ;
wall thick, differentiated into layers, and usually by contraction of the inner layer
giving a double contour to the spore as seen in section. Mature resting-spores show
considerable variation in size, but average about 240 n by 210 m. Occurs in the
more or less decayed tissues of Asteroxylon, Rhynia, and Jlornea.

Locality. — Muir of Rhynie, Aberdeenshire.

Horizon. — Old Red Sandstone. (Not younger than the Middle Division of the
Old Red Sandstone of Scotland.)

Palaeomyces Gordoni var. major, Kidston and Lang. (PI. II, figs. 12 and 13.

Slides Nos. 2527, 2479.)

Spherical resting-spores ; variable in size ; up to 550 n in diameter ; wall about
50 m in thickness, stratified ; borne on stalk-like portion of hypha about 50-60 n
in diameter.

Occurs in the more or less decayed tissues of Rhynia Gwynne-Vaughani and
other plants, and sometimes isolated in the peaty matrix.

' Locality. — Muir of Rhynie, Aberdeenshire.

Horizon. — Old Red Sandstone. (Not younger than the Middle Division of the
Old Red Sandstone of Scotland.)

* Seward, loc. cit.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 869

Palaeomyces Asteroxyli, Kidston and Lang, n.sp. (Pis. II and III, figs. 20-28.

Slide No. 2525.)

Non-septate, branched hyphse ; branches sometimes recurved ; ranging in
diameter from 4 m to 20 n. Vesicles borne terminally on some of the branches ;
not separated from the hypha by a wall ; attaining a diameter of about 80 m.

Occurs between the cells of the inner cortex of the rhizome and of the axes
intermediate between rhizomes and shoots of Asteroxylon Mackiei. Also in the
more or less decayed tissues of the cortex, and occasionally on the surface of the
rhizome in the matrix.

Locality. — Muir of Rhynie, Aberdeenshire.

Horizon. — Old Red Sandstone. (Not younger than the Middle Division of the
Old Red Sandstone of Scotland.)

Palaeomyces Horneae, Kidston and Lang, n.sp. (PI. Ill, figs. 29-35.

Slides Nos. 2531, 2532.)

Non-septate, branched hyphse often with brown walls, attaining a diameter of
10-15 n ; present in the intercellular spaces of the tissues and also in the adjoin-
ing matrix, from which entering hyphse may be traced. Spherical or oval resting-
spores ; about 100 n in diameter ; with, at maturity, moderately thick, brown
walls not differentiated into layers.

Occurs in relation to the surface, and between the cells of some rhizomes and
the basal regions of stems of Hornea Lignieri.

Locality. — Muir of Rhynie, Aberdeenshire.

Horizon. — Old Red Sandstone. (Not younger than the Middle Division of the
Old Red Sandstone of Scotland.)

Palaeomyces vestita, Kidston and Lang, n.sp. (PI. IV, figs. 36-38.

Slide No. 2454.)

Fine non-septate hyphse about 3 ^ in diameter ; bearing spherical resting -spores
about 45 m in diameter ; wall of spore thin, but frequently invested by a close
covering of the slender hyphse. Similar spores of larger size are also met with.

Occurs in decaying tissues, especially of stems of Asteroxylon.

Locality. — Muir of Rhynie, Aberdeenshire.

Horizon.- — -Old Red Sandstone. (Not younger than the Middle Division of the
Old Red Sandstone of Scotland.)

Palaeomyces Simpsoni, Kidston and Lang, n.sp. (PI. IV, figs. 45-4 7.

Slide No. 2568.)

Characterised by the peculiar grouping of the stalked vesicles, which burst
through the surface of a partially decayed stem in oval patches ; vesicles widen out

870 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

gradually from their hyphal stalks. Diameter of hyphal stalk, up to 24 n. Trans-
verse diameter of vesicle, up to 100 n.

Occurred in a decayed portion of a stem of Rhynia Gwynne-Vaughani.

Locality. — Muir of Rhyme, Aberdeenshire.

Horizon. — Old Red Sandstone. (Not younger than the Middle Division of the
Old Red Sandstone of Scotland.)

Palseomyces agglomerata, Kidston and Lang, n.sp. (PI. IY, figs. 39-41.

Slide No. 2533.)

Fine hyphse attaining a diameter of 5 m, with terminal or intercalary dilatations
giving rise to spherical thin-walled resting-spores 40-50 m in diameter, that often
occur in irregular masses.

Occurs in the tissues of all the vascular plants of the Rhynie deposit, but
especially in stems of Rliynia Gwynne-Vaughani and in the matrix in connection
with them.

Locality. — Muir of Rhynie, Aberdeenshire.

Horizon. — Old Red Sandstone. (Not younger than the Middle Division of the
Old Red Sandstone of Scotland.)

From the systematic point of view the fungi which have been described above,
and to some of which specific names have been given, add little to our knowledge.
While providing beautifully preserved and abundant examples of mycelium and
resting-spores in the decaying parts of plants in the silicified peat, their type of
construction is, on the whole, similar to those fungi already known from Carboni-
ferous rocks. Their Early Old Red Sandstone age, however, gives them an
increased interest.

The general characters of these fungi suggest a systematic position with the
Phycomycetes, and probably (though there is no definite evidence for this) with
the Oomycetes. The mycelium seems to have been typically non-septate. The
one example in which transverse septa were fairly frequent and well marked
appears exceptional among the numerous specimens examined, and this mycelium
resembles in other respects the non-septate mycelium found in a corresponding
position. The formation of septa, in addition to those formed in the development
of reproductive organs, is well known to occur among existing Phycomycetes, and
we do not feel justified in laying stress on this character by itself. While
characteristic asexual or sexual reproductive organs are wanting, the type of
resting-spores (chlamydospores), which, when immature, often appear as terminal
vesicles, can be. best paralleled among existing Phycomycetous fungi.

The only systematic conclusion that we appear to be warranted in drawing
from the rich assemblage of fungal mycelia, vesicles, and resting-spores in the
Rhynie peat is that the occurrence of Phycomycetes already established for the

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. -871

Permo-Carboniferous period can be extended back to the Early Devonian. There
is no satisfactory evidence of the existence of Eumycetes, although, in the light
of the demonstration of septate mycelium, the question of their presence must be
left to some extent open.

Since the fungi of the Rhynie peat show only hyphse, vesicles, and resting-
spores, any exhaustive comparison with other fossil and living forms, would be of
little value. It is only necessary to make clear some general lines of comparison.

As already pointed out, these fungi, though more abundant and perfectly pre-
served, resemble the fungal remains known from Permo-Carboniferous rocks.
Some of these have had names given to them, while the features of others are
sufficiently known by descriptions and figures. Attention has been directed by
a number of investigators to the hyphse and resting-spores found in connection
with decaying vegetable remains in the coal-balls. The Rhynie fungi are similar
in general type to these. Thus the groups of spores developed from the intrusive
mycelium in the large resting-spores of Palaeomyces Gordoni and Palaeomyces
Gordoni var. major may be compared with the fungal spores filling certain mega-
spores of Lepidodendron as described by Williamson.* A portion of hypha
attached to one of the spores of the intrusive fungus is indeed shown in
Williamson’s drawing.!

A number of forms of palaeozoic fungi which have been described and named
by various observers are collected by Meschinelli,^ to whose work it will be
sufficient for our purposes to refer. Leaving aside more problematical remains,
the general similarity of the forms described by Meschinelli as Phycomycetes, under
the names of P alseomy cites gracilis, Palseomycites majus , Peronosporites antiquarius,
Protomycites protogenes, and Oochytrium Lepidodendy'i, to some of the Old Red
Sandstone fungi described above will be evident. In particular, we may point out
the resemblance between the hyphse and vesicles or spores of Peronosporites
antiquarius and those of Palaeomyces Asteroxyli and Palaeomyces Gordoni ; between
Palseomycites protogenes and Palaeomyces Gordoni, as regards the double contour
of the spore ; and between the fine hyphse attached to small spores of Oochytrium
Lepidodendri and the intrusive mycelium and spores of some of the fungi described
under No. 14 above. The remains of fine hyphse forming appendages to the spores
are characteristically similar in the latter case.

It is not easy, in view of the merely vegetative growth-forms in which the fungi
of the Rhynie peat occur, to go far in comparing them with existing fungi. As
already pointed out, it is impossible in the absence of distinctive reproductive organs
even to refer them to a particular group of the Phycomycetes. The characters of
the mycelium and resting-spores suggest comparisons with Saprolegniacese, Pythiacese,
and Peronosporacese. In various forms of these (and in other fungi) the mycelium is

* Annals of Botany, vol. i (1888), p. 315. t Log. cit., p]. xviii, fig. 15.

t Fungorum Fossilium Omnium , Iconographia, Vicetiae, mcmii.

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 33).

134

872 DR R. KIDSTON AND PROF. W. H. LAND ON OLD RED SANDSTONE PLANTS

known to develop vesicular dilatations, and thin or thick-walled gemmse or resting-
spores, though rarely in the profusion shown by these extinct fungi.

A specially instructive comparison may be made with the fungus distinguished
by C. West as Stigeosporium* which occurs in the living roots of a number of
Marattiacese. The non-septate hyphse of varying degrees of thickness, the formation
of vesicles and their transformation into thick-walled resting-spores, the differentia-
tion of the wall of which shows an approach to a double contour, are points of
resemblance. West infers, although only vegetative mycelium and chlamydospores
were available, a position in the Oomycetes, and compares the resting-spores of
Stigeosporium with those described for some species of Phytophthora.

Terminal or intercalary vesicles and thick-walled resting-spores borne on non-
septate hyphse are also met with in a number of other, fungi which are only known
as endotrophic mycorrhiza.f The further fate of such vesicles or chlamydospores is
unknown, even in the case of these existing forms, while in them also other repro-
ductive organs are wanting. It is, therefore, unnecessary to enter into detailed
comparisons, and a general reference to such mycorrhizal fungi will be sufficient.

The comparisons last made naturally lead to the consideration of the question
whether there is any evidence of an association between the fungi and the Vascular
Cryptogams in the Rhynie peat suggesting mutualistic symbiosis or mycorrhiza.
This possibility cannot be dismissed, although it is difficult to obtain evidence that is
not open to other interpretations in dealing with the more or less decayed plant-
remains in the deposit.

It appears to be beyond doubt that many of the fungi were living as saprophytes
and are associated with the decay of the tissues in which they are found. This is
clearly the case with the strands of hyphse replacing the phloem of Asteroxylon and
with such fungi as Palseomyces Gordoni that occur in various parts of the different
vascular plants. It also holds for others like Palseomyces agglomerata, Palseomyces
vestita, and most of those to which names have not been given, although they may
not be so generally distributed. There remains the possibility, however, that some
of the plants contained a mycorrhizal fungus when they were alive, and further that
some of the fungi that were evidently living as saprophytes in the peat might have
come from mycorrhizal fungi taking on a saprophytic mode of life after the death of
the plant-tissues.

Certain fungal types are regularly found in association with particular vascular
plants. Thus Palseomyces Hornese is found in a number of specimens of the rhizomes
and the basal region of the stems of Hornea. The hyphse are intercellular, and
their connection with fungus outside the rhizome has been followed. The fact that
other well-preserved rhizomes of Hornea are without any trace of this fungus is,

* Ann. Bot., vol. xxxi (1917), p. 77.

t For numerous figures cf. Janse, Ann. Jard. Bot. Buitenzorg, xiv (1897), p. 53, and various descriptions of
mycorrhizal fungi, especially those occurring in Lycopodiaceae and Ophioglossacese.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 873

however, against interpreting the relation as mycorrhizal. The evidence in this case
is on the whole more consistent with a saprophytic fungus selecting for its growth
certain tissues or regions of a plant.

The fungus that has been named Palseomyces Asteroxyli is very regularly found
in the inner cortex of rhizomes of Asteroxylon and of the basal regions of stems with
the transition structure. Some rhizomes, especially finer branches, are, however, met
with apparently free from the fungus. The localisation of the fungus to the inter-
cellular spaces of the broad inner cortex, leaving both the outer cortex and the
phloem free, is in support of the view that it was mycorrhizal. It must be borne in
mind, however, that a similar regional preference has been seen in the case of other
fungi in the inner cortex of stems of Rhynia Gwynne-Vaughani and R. major,
where these occasional inhabitants were clearly saprophytic. The regularity of the
association in Asteroxylon and its occurrence in the subterranean regions of the plant
appear to justify us in keeping open the question of its having been normally asso-
ciated with the plant when alive. In this case also, although the evidence of a
mycorrhizal relation is considerably stronger, it does not amount to proof, and it is
clear that the fungus is often associated with the breaking-down of the inner cortex
of the rhizomes ; its original intercellular distribution is then, of course, lost.

The case of the two species of Rhynia is of a rather different nature.* There is
no doubt that the active growth of various fungi with evident mycelium bearing
vesicles or resting-sphores was here saprophytic. The distribution and appearance
of the layer of cells with very persistent dark contents immediately below the outer
cortex suggests, however, the possibility that this region might have contained a
symbiotic organism. The distribution of this layer is shown for Rhynia major in
Part I, figs. 21 and 22, and for R. Gwynne-Vaughani in Part I, fig. 59. Against
such an interpretation of the contents is the fact that this layer is marked in the
stouter stems, and not in the rhizomes. The contents are not evidently mycelial
when carefully examined. In some cases, however, appearances have been noted
that would be consistent with the presence of a very fine and closely packed
mycelium. Figs. 82-84 are of such a specimen, and to some extent indicate the
appearance and the difficulty of coming to a definite decision on the question. It is
suggestive that such fungal growths as that shown in fig. 43 often, though not
always, originate in the position of this layer of cells (cf p. 862) ; this would be equally
consistent with the evident fungus being a further development of a mycorrhizal
organism, or being a purely saprophytic fungus selecting for its growth this peculiar
layer of cells. Thus in the case of Rhynia also the only conclusion at present seems
to be that proof of the existence of mycorrhiza is wanting, though there are grounds
for further enquiry into the question.

As already mentioned, the inference that most of the fungi in the deposit were
living as saprophytes seems to be well justified. It is difficult to draw any sharp line

* Hornea showed corresponding structures to those to be described, but the preservation was poor.

874 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

between the occurrence of these fungi and the other cases which at first sight appear
open to the interpretation of a mycorrhizal association. In all of the latter the
evidence appears to be much weaker than was the case for the relation of some
palaeozoic plants to fungi w^hich has been interpreted as probably symbiotic.* Only
convincing evidence from structural relations is admissible in a physiological question
such as that under discussion. The appearances in the Rhynie plants do not afford
such evidence, although they leave open the possibility that a mycorrhiza may have
been present in the actively living plant.

2. SCHIZOPHYTA.

A few remains that belong to either the Bacteria or the Cyanophycese are placed
together in this section, on account of the impossibility of determining whether
certain filamentous forms were green or colourless during life.

Schizophyta No. 1. Unicellular Bacteria. (PI. VII, fig. 85.)

There is no reason to doubt that unicellular Bacteria were present, often in
abundance, throughout the partially decayed mass that is preserved in the Rhynie
chert. The finely granular nature of the matrix, however, renders their recognition
problematical. This applies even to appearances strongly suggesting colonial masses
of Bacteria that occur free or coating vegetable fragments or fungal hyphse in the
peaty matrix. Such appearances closely resemble the mode of occurrence of
Bacteria in some processes of decay at the present time.

One illustrative example is represented in fig. 85. The irregularly lobed masses
there shown occurred in groups in the matrix of a block of the chert that contained
the algal remains to be described in the following section. They are so like the
well-defined bacterial colonies or zoogloeal masses that appear when Algae are decay-
ing in fresh or salt water that they may reasonably be regarded as of this nature.
The granular character would be due to the unicellular bacteria. For comparison
unstained colonies of such a recent bacterium are represented to the same scale
in fig. 86.

Schizophyta No. 2. (PI. VIII, figs. 87 and 88.)

The organism represented in figs. 87 and 88 occurred most abundantly in certain
blocks of the chert at the upper limit of the Rliynia Gwynne-V aughani peat that
formed the base of Bed A'T. It there formed a loose felt of fine tubular filaments,
about 2 n in diameter. Groups of fungal spores are entangled in the meshes of the
felted mass, but seem to have nothing to do with the organism in question. The
filaments might be slender fungal hyphee, but they have not been observed to branch,
and their curvature and arrangement is much more suggestive of the growth of one
of the Trichobacteria than of a fungus. They may represent the sheaths of one of
* Weiss, Ann. Bot., vol. xviii (1904), p. 255 ; Osborn, ibid., vol. xxiii (1909), p. 605.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 875

the Filamentous Bacteria. No satisfactory evidence of septation or of the presence
of bacterial cells within the tubular sheath has been obtained. Under certain con-
ditions of preservation the latter contains dark granules, but the arrangement of
these is too irregular to afford a safe indication of any original structure.

The organism is so well marked in appearance and mode of occurrence that it is
recorded and figured here, but, as pointed out above, its systematic position must be
regarded as somewhat uncertain. It is. suggested below (p. 876) that it may be
referred to provisionally by the name Archseothrix contexta.

Schizophyta No. 3. Archseothrix oscillatoriformis. (PI. VIII, figs. 89 and 90.)

A remarkably well-preserved organism that can only be placed in the Schizophyta
is represented in figs. 89 and 90. It was found in a large cavity due to decay in a
stem of Rhynia Gwynne-Vaughani in the form of a number of slender unbranched
filaments. Two groups of these are seen at a and b in fig 89, and above group a
a long undulating filament is very clearly shown. A portion of this latter filament
is represented more highly magnified in fig. 90. The way in which some of the
filaments were curved back on themselves contrasted with the behaviour of fungal
hyphse, some of which were also present in the cavity. The filaments (fig. 90) were
divided into numerous short discoid cells, the protoplasts of which have persisted
and been preserved.

The structure of this remarkable organism may be most closely compared with
such existing plants as slender forms of Oscillatoria ; it agrees with these in the
short disc-like cells and in the curvature of the unbranched filaments. It is, how-
ever, also comparable with the slender colourless filaments of such existing Bacteria
as Beggiatoa, and it is generally recognised that there is much in common between
the Oscillatoriacese and the Beggiatoaceae. Though it is, of course, impossible to tell
whether the fossil plant was green or colourless when alive, its systematic position
appears to be clear. It is noteworthy that in this specimen it is the protoplasts or
cell-contents that have persisted — a mode of preservation that is unlike that of the
fungi in the same bed of the deposit. This feature as well as the general appear-
ance seems to be in favour of regarding this organism as one of the Cyanophycese.

Archseothrix , n.g., Kidston and Lang.

Slender unbranched filaments suggesting close comparison with Filamentous
Cyanophycese or Filamentous Bacteria.

Archseothrix oscillatoriformis, n.sp., Kidston and Lang. (PI. VIII, figs. 89 and 90.

Slide No. 2542.)

Unbranched filaments, often curved ; about 3-4 v- in diameter. Within a sheath-
like outer wall are the preserved protoplasts of the small discoid cells.

876 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

Occurs in a cavity of a decayed stem of Rhynia Gwynne-Vaughani in the lowest
bed (A'T) of the deposit.

Locality. — Muir of Rhynie, Aberdeenshire.

Horizon. — Old Red Sandstone. (Not younger than the Middle Division of the
Old Red Sandstone of Scotland.)

The filamentous organism described as Schizophyta No. 2 above, may provisionally
be placed under the same generic name as Archasothrix contexta.

ALGITBS, Seward.

1894. Algites, Seward, Catal. Mesoz. Plants in British Museum, pt. i, p. 2.

3.- Algites (Palaionitella) Cranii, n.sp. (PI. IX, figs. 98-104.)

We give this name to some fragmentary remains of a plant that was clearly an
Alga. While, in the absence of distinctive reproductive organs, it is impossible to
determine its systematic position, certain characters of the vegetative organs will
be seen to present resemblances to those found in existing Characese.

The name is applied in the first instance and the species is founded on specimens
of septate filaments, occasionally with evidence of the existence of nodal discs of
small cells, which have been met with in a few blocks of the chert. These remains
are of the type represented in figs. 98-104 and text-figs. 6-11.

We associate with these remains some remarkable specimens of more uncertain
nature found in a loose block of the chert by the Rev. Mr Cran of Skene in prepara-
tions he had made while studying the animal remains in the chert. Mr Cran
generously handed his preparations and drawings to us for description, and we have
great pleasure in naming after him the undoubted algal remains mentioned in the
preceding paragraph. The actual specimens he discovered (figs. 91-97 and text-figs.
1-5) are not treated as the type but, for the present at least, are regarded as to
some extent incertae sedis ; they are, however, with some confidence associated with
Algites Cranii.

In the same way we provisionally place in association with this species some
peculiar globular or oval vesicles and filaments accompanying these (figs. 105-108).
These are of interest and require to be recorded, but their nature is still uncertain.
They were obtained from blocks removed from the bed in situ.

The type specimens will first be described, and then these specimens of more or
less uncertain affinity, but of morphological interest, will be dealt with.

The general appearance of a portion of a section showing a characteristic group
of the remains of Algites Cranii in the peaty matrix is represented in fig. 98. The
prominent structures are stout septate filaments with dark contracted remains of the
protoplasmic contents of the cells. Some of the filaments are more highly magnified
in fig. 99. The tip of a filament is sometimes seen as at a in fig. 98 ; it ends in a

SHOWING STRUCTURE. FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 877

dome-shaped cell, and was evidently of limited growth (cf figs. 99 and 100 and text-
fig. 6). No branching of the simply septate filaments has been seen. In some
specimens, however, as at b in. fig. 98 (cf fig. 101 and text-fig. 7) a node of small
cells is found in the course of a filament, and there are indications of the nodal cells

Figs. 6-11 show details of specimens of Algites ( Palxonitella ) Cranii ; Figs. 1-5 are of some specimens received from
Mr Cran which we place with this plant. Cell-walls shown in another focal plane are dotted.

1. Outline of the verticillately branched axis represented in figs. 91-93. After Mr Cran’s drawing, x 155.

2. Outline of the specimen represented in fig. 95. After Mr Gran’s drawing, x 125.

3. Outline drawing of portion of the specimen represented in fig. 96. X 165.

4. Outline drawing of the specimen represented in fig. 94. xl6f>.

5. A specimen that may be a longitudinal section of a node similar to those in text-figs. 1 and 4. After

Mr Oran’s drawing. xll5. Slide 2573.

6. Apex of filament shown in fig. 98 at (a), x 165.

7. Outline drawing of nodal disc shown in fig. 98 at ( b ) and in fig. 101. x 165.

8. Outline drawing of another node from the same slide, x 165.

9. Outline drawing of the rhizoid node shown in fig. 103. x 165.

10. Outline drawing of the rhizoid node shown in fig. 104. x 165.

11. Outline drawing of another rhizoid node. X165. Slide 2546.

growing out as branches. In text-fig. 8 a second example of a shallow nodal disc
separating two of the large cells is represented. Another case of the existence of a
node of small cells accompanied by branching is seen at c in fig. 98 ( cf. fig. 102) ;
there are certain resemblances in this specimen to the rhizoids to be described

878 DR R. KTDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

below. Lastly, at d in fig. 98 there is a thin-walled globular vesicle, unfortunately
ill-preserved, but apparently attached to a filament. Attention is directed to it,
since similar structures occur in the neighbourhood of filaments of the Alga in
several blocks of the chert from different levels in the deposit, and globular vesicles
are found in the groups of remains described below.

Traversing the' peaty matrix of the block from which fig. 98 was obtained, several
examples of structures which we can only interpret as rhizoids were met with. Two
of them are represented in figs. 103 and 104 and text-figs. 9 and 10. The long
tubular cells are separated from one another by an oblique septum, the region of
junction being more or less enlarged. In relation to the septum an irregular node
of small cells was usually developed (fig. 104). These structures present a general
resemblance to the peculiar rhizoids of the Characese. Another example is
represented in text-fig. 11.

We feel no hesitation as to the vegetable nature of the remains described above,
nor as to their belonging to an Alga. The simple filaments might be comparable to
a number of Algse, but the presence of occasional nodal discs and the appearance of
the rhizoids strongly suggest more special comparison with the Characeae.

Remains 'provisionally associated with Algites Cranii.

(PI. VIII, figs. 91-97 ; PI. IX, figs. 105-108.)

As mentioned above, we received certain remarkable specimens from the Eev.
W. Cran, B.D., of Skene. These occurred in the cloudy matrix of a loose block of
the chert containing Rhynia Gwynne-Vaughani and animal remains.

The first specimen shown to us by Mr Cran consisted of a small axis embedded
in the matrix of a rather thick section. The thickness had the advantage of -leaving
the small “ shoot ” intact, but made observation and figuring difficult. The morpho-
logy of the specimen is well shown in Mr Cran’s drawing reproduced in text-fig. 1.
This shows an axis formed of four internodal cells separating as many nodes ; these
are marked by the whorls of more slender and sometimes septate appendages borne
at each. In relation to, and possibly in continuity with, the upper end of the
specimen is a peculiar row of cell-like structures ; this is one of the difficulties in the
interpretation of this specimen.

The appearance of the specimen is further shown in figs. 91 and 92 on PI. VIII.
These are magnified 100 diameters, and show the small specimen as it lies in
the cloudy matrix. The plane of focus in fig. 91 brings out the internodal cells,
while fig. 92 shows the whorled appendages at the nodes better. A portion of the
specimen showing the appendages at two of the nodes is more highly magnified in
fig. 93.

Another very instructive specimen discovered by Mr Cran is represented on
PI. VIII, fig. 94, and text-fig. 4. This can be readily placed in relation to the
preceding specimen as a node with the bases of the verticillate branches as seen in

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 879

cross-section. The specimen is incomplete on the upper side in the figure, but shows
all the essential parts. Centrally is the outline of the main axis with a diameter of
140 m. This is clothed by a zone of small cells, in relation to which the bases of a
number of branches can be seen.

It is tempting to recognise in these specimens parts of small axes with a con-
struction like the existing Characete. This is not improbably their nature, but some
suspension of judgment is perhaps desirable until further specimens are found. This
also applies to another specimen represented from Mr Cran’s drawing in text-fig. 5.
It may perhaps be the longitudinal section of a node similar to those in the two
preceding specimens, but its nature does not appear to us to be sufficiently clear to
do more than record it.

Some other remains now to be described were found in Mr Cran’s sections in
close association with the above specimens, and probably belong to the same organism,
though there is no direct proof of this.

The specimen shown in fig. 95 had penetrated, or at least was situated beneath,
the surface of a decayed stem of Rhynia Givynne-Vaughani. Its structure, which
is not shown in any one plane of focus, is brought out by Mr Cran’s drawing in text-
fig. 2. There appears to have been an elongated internodal cell widening out to a
nodal disc of small cells. The axis possibly continued beyond this, while laterally
from the nodal disc there arose two wider appendages, one of which was swollen
distally. We tentatively suggest a comparison of this specimen with the node of a
Characeous rhizoid which bears bulbils.

The specimen in fig. 96 may be placed in relation to this. Here the continuity
of the slender tubular axis of the rhizoid (?) can be traced. The segment above the
oblique septum bulges out, and appears to have formed several cells. It continues
into the large globular vesicle (bulbil ?), the distal portion of which is wanting at
the edge of the section. A shallow nodal disc, probably the basal group of cells in
optical section, is evident at the base of this vesicle, which we thus tentatively
interpret as a bulbil attached to a rhizoid. The details of the region of attachment
of the vesicle are shown in text-fig. 3.

In fig. 97 a similar globular structure is seen lying free in the matrix of the
chert. The thickness of the preparation allows it to be viewed as a solid object.
To the right a node-like disc of cells is attached to the globular vesicle, and from
some of the cells of this nodal disc slender tubular filaments appear to. have grown
out. It is possible to regard the vesicle as borne laterally on the filament, but the
above interpretation seems more probable. In that case the remarkable resemblance
which the specimen presents to a germinating bulbil of Char a aspera or Lampro-
thamnus is of interest in considering a group of remains which we appear to be
forced to compare with Characese.

In connection with the specimens last described which came from Mr Cran’s
block, the assemblage of imperfectly preserved remains of uncertain nature shown in

TRANS. ROY. SOC. EDIN., VOL LII, PART 17 (NO. 33). 135

880 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

fig. 105 is of interest. This is a fair sample of what has been met with at certain
spots in the upper region of the Rliynia Gwynne-Vaugliani peat in Bed A”l. Stout
wide tubes with contracted contents are found in transverse and longitudinal section.
The most prominent objects, however, are large oval or globular vesicles that are
usually isolated in the matrix, but have been seen borne on, or connected with, some
of the tubes. The vesicles are often single, but may have outgrowths from the base,
and so be associated in pairs (fig. 107). In other cases they form rows of two or
three (figs. 105 and 106), and sometimes indications of a node separating the vesicles
have been seen. The thin wall of the vesicle is sometimes smooth (fig. 107), but in
other specimens appears papillate (fig. 106, to the right.)

The specimen shown in fig. 108 was found in the peat among such vesicles. In
spite of its advanced decay, rendering the outline faint, an organisation of the
tubular structures into an internode separating two nodes with verticillately
arranged appendages can be clearly traced. This specimen also shows the tubular
filaments with their contracted contents.

It seems significant that globular, vesicle-like structures were met with in relation
to the specimens we have described as Algites Cranii ( cf. fig. 98), and also in asso-
ciation with this plant in blocks of the chert from other beds. In addition, we have
the globular vesicles in association with Mr Cran’s specimen (figs. 96 and 97).
While it cannot be stated definitely that these bodies are all of the same nature, it
is a possible tentative interpretation that they all belonged to the same type of plant
and were bulbils serving for vegetative reproduction like those in some Characeae.

While the various remains described above differ from one another, they do so in
a way that is consistent with their all being parts of the same plant. Their asso-
ciation is in favour of such an interpretation, and this is supported by the fact that
we have in one group of existing Algae much the same sort of range in morphological
construction in different regions of the plant. The verticillately branched axes in
figs. 93 and 108 and the nodal arrangement in fig. 94 resemble broadly what is found
in the “shoots” of the Characeae. The simply septate filaments in figs. 98 and 99
with occasional nodes (fig. 101) can be paralleled with the “leaves” of some
Characeae. The rhizoid structure (figs. 103 and 104) is also only to be compared
with that in the same group. Lastly, the large globular unicellular structures,
including the specimen with an attached node in fig. 97 that appears to be germinat-
ing, find a general but fairly close parallel in the type of bulbil present in Chara
aspera * and LamprothamnusA

These resemblances taken collectively seem to us of real significance, and are
difficult to explain without assuming that they indicate relationship. Since, how-
ever, no specimen of the characteristic reproductive organs of the Characeae has as
yet been found in the Bhynie deposit, the question of the systematic position of
Algites Cranii , and the other remains we have provisionally associated with it, is

* Cf. Giesenhagen, Flora, lxxxii (1896), p. 381. t Cf. M‘Nicol, Ann. Bot., vol. xxi (1907), p. 61.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 881

better for the present regarded as an open one. That, however, in a number of
respects these remains resemble the isolated existing group of the Characeee has been
made evident in the description and discussion above. In this connection it should
be borne in mind that structures suggestively like the oogonia * of a Characeous
plant have been recorded from Devonian rocks in America.

We have indicated the uncertainty that must still exist as to the systematic
position of the remains under consideration by placing them for the present in the
comprehensive genus Algites. On the ground, however, of various features of the
vegetative structure that suggest that we are probably dealing with a Characeous
plant with uncorticated internodal cells, we provisionally suggest the generic name
Palseonitella. We thus for the present name the plant Algites ( Palseonitella )
Cranii ; the name Palseonitella may be definitely adopted if the systematic position
of the plant is confirmed by the discovery of reproductive organs, or abandoned
should these show that the affinities of the plant are not with the Characeae.

Algites ( Palseonitella ) Cranii, Kidston and Lang, n.sp. (PI. IX, figs. 98-104,
and text-figs. 6-11).

Septate filaments composed of cells with fairly thick walls and contracted dark
contents. Tip of filament ending in a dome-shaped cell. Sometimes a short node
of small cells occurs separating two of the large cells. Branching from cells of
the node. Rhizoids (?) with elongated tubular cells separated by oblique septa, in
relation to which there is an irregular node of small cells. Large vesicles associated
with the remains.

Occurs in the matrix of the peat in certain beds of the Rhyme deposit. The
rhizoid-like structures may be within decayed tissues of plant-fragments.

Locality. — Muir of Rhynie, Aberdeenshire.

Horizon. — Old Red Sandstone. (Not younger than the Middle Division of the
Old Red Sandstone of Scotland.)

Incertse sedis, but associated with Algites Cranii as probably belonging to it.

(PI. VIII, figs. 91-97, and text-figs. 1-5.)

Small axis with internodal cells and whorls of appendages at the nodes.
Rhizoid-like structures and spherical bulbils (?).

(PI. IX, figs. 105-108.)

Tubular filaments with contracted contents ; sometimes showing an internodal
cell separating nodes with verticillate branches. Globular or oval vesicles (bulbils ?).

* Knowlton, Amer. Journ. Science, vol. xxxvii, p. 202 ; Seward, Fossil Plants, vol. i, pp. 225-226.

882 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

NEMATOPHYTON, Penhallow.

1859. Prototaxites, Dawson, Quart. Journ. Geol. Soc., vol. xv, p. 484.

1871. Nematoxylon , Dawson, Geol. Survey of Canada, “Fossil Plants of Devonian and Upper Silurian

Formations,” p. 20.

1872. Nematophycus, Carruthers, Monthly Micros. Journ., p. 172.

1888. Nematophyton, Penhallow, Trans. Roy. Soc. Canada, vol. vi, section iv, p. 36.

4. Nematophyton Taiti, n.sp., Kidston and Lang.

The discovery in the Rhynie chert of a fragment of tissue, the structure of which
agrees closely with that of plant-remains known from the Silurian and Devonian
formations under the names Prototaxites , Nematoxylon, Nematophycus, or Nema-
tophyton adds to the interest of the deposit. This small fragment only appears in
two successive sections of a block from Bed Affl ; it measures only 3x2 mm. The
characteristic structure is, however, well shown, the tissue being composed of
systems of wide and of narrow tubes ; the differences from the specimens on which other
species have been founded are merely in details, such as those on which these species
are distinguished. This .is, we believe, the first record of Nematophyton from the
Old Red Sandstone of Scotland We have pleasure in naming this species Nema-
tophyton Taiti, after Mr David Tait, who superintended the excavations at the
Muir of Rhynie that resulted in the discovery of the silicified peat-bed in situ.

We associate with this fragment, without giving it a separate name, another
slightly larger portion of a plant which occurred at the same level in the same block,
but was cut in two other sections of the series. This second fragment, which was
only about 4 cm. distant from the fragment showing the typical structure of Nema-
tophyton, is a portion of the peripheral region of a cylindrical structure. The
structural differences it exhibits from the other fragment can reasonably be ex-
plained on this ground. Though there is no proof of continuity, we feel justified
in placing it with the first specimen under the same name. The two specimens will,
however, be described apart, and the figures will enable the reader to form his own
opinion as to their belonging to the same plant.

If, as we beiieve, the second fragment shows the peripheral region of a plant
the central portion of which had the typical Nematophyton structure, it is a welcome
addition to our knowledge of this remarkable genus. It is in any case a portion
of the peripheral region of a plant of a similar type of construction.

The two specimens may now be described in detail, taking them in the order in
which they are mentioned above.

One of the two sections with the typical structure of Nematophyton is slightly
larger than the other and shows the tissue better preserved. The irregular edges
of the fragment are decomposed and covered with a granular mass that is possibly
of bacterial origin. The greater part of the better section is shown magnified 50
diameters in fig. 109. This figure shows that the wide tubes making up the bulk

SHOWING STRUCTURE, FROM THE RIIYNIE CHERT BED, ABERDEENSHIRE. 883

of the tissue are on the whole cut transversely in this section, which may be pre-
sumed to be transverse to the long axis of the portion of the plant from which the
fragment comes. A number of approximately circular areas from which the wide
tubes are absent are shown; for these the term “medullary spots”* may be
employed. A portion of the section is magnified 100 diameters in fig. 110, and two
other portions are magnified 250 diameters in figs. Ill and 112. The last-named
figure shows a portion of one.of the medullary spots surrounded by the tissue formed
of large tubes.

The relatively large tubes show considerable range in size, but those of different
diameter are mixed together and not arranged in concentric zones as in Nema-
tophyton Logani. The largest have a diameter of 52 m, while the smaller ones
measure only 20 m. They are frequently contiguous, often forming rows or small
groups, while at other places they are distinctly separated from one another. The
intervening spaces, like the medullary spots, were occupied by the system of fine
tubes or hyphae to be described below.

The walls of the larger tubes are usually very thick and often appear thicker at
one side than at the other. Thus in a tube 34 m in diameter the thickness of the
wall ranged from 4 m to 7 m. Another tube 48 m in diameter had the wall 14 m thick.
Other tubes had thinner walls, the wall of one with a diameter of 20 m being only
2 m thick. The tubes thus vary from 20-52 m in diameter, with a wall-thickness
varying from 2 m to 14 m or more. The apparent thickness of the wall may have
been affected by the state of preservation of the tube.

The tubes so far described are cut transversely, but others lie obliquely and are
therefore seen from the side or in longitudinal section as they lie between the other
tubes. These obliquely running tubes are sometimes relatively narrow, down to
12 m, but belong to the system of relatively wide tubes. Some of them can be
traced into the medullary spots.

Where the relatively large tubes are not in close contact, and especially in the
areas of the medullary spots, there is evidence of the presence of a second system
of delicate tubes of about 10 m in diameter. In most places between the wide tubes
and in most of the medullary spots the tissue formed by the narrow tubes has lost its
structure and is either amorphous or replaced by a pseudo-cellular structure. In
one or two of the medullary spots, however, the delicate tubes could be more or less
clearly followed, and we have no doubt that they were originally present in the
corresponding situations from which they have decayed.

It will be seen that, small as this specimen is, it shows all the types of tissue
described for other species of Nematophyton.

It is not necessary or desirable to press the description of the tissues of this
small fragment of Nematophyton further. It is sufficient to be able to recognise
that it agrees with other specimens of this genus in the absence of true tissues such
* Penhallow, Ann. Bot.,v ol. x, p. 144, 1896.

884 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

as result from cell-division and in being largely composed of relatively wide tubular
elements, between which, and most characteristically in the medullary spots, there is
a system of delicate, narrow, interlacing tubes.

The Rhynie species differs * from N. Logani in the absence of radial bands and
the presence of “ medullary spots.” In N. Storriei, although there are “ medullary
spots,” the large tubes are always separated by a slight space, are smaller, and have
thinner walls. With N. Ortoni it agrees in the presence of “ medullary spots,” but
the wide tubes of that species are larger, attaining a maximum diameter of 67 m,
and are more uniform in size. The thick walls of the Rhyme species show some
similarity to those of AT. crassum, but the wide tubes are larger in the Scottish
species and the cavity is always prominent.

From N. laxum and N. tenue , as described by Penhallow, it appears to be dis-
tinguished by several characters. The preservation of these forms and of N. Hiclcsii
and N. Dechenianum is, however, too imperfect to admit of detailed comparisons.

Such specific distinctness, interesting as it is, is secondary in importance to the
recognition in this deposit of an example showing the characteristic construction of
this type of plant. The diagnosis will be given after the description of the second
fragment.

Of the second fragment there were also two sections, one of which shows the
tissues rather better preserved than the other, though both are instructive. The
greater part of the better section is represented, magnified 30 diameters, in fig. 113.
As this shows, the fragment includes a portion of the outer surface of the plant
and the tissue lying beneath this. The curvature of the surface indicates that the
part of the plant was cylindrical (or less probably spherical). If we assume it to
have been cylindrical, the curvature indicates a diameter of about 1 ’2 cm.

The following structural regions can be distinguished proceeding from the outer
surface inwards : (a) a narrow, clear, structureless zone ; ( b ) a zone about 1 mm. in
depth, composed of tubular filaments standing at right angles to the surface ; (c)
an inner region in which the tubes or filaments run irregularly ; these tubes are
continuous with those of zone b. A number of small areas of different texture, that
appear as dark spots in the figure, are distributed through this region. The same
zones are shown rather more clearly in fig. 114.

The chief structural component of the inner tissue (zone c ) has the form of
branched tubes or filaments running irregularly in all directions and crossing one
another so as to give the effect of a reticulate structure (figs. 115 and 116). These
filaments when best preserved stand out by reason of their brown colour and homo-
geneous appearance. This seems referable to a preservation of the contents rather

* It will be sufficient to give a general reference to the diagnoses of N. Logani, N. Hicksii, N. crassum, N. laxum,
and N. tenue in Penhallow’s paper {Trans. Canad. Roy. Soc., vol. vii, 1889, p. 28) ; to that of N. Ortoni, in a paper
by the same investigator (Ann. Bot., vol. x, 1896, p. 41) ; and to that of N. Dechenianum, in the paper by Solms
(Jahrb. d. k. preuss. Geol. Landesanstalt, 1894).

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 885

than of the walls of the tubes.* The cylindrical brown threads may appear as if
cracked across, but there is no satisfactory evidence of the filament having been
septate. At places a beaded appearance, possibly due to contraction of the contents,
is found. Careful search reveals tubes in which the wall is preserved and which,
therefore, appear wider than the brown threads. There has been much breaking-
down and rearrangement of the substance during preservation, and the clear
intervals or matrix between the tubes often exhibits pseudo-cellular structure ; this
is most marked where the decay is most advanced and the tubular structure has
disappeared.

A feature of considerable interest in the preservation of some of the tubes is the
appearance of a fine spiral marking. This runs round the inside of the thick wall
of the tube. In favourable examples (fig. 118) the thickness of the wall can be
traced, and the spiral marking seen in section to be due to a regular ridging of the
innermost layer of the wall, f In other specimens only the spiral is evident. This
mode of preservation, revealing a potential structure, is also found in the narrower
tubes of zone b (fig. 117).

Certain oval or spherical thin-walled vesicles are met with at places in the regions
between the brown tubes— one of them is seen below the tube in fig. 118. They
suggest comparison with some of the fungal types already described, and in some
cases have been observed to be borne on slender branched hyphse. In the light
of the common presence of such fungi in the portions of all the plants in this deposit,
we are probably justified in regarding these vesicles as belonging to a saprophytic
fungus, and not as a part of the structure of the organism under consideration.
Their presence is noted, however, since in such an anomalous organism as Nema-
tophyton they might prove to be of importance.

The small areas of different texture dotted through the inner region are fre-
quently ill preserved and difficult to analyse. In fig. 116 they appear as dark
masses owing to their greater opacity, the only structural feature that is indicated
being the entrance of some of the wide tubes into the dark mass.

Careful examination of the best preserved examples shows that these wider
tubes, which may have their rather thick wall preserved, resemble in cross-section
the large tubes of the other specimen of Nematophyton. These wide tubes are
accompanied by others of moderate fineness, and the bulk of the area seems to be
made up of numerous still finer tubes closely entangled. The general resemblance
in construction between these areas and the “medullary spots” in the other
specimen is striking, and we have no doubt as to their being corresponding
structures.

The region (figs. 113, 114, b) with the tubes arranged parallel to one another,

* A similar mode of preservation was met witli in N. Hiclcsii ( cf . Penhallow, loc. cit., 1889, p. 20).
f A very similar spiral marking, due to contraction and ridging of the inner layer of the cell-wall, is often met
with in the cells of Laminaria. It has there been interpreted as an artefact due to drying or dehydration (cf.
Thoday, Mrs M. G., New Pliytologist, x (1911), p. 69).

886 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

and at right angles to the surface, shows similar elements to the inner zone
differently arranged; the “medullary spots” are not, of course, represented.
The vesicles and fine hyphse mentioned above and regarded as a saprophytic
fungus in connection with the inner zone, occur between the tubes of the
outer zone also.

The tubes in the outer region are represented by their homogeneous brown
contents, and can be traced from the irregularly running tubes of the inner region
which bend outwards and branch. The free ends form the parallel tubes of the
outer zone, which vary in thickness. They tend to widen slightly at their outer
ends, where their tips stand at about the same level. Some of these tubes showed
the same spiral marking as was described for the wider tubes of the inner tissue
(fig. 117).

Between the vertically running tubes, and at places appearing to separate them .
into bundles, a granular mass is such a regular constituent of this outer zone as
to suggest that it might have been based on a real structure, but the appearance
is never decisive. When the clear evidence of breaking down and rearrangement
afforded by the pseudo-cellular structure is taken into consideration, we are not
prepared to attach significance to the brown granular masses in the outer zone,
though attention may be directed to them.

The narrow, clear outer zone (figs. 113, 114, a) above the summits of the vertical
tubes is structureless, but the pseudo-cellular structure, that can be traced through-
out much of the fragment, comes out very clearly here. It seems reasonable, in
the light of the rest of the structure, to interpret this outermost zone as having
been a structureless,- and perhaps mucilaginous, layer during life. It has no
sharply defined outer limit, though this is marked as a dark line owing to a
granular deposit.

Although there is no continuity and the condition of decay or preservation is
somewhat different, the second fragment described above has been treated as the
peripheral region of Nematophyton. This peripheral region in previously known
specimens of the genus has, at best, been represented by a thin coaly layer.* The
demonstration of the general structure of the peripheral region is the main addition
to our knowledge of Nematophyton that is afforded by these specimens.

The affinities of Nematophyton with its remarkable and very consistent type of
construction must still be regarded as an open question. The general opinion
since Carruthers’ paper t has been that they were to be sought for in the Algae,
although there has always been difficulty in recognising a clear relationship to any
algal group. There are difficulties in the way of a close comparison with Lamin-
ariacese, although in some respects the Bhynie specimens afford additional points

* This layer is described as “ very thin, hardly exceeding 3 mm.” for N. Logani (Penhallow, Trans. Canad. Roy.
Soc., vol. vi, 1888, p. 38), and as l-5 mm. in iV. Ortoni (Penhallow, Ann. Bot., vol. x, 1896, p. 42).
f Monthly Microscopical Journal , vol, viii, 1872.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 887

of comparison with this group. The wide non-septate tubes of Nematophyton
have long been recognised as being more readily comparable with the tissues of
such Algse as the Udotese than with the septate cell-rows of the tissue derived
by cell-division that make up the thallus of the Lgminariaceie ,* although the size
attained by the fossil plant suggested comparisons with the latter.

While the general structure makes it somewhat difficult to think of Nematophyton
as a land-plant, this possibility gains prominence in the light of the occurrence
of the plant in the Rhynie deposit. Dawson’s view of Nematophyton Logani in
the Lower Devonian rocks of Gasp6 was that its large trunks had grown rooted
in soils which were covered with a growth of Psilophyton and Arthrostigma,
and that, while probably a marsh-plaut, it was not marine. f The presence of

Nematophyton in the terrestrial Rhynie deposit supports the conclusion drawn
from the mode of occurrence of N Logani at Gasp6. Such a terrestrial mode
of life of the plant would not be inconsistent with fragmentary remains being
preserved in what are more probably marine deposits, while it is difficult to
account for the repeated occurrence of a marine plant in terrestrial deposits.
The facts so far known, although they involve difficulties, would thus appear to
be in favour of Nematophyton , whatever its systematic position may prove to
be, having been a marsh- or land-plant of Silurian and Devonian times.

Nematophyton Taiti, n.sp., Kidston and Lang.

Small fragment showing a construction of wide and narrow tubes characteristic
of the genus. Tubes of wider system showing a range in size from 20 n or less
up to 52 n ; mostly about 40 n. Thickness of wall varying from 2 n to 14 n or
more, probably in relation to condition of preservation. Some narrow tubes of this

* It may be noted that a recent suggestion, while emphatically dismissing comparison with the Laminar icex,
treats Nematophyton in connection with massive Fungi (Church, Thalassiophyta, p. 49 and footnote).

t Dawson’s views on this -point are expressed in a number of his publications (e.g., The Fossil Plants of the
Devonian and Upper Silurian Formations of Canada, 1871, and The Geological History of Plants, 1888), but most
clearly and recently in Penhallow and Dawson, Trans. Canad. Roy. Soc., vol. vi, 1888, pp. 27-36. It will be
remembered that these views were founded on study of the mode of occurrence of the plant as shown in the rock-
sections exposed at Gaspe and the Bay des Ohaleurs. Their nature may fairly be given by some extracts from the
last-named paper, though the context should be consulted. “ The mode of occurrence and state of preservation
of the specimens seemed to make it certain that they had belonged to land-plants ” (p. 27). “ I also found stumps
with branching roots apparently rooted in situ in the shales and argillaceous sandstones of the locality ” (p. 28).
“ I also ascertained that these remarkable plants had probably grown in the clays and sands in which Psilophyton
and other plants had been rooted, and consequently, that though probably marsh-plants they were not marine.
They must have grown on low flats, probably often inundated, though whether this was with salt or fresh water
is indicated merely by the negative fact that no properly marine organisms occur in the containing beds ” (p. 28).
“It was further found that Psilophyton princeps, P. robustins, Arthrostigma gracile, and Cordaites angustifolia were
constant associates of these plants” (p. 29). In connection with the last quotation, it is to be noted that the
remains referred to as Cordaites angustifolia were regarded by Dawson as of quite uncertain nature, and are even
discussed as possible foliage of Nematophyton ( loc . cit., p. 34). Fig. 3 in the paper from which the above extracts
have been made shows a trunk of Nematophyton Logani resting on an underclay filled with Psilophyton. Fig. 4
shows another trunk embedded in a bed of shale containing Psilophyton, the bed being underlain and overlain
by sandstone.

TRANS. ROY. SOC. EDIN., VOL. LII, PART 1Y (NO. 33). 136

888 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

system can be traced into the medullary spots. These “ spots ” were about
'2 to "3 mm. in diameter, and were originally occupied by narrow tubes about 10 m
in diameter.

Another fragment is regarded as the peripheral region of the plant. Internally
it showTs interlacing tubes and small medullary spots (T5 mm.). The tubes bend
outwards, and in the superficial layer, which has a thickness of about 1 mm., stand
parallel and at right angles to the surface. Externally there is a narrow amorphous
layer above the free terminations of the tubes.

Locality. — Muir of Rhynie, Aberdeenshire.

Horizon. — Old Red Sandstone. (Not younger than the Middle Division of the
Old Red Sandstone of Scotland.)

5. The Succession of the Plants as shown in one Complete Vertical
Section of the Bed.

The plant-containing chert was discovered by Dr Mackie, as has been men-
tioned in previous papers, in the form of isolated blocks. The bed from which these
came has since been successfully exposed by the Ceological Survey. An account
of the excavations made for this purpose is given in a Report * of a Committee
of the British Association. A set of blocks forming a complete vertical section
from Trench No. 1 has been available for our investigation, and a general account
of this section was included in Part I of this series of papers. Since then thin
sections for microscopical examination of all the plant-containing beds of this
vertical section have been studied. We are thus able to extend and qualify our
earlier account, and to indicate more precisely the distribution of the plants
throughout the chert-band at this one spot, where it was about eight feet in
thickness.

In the following description the numbering and lettering of the beds t as given
in the section represented on pp. 762-3 of Part I will be adhered to, and the
composition of the beds followed from below upwards. While the description
will be based on the actual section in one vertical plane, variations in the
succession of the plants, especially in the critical lower region of the bed, must
be taken into consideration. The information afforded by certain loose blocks
referable to this region can suitably be included in the description.

Above the bed of clay passing down into bedded shales the eight-foot section of

* Brit. Assoc. Rep., 1916 (Newcastle), p. 206.

+ The following slides in the Kidston Collection give a complete survey of this vertical section : —

N 20, 2561
N 19, 2429
M 18a, 2560
M 186, 2559
L 17, 2558

L 16, 2430

IC 15, 2562
I 14, 2435

H 13, 2563
G 12, 2564
F 11a, 2565
F 116, 2556

When a number of slides are mentioned, as in B 6 and

E 10, 2566
E 9, 2518
D 8, 2543

B 6, 2449, 2567, 2458, 2487
A"2, 2499

A'T, 2492, 2447, 2444, 2533, 2514, 2423.

A'T, they are in descending order through the bed.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 889

the chert-band in Trench No. 1 commences with the bed marked A". This bottom
region of the chert is of special interest, since all the species of vascular plants and
practically all the types of Thallophyta found in the deposit have been met with in
its thickness of 1 foot 3 inches, and indeed in the lower portion Bed A"l. The basal
region consists of a pure accumulation of Rhynia Gwynne-Vaughani, usually in a
clear peaty matrix. Occasionally sporangia of R. major have been met with in the
peat thus formed. Above this came a peat formed of more or less decayed Hornea
Lignieri, rhizomes, stems, and sporangia of this plant being present. In the upper
portion of this region Rhynia major is associated with the Hornea (Part III, fig. 2).
Overlying this Hornea peat came a sandy bed enclosing remains of Asteroxylon
and some Hornea. The Asteroxylon was best preserved immediately above the
Hornea peat, and its rhizomes extended down into the latter, penetrating portions
of the plants (Part III, fig. 2 ; Part IV, fig. 24). On passing further upwards into
the dark sandy Bed A"2 the preservation of the stems of Asteroxylon embedded in
the sandy matrix is less good.

While this is the succession shown in one block at this particular spot, other blocks
known to have come from the same level at no great distance show that it was not
invariable. Thus in one block a clear peaty matrix with stems and sporangia of
Hornea mixed with rhizomes, stems, and intermediate regions of Asteroxylon came
in place of the pure Hornea peat. Another block had the typical Rhynia Gwynne-
Vaughani peat below, and above this well-preserved Asteroxylon in a clear peaty
matrix. A very similar block had the Rhynia Gwynne-Vaughani peat, below
followed by a zone with Asteroxylon only, while another thin layer of Rhynia
Gwynne-Vaughani came above this. With the exception of this last specimen, peat
composed of R. Gwynne-Vaughani has only been seen at the extreme base of this
bed. It was not met with above Bed AM in the section studied.

In view of the limitation of R. Gwynne-Vaughani to the base of the section of
the bed as exposed in situ, it is convenient to refer in this place to two loose blocks
of particular interest. One of these is the cubical block, from the weathered face of
which fig. 2 in Part I was taken. This block, which is now in the Natural History
(British) Museum, proves that at some places in the bed there was an accumulation
of pure R. Gwynne-Vaughani peat to a depth of at least 12 inches. The other is the
block represented in Part I, fig. 5, in which the closely associated, tapering stems of
R. Gwynne-Vaughani stood practically erect, as they had grown above a peat
composed of the remains of the same plant.

Returning to the description of the section from Trench I, it is to be noted that
following on Bed A "2 came a definite band of cherty sandstone of some thickness
(A"3, A'). This had streaks of carbonaceous matter, but, owing to the absence of
plant-remains showing structure, it was not studied in detail.

The thick Bed B 6 again exhibits peaty structure, being composed throughout of
plant-remains with their form and structure fairly well preserved. These lay in a

890 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

peaty matrix interrupted by irregular sandy layers. One block including the whole
thickness of this bed was divided into four by horizontal planes, and the distribution
of the plants followed in vertical sections of the. four portions. The peat of the
lowest quarter of the bed contrasts sharply with the underlying sandy band in the
absence of inorganic constituents. It has a peaty matrix with spores and fragments
of plant-tissues arid, embedded in this, fairly well-preserved rhizomes, stems, and
sporangia of Hornea ■ The sections studied show a practically pure Hornea peat
with a very occasional stem that may belong to Rhynia. The division of Bed B
above this shows a sandy matrix enclosing remains of Hornea and Asteroxylon,
and overlying this a Hornea peat similar to that just described. The succeeding
region has the peat much interrupted and mixed with sand ; the plant-remains are
Hornea and Asteroxylon. The uppermost region of B 6 shows the same mixture of
peat and sand ; the plants are Hornea , Asteroxylon, and Rhynia major.

It is characteristic of Bed B 6, in contrast to A*l, that the peat is more
interrupted by irregular sandy layers. The preservation of the plant-remains also is
different, their tissues tending to become amorphous and yellow. The indications of
accumulation of the peat by growth in situ of the vascular plants are less definite
than in the case of the beds below ; the appearances are equally consistent with the
view that the plant-remains might have been carried a short distance from adjacent
growths of the plants.

Above Bed B 6 came the second definite bed of fine siliceous sandstone (C) ; this
was not examined in detail, owing to the obvious absence of plant-remains.

The succeeding three and a half feet of the chert-bed in the section from Trench
No. 1 contained, so far as we have observed, only Rhynia major. The main peaty
beds are separated by sandy layers, and similar more irregular bands of dark sandy
matrix interrupt the thicker masses of peat. This composition is well illustrated by
the block from Bed E 10 that was represented natural size in Part I, fig. 1*

The condition of preservation of the remains of Rhynia major may be noted
very briefly, taking the beds in order. In Bed D the peaty mass is yellow and
almost amorphous, only a few of the stems showing recognisable structure. In
contrast to this type of preservation the stems in Bed E, which are often of large
size, retain their cellular structure, although they may be more or less decayed and
broken down. Rhizomes of Rhynia major have been found in this bed. In the
lower portion of the following bed (F 116) the peat is composed of greatly altered
plant-remains, which are shown to be Rhynia major by a few recognisable stems ;
the general condition of this bed is similar to that of Bed I). An isolated tracheide
of Asteroxylon was found in the matrix. In the upper portion (F 11a) the stems in
various conditions of decay retain the cellular structure in the more usual way ;
some resemble rhizomes, although rhizoids have not been seen. Beds G 12 and H 13
contained more or less decayed stems of R. major, sometimes associated with

* This block of the chert is preserved in the British Museum (Natural History).

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 891

sporangia. The stems in Bed 1 14, though somewhat (decayed, are hardly compressed
and lie loosely in the peaty matrix which separates them. The smaller stems have
a stellate outline, due to unequal contraction with ridging of the surface ; they were
evidently less firm and rigid in life than the larger stems. The section of Bed K
shows that the lower portion consists mainly of peaty and sandy matrix, above which
comes a narrow layer of peat with uncompressed stems that are almost amorphous.
In the lower (L 16) and upper (LIT) portions of the following Bed L, which still
contains Rliynia major only, the preservation of the tissues is of the usual type.
Very well-preserved stems, which may even show the contracted protoplasmic
utricles of the cells, occur alongside others that are more or less decayed or broken
down. In Bed M a change in the composition of the peat is met with. Stems of
Rhynia major are found as in the preceding beds ; mixed with these in M 186 and
throughout the upper portion (M 18a) forming a pure peat, there are remains of
Hornea -stems, rhizomes, and sporangia being present. Hornea having thus
reappeared in this section of the chert-band, persists in the two levels of the last
peat-bed N. In N 19 the prominent remains are large and very perfectly preserved
stems of Rhynia major , but stems and sporangia of Hornea are associated with
them. In N 20, the uppermost bed of the chert-band showing a peat formed of
plant-remains, this is composed entirely of the parts of Hornea with their usual
rather imperfect preservation. The last bed of the section, marked 0, which
consisted of siliceous sandstone with carbonaceous remains, was not examined in
detail.

Having thus traced the vertical distribution of the vascular plants through this
section of the deposit, the general distribution of the Thallophyta remains to be
indicated. This need not be treated in detail, for there is evidence that some
forms of fungi are present in the lowest and highest beds. Their nature and
abundance in the various beds show a general relation to the remains of vascular
plants composing the peat and their mode of preservation. Practically all the
Thallophyta that have been described occurred in Bed A"l at the base of. the
deposit ; it has been shown above that all the vascular plants were present in
this region also. Some of the fungi have only been found in this lower region
(e.g. Palseomyces Asteroxyli, P. Horneae). Others, such as P. agglomerata and
the intrusive fungus described as Fungus No. 14, occurred here and are met with
again in the Hornea peat of the uppermost beds. The most widely distributed
fungus was Paldeomyces Gordoni, often with intrusive forms of the types described
under Fungus No. 15.

Of other Thallophyta, unicellular Bacteria were probably widespread, but, for
reasons that have been given, are difficult to identify. The filamentous form described
as Schizophyta No. 2 ( Archaeothrix contexta) was typically present as a felt above
the Rhynia Gwynne-Vaughani peat at certain places in Bed A"l. It also occurred
around decaying fragments of R. Gwynne-Vaughani , and is met with again in Bed

892 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

F 116. It is of interest, as indicating a recurrence of similar, probably submerged,
conditions at these levels, that the algal remains described under the name Algites
Cranii occur at the same two levels in the chert-band.

The fragments of Nematophyton have only been met with at one spot in the clear
peaty matrix of Bed A"l, some little distance above the level where the felt of
Schizophyta No. 2'and the remains of Algites Cranii occurred.

It is noteworthy that Crustacean remains were most abundant in the Rhynia
Gwynne-Vaughani peat of the lower region of Bed A"l, where the last-named plants
were found.

6. Remarks on the Nature, Conditions of Accumulation, and
Preservation of the Rhynie Deposit.

In the first of this series of Memoirs (Part I, p. 764) some remarks were made
and a general view expressed regarding the nature and origin of the silicified chert-
band found at Rhynie. The accumulation of beds rich in plant-remains and free
from admixture of mineral particles was compared to the growth of a modern peat.
The fact that the peat is interrupted at intervals by more or less regular sandy
layers was interpreted as indicating frequent flooding of the surface. The silicifica-
tion was regarded as due to a final saturation of the peat-bed, or rather the system
of peat-beds and sandy layers, by water with silica in solution, derived from
fumaroles or geysers.

As a general hypothesis which covers the main facts this may stand. Now that
the plant-remains have been examined throughout a complete vertical section, it is
possible, however, to amend and amplify this view in some points of detail.

The outstanding features of the Rhynie bed as shown in the section described
above may, in the first place, be summarised, since it is from a consideration of them
that any more detailed indications as to the conditions of accumulation of the deposit
can be expected. It has been seen that —

(а) The structural characters of Rhynia and Asteroxylon , especially the presence
of stomata in the epidermis from the base of the stem upwards, indicate that the
growth of these plants was not truly aquatic. The soil may, of course, have been
swampy or saturated with water. On the other hand, Hornea might have grown in
shallow water, though there is no direct evidence for this.

(б) There are indications of growth in situ of Rhynia Gwynne-Vaughani in the
basal bed, of Asteroxylon in the sandy layer covering the Hornea peat of Bed A"l,
and of the Hornea in the same region. There is also evidence for the growth in
situ of Rhynia major in the peat of some of the upper beds. These various indica-
tions justify us in regarding the peat-layers as largely formed by the growth of
plants on the spot ; the carriage of remains from the near neighbourhood and
their accumulation, possibly under water, is not excluded in the case of some of
the beds.-

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 893

(c) Rhyviia Gwynne-Vaughani has only been met with at, and slightly above,
the base of Bed A'T, i.e. at the extreme base of the deposit. Asteroxylon , except for
an isolated tracheide in Bed F 116, has not been met with above Bed B 6. Hornea
occurs in the lower beds up to Bed B 6, and reappears in the uppermost Beds M and
N. Rhynia major is only an occasional and relatively unimportant component of
the beds up to B 6. Above this from D to L it forms the whole deposit, being then
gradually replaced by Hornea in the uppermost beds.

( d ) The change in composition of the plant-remains above the Rhynia Gwynne-
Vaughani zone at the base takes place without any intervening sandy layer. The
succeeding peat may be composed of Hornea with some Rhynia major, but this in
other blocks is represented by loosely packed Hornea and Asteroxylon, or by Aster-
oxylon only.

(e) The peat-beds are separated and interrupted by sandy layers of two sorts :
(l) There is a very definite bed of sandstone separating the characteristic remains of
A"1 and A"2 from the different type of peaty deposit in B 6. Another definite sand-
stone layer occurred at C, and was followed by the succession of beds with pure
Rhynia major. (2) The separation of the other peat-beds is due to an accumulation
of sand mixed with abundant plant-remains, the decay of which has been more or
less complete. The passage of the plant-remains into thin carbonaceous partings
can be followed at places. The less marked interruption of the peat-beds is due to
similar irregular accumulations of sand and, associated with this, more complete decay
of the vegetable tissues. Occasional isolated grains of sand in the peat itself indicate
the other extreme of the association of sand with the peaty mass.

(/) There are indications in the different nature of the preservation of the stems
and from associated algal remains of more aquatic conditions at certain levels. This
can be regarded as probable for the region above Rhynia Gwynne-Vaughani in A'T,
and for Beds D and F 116. Possibly the presence of Hornea is another indication
of moister conditions : in connection with this the reappearance of this plant in
Beds M and N is to be noted.

(g) The loose packing and excellent preservation of the plant-remains in the
basal beds is a remarkable feature of the deposit. There is no regular increase in
decay and compression on passing from the upper peat-beds downwards.

Certainty as to the exact conditions of accumulation of this ancient deposit is
probably unattainable. Our knowledge of the details of accumulation of various
deposits at the present day, which might afford further help, is only sufficient to
serve for the institution of general comparisons. A consideration of the facts
summarised above with some others appears, however, to justify a further interpre-
tation of the story to which the following remarks are a contribution.

The more detailed examination of the peat-like bands themselves has supported
the interpretation of their nature which is expressed in the term “ silicified peat” used
throughout these papers. In some cases it is clear that the plants grew in the

894 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

position in which their remains are found ; it is possible, though not certain, that in
other cases the remains may have accumulated in water from plants growing close
by. The structure of the resulting mass, excluding the sandy bands, agrees with
that of recent peats in the absence of mineral particles and the presence of a matrix
embedding more or less well-preserved portions of plants. The matrix, though in
great part amorphous, must have had sufficient consistency to support the fragments
of tissues, spores, etc., evenly distributed through it. The decay of the plant-
remains is to be associated with the presence of saprophytic fungi in them, and
doubtless also with the action of bacteria. It seems to have taken various forms and
to have been arrested at various stages. The arrest may have been related to
absence of oxygen ; the more thorough decay in relation to the admixture of sand
supports this view. Hyphse and resting-spores of fungi are chiefly found in relation
to the plant-remains, and when they occur in the matrix were probably in great part,
though not entirely, set free by the decay of the plants.

All these features find close parallels in recent peats, especially when the
structure of the latter is examined in microscopical sections. The component plants
are, of course, quite different, and the fungi do not belong to the same types.
Persistent dark hyphse and resting-spores and sclerotia occur abundantly in the
plant-remains and the amorphous matrix of recent peats. Recent peats, however,
become more amorphous on passing downwards, and the difference in this respect of
the Rliynie deposit will require to be considered below.

The presence of numerous layers of sand interrupting the peat in the Rhynie
deposit is a feature in which it differs from most recent peats. We have no definite
evidence showing that this sand was wind-borne, and therefore suggested frequent
inundations to account for the irregular layers of sand, and more prolonged sub-
mergence to account for the more definite sandstone beds. The possibility of some
of the sand having been wind-borne must, however, be kept in mind. This applies
especially to the irregular sandy layers and the scattered grains in the matrix of the
peat. The dark colour of these regions appears to be due to the altered vegetable
matter ; decomposition seems to have been complete and rapid in the sandy layers.
The vegetable remains thus represented in the layers of sandy matrix must corre-
spond to a considerable thickness of the peat, any remains of stems, etc., that persist
being almost completely flattened.

Since the Rhynie deposit occurred in what is known to have been an actively
volcanic region, it is reasonable to regard the silicification of the bed as in some
way due to water containing silica in solution, possibly derived from fumaroles.
The separation .of silica in the colloidal form may have been determined, in part
at least, by the decaying plant-remains. Such a deposition of silica is known to
follow from the growth of Cyanophycese and Bacteria, and around decaying organic
material in hot springs. The Rhynie chert may have been the result of a some-
what similar process.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 895

When we take into consideration the composition of the water of siliceous
springs at the present day, it further seems quite possible that plants like Rhynia
may have grown in a substratum saturated with the siliceous solution. In the
light of what is known as to the vegetation found around siliceous springs at the
present day, the xerophytic construction of the Rhynie plants may be of signifi-
cance. In the case of the vegetation around fumaroles in Java * the plants are
xerophytic, rooted in the hot acid soil, and with their foliage exposed to hot
sulphurous vapours. The soil consists of a white siliceous clay. The clay under-
lying the Rhynie deposit might possibly be regarded as of this nature, while
corresponding in its effects on the vegetable remains above to the clay beneath
recent peats.

The possibility that the water saturating the Rhynie plant-remains might have
been siliceous throughout their accumulation may further explain a most striking
difference between the preservation of this bed, considered as a whole, from what
is found in recent peats. In the case of the latter the best preservation is
met with in the upper region, decay and compression being marked in the lower
part of the bed. In the Rhynie deposit, on the other hand, the preservation is
as good in the lowest beds of the eisdit-foot section as it is above. Further, the
plant-remains at the base of the deposit are not compressed, and are often loosely
placed in the matrix. It was suggested above that the water around the growing
plants might have been siliceous, and that the decay going on in the accumulating
remains might have determined the separation of colloidal silica ; there are
indications from the microscopic sections that this separation did occur round the
plants in the matrix in the first instance. On this view the silicification, or at
least its early stages, would have been progressive as the bed was accumulating,
instead of the thick peat-bed being first formed and then at a later date silicified
as a whole. Such an explanation would account for the loose packing and perfect
preservation of the plants in the basal beds where, on the analogy of recent peats,
the greatest alteration and decay would have been anticipated.

The peculiar changes which are involved in the development of hemispherical
projections in Rhynia Gwynne-Vaughani and the necrosis areas, wound-reactions
and unequal enlargement of cells in the stems of this species and of R. major (see
Part IV, pp. 835-6), must be recalled here. These various developments suggested
a reaction to some prolonged external stimulus. The possibility that this was
connected with the volcanic conditions that accompanied the supply of the siliceous
water appears to be a justifiable speculation to raise. The intumescences and other
reactions that are known to follow on the stimulus of irritating vapours in the
case of existing plants (Part IV, p. 834) should be borne in mind in considering this.

The characters of the deposit as a whole appear to indicate a steady growth
and accumulation on a subsiding surface. This might have been a swamp beside

* Of. Schimper, Plant Geography , p. 386, and literature there cited.

TRANS. ROY. SOC. EDIN., VOL. LII, PART IV (NO. 33).

137

896 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

a stream or the edge of a small lake or pond, perhaps fed by the water from a
fumarole or siliceous spring. There appears to have been a primary vegetation of
Rhynia Gwynne-Vaughani over the moist clayey surface. This led at places to
a growth of peat formed in situ from the perished remains of this plant. Occasional
sporangia from plants of Rhynia major growing in the neighbourhood fell into
the peat as it formed. In the region represented by the main section of the bed,
this primary covering of Rhynia Gwynne-Vaughani was arrested, never to recur,
by the subsidence bringing water, probably charged with silica, over the surface.
The necroses and associated wound-reactions are especially marked at this level.
The abundance of remains of the small Crustacea and the presence of Algites Cranii
appear to further indicate the more aquatic conditions that supervened. It seems
probable that this provided the suitable environment for Hornea, the remains of
which were accumulated in situ as the Hornea peat of A"l. Aster oxylon appears
to have occurred in situ with its rhizomes penetrating this Hornea peat as drier
conditions recurred. Then followed the first definite sandstone band, perhaps mark-
ing a further period of subsidence and sedimentation. This is consistent with the
presence of Hornea (mixed with the remains of Asteroxylon ) throughout Bed B 6,
and the deposition of another definite sandstone layer above this. From this period'
onwards the growth of Rhynia major appears to have kept pace with the sub-
sidence, thus leading to the accumulation of a thick mass of peat. The sandy layers
interrupting this may indicate either floodings or accumulations of blown sand —
perhaps both. There are occasional indications of more aquatic conditions in the
course of the accumulation of this mass of Rhynia major peat. Finally the sub-
sidence seems to have been more active and more aquatic conditions returned,
bringing in Hornea, which replaced Rhynia major. With increasing depth, sediments
without plant-remains were deposited over the peat-bed.

While this interpretation of the changes of conditions indicated by the section of
the peat-bed studied is speculative, and it would be dangerous to press the com-
parisons into detail, some such way of looking at the facts seems justified. It enables
us to comprehend the mode of accumulation of the deposit in a gradually subsiding
marsh or small lake in much the same way as such regions get filled up with peat
at the present day. The differences from modern peats, as well as the resemblances,
must of course be borne in mind.

The interesting and puzzling fact of the occurrence of fragments of Nematojphyton
in the deposit has been regarded above as supporting the view that this anomalous
plant grew on land. How the fragments got into the peat, and what is the
geological and botanical interest of their presence, must remain to some extent open
questions pending further knowledge of this remarkable plant and of the Rhynie
deposit as a whole.

The accumulation of vascular plants in the Rhynie chert-band was clearly
derived from a continued growth on a land-surface that was at times submerged.

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 897

The abundance of fungi and the presence of an alga suggesting comparison with
Characese are consistent with this. No evidence of marine conditions has been
met with in the study of the plant-remains. The plant-containing chert discovered
at Rhynie thus provides a special, but quite definite, record of a land surface covered
with vegetation in the Early Old Red Sandstone of Scotland.

EXPLANATION OF PLATES.

(All the figures are from untouched photographs.)

Plate I.

Fungus No. 1.

Fig. 1. A strand of stout septate hyphse from a longitudinal section of a partially decayed stem of
Asteroxylon. x 100. (No. 2521.)

Fig. 2. Strand of hyphse from the same stem with only very occasional septa ; at places the branched
hyphse form irregular knots, x 100. (No. 2521.)

Fig. 3. Septate hyphse from the strand in fig. 1. x 250. (No. 2521.)

Fig. 4. Non-septate hyphse from the strand in fig. 2. x 250. (No. 2521.)

Fungus No. 2. ( Palxomyces Gordoni.)

Fig. 5. Portion of a longitudinal section of a stem of Asteroxylon. The soft tissues (phloem) of the
stele are replaced by strands of stout, longitudinally running hyphse similar to those in the preceding
figures. In the adjacent cortex are three large resting-spores of Fungus No. 2. x 34. (No. 2494.)

Fig. 6. Portion of a longitudinal tangential section of a partially decayed stem of Asteroxylon showing
hyphse and resting-spores of Fungus No. 2 in the middle cortex. x 50. (No. 2522.)

Fig. 7. Portion of same section as fig. 6. x 100. (No. 2522.)

Fig. 8. Some of the finer hyphse and collapsed vesicles in the same section. x 250. (No. 2522.)

Fig. 9. Single resting-spore of Fungus No. 2 borne on a stout lateral branch of a more slender hypha.
x 250. (No. 2522.)

Fig. 10. Mature resting-spore of Fungus No. 2 showing the double contour due to the inner layer of the
wall having separated from the outer layer, x 250. (No. 2522.)

Fig. 11. Another resting-spore showing fine threads of wall substance between the outer and inner layers
of the wall. x 210. (No. 2522.)

Plate II.

Fungus No. 3. ( Palxomyces Gordoni var. major.)

Fig. 12. Large thick-walled resting-spores of Fungus No. 3 with, to the right, a resting-spore inter-
mediate in size between this and the spores of Fungus No. 2. x 100. (No. 2527.) <,

Fig. 13. Another specimen showing the stout hyphal stalk on which the resting-spore is borne, x 100.
(No 2479.)

Fig. 14. Portion of the wall of the resting-spore of Fungus No. 3 showing an outer layer, a broad
middle layer, and an inner layer ; within the latter a thin layer bounding the contents is contracted away,
x 250. (No. 2527.)

898 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

Fungus No. 4.

Fig. 15. Group of resting-spores of Fungus No. 4 from a decayed stem of Asteroxylon. x 100.
(No. 2528.)

Fig. 16. Three of the resting-spores from fig. 15. x 250. (No. 2528.)

Fungus No. 5.

Fig. 17. Small, oval, thick- walled resting-spores of Fungus No. 5 from a decayed stem of Rhynia major.
x 100. (No. 2529.)

Fig. 18. Some of the resting-spores from fig. 17. x 250. (No. 2529.)

Fungus No. 6.

Fig. 19. Hyphse bearing oval and spherical vesicles from a partially decayed stem of Rhynia major.
x 250. (No. 2393.)

Fungus No. 7. ( Palxomyces Asteroxyli.)

Fig. 20. Hyphse of various diameters and vesicles of Fungus No. 7 from the decayed inner cortex of
a rhizome of Asteroxylon. x 100. (No. 2439.)

Fig. 21. Stout hyphse of Fungus No. 7, one of which is branching, x 250. (No. 2525.)

Fig. 22. Two longitudinally running hyphse of Fungus No. 7, showing a transverse, anastomosing

branch. x 250. (No. 2525 )

Fig. 23. Fine hyphse of Fungus No. 7 dilating into oval vesicles, x 250. (No. 2530.)

Fig. 24. Fine hypha of Fungus No. 7 dilating into a vesicle, x 250. (No. 2439.)

Plate III.

Fig. 25. Longitudinal section of outer portion of a rhizome of Asteroxylon showing the two layers of
outer cortex practically free from the fungus, the hyphse and vesicles of which are present between the
cells of the inner cortex, x 100. (No. 2525.)

Fig. 26. Portion of tangential section of the inner cortex of another rhizome of Asteroxylon showing
the intercellular position of the fungus. x 100. (No. 2468.)

Fig. 27. More broken-down portion of inner cortex of a rhizome showing the hyphse and vesicles of
Fungus No. 7. x 100. (No. 2530.)

Fig. 28. Two cells of the outer cortex of the rhizome of Asteroxylon shown in fig. 25. The figure shows
hyphse on the surface of the cortex with, to the left, a recurved, hook-like branch (a) ; on the right the
fungus with a vesicle (6) appears to be within a superficial cell, x 250. (No. 2525.)

Fungus No. 8. ( Palxomyces Hornex.)

Fig. 29. Portion of a section of the rhizome of Hornea showing hyphse and resting-spores of Fungus
No. 8 between the cells of the tissue. x 100 (No. 2531.)

Fig. 30. Outer portion of a rhizome of Hornea showing strands of stout hyphse and a hyphal tangle,
x 75. (No. 2531.)

Fig. 31. Vertical section of another rhizome of Hornea showing a hypha entering from the hyphal mass
outside and ramifying between the cells of the rhizome. x 75. (No. 2532.)

Fig. 32. Hypha of Fungus No. 8 between cells of rhizome. x 250. (No. 2531.)

Fig. 33. Portion of the hyphal knot shown in fig. 30. x 250. (No. 2531.)

Fig. 34. Resting-spore and hyphse of Fungus No. 8 in rhizome of Hornea. x 250. (No. 2531.)

Fig. 35. Group of mature resting-spores, presumably of Fungus No. 8, in relation to decayed rhizomes
and stem-bases of Hornea. x 100. (No. 2443.)

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 899

Plate IV.

Fungus No. 9. (Paheomyces vestita.)

Fig. 36. Longitudinal section of the cortex of the stem of Asteroxylon showing the resting-spores of
Fungus No. 9 distributed through the decayed tissues, and also within a large resting-spore of Fungus No. 2.
x 100. (No. 2454.)

Fig. 37. Resting-spores of Fungus No. 9 showing their investment by fine hyphse. x 250. (No. 2473.)
Fig. 38. Hyphse and resting-spores of Fungus No. 9 ; some of the latter are invested by the fine hyphse.
x 250. (No. 2454.)

Fungus No. 10. ( Palmomyces agglomerate/,.)

Fig. 39. Group of hyphse and resting-spores of Fungus No. 10 in relation to a decayed stem of Rhynia
Gwynne-Vaughani. x 100. (No. 2533.)

Fig. 40. Portion of group in fig. 39 showing hyphse and spores. x 250. (No. 2533.)

Fig. 41. Portion of group in fig. 39. Intercalary enlargements to form spores are seen in some hyphse.
x 250. (No. 2533.)

Fungus No. 1 1 .

Fig. 42. Hyphse and small vesicles of Fungus No. 11 breaking out from a tissue cell of Rhynia Gwynne-
Vaughani. x 250. (No. 2534.)

Fig. 43. Hyphse and small vesicles of Fungus No. 11 in the position of the dark layer of cells below the
outer cortex of a decayed stem of Rhynia Gwynne-Vaughani. x 500. (No. 2527.)

Fungus No. 12.

Fig. 44. Cortex of Rhynia Gwynne-Vaughani containing hyphse and vesicles of Fungus No. 12. x 100.
(No. 2535.)

Fungus No. 13. ( Palaeomyces Simpsoni.)

Fig. 45. Surface of a decayed stem of Rhynia Gwynne- Vaughani in a section from Mr Simpson, showing
an oval break in the surface beneath which are the crowded young vesicles of Fungus No 13. x 100.
(No. 2568.)

Fig. 46. Group of stalked vesicles of Fungus No. 13 proiecting from the surface of the same stem,
x 100. (No. 2568.)

Fig. 47. Some of the vesicles in fig. 46. x 250. (No. 2568.)

Plate V.

Fungus No. 14.

Fig. 48. Globular mass of fine hyphse and resting-spores and short stalk composed of associated fine
hyphse of Fungus No. 14. x 100. (No. 2533.)

Fig. 49. Resting-spores from specimen in fig. 48. x 250. (No. 2533.)

Fig. 50. Mass of hyphse and resting-spores of Fungus No. 14 enclosed by the thick wall of the resting
spore of Fungus No. 3. x 100. (No. 2536.)

Fig. 51. Fine hyphse of Fungus No. 14 filling the cavity of a partially decayed resting-spore of Fungus
No. 3 and its hyphal stalk. x 100. (No. 2479.)

Fig. 52. Resting-spore of Fungus No. 3 with groups of vesicles and hyphse of Fungus No. 14 on the
surface, x 100. (No. 2537.)

Fig. 53. Same specimen seen from other side in section. x 100. (No. 2537.)

Fig. 54. Another specimen of Fungus No. 3 showing vesicles and hyphse of Fungus No. 14 on the
surface, x 100. (No. 2428.)

900 DR R. KIDSTON AND PROF. W. H. LANG ON OLD RED SANDSTONE PLANTS

Fig. 55. Medium-sized, thick-walled, resting-spore of Fungus No. 3 containing the resting-spores formed
on hyphse of the intrusive Fungus No. 14. x 100. (No. 2443.)

Fig. 56. Spores within the thick boundary wall from the preceding figure. x 250. (No. 2443.)

Fungus No. 15.

Fig. 57. Group of thick-walled resting-spores of Fungus No. 2 ; two of these contain the intrusive
Fungus No. 15. x 100. (No. 2538.)

Fig. 58. One of the resting-spores in fig. 57 showing the spores of the intrusive Fungus No. 15. x 250.
(No. 2538.)

Plate VI.

Fig. 59. Resting-spore of Fungus No. 2 containing small spores of Fungus No. 15 within the main
cavity and also between the layers of the wall. x 100. (No. 2539.)

Fig. 60. Large resting-spore containing spores of the intrusive Fungus No. 15. x 250. (No. 2432.)

Fig. 61. Resting-spore of Fungus No. 2 containing well-preserved spores of the intrusive Fungus
No. 15. x 100. (No. 2427.)

Fig. 62. Spores from the specimen in fig. 61. x 250. (No. 2427.)

Fig. 63. Group of resting-spores of Fungus No. 2, two of which contain the intrusive Fungus No. 15.
x 100. (No. 2427.)

Figs. 64 and 65. Two of the specimens in fig. 63 more highly magnified to show the remains of fine
hyphse attached to the spores of the intrusive fungus, x 250. (No. 2427.)

Fig. 66. Spores of an intrusive fungus from another large resting-spore showing remains of the fine
hyphse. x 500. (No. 2540.)

Fig. 67. Resting-spore of Fungus No. 2 enclosing large thin-walled resting-spores of an intrusive
fungus, some of which in turn are occupied by a smaller fungus, x 250. (No. 2435.)

Fig. 68. One of the enclosed spores from fig. 67 showing the smaller spores and stalk-like remains of
hyphse within it. x 500. (No. 2435.)

Fig. 69. Resting-spore of Fungus No. 2, packed with spores of the intrusive fungus, which at one point
have burst through the wall. x 100. (No. 2541.)

Fig. 70. Portion of the preceding specimen showing the hyphse and vesicles bursting the enclosing wall,
x 250. (No. 2541.)

Fig. 71. Resting spore of Fungus No. 2 with the thin-walled spores of the intrusive fungus on its
exterior. x 225. (No. 2542.)

Fungi not sorted into Form-types.

Fig. 72. Spherical spores distributed through a decayed stem of Rhynia Gwynne-Vaughani. x 100.
(No. 2391.)

Fig. 73. Two of these spores, x 250. (No. 2391.)

Plate VII.

Fig. 74. Small fungal spores distributed through a decayed stem of Rhynia Gwynne-Vaughani. A large
thin- walled spore (c/. fig. 76) is present with them. x 100. (No. 2543.)

Fig. 75. Some of the small spores in the preceding figure. x 250. (No. 2543.)

Fig. 76. One of the large thin-walled spores accompanying the small spores (cf. fig. 74) containing
four well-defined spores. x 250. (No. 2543.)

Fig. 77. Small spores (?), possibly fungal, but with appearances suggesting algal cells. x 500.
(No. 2544.)

Fig. 78. The same, x 1000. (No. 2544.)

Fig. 79. Two spherical groups or colonies of dark cells of uncertain nature. x 100. (No. 2558.)

Figs. 80a, 80b. Two views of a resting-spore of a fungus with a peculiar covering of delicate radiating
tubes, x 500. (No. 2545.)

SHOWING STRUCTURE, FROM THE RHYNIE CHERT BED, ABERDEENSHIRE. 901

Fig. 81. Well-preserved fungal resting-spores borne on hyphae free in the matrix. x 210. (No. 2556.)

Fig. 82. Portion of peripheral region of a stem of Rhynia major, ep., epidermis; ox., outer cortex; x.,
layer of cells with problematical dark contents; i.c., inner cortex. x 100. (No. 2407.)

Figs. 83, 84. Cells of the layer with problematical dark contents in the preceding figure, showing some
indications of fine hyphal nature of the contents. x 250. (No. 2407.)

Schizophyta.

Fig. 85. Colonial masses of unicellular bacteria from matrix of the peat, x 100. (No. 2546.)

Fig. 86. Colonies of a recent bacterium, unstained, for comparison with fig. 85. x 100.

Plate VIII.

Fig. 87. Portion of the peaty matrix showing the felted arrangement of the fine filaments of a possible
filamentous bacterium. (Schizophyta No. 2.) x 100. (No. 2479.)

Fig. 88. Portion of preceding specimen. x 250. (No. 2479.)

Fig. 89. General arrangement of two groups of Archxothrix oscillator ij 'ormis in a cavity of a stem of
Rhynia Qwynne- Vaughani. x 100. (No. 2542.)

Fig. 90. Single filament of Archxothrix oscillatoriformis from the preceding figure, showing the small,
discoid cells, x 650. (No. 2542.)

Specimens received from Rev. Mr Cran, placed with Algites ( Palxonitella ) Cranii.

Figs. 91, 92. Views at two depths of focus of a small verticillately branched axis, x 100. (No. 2570.)

Fig. 93. Portion of same specimen showing three nodes with the whorled appendages, x 150. (No. 2570.)

Fig. 94. Transverse section of a node showing corticating cells and whorled appendages. x 200.

No. 2569.)

Fig. 95. Tubular rhizoid (?) with cellular node and two swollen cells borne at this, x 100. (No. 257 1.)

Fig. 96. Tubular rhizoid (?) with short branch bearing in relation to a septum a large globular cell

(bulbil?) with a shallow nodal disc at its base. x 100. (No. 2572.)

Fig. 97. Globular cell (bulbil?) with a nodal disc of small cells, some of which have filamentous growths
attached, x 100. (No. 2573.)

Plate IX.

Algites ( Palxonitella ) Cranii.

Fig. 98. Portion of peaty matrix giving a general view of the appearance of the algal remains, (a)
filament showing tip ; ( b , c) filaments with nodal discs ; (d) vesicle-like structure, x 50. (No. 2546.)

Fig. 99. Portion of same showing septate filaments with contracted dark contents. x 100. (No. 2546.)

Fig. 100. Apex of a filament with small terminal cell ; this is also seen to the left in fig. 99. x 100.
(No. 2546.)

Fig. 101. The node with short branch-like protrusions shown at ( b ) in fig. 98. The pointer indicates the
node, x 150. (No. 2546.)

Fig. 102. The node with rhizoid-like branches shown at (c) in fig. 98. ' The pointer indicates the node,
x 150. (No. 2546.)

Fig. 103. Rhizoid showing junction of two of the elongated cells, x 150. (No. 2388.)

Fig. 104. Another rhizoid showing irregular node of small cells at the junction of two elongated cells,
x 150. (No. 2388.)

Remains placed with Algites ( Palxonitella ) Cranii.

Fig. 105. Portion of peaty matrix showing numerous large thin-walled vesicles (? bulbils) associated with
tubular elements with contracted contents. x 12. (No. 2475.)

Fig. 106. Two of these vesicles connected together by a nodal cell. x 65. (No. 2476.)

902 OLD RED SANDSTONE PLANTS FROM RHYNIE CHERT BED, ABERDEENSHIRE.

Fig. 107. Immature vesicle (bulbil?), beside another only partially shown, with short protrusions from a
possible basal node. x 100. (No. 2475.)

Fig. 108. Specimen associated with vesicles like those in the preceding figures; it shows an internodal
cell and two whorls of branches at the lower and upper nodes, x 100. (No. 2476.)

Plate X.

Nematophyton Taiti.

Fig. 109. Portion of a fragment of Nematophyton Taiti showing the wide tubular elements and two of
the “medullary spots” (m.s.). x 50. (No. 2523.)

Fig. 110. View of a portion of the preceding specimen, x 100. (No. 2523.)

Fig. 111. A portion showing cross sections of the large tubes. x 250. (No. 2523.)

Fig. 112. Large tubes bounding a “medullary spot” (m.s.) packed with the narrow tubes.' x 250.
(No. 2523.)

Fig. 113. View of one of the sections of the second fragment which shows the peripheral tissues. x 30.
(No. 2525.)

Fig. 114. Portion of the other section of this specimen. x 50. (No. 2526.)

Fig. 115. Portion of fig. 113 showing the inner tubes and the base of the peripheral zone, x 50.
(No. 2525.)

Fig 116. Portion of the preceding section showing the inner tubes and their relation to medullary spots,
x 100. (No. 2525.)

Fig. 117. Some of the fine vertical tubes showing the spiral marking of the inner layer of the wall,
x 250. (No. 2523.)

Fig. 118. One of the inner tubes showing the spiral marking clearly and at places its relation to the
outer thick portion of the wall of the tube that has not wholly disappeared, x 250. (No. 2525.)

We again gratefully acknowledge our indebtedness to the Executive Committee of the Carnegie Trust
for a grant to defray the cost of the plates illustrating this Memoir.

(All the figured specimens in this series of Memoirs are in the collection of Dr R. Kidston.)

Vol: LI 1.

Trans: Roy: Soc: Edi

Kidston and Lang -

Plate I

Kidston and Lang. Photo:

Zit

Collotype Co., Edinburgh.

Trans: Roy: Soc: EdinT

Vol: LI 1 .

Kidston and Lang --Plait H

Kl DSTON AND LANG PHOTO :

Collotype Co., Edinburgh.

Trans: Roy: Soc: Edinf

Kidston and Lang-: Plait III

Vol: LI!.

Kidston and Lang Photo:

Zinco Collotype Co.. Edinburgh

Trans: Roy: Soc: Edin!

Vol: LI 1.

Kidston and Lang -:Pla'ie H

Kl DSTON AND LANG PHOTO :

Collotype Co.. Edinburgh.

Trans: Rov: Soc: Edinf

Kidston and Lang -? Plate Y

Vol: LI 1 .

Zinco Collotype: Co., Eoinburqm.

Kidston and Lang Photo:

S3

Trans: Roy: Soc: Edinf

Kidston and Lang -e Plate M

Vol: LI 1 .

Kidston and Lang Photo :

Zinco Collotype Co.. Edinburgh.

Trans: Roy: Soc: EdinT

Kidston and Lang-: Plate W

Vol: LI 1.

Kl DSTON AND LANG PHOTO :

Zinco Collotype Co., Edinburgh.

Trans: Rov: Soc: EdinF

Kidston and Lang-: Plate MI

Vol: LI 1.

Kl DSTON AND LaNG PHOT' I :

Zinco Collotype Co.. Edinburg

Trans: Roy: Soc: Edinf

Kidston and Lang-: Plate IX

Vol: LI 1.

Kidston and Lang Photo:

Zinco Collotype Co., Edinburgh

Trans: Rov: Soc: Edinf

Kidston and Lang-: Plate X

Vol: Lll.

Kidston and Lang Photo:

Zinco Collotype Co.. Edinburgh.

( 903 )

INDEX.

A

Aberfoyle District , The Highland Border Rocks of
the. By T. J. Jehu and Robert Campbell,
175-212.

Amphicheiral Knots. By Mary Gertrude Hase-
man, 597-602.

Amphicheiral s icith Twelve Crossings, On Knots,
with a Census of the. By Mary Gertrude
Haseman, 235-255.

Anatomy and Affinity of certain Rare and Primitive
Ferns. By J. M‘L. Thompson, 363-397.

Anatomy of the Lower Dicotyledons : Stem of the
Berburidaceae. By R. J. Harvey-Gibson and
Elsie Horsman, 501-515.

Stem of the Calycanthaceae. By Christine

E. Quinlan, 517-529.

Antarctic Expedition, Scottish Rational, 1902-1 904 :
Cambrian Organic Remains from Weddell-Sea.
By W. T. Gordon, 681-714.

Shackleton, 1914-1917 : Depths and De-
posits of Weddell Sea. By J. M. Wordie,
781-793.

Shackleton, 1914-1917 : The Natural His-
tory of Pack-Ice in the Weddell Sea. By
J. M. Wordie, 795-829.

Assort ative Mating in Mendelian System, Statistical
Effects of. By R. A. Fisher, 399-433.

Asteroxylon Mackiei , Kidston and Lang. By R.
Kidston and W. H. Lang, 643-680, 836.

B

Baldwin (It.). See Doddson, A. T., and R. M.
Carey.

Bees, Isle of Wight Disease in Hive. By John
Rennie, P. B. White, and Elsie J. Harvey,
737-779.

Berberidace x, Anatomy of Stem of the. By R. J.
Harvey-Gibson and Elsie Horsman, 501-515.

Bionomics, Structure, and Forest Importance of
Myelophilvs minor Hart. By Walter
Ritchie, 213-234.

Border Rocks, Highland, of the Aberfoyle District.
By T. J. Jehu and Robert Campbell, 175—

0

Calciferous Glands of Earthworms. By J. Stephen-
son and Baini Prashad, 455-485.

Calycanthaceae, Anatomy of Stem of the. By
Christine E. Quinlan, 517-529.

Cambrian Organic Remains from the Weddell Sea :
Scottish National Antarctic Expedition, 1902-
1904. By W. T. Gordon, 681-714.

Cameron (A. E.). The Insect Association of a Local
Environmental Complex in the District of
Holmes Chapel, Cheshire, 37-78.

Campbell (Robert). See J ehu, T. J.

Carey (R. M.). See Doodson, A. T., and R.
Baldwin.

Ceratopteris, R. Br., Sporangial-forin and spore-
counts of. By J. M‘L. Thompson, 363-397.

Cheilanthes, Sporangial-form and spore-coxrnts of.
By J. M‘L. Thompson, 363-397.

Correlation between Relatives on the supposition of
Mendelian Inheritance. By R. A. Fisher,
399-433.

Cryptogramme crispa (L.), R. Br., Sporangial-form
and spore-counts of. By J. M‘L. Thompson,
363-397.

D

Darneli.-Smith (G. P.). The Gametophyte of
Psilotum, 79-91.

Davie (R. C.). On the Leaf-trace in some Pinnate
Leaves, 1-36.

Depths and Deposits of the Weddell Sea. By J. M.
Wordie, 781-793.

Development of Heart in Man. By D. Waterston,
257 -302.

Dicotyledons, Lower, Anatomy of Stem of the
Berberidacese. By R. J. Harvey-Gibson and
Elsie Horsman, 501-515.

Anatomy of Stem of the Calycanthaceae. By

Christine E. Quinlan, 517-529.

Dominance, Statistical Effects of. By R. A. Fisher,
399-433.

Doodson (A. T.), R. M. Carey, and R. Baldwin.
Theoretical Determination of the Longitudinal
Seiches of Lake Geneva, 629—642.

212.

TRANS. ROY. SOC. EDIN., VOL. LI I, PART IY.

138

904

INDEX.

E

Earthworms , Calciferous Glands of. By J.
Stephenson and Baini Prashad, 455-485.

Family Megascolecidee, Prostate Glands of.

By J. Stephenson and Haru Ram, 435-453.
Elasmohranch Fore-brain, Morphology of a type of.

By J. Stuart Thomson, 487-500.

Environment , Contribution of, to Statistios of
Human Measurements. By R. A. Fisher,
399-433.

Evolution of Vascular System in Ferns. By J.

M‘L. Thompson, 715-735.

Ewart ( J. Cossar) and Dorothy Mackenzie. The
Moulting of the King Penguin, 115-132.

F

Factorials and Allied Products with their Logarithms.

By Frank Robbins, 167-174.

Ferns , New Stelar Facts, and their Bearing on
Stelar Theories. By J. M‘L. Thompson, 715—
735.

Rare and Primitive, Anatomy and Affinity of

certain. By J. M‘L. Thompson, 363-397.
Ferret-Polecat Hybrids, The Formation, Rupture,
and Closure of Ovarian Follicles in. By A.
Robinson, 303-362.

Ferrets, Formation, Rupture, and Closure of Ovarian
Follicles in. By A. Robinson, 303-362.

Fin of Fishes, Shoulder Girdle and Pectoral. By I
E. W. Shann, 531-570.

Fisher (R. A.). On the Correlation between Rela-
tives on the supposition of Mendelian Inherit-
ance, 399-433.

Fishes, Shoulder Girdle and Pectoral Fin of. By
E. W. Shann, 531-570.

Forest Importance, Structure, and Bionomics of
Myelophilus minor Hart, By Walter Ritchie,
213-234.

G

Gametophyte Generation of the Psilotacese. By
A. Anstruther Lawson, 93-113.

of Psilotum. By G. P. Darnell-Smith, 79-91.

Generation, Gametophyte, of the Psilotacese. By’
A. Anstruther Lawson, 93-113.

Glands, Calciferous, of Earthworms. By J.
Stephenson and Baini Prashad, 455-485.

Prostate, of Earthworms. By J. Stephenson

and Haru Ram, 435-453.

Gordon (W. T.). Cambrian Organic Remains from
a Dredging in the Weddell Sea : Scottish
National Antarctic Expedition, 1902-1904,
681-714.

Gymnogramme japonica, Desv., Stelar Anatomy of.
By J. M‘L. Thompson, 363-397.

H

Harvey (Elsie J.). Experiments in Infection of
Hive Bees with Tarsonemus woodi, n. sp.,
765-767.

Harvey-Gibson (R. J.) and Elsie Horsman. Con-
tributions towards a Knowledge of the
Anatomy of the Lower Dicotyledons. II.
The Anatomy of the Stem of the Berberidacese,
501-515.

Haseman (Mary G.). On Knots, with a Census of
the Amphicheirals with Twelve Crossings,
235-255.

Am phieheiral Knots, 597-602.

Heart in Man, Development of. By D. Water-
ston, 257-302.

Highland Border Rocks of the Aberfoyle District.
By T. J. Jehu and Robert Campbell, 175-212.

Homogamy, Statistical Effects of, in Mendelian
System. By R. A. Fisher, 399-433.

Hornea Lignieri, n.g., n.sp. By R. Kidston and
W. H. Lang, 603-627, 836,

Horsman (Elsie). See Harvey-Gibson (R. J.).

I

Insects of a Local Environmental Complex. By
A. E. Cameron, 37-78.

Isle of Wight Disease in Hive Bees. By John
Rennie, P. B. White, and Elsie J. Harvey,
737-779.

J

Jamesonia, Hk. Gr., Anatomy and Affinity of. By
J. M‘L, Thompson, 363-397.

Jehu (T. J.) and Robert Campbell. The High-
land Border Rocks of the Aberfoyle District,
175-212.

K

Kidston (R.) and W. H. Lang. On Old Red
Sandstone Plants showing Structure, from the
Rhynie Chert Bed, Aberdeenshire. Part II.
Additional Notes on Rhynia Gwynne-Vaughani,
Kidston and Lang ; with Descriptions of
Rhynia major, n.sp., and Hornea Lignieri,
n.g., n.sp., 603-627.

Part III. Asteroxylon Mackiei, Kidston and

Lang, 643-680.

Part IY. Restorations of the Vascular

Cryptogams, and Discussion of their Bearing
on the General Morphology of the Pteri-
dophyta and the Origin of the Organisation
of Land Plants, 831-854.

Part V. The Thallophyta occurring in the

INDEX.

905

Peat Bed ; the Succession of the Plants
throughout a Vertical Section of the Bed ; and
the Conditions of Accumulation and Preserva-
tion of the Deposit, 855-902.

Knots, On, with a Census of the Amphicheirals
with Twelve Crossings. By Mary Gertrude
Haseman, 235-255.

Amphicheir,il. By Mary Gertrude Hase-
man, 597-602.

L

Lang (W. H.). See Kidston (B>.).

Lawson (A. Anstruther). Gametophyte Genera-
tion of the Psilotacese, 93-113.

Leaf-trace in some Pinnate Leaves. By R. C.

Davie, 1-36.

Leaves, Pinnate, Vascular Anatomy of. By R. C.
Davie, 1-36.

Llavea, Lagasca, Anatomy and Affinity of. By
J. M‘L. Thompson, 363-397.

Logarithms of Factorials and Allied Products. By
Frank Robbins, 167-174.

M

Mackenzie (Dorothy). See Ewart (■!. Cossar).

Man, Development of Heart in. By D. Waterston,
257-302.

Megascolecidx, Family, Prostate Glands of Earth-
worms of. By J. Stephenson and Haru
Ram, 435-453.

Mendelism, Correlation between Relatives under.
By R. A. Fisher, 399-433.

Morphology of the Stele of Platyzoma micro-
phyllum, R. Br. By J. M‘L. Thompson,
571-595.

Moulting of the King Penguin. By J. Cossar
Ewart and Dorothy Mackenzie, 115-132.

Multiple Allelomorphism, Effects of, upon Correla-
tions of Relatives. By R. A. Fisher, 399-
433.

Myelophilus minor Hart, The Structure, Bionomics,
and Forest Importance of. By Walter
Ritchie, 213-234.

Myology, Comparative, of Shoulder Girdle and
Pectoral Fin of Fishes. By E. W. Shann,
531-570.

N

Nerve-fibre Tracts in the Prosencephalon of Spinax.
By J. Stuart Thomson, 487-500.

Neurone Areas in the Prosencephalon of Spinax.
By J. Stuart Thomson, 487-500.

Nothochlxna, Broug., Sporangial-form and spore-
counts of. By J. M‘L. Thompson, 363-397.

O

Old Red Sandstone Plants showing Structure. Part
II. By R. Kidston and W. H. Lang, 603-627.

Part III. By R. Kidston and W. H. Lang,

643-680.

Part IV. By R. Kidston and W. H. Lang

831r854.

Part V. By R. Kidston and W. H. Lang,

855-902.

Origin of Pith in Ferns. By J. M‘L. Thompson,
715-735.

Ovarian Follicles, Formation, Rupturej and Closure
of, in Ferrets and Ferret-Polecat Hybrids. By
A. Robinson, 303-362.

P

Pack-Ice in Weddell Sea, Natural History of. By
J. M. Wordie, 795-829.

Penguin, Moulting of the King. By J. Cossar
Ewart and Dorothy Mackenzie, 115-132.

Physiography and Topography in relation to an
Insect Association. By A. E. Cameron,
37-78.

Plant Environment in relation to Insects. By A. E.
Cameron, 37-78.

Platyzoma microphyllum, R. Br., Sporangial- and
Spore-Differences in. By J. M‘L. ' Thompson,
157-165.

Morphology of Stele of. By J. M‘L. Thomp-
son, 571-595.

Prashad (Baini). See Stephenson, J.

Products, Factorials and Allied, with their Log-
arithms By Frank Robbins, 167-174.

Prosencephalon of Spinax, Morphology of. By J.
Stuart Thomson, 487-500.

Prostate Glands of the Earthworms of the Family
Megascolecidse. By J. Stephenson and Haru
Ram, 435-453.

Psilotacese, Gametophyte Generation of the. By
A. Anstruther Lawson, 93-113.

Psilotum, the Gametophyte of. By G. P. Darnell-
Smith, 79-91.

Q

Quinlan (Christine E.). Contributions towards a
Knowledge of the Anatomy of ,'the Lower
Dicotyledons. Part III. The Anatomy of the
Stem of the Calycanthaceae, 517-529.

R

Ram (Haru). See Stephenson, J.

Relatives, Correlation between, in Mendelian
System. By R. A. Fisher, 399-433.

906

INDEX.

Bennie (John). Tarsonemus ivoodi, n. sp. : tlie I
Causal Organism in Isle of Wight Disease in
Hive Bees, 768-779.

Philip Bruce AVhite, and Elsie J. Harvey. J

The Etiology of Isle of Wight Disease in Hive
Bees, 737-754.

Rhynia Gwynne-Vaughani, Kidston and Lang. By
R. Kidston and W. H. Lang. 603-627, 832.

Rhynia major, n.sp. By R. Kidston and W. H.
Lang, 603-627, 835.

Rhynie Oliert Bed, Aberdeenshire, On Old Red
Sandstone Plants showing Structure, from the.
Part II. By R. Kidston and W. H. Lang,
603-627.

Part III By R. Kidston and W. H. Lang,

643-680.

Part IV. By R. Kidston and W. H. Lang,

831-854.

Part Y. By R. Kidston and W. H. Lang,

855-902.

Ritchie (Walter). The Structure, Bionomics,
and Porest Importance of Myelophilus minor
Hart, 213-234.

Robbins (Frank). Factorials and Allied Products
with their Logarithms, 167—174.

Robinson (A.). The Formation, Rupture, and
Closure of Ovarian Follicles in Ferrets and
Ferret-Polecat Hybrids, and some Associated
Phenomena, 303-362.

S

Scottish Notional Antarctic Expedition, 1902-1904:
Cambrian Organic Remains from the Weddell
Sea. By W. T. Gordon, 681-714.

Sea-Ice: Natural History of Pack-Ice in the
Weddell Sea. By J. M. Wordie, 795-829.

Seiches of Lake Geneva, Theoretical Determination j
of the Longitudinal. By A. T. Doodson,
R. M. Carey, and R. Baldwin, 629-642.

Sliackteton Antarctic Expedition, 1914-1917 :
Depths and Deposits of the Weddell Sea. By
J. M. Wordie, 781-793.

Natural History of Pack-Ice. By J. M.

Wordie, 795-829.

Shann (E. W.). The Comparative Myology of the
Shoulder Girdle and Pectoral Fin of Fishes,
531-570.

Shoulder Girdle and Pectoral Fin of Fishes. By
E. W. Shann, 531-570.

Soil-insect Census. By A. E. Cameron, 37-78.

Spihax, Morphology of the Prosencephalon of. By
J. Stuart Thomson, 487-500.

Stele of Platyzoma microphyllum, R. Br. By
J. M‘L. Thompson, 571-595.

Stephenson (J.) and Baini Prashad. The Calci-
ferous Glands of Earthworms, 455-485.

Stephenson (J.) and Haru Ram. The Prostate
Glands of the Earthworms of the Family
Megascolecidse, 435-453.

Stromatopteris moniliformis, Mett., Anatomy and
Affinity of. By J. M‘L. Thompson, 133-156.

Structure, Bionomics, and Forest Importance of
Myelophilus minor Hart. By Walter Ritchie,
213-234.

T

Tarsonemus woodi, li. sp. : Isle of Wight Disease
in Bees. By Elsie J. Harvey, 765-767.

By John Rennie, 768-779.

Thallophyta occurring in Rhynie Chert Bed,
Aberdeenshire. By R. Kidston and W. H.
Lang, 855-902.

Thompson (J. M‘L.). The Anatomy aud Affinity
of Stromatopteris moniliformis, Mett., 133-156.

— — A Further Contribution to the Knowledge of
Platyzoma microphyllum, R. Br., 157-165.

The Anatomy and Affinity of certain Rare

and Primitive Ferns, 363-397.

The Morphology of the Stele of Platyzoma

microphyllum, R. Br., 571-595.

— — New Stelar Facts, and their Bearing on
Stelar Theories for the Ferns, 715-735.

Thomson (J. Stuart). The Morphology of the
Prosencephalon of Spinax as a type of Elasmo-
branch Fore-brain, 487-500.

Trismeria, Fee, Anatomy and Affinity of. By
J. M‘L. Thompson, 363-397.

W

Waterston (D). The Development of the Heart
in Man, 257-302.

Weddell Sea, Cambrian Organic Remains from ;
Scottish National Antarctic Expedition, 1902-
1904. By W. T. Gordon, 681-714.

Depths and Deposits. By J. M. "Wordie,

781-793.

■ Natural History of Pack-Ice in. By J. M.

Wordie, 795-829.

White (Philip Bruce). Pathology of Isle of
Wight Disease in Hive Bees, 755-764.

Wordie (J. M). Shackleton Antarctic Expedition,
1914-1917: Depths and Deposits of the
Weddell Sea, 781-793.

■ Natural History of Pack-Ice in the Weddell

Sea: Shackleton Antarctic Expedition, 1914-
1917, 795-829.

PRINTED

GREAT BRITAIN BY NEILL

CO., LTD., EDINBURGH.

The Transactions of the Royal Society of Edinburgh will in future be Sold at the

following reduced Prices : —

Vol.

Price
to the
Public.

Price

to

Fellows.

Yol.

Price
to the
Public.

- Price
to

Fellows.

Vol.

Price
to the
Public.

Price

to

Fellows.

£

s.

d.

£

s.

d.

£

s.

d.

£

s.

d.

£

pfe.j i

d.

£

s.

d.

VI.

0

11

6

0

9

6

XXXV.* Pt. 1

2

2

0

1

11

0

XLVI. Pt. 1

1

1

10

0

16

6

VII.

0

18

0

0

15

0

„ Pt. 2

1

11

0

1

3

6

„ Pt. 2

1

5

8

0

19

4

VIII.

0

17

0

0

14

0

„ Pt. 3

2

2

0

1

11

0

„ Pt. 3

1

7

3

1

0

11

IX.

1

0

0

0

17

0

„ Pt. 4

1

1

0

0

16

0

XL VII. Pt. 1

0

19

9

0

15

0

X.

0

19

0

0

16

0

XXXVI. Pt. 1

1

1

0

0

16

0

„ Pt. 2

1

3

0

0

17

4

XI.

0

14

6

0

12

0

„ Pt. 2

1

16

6'

1

7

6

„ Pt. 3

1

0

10

0

15

8

XII.

0

14

6

0

12

0

„ Pt. 3

1

0

0

0

16

0

„ Pt, 4

1

7

7

1

0

9

XIII.

0

18

0

0

15

0

XXXVII. Pt. 1

1

14

6

1

5

6

XLVIII. Pt. 1

1

2

9

0

17

2

XIV.

1

5

0

1

1

0

„ Pt. 2

1

1

0

0

16

0

„ Pt. 2

1

9

6

1

2

5

XV.

1

11

0

1

6

o|

„ Pt. 3

0

16

0

0

12

0

„ Pt. 3

1

11

0

1

3

3

XX. Pt. 1

0

18

0

0

14

0

„ Pt. 4

0

7

6

0

5

8

„ Pt. 4

0

16

8

0

12

4

XXII. Pt. 2

0

10

0

0

7

6

XXXVIII. Pt. 1

2

0

0

1

10

0

XLIX. Pt. 1

0

7

6

0

7

6

„ Pt. 3

1

5

0

1

1

0

„ Pt. 2

1

5

0

0

19

0

„ Pt.2

1

12

6

1

4

6

XXVII. Pt. 1

0

16

0

0

12

0

„ Pt. 3

1

10

0

1

3

0

„ Pt. 3

1

10

0

1

2

5

„ Pt. 2

0

6

0

0

4

6

„ Pt. 4

0

7

6

0

5

8

„ Pt. 4

1

1

9

0

16

0

„ Pt. 4

1

0

0

0

16

o

XXXIX. Pt. 1

1

10

0

1

3

0

L. Pt. 1

1

5

9

0

18

9

XXVIII. Pt. 1

1

5

0

1

1

0

„ Pt. 2

0

19

0

0

14

6

„ Pt. 2

1

7

0

1

0

0

„ Pt. 2

1

5

0

1

1

0

„ Pt. 3

2

3

0

1

11

0

„ Pt. 3

1

7

0

1

0

0

,. Pt. 3

0

18

0

0

13

6

„ Pt. 4

0

9

0

0

7

0

„ Pt. 4

1

2

0

0

17

0

XXIX. Pt. 1

1

12

0

1

6

0

XL. Pt. 1

1

5

0

0

19

0

LI. Pt. 1

1

5

0

0

19

0

„ Pt- 2

0

16

0

0

12

0

Pt. 2

1

12

6

1

5

6

„ Pt. 2

1

13

0

1

5

0

XXX. Pt. 1

1

12

0

1

6

0

„ Pt. 3

1

6

0

0

19

6

„ Pt. 3

1

15

0

1

6

0

„ Pt. 2

0

16

0

0

12

0

„ Pt. 4

1

0

0

0

16

0

„ Pt. 4

1

11

6

1

3

6

„ Pt. 3

0

5

0

0

4

0

XLI. Pt. 1

1

1

0

0

15

9

LII. Pt. 1

1

7

0

1

0

6

„ Pt. 4

0

7

6

0

5

8

„ Pt. 2

1

9

6

1

2

0

„ Pt. 2

1

10

0

1

2

0

XXXI.

4

4

0

3

3

0

„ Pb 3

2

5

0

1

13

6

„ Pt. 3

2

3

0

1

11

6

XXXII. Pt. 1

1

0

0

0

16

o1

/ XLII.

2

2

0

1

11

0

„ Pt. 4

2

13

0

2

0

0

„ Pt. 2

0

18

0

0

13

6

1 1 .

) XLIII.

2

2

0

I

11

0

„ Pt. 3

2

10

0

1

17

6

1 XLIV.Pt.l

1

18

6

1

9

0

„ Pt. 4

0

5

0

0

4

0

( „ Pt.2

1

1

0

0

15

9

XXXIII. Pt. 1

1

1

0

0

16

0

XLV. Pt. 1

1

9

0

1

2

0

„ Pt. 2

2

2

0

1

11

0

„ Pt. 2

1

7

0

1

0

0

„ Pt. 3

0

12

0

0

9

6

„ Pt. 3

1

13

9

1

5 .

3

xxxiv.

2

2

0

1

11

0

„ Pt. 4 !

0

4

6

0

3

6

* Yol. XXXV, and those which follow, may be had in Numbers, each Number containing a complete Paper.

January 1921. — Volumes or parts of volumes not mentioned in the above list are not for the present on sale
to the public. Fellows or others who may specially desire to obtain them must apply direct to the' Society. As
the Society reprints from time to time parts of its publications which have become scarce, the absolute correctness
of this list cannot be guaranteed beyond this date.

The papers published in this Part of the Transactions may be had on application to the Publishers,

at the following prices: —

No. XXVII,
No. XXVIII,
No. XXIX,
No. XXX,.

Price to Public.
9s. Od.

5s. 6d.

9s. Od.

2s. Od.

Price to Fellows.

7s. Od.

4s. 3d.

6s. 9d.

Is. 6d.

No. XXXI,
No. XXXII,
No. XXXIII,

Price to Public. Price to Fellows.
8s. Od. 6s. Od.

6s. 6d. 5s. Od.

. 13s. Od. 9s. 9d.