ТНЕ
PRANSACTLONS
THE LINNEAN SOCIETY

LONDON.

SECOND SERIES.—VOLUME VIII.

BOTANY.

LON DON:
PRINTED BY TAYLOR AND FRANCIS, RED LION COURT, FLEET STREET,

SOLD AT THE SOCIETY’S APARTMENTS, BURLINGTON HOUSE, PICCADILLY, W. 1.

AND BY LONGMANS, GREEN, AND CO., PATERNOSTER ROW, E.C. 4.

1913-1922.

pe

ПІ.

en

Liv.

<

CONTENTS.

PART I.—Janvary, 1913.

A Contribution to а Knowledge of the Mutating Ginotheras. Ву REGINALD В.
GATES, M.A., Ph.D:, Royal College of Science, South Kensington. (Сот-
municated by Prof. J. BRETLAND Farmer, F:R.S., F.L.S.) (Plates 1-6.)

pages 1—68

PART II.—May, 1913.

. 1— Оп some Indian Jurassic Gymnosperms (Plates 7-9), and 2.— Rhexoxylon

africanum, а new Medullosean Stem (Plates 10 & 11). Ву NELLIE BANCROFT,
B.Sc. (Lond.), F.L.S., 1851 Research Scholar, Newnham College, Cambridge.
69-103

PART ШІ.--Остовев, 1913.

Marine Alge, Rhodophycee, of the ‘Sealark’ Expedition, collected by Mr. dix
Stanley Gardiner, М.А. Ву Mrs. A. WEBER-VAN Bosse, Ph.D. (Сот-
municated by Prof. J. STANLEY GARDINER, M.A., F.R.S. FDS.) (Plates
19-14 поа 1 Texbfügnre) . Q . . + — ; i . . . 105-142

PART IV.—DEGEMBER, 1913.

Observations on the Morphology and Anatomy of the Genus Mystropetalon, Harv.
Ву В. J. Harvey Ствзох, M.A., F.L.S., Professor of T in " University
of Liverpool. (Plates 15 & 16) ... _ 22. 148-154

PART V.—DECEMBER, 1913.

. On the Cuticles of some Recent and Fossil Cycadean Fronds. Ву Н. HAMSHAW

Tuomas, M.A., F.G.S., Curator of the Botanical Museum, Cambridge, and
NELLIE Bancrort, B.Sc., F.L.S., 1851 Research Scholar, Newnham College,
Cambridge. (Plates 17-20 and 32 Text-figures.) . . + + + + + 155-204

PART VI.—Jvury, 1914.
- VI. An Anatomical Study of the Paleozoic Cone-genus Lepidostrobus. By AGNES
ARBER, D.Sc., F.L.S., Fellow x Newnham College, Cambridge. (Plates 21-27
and 4 Text-figures.) . . 2... bee . . 205-238

PART VII.—DECEMBER, 1914.

VII. A Contribution to the Study of the Evolution of the Flower, with special reference
to the Hamamelidaces, Caprifoliaceze, and Cornacez. By ARTHUR SAMUEL
Hornet, B.Sc., F.L.S., F.G.S., Lecturer on Botany in the Royal Horticultural

Society's School of Horticulture. (Plates 28-30 and 13 Text-figures.) . 239-309

PART VIII.—DECEMBER, 1915.

уш. Notes on the Morphology of certain Structures concerned in Reproduction in the
Genus Gnetum. By Н. Н. W. Pearson, Sc.D., F.L.S., Henry Bolus Professor

of Botany in the South African College. (Plates SP & 32.) . . . 311-222
` Ow Isoëtes japonica, 4. Br. Ву OvgiL West, FL S., and Н. TAKEDA, D.I.C.
(Plates 33-40 and 20 Text-figures.) . . dc с со 888-376

PART IX.—SEPTEMBER, 1922.
Пера Contente, and ет, ее. ват

2nd Ser. BOTANY.) 52270 UL VI DAT E

NTRIBUTION TO A KNOWLEDGE OF THE

_ REGINALD R.

LRANSAGTIIONS

OF

TUS INN GAN SOCIETY.

I. A Contribution to a Knowledge of the Mutating Oenotheras. By REGINALD В.
GATES, M.A., Ph.D.; Royal College of Science, South Kensington. —
by Prof. J. BRETLAND Farmer, F.R.S., F.L.S.)

(Plates 1-6.)

Read 7th November, 1912.

ConTENTS. Page
I. Introduction | aif ces nd bs eere ica ыы eas csv i sss Cie ет V VI sys S 1
II. Conspectus of Species and Mutations of the Onagra group of Oenothera........ 11
Key to Oenothera species of the Onagra group .......................... 12
III. Relationships of Species ............................................. 14
IV. Variations and Mutations of О. Lamarckiana, Ber. ........................ 20
O. grandiflora, Solander ............................ 38
О. бюз, LIB. атс 222. ве u y l u еен Vi WE 89
dA si din, es l... аа, 40
V. (A) Cultures of Oenotheras from Lancashire......... esea ey en 43
(В) Cultures from Botanical Gardens .................................. 47
VI. Hybridization in Oenothera aserat. isune ee lerra ee eer eie 48
(а) Crosses involving O. Lamarckiana and its Mutants ....... Led у 48
(b uo. О ре нан нн. уле Мева Ту о бв ае сенн г 50
(с) „ between O. biennis and О. Lamarckiana and its derivatives .... 52
ОТИШ еее. 55
УП. General Considerations and Summary .................................. 58
ҮШІ. Bibliography ...... Аа ари л нь 62
ІХ. Explanation of the Plates ....... TNT icc л ciki: 64

I. INTRODUCTION.

THE purpose of the present paper is to bring together and organize certain results of

experimental work which I have carried on with the Oenotheras during the last six

years. My cytological work with these m which is already well known, will on
SECOND шише NY, VOL. VIII. O5

2 DR. В. В, GATES—CONTRIBUTION ТО А

be referred to incidentally, the main purpose being the presentation of certain results of
studies in variation and hybridization with these forms, and a discussion of their general
bearing on our view-point with regard to the mutation problem in Oenothera.

I may say at the outset that my purpose has been to study these mutation phenomena
from as many points of view as possible, in the belief that all the data from every
source could finally be brought together and applied in the explanation of the
evolutionary significance of this type of behaviour. The greater the amount of
information already at hand with regard to any group of forms, the wider is the basis
of interpretation for all additional facts, and I have therefore felt that more could be
gained by attempting to consider a single group of organisms from every angle than by
studying forms belonging to many genera from any given point of view—e. g., that of
cytology or hybridization. This is my apology for continuing to devote my attention to
one somewhat limited group. I believe this intensive method is a valuable one, and in
the present state of our knowledge is to be recommended for other genera as well.

I have therefore studied the Oenotheras, not only as regards their cytology, but in
the details of their external characters, their variability, and hereditary behaviour.
This has involved also a study of the history of the mutating and related species in
cultivation, and as naturalized in Europe and other parts of the world; as well as
numerous cultures of wild forms from many parts of the North American continent, to
determine the relationships and distribution of the many geographic races. The
results are of course very incomplete, and even fragmentary, but I think the method is
amply justified by certain results already obtained: for example, the way in which the
cytological structure of O. gigas and its peculiar behaviour in crossing are of mutual
help in explaining each other; and the manner in which cultures of certain wild forms
of O. muricata from Canada have thrown light on the probable origin and significance
of the twin types in certain hybrids of O. Lamarckiana.

The large quantity of data with regard to the mutating Oenotheras which has
accumulated since the publication of ‘ Die Mutationstheorie’ by DeVries (1901-3) has
led to a number of modifications of the views expressed by the author of that |
monumental work with reference to the mutation phenomena which he first described.

Subsequent studies of Oenothera Lamarckiana by MacDougal (1908, 1905, 1907),
myself, and others have confirmed іп a general way the accuracy of DeVries’s
observations concerning the hereditary behaviour of this species. The general facts
regarding the sudden, occasional, discontinuous origin of aberrant types which breed
true have been amply confirmed. But there remains the all-important question of the
evolutionary significance of this and similar behaviour. For an adequate interpretation
of these phenomena, much fuller data in a variety of directions were required. Some
of these data have now been supplied, and have thrown new light on the nature and
meaning of the process.

A more detailed discussion of the mutation phenomena in Oenothera Lamarckiana
uw will be taken up in later sections of this paper. In this introductory statement it will |

2. be well to view the — theory in certain of its broader aspects, and to ask what X

KNOWLEDGE OF THE MUTATING OENOTHERAS. 3

place can now be assigned to mutation as an evolutionary factor. Without anticipating
the more minute analysis of the mutation process in Oenothera, I may be permitted to
state certain points of view at which I have arrived.

One of the facts which I shall be particular to emphasize in this paper is that,
although a continuous series can be formed between a mutant and its parent in the
case of many characters, yet each of the mutants is actually discontinuous in its origin,
its organs exhibiting a decided break from those of the parent.

If a colony of O. Lamarckiana and its mutants were growing together in large
numbers, any casual observer would probably consider that they represented one general
"widely-variable or polymorphic species. Yet experiments show that each type originates
discontinuously, and that the differences, which are exhibited in every organ, are as
great as are to be seen in elementary species. Some of the forms would very probably
be given specific rank by “splitters” if they were collected wild, while the others
would certainly rank at least as varieties, or forms.

It has to be remembered, however, that in cultures the characters are better displayed
and developed than can usually be the case with wild forms under conditions of
competition with surrounding vegetation. Іп this serse, and also through improvement
in the conditions of nutrition, cultivation brings out differences which would otherwise
escape the observer, but it is probable that there is no other direct effect of cultivation,
at least in early years. There is, however, some rather cogent evidence (cf. Hill, 1912)
for the view commonly held by horticulturists, that after a period of years, cultivation
induces variability. Darwin also believed in the existence of this effect, which he
ascribed to “changed conditions.” Definite experimental proof of such an effect is
greatly needed. It seems probable that change of climate or of soil might account for
all such cases of induced variability, rather than any direct effect (such as increased
nutrition or freedom from competition) of cultivation itself. Many such cases are, of
course, to be explained by crossing.

This brings us to a consideration of the status of О. Lamarckiana. As regards
its characters, it is undoubtedly a “good species,” differing throughout from
O. grandiflora, O. biennis, and O. muricata, its nearest relatives. Its mutants
are grouped like a constellation of satellites around it, and none of them exhibit any
character closely approaching any other known species. Nearly all agree with
О. Lamarckiana in certain bud-characters and in pubescence, though they all differ
markedly in foliage. As regards the cause of this type of ‘ variability,” several reasons
might be ascribed: (1) crossing; (2) change of climate or other effect of cultivation ;
(3) internal mutational changes independent of any of these.

Considering the first of these, it is clear that О. Lamarckiana, like all other open-
pollinated plants, must have at one time undergone crossing. This fact has been
emphasized in another paper (Gates, 1911 4), and here I need only state that whether
О. Lamarckiana originally existed in the wild state as an isolated species or, as seems
more probable, was only one of many freely intercrossing races along with certain forms

of О. grandiflora, is immaterial i if we admit, as we must, that open-pollinated colonies

=

4 DR. R. В. GATES—CONTRIBUTION TO A

of plants freely intercross. This is the condition under which their evolution takes
place. If, as some think, О. Lamarckiana originated in Gardens through crossing
between, е. g., О. grandiflora and О. biennis, then it would probably have also
originated long previously through crossing where these occupied the same area in
Eastern North America. Indeed, this conclusion is difficult to avoid, and I have
previously shown (Gates, 1911 о) that both О. biennis and О. grandiflora survived in the
region of Virginia and Carolina until a century ago.

As discussed later in this paper, a Lancashire colony of Oenotheras which has existed
for certainly over a century, contains O. Lamarckiana as the dominant form. Such
facts as these serve to show that the essential points to decide in determining the
status of mutation as an evolutionary factor are, (1) will O. Lamarckiana survive in
competition with its near relatives, and (2) will the mutants survive in competition with
О. Lamarckiana. The former point has already been settled affirmatively by a study of
the Lancashire colony already mentioned. The only actual evidence yet obtained for
the latter is the fact that DeVries found certain forms, such as the putative mutants
О. brevistylis and О. laevifolia, persisting year after year with О. Lamarckiana, in the
field at Hilversum near Amsterdam where they were originally found. Competition
experiments will be required before it can be known which of the mutants will survive
under various climatic conditions.

DeVries has emphasized the fact that certain of the mutants seem to differ in their
physiological adjustments or adaptations; O. gigas, for instance, being more strongly
biennial than the other forms. But the value of any such differences can only be
determined by actual experiments on the survival of the mutants under various climatic
conditions.

The ingenious theory of DeVries, that new forms appear suddenly, already adapted to
various environments, and are afterwards so distributed that each reaches the environ-
ment for which it is most fitted, in spite of certain obvious difficulties regarding its
survival during the period of its distribution, may account for certain cases of what we
may call climatic adaptation. But such a principle can scarcely be extended to the
more complex cases of adaptation in plants and animals which it is one of the main
duties of evolutionists to explain. I may quote the celebrated passage from Darwin
(° Origin of Species,’ 6th Edition, page 2) in which he cites two such cases :—'* It is pre-
“ posterous to attribute to mere external conditions, the structure, for instance, of the
“ woodpecker, with its feet, tail, beak, and tongue, so admirably adapted to catch insects
“ under the bark of trees. In the case of the mistletoe, which draws its nourishment
“from certain trees, which has seeds which must be transported by certain birds, and
“which has flowers with Separate sexes absolutely requiring the agency of certain
“insects to bring pollen from one flower to the other, it is equally preposterous to
_ “account for the structure of this parasite, with its relations to several distinct organic
| ud ешг » by the effects of external conditions, or of habit, or of the volition of the plant

KNOWLEDGE OF THE MUTATING OENOTHERAS. 5

the origin of such complex biological relationships. Darwin’s principle of natural
selection, although unproven in such cases, still remains the only adequate theoretical
explanation.

To return to the question of crossing, and its effect upon O. Lamarckiana in relation
to the mutation phenomena: as I have stated elsewhere, it seems probable that this
behaviour is an expression of the germinal instability resulting from crossing of races
in the ancestry of the parent form. It is evidently not the re-appearance of forms
which entered into that ancestry, пог is it merely a case of hybrid splitting as that 13.
now understood. DeVries has classified mutations as progressive, retrogressive, and:
degressive, the last being the re-appearance of characters which had previously
disappeared. From my study of the mutants of Oenothera they appear to me to be
(1) positive, due to plus variation in one or more characters, of which О. rubricalyx is
an example; (2) negative, due to the minus variation or loss of one or more characters, ,
e.g., О. nanella; and (3) mutations involving a divergent change in various characters:
toa new condition of equilibrium. Most of the mutants of Oenothera belong to the
last category, and even the positive and negative mutations involve changes in all the
other organs in most cases—e. g., О. gigas, О. lata, О. nanella. Occasional positive or
negative mutants occur in Oenothera, in which only one character is affected—e. g.,
pigmentation in О. rubricalyx. Mutations affecting single characters appear to be much
more common in other genera. But it is not certain that there is a real distinction here
between qualitative and quantitative mutations. In any case, DeVries is doubtless
correct that whether a single character or many are affected, the originating cause, in.
each сазе is a single germinal change. Changes which affect every organ, including
foliage and flower characters, are probably more fundamental than those affecting only
colour characters, or the size of a particular organ.

I have elsewhere pointed out (Gates, 19114, p. 599) that the appearance of the
mutants О. gigas and О. rubricalyx cannot be accounted for on the basis of hybrid
splitting, and also that the mutational changes may occur at different points in the life
cycle. Тһе evidence for the heterozygous character of the original O. rubricalyx
mutant, and hence for the occurrence of this particular change during the reduction
divisions in one spore mother-cell, is particularly clear; while the evidence from the
tetraploid number of chromosomes іп О. gigas (Gates, 1911 с) makes it probable that
the change here may have occurred in the fertilized egg or in an apogamously developed
megaspore*. The theory of DeVries regarding the origin of the mutants has been that
in every case the change occurred in the spore mother-cells, each mutant originating
from the union of the product of a mutated and a non-mutated spore. There
are, however, obvious difficulties with this view, as will be pointed out in a moment.
My cytological work (see Gates, 1908) with the Oenotheras showed the presence of

id Since this was written, Stomps (1912) and Miss Lutz (1912) have observed triploid mutants from Lamarckiana
having 21 chromosomes. This increases the probability that tetraploid mutants may originate from the union of `
two unreduced germ-cells, but. it by no means shows that they do not also 2 in the manner I have |
^6 Er I shall discuss this matter in detail elsewhere. |

6 . DR, В. В. GATES—CONTRIBUTION TO А

occasional irregularities in the distribution of chromosomes during meiosis, which
furnish a possible basis for the origin of negative mutants, provided that the chromo-
somes in Oenothera are qualitatively different. But since the chromosomes of these
plants are morphologically undifferentiated, there is at present no opportunity of
following this possibility further. Chromosome distributions in the germ-cells are
doubtless concerned in the origin of mutants having 15 or 16 chromosomes.

То return to the difficulty regarding the hypothesis that a mutation is of hybrid
nature, arising from the fusion of a mutated with an unaltered cell, let us suppose, for
instance, that O. rubrinervis arises from a mutated germ-cell crossed with an ordinary
O. Lamarckiana germ-cell. If this is so, why does O. rubrinervis breed true when self-
pollinated, yet when crossed with O. Lamarckiana it splits in the F, giving О. Lamarckiana
and О. rubrinervis, both of which remain constant in later generations? It might be
said that the primary “mutated” germ-cell was different from a rubrinervis germ-cell,
but that when erossed with O. Lamarckiana it gave a combination which bred true and
produced the rubrinervis characters. However, if this were admitted, there is no reason
for supposing that our rubrinervis germ-cell when crossed with O. Lamarckiana would
behave in an entirely different manner and give alternative inheritance. То explain the
origin of a mutation on this basis, one must assume a change in the hereditary
behaviour of the germinal material in two successive generations where one would
expect it to be the same. Why should a mutated germ-cell, crossed with a non-
mutated one, give a blend which never splits, while the germinal material derived from
this blend, when crossed with another non-mutated germ-cell of O. Lamarckiana, always
exhibits alternative inheritance? Such a condition of affairs is conceivable but not
probable. It involves a paradox which has not yet been explained away.

Leclere du Sablon (1910) has formulated the ingenious hypothesis that the phenomena
of mutation in O. Lamarckiana are to be explained as a case of Mendelian splitting in
which there is a very high degree of “ coupling" between the characters. But this
seems to be merely putting the facts into another notation without really adding to our
knowledge of why the new forms appear. Moreover, it fails to take account of the
change in chromosome number in O. gigas. This nuclear change places O. gigas in a
different category from the other mutants.

I have already mentioned the hypothesis that the mutants may be due to occasional
irregularities in the distribution of the meiotic chromosomes. Another, and possibly
‚ more probable, explanation is that though the mutations are internal in origin, yet they are
released by an external stimulus which affects the cytoplasm, and indirectly the nuclear
‘Structure. Ifthis isan approach to the true explanation, the releasing stimuli must be very
slight. This view is difficult to harmonize with the fact that the mutants seem to appear
with about the same frequency wherever O. Lamarckiana is cultivated. However,
comparatively little is known concerning such environmental responses іп organisms.

72 "ir ally, we may agree that mutation,in so far as it is distinguished from hybrid-
tting (seo р. 47), probably accounts for much species-formation. Many polymorphic
are probably indebte _ to such à process for the large number of closely

KNOWLEDGE OF THE MUTATING OENOTHERAS. 7

related forms they contain. It seems evident that in many cases the differences
between these elementary species (using the term in DeVries’s sense) are not of value
for selection. .The limits of natural selection as an eliminating factor are probably
much more circumscribed than extreme selectionists have supposed.

If we refer briefly to other factors which are evidently concerned in species-origin
and evolution, we must compare mutation, as an evolutionary factor, with natural
selection. If there is any one point in Darwinian natural selection which has not
been fully appreciated in post-Darwinian discussions, it is, perhaps, that Darwin
assumed, at least in most cases, an original environmental change for the organism,
which led to the necessity for its readjustment to its environment, modification of the
species being produced during this period of readjustment. Thus the change in the
climate of а given area or the introduction of a species into а new area with different
environmental conditions are continually referred to by Darwin as the initial changes
leading to modification of particular species. When no environmental changes were
taking place, natural selection was looked upon as maintaining the nicely-balanced
equilibrium between the species. It is always a climatic change, or the entrance of
new organisms, which disturbs that equilibrium, and natural selection then becomes a
factor producing modification. True, Darwin often emphasizes the keenness of the
struggle for existence, and its effect in bringing and maintaining all the organs of the
competing organisms at their highest degree of efficiency. But when some new
departure is to be made in the structure of any organism, а primary environmental
change is first supposed to throw the organism to one side of the centre of its
equilibrium, this being followed by progressive change through the effects of the
selection of variations, until the organism is once more adjusted to the mean of its
organic and inorganie environment.

Two distinct effects of natural selection were therefore recognized by Darwin:
(1) а conservative effect, maintaining the race at its highest degree of efficiency while
the environmental conditions remain unchanged; and (2) a modificational effect,
causing the selection of new variations and the final establishment of a new equilibrium
when the previous equilibrium has been disturbed by an organic or an inorganic
environmental change.

Darwin’s distinction between these two independent effects of natural selection is by
no means always clear. But recent experimental work, particularly, has emphasized
the importance of natural selection as a conservative force, maintaining the condition of
equilibrium which exists between any organism and its whole environment. It is a
condition to which organisms have been almost universally subjected during the whole
period of evolution. But it is not a force, and it is not as constant and unintermittent
in its effect, as, e. g., gravitation. Тһе modificational effect of selection has, on the other
hand, thus far remained unproven, notwithstanding the experimental researches of the
last decade. As a result of the negative evidence from experimental work, we are now
in much the same position with regard to both selection and the inheritance of
acquired modifications. "They are still unproven, though there is a certain amount of

8 DR. В. В. GATES—CONTRIBUTION ТО A

rather incomplete and not-fully-convineing evidence for both. The Lamarckian factor,
which has always received support from anatomists and paleontologists, has in recent
years experienced a distinct renaissance, particularly in the attempts of Semon,
Rignano, and others to eliminate its main weakness by formulating а biological
mechanism by means of which the transmission of acquired modifications could
take place. |

То return to mutation and selection, the latter at least, and perhaps also the former,
requires “changed conditions” in order to produce a modificational effect. But the
vicissitudes of changing climates and distributions cannot be sufficient to produce the
orderly phylogenies which frequently appear when we view larger groups of organisms
as a whole, and especially when we consider the broader outlines of the palzeontological
record. Hence, at this point both mutation and selection appear to break down as
evolutionary factors. The larger simultaneous evolutionary trends exhibited in the
history of organisms appear to present problems which are more or less apart from
mere questions of species-formation in each group. These problems, together with
those of the origin of phyla, have been comparatively little realized in the literature of
experimental evolution. On the other hand, the experimental researches of the last
decade have made it more plain how closely is species-origin in each group bound up
with the biological inter-relationships of the organisms composing that group. By such
means it may be possible ultimately to distinguish between factors leading to the
production of more or less incidental and evanescent species, and those which have been
more vitally concerned in the phylogeny of the group. It seems at present that
mutations such as occur in Oenothera Lamarckiana are chiefly incidental, rather than
phylogenetic in value, though the lack of perspective in the case of such recent pro-
ductions makes it difficult to judge of their real value. The only final way to answer
this question is by determining whether the mutants will survive in competition and
leave descendants. |

Perhaps we шау sum up the present status of evolutionary investigation with the
statement that recent researches have resulted chiefly in revealing the inadequacy of
single factors, such as natural selection and mutation, to account for all evolution.
Both these factors have doubtless played their part, however, together with the neo-
Lamarckian factor, orthogenesis, and other factors, some of which have doubtless not
yet been perceived. New organisms do not arise by any single method, but evolution is
a multifarious process, and the deeper currents of evolutionary progress are still very
little understood. x

The greatest advances since Darwin have been made in our experimental knowledge
_ of heredity and its structural basis, and in the laws of variability. The foundations
for the laws of inheritance laid by the experiments of Mendel, and so rapidly developed.
and amplified by Bateson and the Mendelian school with active workers in several
countries, have shown at least one fundamental fact—namely, the alternative expression
of many characters on crossing, in a wide range of plants and animals. It remains to

KNOWLEDGE OF THE MUTATING OENOTHERAS. 9

determine how the segregational type of behaviour is related to blended inheritance,
and what features the two have in common.

As.regards the laws of variability, we have been enabled to draw a sharp distinction
between heritable and non-heritable variations, although, since character-changes having
the same external appearance may belong in either category, we cannot safely predict
what the behaviour of a variation will be in any case until it is experimentally
tested. Notwithstanding the attention which has been focussed on discontinuous
variations or mutations in recent years, particularly by the work of DeVries, even those
which are evidently due to the simple loss of characters are still largely beyond the
range of our explanation. But it may reasonably be hoped that with further knowledge
‘of their origin and causes they may be brought, at least to a limited extent, under
experimental control. Any further advances in this direction will be of great theoretical

as well as practical value.

These investigations with Oenothera have been carried out at the University of Chicago,
the Marine Biological Laboratory, Woods Hole, Mass., and the Missouri Botanical Garden.
They were continued last year at the J ohm Innes Horticultural Institution, Merton,
Surrey. То the Directors of all these institutions I am very greatly indebted, for the
facilities afforded for carrying on my cultures, and keeping photographie and other
records. Таш also indebted to the Carnegie Institution for a grant in aid of expenses
in 1908. І have further to thank Professor Farmer for suggestions in connection with
the preparation of this paper, and a Grant from the Royal Society has defrayed the
expenses of publishing the plates. Numerous botanists from many parts of the world
have very kindly sent me seeds of Oenothera species, and I shall be most grateful to
any botanists who will send me further seeds of wild or naturalized species from any
country. The study of naturalized colonies of Oenothera will furnish an important part
of the data required.

In connection with these cultures, as the variety of races under observation has
multiplied, and the number of new forms recognized has increased, the difficulties in
connection with the adoption of a method of nomenclature which would be serviceable
to experimental evolutionists and at the same time acceptable to systematists have
become increasingly pressing. After much thought and discussion on the subject, and
consultation with Dr. O. Stapf, F.R.S., Keeper Royal Herbarium, Kew, and Dr. A.
B. Rendle, F.R.S., British Museum (Natural History), I venture to suggest the
following usages.

The names of forms which are (1) known only in cultures, or (2) which have originated
in cultivation, or (3) the description of the behaviour of which is based on material long
in cultivation, shall be followed by the abbreviation сий., to distinguish them from forms
derived directly from wild sources.

This usage is already followed partly by systematists, and may perhaps be adopted
with advantage by breeders. Like every rule, however, there will be plenty of

SECOND SERIES.— BOTANY, VOL. VIII. с

10 DR. В. В. GATES—CONTRIBUTION TO А

difficulties in its application, and judgment will be necessary in deciding each case.
The greatest difficulty will probably be in connection with (3), in determining how
many years of culture may elapse before а given form must be considered “іп culti-
vation.” It has always been the case, and rightly so, that many species were described
by systematists from the growing plants; and the modern tendency is to describe species
as far as possible from their living characters. Obviously, such forms are not to be
classed as cultivated species. But when new forms are recognized and segregated
after many years of culture they may perhaps fairly be considered as “cultivated”
forms. The fact that such species were first recognized in cultures will not, however,
justify the assumption that they do not also occur wild. Also, the use of the term,
with, e. g., O. Lamarckiana, Ser., does not necessarily mean that the species originated
in eultivation, but merely that the material under discussion was derived from a race
long in cultivation. Thus, the same species may be referred to in one connection as a
wild, and at another time as a cultivated form.

Similarly, mutations or forms which have originated through germinal changes under
experimental control, so that the history of their origin is known, should be preceded
by the abbreviation mut.

The greatest value of the usage suggested above, and, indeed, the reason for this
suggestion, lies in the fáct that it will enable breeders to use the convenient binomial
terminology for forms which are in many cases nof strictly comparable to wild species,
without the danger of their being mistaken for such by systematists. Thus I have in
my cultures at the present time numerous forms, some of which have originated through
crossing, yet which breed fairly true and differ from each other quite as much as do the
ordinarily accepted species in many genera, others which differ to a much slighter
extent or even in single characters. For breeding purposes many of these various
races urgently require description and a name by which one сап refer to them. Such
races should be grouped around the recognized species to which they are most nearly '
allied. If cult. be written after the names of such races, this will avoid the danger of
seriously confusing them with the wild forms, while leaving open the question whether
their differences are of specific, varietal, or lesser value. At the same time, this will not
relieve systematists from the duty of examining all such descriptions, and according
specific rank and treatment to any forms which, by their distinctness, deserve it,
regardless of whether they originated in cultivation or elsewhere.

It will be easy to raise numerous objections to these suggestions, but it is believed
| that they will at least serve the purpose of making clearer the status of the many forms `
Which must perforce be described by experimental workers before any permanent record |
of such experiments can be made.

adoption of the term mut. as a designation for mutations
mentally known, will, I believe, be particularly useful.

whose origin is

KNOWLEDGE OF THE MUTATING OENOTHERAS. 11

II. CONSPECTUS OF SPECIES AND MUTATIONS OF THE ONAGRA GROUP OF OENOTHERA.

О. grandiflora, Solander* -------- О. Tracyi, Bartlett.
О. MacBridee, Nels.
O. macrosceles, A. Gray.

cognata, Hort. Panorm. cult = зед.

mut. gigas, H. DeVr. > О. mut. gigas nanella.
mut. rubrinervis, Н. DeVr. —— О. mut. rubricalyz, Gates.
mut. /ata, Н. DeVr. » O. mut. lata nanella.
mut. semilata, H. DeVr.

mut. oblonga, H. DeVr.

mut. scintillans, Н. DeVr. ——— 0. mut. scintillans nanella.
mut. elliptica, H. DeVr.

mut. albida, H. DeVr.

mut. leptocarpa, H. DeVr.

mut. fatua, H. DeVr.

mut. spathulata H. DeVr.

mut. sublinearis, H. DeVr.

mut. subovata, H. DeVr.

О. Lamarckiana, Ser. cult. —— <

> > Е Ss

'

'

i

i

i

!

yar. cruciata. 0. mut. nanella oblonga.

O. mut. nanella elliptica.

O. mut. nanella lata.

O. mut. nanella scintillans.

O. mut. scintillans.

O. mut. laevifolia salicifolia.
О. mut. laevifolia nanella.

| О. mut. rubrinervis ?

О. mut. laevifolia, H. ПеУт.-->

(
О. mut. пале а, H. DeVr. ———> и
5

. | О. mut. brevistylis, Н. DeVr.
O. Jamesii, Torr. & Gray.
‚ О. argillicola, MacK.
O. Hookeri, Torr. & Gray.
O. Drummondii, Hook.
O. Simsiana, Ser.
O. ornata, Nels.
O. biennis, Linn.
O. canovirens, Steele.
0. Oakesiana, Robbins.
_ О. rhombipetala, Nutt.
0. strigosa, МасК. & Bush.
-Q cheradophila, Sec
_ O. muricata, Linn. |
(——— var. canescens, Robinson.
0. ammophila, due (an. ‘P. >-

12 DR. В. В. GATES—CONTRIBUTION TO А

Key to Oenothera species of the Onagra group.*
Petals cruciate, narrow and linear.
Cruciate varieties of O. Lamarckiana and O. biennis are described in DeVries,
‘Die Mutationstheorie,’ ii. pp. 593-633.
Petals broad ; flowers large (petals 30-60 mm. long).
Style short, so that the stamens surround the stigma in the bud ; flowers about
7 ст. in diameter, petals 4 cm. long
Style long, so that the flowers are not м жыл тү in i the bad:

Stems decumbent; plants clothed with soft pubescence, or canescently.

strigose.
Leaves ovate-elliptical or oblong, rather obtuse .
Leaves oblong-lanceolate, repandly denticulate, acute . ‘
Stems erect, more or less muricated ; flower-buds nearly or qiite жй,
bearing few or по long muricate hairs. Leaves usually more or less
ciliate and pubescent.
Rosette-leaves narrowly oblanceolate, tapering above the acute apex.
Stem-leaves smaller, passing into sessile bracts б сз;
Rosette-leaves spatulate-lanceolate with x ТЯ Stem-leaves
subsessile, oblong-lanceolate
Rosette-leaves oblong-spatulate, шген at the я pina ti d:
toothed base into a long, broad petiole. Stem-leaves ovate-lanceolate,
tapering at both ends, denticulate, blade passing below into a short
margined petiole . eov we
Rosette-leaves broadly көзі to э Habana panies bite. Stem-leaves
undulate or repand-denticulate, linear to linear-oblong .
Stems erect, muricated. Flowers not glabrous, but rather qawsely —
with muricated hairs.
Leaves canescently pubescent, plane, rather narrowly lanceolate, sessile,
obscurely denticulate. . ‘ «
Leaves and usually buds less або (бет ыы har)
Petals broad ; flowers smaller (petals about 12-30 mm. long).
Style exceeding the stamens. Leaves narrowly oblong-lanceolate to linear-
lanceolate. Calyx densely white hirsute-pubescent
Style short, not reaching beyond the anthers in the bud.
Sepal tips spreading ; rosette-leaves narrowly lanceolate, approximately
toothed at the acute apex, more deeply toothed аб the slender,
tapering base. Seeds large

LI LI .

Sepal tips appressed ; plant annual, with с rosette of a few и;
red-spotted leaves, closely resembling О. grandiflora. Stem-leaves

... lanceolate, remotely denticulate, acute at both ends í ;
Sepal tips appressed ; plant biennial. Lower stem-leaves 15-20 ‹ cm. Каш
|. oe to elliptic-lanceolate, shallowly toothed, and often

x | -toothed near base, acuminate. Free а жык э tips about 2 mm.

` . LI ` . Ld ° ` . .

О. Simsiana.

O. Drummondii.
О. Jamesi.

О. MacBridee.

О. macrosceles. |

0. grandiflora.

O. argillicola.

O. Hookeri.
O. Lamarckiana and
[its segregates.

O. ornata.

0. Oakesiana.

O. Tracyi.

0. Dentist. а стен

seem to ће: — Kewensis.’

F.R.S., Director of the Royal Botanic Gardens, Kew, for permision u to
(unpublished) su

KNOWLEDGE OF THE MUTATING OENOTHERAS. 13

Sepal tips appressed ; plant biennial. Stem-leaves numerous and crowded,
excurved-ascending, lower leaves 9-10 cm. long, lanceolate-oblong, tip
very acute; petiole and midvein broad, the latter somewhat succulent.

Leaves densely cinereous-pubescent, blue-green . . . О. canovirens.
Petals rhombic-ovate, 12-25 mm. long. Stem-leaves ойу КЕБЕ
acuminate, denticulate or subentire . i . О. rhombipecala.
Petals 15-20 mm. long, light yellow; бара: tips 3 mm. “tgs Plants
strigose and somewhat villous ; stigma-lobes flattish, spreading . . 0. strigosa.
Petals 9-15 mm. long ; leaves narrow in most races but broad in certain
Taboss {ла eub o seg muricata:
Whole plant foi: ints or nel uy v u s xe dicus DU. Vr Тағ, балейсейы

Petals about 8 mm. long.
Sepal tips about 3 mm. long, style short, petals broad (9 mm.) . . . О. parviflora.

Sepal tips about 3 mm. long, style short, petals cruciate . . var. cruciata.

Sepal tips less than 1 mm. long, style short, petals broad ; т"
оо ee S ре г Огейтайортш.

Petals lanceolate, style long, foliage soft-pubescent . . . . . . + О. Heribaudi.

The following salient features of these species, as far as they сап be judged from
present knowledge, may be pointed out :—

О. Simsiana is sharply distinguished by its large flowers with short style.

O. Drummondii and О. Jamesii have more or less decumbent stems, the former being
very distinct with its short, ovate-elliptical leaves; the latter nearly agreeing with
О. Lamarckiana in foliage characters, but ee differing chiefly in its decumbent
stem and greater pubescence.

O. MacBridee is apparently nearest to O. grandiflora, from which it differs chiefly
in its narrower leaves and somewhat shorter style %

О. macrosceles comes nearest О. Jamesii and О. Lamarckiana in foliage, but differs
from the former in its upright stem, and from both in being nearly glabrous.

O. argillicola is — set apart from O. grandiflora by its very narrow, linear-oblong
leaves.

ж Since the above was written I have been able, through the kindness of Professor Nelson, who sent me seeds, to
grow this species and study its characters. The following notes are appended from my observations on this species.

Cotyledons very long and narrow, twice the length of related forms. Rosettes of characteristic rather broadly
lanceolate-oblong leaves. Central stem erect, surrounded by many long. spreading branches from the rosette.
Cauline leaves long and narrow, pointed, pale green, midrib white, surface smooth, margin repand-denticulate.
Hypanthium, bud-cone, and ovary closely set with long hairs arising from faint red papilla. Sepals pale pink with
touches of red on the hypanthium, вера! tips short and closely appressed. Petals 44 mm. long by 52 mm. broad,
emarginate, length of bud-cone 40 mm., diameter of cone at base 9 mm., length of hypanthium 48 mm., ovary
19 mm., вера! tips 4 mm., thickness of ovary 8-3-5 mm., thickness of hypanthium 3 mm. Base of stigma usually
some distance above the stamens.

Тһе buds of these plants were apparently not so nearly glabrous as one might judge from Nelson’s description.
The plants do not resemble О. grandiflora closely in buds or in foliage, nor do they come in the biennis series, since
ed have eri flowers and mesa сете

14 DR. В. В. GATES—CONTRIBUTION ТО А

Oenothera Hookeri differs from О. Lamarckiana and its mutants in its canescent
pubescence and the absence of crinkling in the leaves.

О. ornata apparently forms a transition to the smaller-flowered O. biennis forms,
though its style exceeds the stamens. It appears to differ from the latter in its
denser hirsute-pubescence and somewhat larger flowers with longer style.

O. Tracyi in general characters agrees closely with O. grandiflora, from which it

differs chiefly in its smaller flowers with short style.
~ O. biennis as now understood has leaves long, elliptic-oblanceolate to elliptic-
lanceolate, shallowly toothed and frequently with jagged teeth near the base. The
petals vary from 12-25 mm. long.

О. Oakesiana differs from О. biennis in foliage, but also in having spreading sepai
tips and large seeds.

О. canovirens, another segregate from О. biennis, differs from that form markedly
in foliage, the petioles being somewhat succulent, the leaves much shorter, densely
cinereous-pubescent, and blue-green.

О. rhombipetala differs from О. biennis in foliage, but chiefly in having rhombic-ovate
petals.

О. strigosa differs from О. biennis chiefly in being strigose, the foliage variable.

О. cheradophila differs from О. strigosa, which it most resembles, chiefly in having
much smaller flowers (petals 8 mm. or less) with shorter sepal tips.

О. muricata differs constantly from the О. biennis series of forms in having smaller
flowers (petals 9-15 mm. long) Most of {Ше races also have narrow leaves, but
certain ones have broad leaves. |

О. parviflora has still smaller flowers than O. muricata (petals 8 mm. long). 1% also
differs in foliage characters.

О. Heribaudi agrees with О. parviflora in flower-size. It has the habit of О. biennis,
but has а long exserted style, and its foliage resembles that of О. sinwata, a form
which does not come strictly within the Onagra group.

It is а matter of opinion how far О. Oakesiana, О. canovirens, 0. rhombipetala,
O. strigosa, O. cheradophila and similar segregates from O. biennis should be recognized
as distinct species. ‘‘ Splitters ” will recognize many at least as such, while others, such
as Lévéillé in his monograph of the genus, will “lump” them all as races of one species. -

That these different races will breed true there can be no doubt. Indeed, even smaller
_ differences remain constant in cultures.

ІІІ. RELATIONSHIPS OF SPECIES.

General Distribution —The genus Oenothera is well-known to be exclusively American

. in origin, although in the last three centuries, since its introduction into European a

Botanical it has escaped and is now naturalized and growing wild in шапу =
p ts of the world. | in es large-flowered forms related to О. Lamarckiana are | 2
n to have been | rowing in great nibon on — — coast since at 2

KNOWLEDGE OF THE MUTATING OENOTHERAS. 15

pl. 1534.) They are also found in various other parts of England, and are widely
distributed, particularly on sand-dunes and along railway-lines in France, Germany,
Holland, and in fact all over Europe. They also occur in such regions as the Azores,
South Africa, Australia, Japan, &с., and I have obtained new races with unexpected
characters in cultures of seeds from some of these localities. "The origin of such races is
a problem of much interest, and may furnish important data on the question of the
effects of climate in the origin of races. It is very desirable that detailed studies of
naturalized colonies of Oenotheras in various parts of the world should be undertaken.

Specific Differences.—The species with which the students of the mutating Oenotheras
have been chiefly concerned are О. Lamarckiana, О. grandiflora, О. biennis, and
О. muricata. An analytical key (Gates, 1909 с) gives the main distinguishing features
of these species and their segregates and relatives, and subsequent papers (Gates,
1911 а, 1911 4) have traced the history of the origin of these names as they were
applied to races introduced into European Gardens. It seems desirable to point
out that after long familiarity with these forms I find it necessary to conclude
that the limits between the species are purely arbitrary. Fixed limits can only be
maintained as a matter of convenience in dealing with a host of forms. For races are
found which overstep any boundary-line which can be drawn between each species and
those to which it is most closely related. Тһе type of each species is only one of many
related races which have come to be included with the type in systematic descriptions.
Which of these races originally received the name of a given species was purely a matter
of chance, because it was the first to come under the botanist's notice. But while these
types furnish useful landmarks, yet, when whole series of additional races are observed,
they must either be given new specific names or classed as races under certain species
already described. Тһе advisability of giving specific rather than racial names to these
newer-discovered forms will not be discussed here, but it is sufficiently clear that names
of some sort are necessary in dealing with them, either from the standpoint of a more
complete knowledge of the flora of any country, or from that of the evolutionary
experimentalist who must have names to designate the various forms with which he
works. When these forms differ in single characters, such as flower-colour, which can
be described in one word, that word is all that is necessary for a name. But when such
geographic races differ slightly in many characters affecting leaf, stem, flower, and fruit,
as in Oenothera, the necessity for an accurate description and а name for each such race
is sufficiently evident, and is highly desirable both from the systematic and the breeding
point of view.

Since, therefore, more or less arbitrary lines of distinction between the species are
necessary I have, for convenience, adopted the following :— 0. Lamarckiana and
O. grandiflora forms nearly agree in flower-size, but the latter have no long muricated
hairs, or very few, on their sepals, while the former possess at least a considerable
number. Тһе O. grandiflora races are also characterized by possessing rosette-leaves
with deep basal lobes, while these do not occur in the known O. Lamarckiana races.
. The O. biennis races differ from the two previous groups only in having smaller flowers |

16 DR. В. В. GATES—CONTRIBUTION ТО А

(diameter about 3 cm.) with short style, so that the anthers surround the stigma and
self-pollination takes place in the bud. The many wild races of О. biennis differ widely
in leaf-characters and vary somewhat in size of flower. The races of O. muricata have
smaller flowers with usually shorter style than O. biennis. This is the only distinction
which can be drawn which will apply to all known races. Many narrow-leaved races
of O. muricata occur, but certain races have broad leaves as well as small flowers, so that
the latter character alone is retained as delimiting this group of races (see Gates, 1911 d,
p. 580).

Below are represented the lines of difference which may be drawn between these four
species and their relatives and segregates :—

1. Flowers large (length of petals 35-50 mm.). Style long and flowers open-
pollinated (except in О, brevistylis, О. Simsiana, and certain other

forms).
a. Buds smooth or bearing only short hairs, rosette-leaves usually having
basallobes . . . 12.2. 0. grandiflora races.
6. Buds bearing also long, Jointed ЖАН, eid arise ient papillae 22.02. О. Lamarckiana races.
2. Flowers small (length of petals 15-25 ari Style short, causing self-
pollination . . vc «ey alo wi wow «© ап Тасев,

3. Flowers smaller бома o of petals 9-15: mm.). ^ Бей short, flowers self-

pollinated . . . хе соо (0, Willricala races.

Before proceeding to an examination of their characters and behaviour let me outline
the history of some of these races and their present distribution in America.

History of Races.—1 have devoted considerable attention (Gates, 1911 а and 1911 d)
to the history of the introduction of Oenotheras from America into Europe, traciug the
records of the individual races since the first introduction in 1614, as far as possible
from the early names, plates, herbarium specimens, and descriptions. Some of the main
conclusions may be mentioned here in condensed form. Тһе earliest Oenothera, of
which there is any record, was brought to Europe probably in 1614 and grown in the
Garden at Padua. In 1619 C. Bauhin grew this race from seeds obtained from Padua,
and the species is described in the Appendix to Bauhin's *Pinax, page 520 (1623).
Bauhin's original specimen no longer exists, but the evidence from all sources, though
inconclusive, indicates that this first introduction, which Bauhin named .Lysimachia
lutea corniculata, was a type of О. biennis, now wild in Europe, and having
larger flowers than the American races of О. biennis now known, but smaller than
O. Lamarckiana or O. grandiflora. Afterwards various independent introductions and
descriptions of Oenotheras were made from time to time, and some of the data regarding
their names, origin, and probable identity are given in the following table. In some
cases it seems impossible to be certain regarding the identity of the plant described or

illustrated. But in a few instances comparisons are made or other data provided which

| жшке {һе determination at least very highly probable. Thus the figure and description
Т тасма Americana by Columna leaves — doubt regarding its. identity,
22 оца the plant -— perhaps 1 have come from the same source as ime Hy cy

KNOWLEDGE OF THE MUTATING OENOTHERAS.

TABLE I.— Early references to Oenothera.

17

Date Place. Name. Author. Work. Probable identity
1614. | Padua?, | Seeds from Virginia.
1619. | Basil. Lysimachia lutea corniculata,* C. Bauhin. | Pinax, рр. 245, 520. | Description. | “European O. bi- |
епті,”
1627. | Venice. | Hyoscyamus Virginianus. + Alpino. De Plantis Exoticis. Description W 4
and plate.
1628. | Rome. | Lysimachia Americana. t Columna in| Nova Plant., Anim. et i 9 б ^
Hernandez. Miner. Mexicanorum.|
1629. | London. | Lysimachia lutea siliquosa Virgi- | Parkinson. | Paradisus. » » » »
niana. $
1640. 5 % Ӛ е ^ Theatrum Botanicum. n y; А Р
1669. » Lysimachia corniculata minor | Morison. | Hort. Reg. Blesensis. | Description. | О. muricata.
lutea Canadensis.
1669. н Lysimachia lutea flore globoso, j di 5 в О. fruticosa.
Park. Ger.
1680. | Oxford. | Lysimachia lutea corniculata non " Plant. Hist. Univ. я ате
рарроѕа Virginiana majo Oxon., ii.
тасма lutea corniculata non Уу Š re $i —
рарроѕа Virginiana minor.
1680. ” Lysimachia Mo sorte latifolia г » » Plate. 0. biennis.
lutea corniculata
1680. » Lysimachia deri angusti- бі я % в 0. muricata.
folia corniculata
1686. | London. | Lysimachia, lutea Virginiana. Ray Historia Plantarum, i. | Description. | European 0. bi-
| ennis.
1686. » simachia Virginiana altera, á ‘ 5 ^ 0. grandiflora
foliis latioribus, floribus luteis, or О. La-
majoribus. ma na.
1694. | Paris Onagra latifolia. Tournefort. Élém. de Botanique. | Listed. European O. bi-
; ennis.
1694.| ,, Onagra angustifolia. а ii > » О. biennis.
1694.| ,, Onagra angustifolia, caule rubro, У " $5 s О. muricata.
е minori.
HUI a Onagra latifolia, floribus amplis. а gu a Rei Her- » 0. grandiflora.
1714.| , Lysimachia latifolia, spicata, lutea | Barrelier. Бани Gall, Hisp. | Plate 989. |0. biennis.
Lusitanica. et Italiam observatæ.
1714. $ Lysimachia angustifolia, spicata, А A » Plate 990. О. muricata.
lutea Lusitanica.
ITE Lysimachia lutea, corniculata, » » » Plate 1232. | 0. grandiflora
latifolia Lusitanica. 0. асан.
1760. | London. | Oenothera, foliis lanceolatis, den- | Miller. Figures of plants de- | Plate. О. muricata.
tatis, caule hispido. seribed in the * Gar-
| dener's Dictionary.’
1760. ý Oenothera foliis ovato-lanceolatis | ,, я " » 0. biennis.
planis. |
* iem бой Padua, + Seeds from an English physician. |
+ » Johannes Ропа іп Verona. $ s ь Yirginis 277

| tie SERIES —BOTANY, VOL. ҮШ. 0

>s

18 DR. В. В. GATES—CONTRIBUTION TO А

Virginianus. But there сап be по doubt that Бау? Lysimachia Virginiana altera,
folius latioribus, floribus luteis majoribus is a large-flowered form, though there are no
data to decide whether it belongs іп O. grandiflora ог О. Lamarckiana. Similarly,
Barrelier іп 1714 gives figures of three species of Oenothera which can be identified
with practical certainty as (1) O. biennis, (2) О. muricata, and (8) О. grandiflora or
O. Lamarckiana, the last figure applying perhaps equally to either of these species.

The facts which emerge are that a large-flowered Virginian form, which must have
belonged in the grandiflora-Lamarckiana series, was recognized as early as 1686 by Ray,
while O. muricata was recognized in 1700 by Tournefort and figured by Barrelier
in 1714, together with О. biennis and the large-flowered form. Curiously enough,
the large-flowered Oenotheras seem to have been lost from English Gardens, for
Bartram in 1778 re-discovered O. grandiflora, this time in Alabama, and it then was
introduced into England and received its modern name. But О. grandiflora is
now known to have survived in Virginia and Carolina as late as 1820, so that there
can be no doubt that the early introduction before 1686 was of a race belongiug to
O. grandiflora or O. Lamarckiana from Virginia. O. grandiflora, therefore, had two
introductions into England, one from Virginia and one from Alabama, unless (as is
equally probable but not provable) the earlier introduction was О. Lamarckiana. In
1796, some years after the introduction from Alabama, О. Lemarckiana was recognized
in Paris as distinct from O. grandiflora, and thus it had its taxonomic origin. It
had long been going under the name О. biennis. Аз late as 1860 O. Lamarckiana
was again introduced into England, this time from Texas, and the plants of DeVries’s
cultures are descended from this source. It seems, therefore, certain that the naturalized
plants on the Lancashire sand-dunes (which contain both O. Lamarckiana and O. gran-
diflora) had an independent and much earlier origin, yet the O. Lamarckiana characters
in both cases agree even to the point of identity.

Of course the question of the origin, introduction into cultivation, and identity of
O. Lamarckiana has furnished the most absorbing part of this historical work. Has
O. Lamarckiana originated in Europe or America? Is it a hybrid produced in
Botanical Gardens by spontaneous crossing, and, if so, what remains of DeVries's
mutation theory? I have pointed out elsewhere (Gates, 1911 d) that all the hereditary
behaviour of O. Lamarckiana cannot be explained as the result of hybrid splitting. I
have also pointed out that if О. Lamarckiana originated as a result of crosses between
О. grandiflora and О. biennis in European gardens, the same thing must have happened
long previously in the Virginia-Carolina region, where both were certainly originally
indigenous. But I have further pointed out that it is of little moment whether
O. Lamarckiana originated from crosses on this side or the other side of the Atlantic,
for it is self-evident that in such open-pollinated forms as О. Lamarckiana and О. gran-
diflora, crossing of races must be continually occurring in nature. ‘This is, in fact, one
of the conditions under which the evolution of any open-pollinated group of forms must
: take place. We must, therefore, assume a mixed ancestry for O. Lamarckiana in any

case, many closely related races having taken part in that ancestry. But there is no

KNOWLEDGE OF THE MUTATING OENOTHERAS. 19

evidence that O. Lamarckiana is in this sense any more “impure” than O. grandiflora.
Further, while there is no indubitable evidence that O. Lamarckiana was introduced
from America previous to 1686 having just the characters it has now, yet I can find no
reason for supposing that the large-flowered plants of Ray (1686), Tournefort (1700), and
Barrelier (1714) belonged any nearer to О. grandiflora than to О. Lamarckiana.

The fact that О. Lamarckiana has not been found wild in North America during the
last half-century cannot be used as a valid argument for its origin as a hybrid in
European Gardens, for several reasons. (1) Certain races of О. biennis also are now
known only in Europe and seem to have disappeared from America. (2) It is evident
that, at least in the Eastern States, which have been longest settled, the open-pollinated
forms such as O. grandiflora have become practically extinct since the introduction of
civilization, while the self-pollinating O. biennis and O. muricata races are still found
abundantly in the East as well as the West. This might be anticipated, for in self-
pollinated plants, flowering is sure to be followed by seed-production, while in open-
pollinated species the flowers are often very imperfectly pollinated. Hence, with the
increasing enemies resulting from the advent of civilization, species such as О. grandi-
flora and O. Lamarckiana, which are open-pollinated, would soon reach a condition in
which their seed-production was insufficient to maintain the race, while plants with the
close-pollinating habit would not be adversely affected. That the close-pollinated species
of Oenothera produce many more seeds than the open-pollinated when left to themselves
under the same conditions, is a matter of direct observation in my own cultures, so that
it is not surprising that in the Eastern States, which have been longest occupied by
civilized man, the open-pollinated species should have become extinct.

On the other hand (3) many open-pollinated forms still remain in the Western States,
which are more recently and less densely populated. As recently as 1867 a specimen
was collected in Colorado (see Gates, 1911 4, p. 589) which appears almost identical
with O. rubrinervis, a mutant from O. Lamarckiana. А new species (О. ornata),
evidently closely related to O. grandiflora, was described from Idaho only last year
(Nelson, 1911). It is evident, then, that forms at least very closely related to
О. Lamarckiana ате still to be found in the Western States, and there remains at present
no sufficient reason for assuming that O. Lamarckiana originated only in cultivation.
Its mixed ancestry, on the other hand, must be admitted by all who carefully
consider the question. But this condition it shares with O. grandiflora, O. Hookeri,
and indeed with all open-pollinated plants in which mongrel races are fertile.

As regards the original distribution of these species in North America, it seems quite
evident that the large-flowered forms, including both O. grandiflora and O. Lamarckiana,
originally extended from the Eastern Atlantic region of *' Virginia” westward to Texas,
and it is certain that О. grandiflora extended at least from Virginia to Alabama. Тһе
physiological behaviour of this species in cultures indicates that the race now surviving
in Alabama is adapted to a somewhat more southern climate than O. Lamarckiana.
It is stated, however (Macoun, 1883, p. 171), that O. grandiflora is common in
Ontario. I have not seen this form to determine its exact characters. From the data

D2

90 DR. В. В. GATES—CONTRIBUTION ТО A

now at hand from numerous collections of wild seeds from all over the continent, as well
as from other facts, it seems that though the three species O. grandiflora, O. biennis,
and O. muricata very largely overlap in their distribution, yet O. grandiflora is on the
whole the most southern and O. muricata the most northern. I have collected identical
races of O. muricata from Nova Scotia, New Brunswick, and Manitoba, so that the
decrease in flower-size in the more northern latitudes may not be devoid of significance.

IV. VARIATIONS AND MUTATIONS OF О. LAMARCKIANA, Ser.

Since the important researches of DeVries, covering many years’ experiments, in
which he showed that a group of divergent types arise from O. Lamarckiana when it
is cultivated in large numbers, the need of a more detailed analysis of this behaviour has
become evident. The idea of recurring “mutation periods” as accounting for this
behaviour has not found universal favour with biologists, and there is no detailed
evidence in its support. It is probable that the premutation hypothesis of DeVries will
also be found untenable. Indeed, the author's own views on some of these subjects
have probably changed since the publication of ‘ Die Mutationstheorie’ in 1901. But a
satisfactory hypothesis as to why the aberrant types appear has not been given. From
the facts presented in previous papers and briefly summarized in the last section, it will
appear that the mutation phenomena in O. Lamarckiana are closely connected with the
crossing which has taken place in its ancestry, but, so far from depriving these
phenomena of significance, they show rather that the species-unit in open-pollinated
plants is much more complex than was previously realized. O. Lamarckiana,
O. grandiflora, and O. Hookeri are all undoubtedly “ hybrids,” if by that is meant forms
whose ancestry has undergone crossing, and has therefore been contributed to by
several rather closely related races. But in this connection several things must be kept
in mind: (1) That the mutants which appear in each generation probably are not
the same as the races which contributed to the ancestry of the species; (2) that
0. Lamarckiana itself shows much more variation in different cultures (races) than has
previously been supposed ; (3) that (as I shall show) the mutants themselves frequently
exhibit a wide range of variation, in some cases much wider than would be supposed
from the previous accounts of this mutation process.

In the following account I shall endeavour to exhibit the range of fluctuating
variation of О. Lamarckiana itself and of its various mutants. It is important to
remember, on the other hand, that, although some of the races frequently overlap in
their variability (so that an apparently continuous series can be formed from them), yet
they are really discontinuous in origin, as evidenced by their behaviour when inbred or
crossed. Another important point, which has been neglected in previous studies of these
forms, is the series of ontogenetic stages passed through by the rosettes in their
. development. Successive leaf types appear, which often differ from each other to
a striking degree, so that different stages of the same rosette may be much more unlike
5. anera Mean ұлды — scs a {о different races. In the сезе) o.

KNOWLEDGE OF THE MUTATING OENOTHERAS. 21

hybrids it is particularly important that the hybrid offspring should be compared
with each other, and with their parents in every stage of their ontogeny, and not merely
in the end stages. 1% is also essential that care be exercised to compare corresponding
stages of the rosettes, for misleading conclusions may arise from a comparison of
rosettes of different ages. Nowhere is the necessity for taking account of all the onto-
genetic stages of organisms more apparent than in such forms as the Oenotheras,
where a number of different types of leaf succeed each other in the development of
the individual. Indeed, in Oenothera, from the first seedling leaf to the fully
developed rosette and the mature flowering plant, there is a constant succession of
changes in leaf-shape. The cotyledons also pass through a regular series of changes in
shape and size after they first appear.

The following series of photographs is offered in proof of some of these statements.
They are selected from a much larger number, which have accumulated in connection
with the records of my experiments. РІ. 1. fig. 1 shows a pan of young seedlings of
O. Lamarchiana. They are very uniform, but are developing at different rates.
Figs. 25, 21, and 15, which show respectively O. rubricalyx, О. brevistylis, and О. laevifolia
at the same age, should be compared with this. The comparison will show that even the
young seedlings are sharply distinct from each other in shape of leaf-blade and length
of petiole. The crinkling of 0. Lamarckiana may appear іп even the earliest leaves
following the cotyledons.

It should be pointed out that O. Lamarckiana is by no means the only Oenothera
having crinkled leaves. An equal amount of crinkling is to be observed in a race of
O. biennis from the Chelsea Physic Garden, which I have grown for several years.
Such mutants as O. lata and О. nanella show as much or more crinkling than
О. Lamarckiana, while nearly all the mutants (even О. /aevifolia) show more or less
of it.

The suggestion of DeVries (1909, 1. р. 309) that the crinkling is due to faulty
correlation between the growth of the veins and of the areas of mesophyll between them,
appears to be correct, and is supported by the fact (referred to elsewhere in this paper)
that in certain dwarf plants the reduction in the dimensions of the vessels of the stem
is more or less independent of the change in size of the cortical cells.

Pl. 1. fig. 2 is a photograph of a rosette of O. Lamarckiana about four months old—
i.e. four months after the seeds were sown. Fig. 3 shows a similar rosette of
O. Lamarckiana at about the same stage of development, grown in heavy clay soil.
Under these conditions the development of the rosette is slow, and the various types of
rosette-leaves can be carefully followed as they appear. In this way it is found that
three distinct types of rosette-leaf successively make their appearance :—(1) Narrow type
of leaf, smooth, entire and oblong, with rounded tip: two leaves of this type are
present in fig. 3; this is followed by (2) a cycle of three leaves, broad, nearly smooth,
with very obtuse tips, narrowed abruptly to the petiole, longer than (1); then after one
‚ог two transition leaves appears (3) the type of leaf of the mature rosette. These are

22 DR. В. В. GATES—CONTRIBUTION TO А

longer and narrower than (2), usually much crinkled, pointed, the blade narrowed
gradually to a petiole with irregularly lobed or jagged margin. In fig. 4, which isa
little further developed, the same three types of leaves are found. Figs. 3 and 4 should
be compared with fig. 18, Pl. 2. figs. 14 and 16, which show that the O. levifolia rosette
passes through the same series of developmental stages as О. Lamarckiana. Indeed,
these three figures are. perhaps, indistinguishable from figs. 3 and 4 of O. Lamarckiana
in photographs, yet the earlier stage (fig. 15) of O. laevifolia is plainly different from the
corresponding stage of О. Lamarckiana (fig. 1). From such facts one learns that types
which are at first divergent may converge during development, and later diverge again
so that the adults are as unlike as the young seedlings. In Oenothera, as the plants
grow, such changes are constantly going on, and they add much to the complexity of the
facts, and greatly increase the necessity for the most detailed observation and comparison
of the individual plants in all the stages of their ontogeny. The necessity for this
becomes particularly strong in dealing with hybrids.

After the third type of rosette-leaf appears there follows a continuous succession of
leaves of this type, the leaves of euch cycle being longer than those of the preceding
cycles (though otherwise the same), until a maximum diameter of the rosette is reached
(РІ. 1. fig. 5). After this, if the rosettes of О. Lamarckiana or its derivatives are kept
under constant conditions of high temperature and high moisture content, leaves of the
same type continue to appear, the older ones dying away below, until a considerable
area of stem may be formed without internodes (see Gates, 1912). The plants will
usually continue to produce rosette-leaves as long as these tropical conditions are
prolonged. Under ordinary conditions of culture, however (in which the seeds are sown
in pans in January, the seedlings afterwards transplanted to pans in the greenhouse, and
the young rosettes planted out-of-doors іп May), a stem soon begins to form. The lower
cauline leaves form a somewhat abrupt transition from the rosette-leaves. As observable
from figs. 7 and 8, the upper stem-leaves are sessile or nearly so, with a broad, more or
less aurate base, and they diminish gradually but strongly in length and width in
passing up the stem. Тһе lowermost stem-leaves have long petioles and elongated
blades like the rosette-leaves, but these quickly change above to the cauline type of leaf
having a very short petiole and shorter, aurate blade. The leaves of the whole organism
therefore form a continuous series, whose continuity is, however, broken by rather sharp
transitions at two places in the rosette development, and also in the passage from radical
to cauline leaves. In Oenotheras of the Lamarckiana group there is, under the
conditions of my cultures, a considerable pause after the maximum diameter of the
rosette is reached, during which further cycles of rosette-leaves of the same type
continue to be produced, before elongation of the stem begins.

In Pl. 1. fig. 6 are shown six leaves selected from typical mature rosettes of
О. Lamarckiana at this time, to show the range of variation in the length, width, and
. pointedness of the leaves at this stage. With this should be compared РІ. 2. figs. 18 and
23, selected respectively from О. laevifolia and О. brevistylis to show their ranges of

variability in the same stage of development. It will be seen that if the leaves of all

KNOWLEDGE OF THE MUTATING OENOTHERAS. 23

three were mixed indiscriminately they would form a perfectly continuous series, both
as regards width of blade and breadth of leaf-tip. In other words, the three series of
leaves clearly overlap in the ordinary range of their variability, for they were not
selected from extreme individuals but from representative members of each type of
mature rosette. Yet it is equally clear that this is not really a case of continuous
variation, for О. laevifolia produces narrower leaves than ever occur in О. Lamarckiana,
and О. brevistylis contains leaves with broader, more obtuse tips than ever occur in
O. Lamarckiana. But of much more importance is the fact that the modal condition for
width of leaf and breadth of tip is different for each species. Shull (1907, p. 22) has
shown statistically that the modal condition of О. Lumarckiana and О. rubrinervis shows
constant differences for several characters.

It is true that there is a certain amount of individual variability within each race—for
example, the leaves of the mature rosette will be, on the average, narrower in certain
individuals than in others ; but it is not often that all the leaves of a rosette are of such
character that the plant cannot be determined with reasonable certainty at this stage,
although occasional cases of transgressive variability oceur between most of the forms,
which render the determination of such individuals indecisive until a later stage of their
development. Such cases occur in which, for instance, rosettes of О. laevifolia or
O. rubrinervis or O. brevistylis cannot be distinguished from O. Lamarckiana, or vice
versa. But it must be said that such cases are not common. That each of these forms
has a different modal condition for its various leaf-characters is perfectly clear and
indisputable, even though transgressive variability or fluctuation occasionally asserts
itself with regard to any of these characters in particular individuals.

РІ. 1. figs. 7 and 8 are photographs of adult plants of О. Lamarckiana, showing the
range of variability in the crinkling of the leaves. Тһе individual in fig. 7 has an
exceptional amount of crinkling, this feature being almost as fully developed as in
О. lata, though the leaves are longer and more pointed than in that mutant. Fig. 8
marks the other extreme, in which the upper stem-leaves are practically devoid of
crinkling. This is much nearer the modal condition of О. Lamarckiana than the
previous figure, for the upper stem-leaves usually have very little crinkling.

These figures give a fair idea of the range of fluctuating variability in cultures of
O. Lamarckiana derived from seeds of DeVries. I have since had experience with races
of Oenothera from other sources, which, although they are undoubtedly to be included in
the species О. Lamarckiana in the strictest sense, yet differ quantitatively in nearly
every character from the О. Lamarckiana race of DeVries’s cultures, even when grown
side by side with them under the same environmental conditions for several generations *.
А study of certain of these races, and a comparison of their characters through several
generations with those of the О. Lamarckiana of DeVries’s cultures, shows that such
quantitative differences, however they may have arisen, continue to be transmitted, so
that the races remain strictly differentiated. I am inclined to believe that this fact of
the strict inheritance of quantitative differences in various races of the same species may

* See also note, pago 67.

24 DR. В. В. GATES—CONTRIBUTION TO А

prove to һе one of the salient features of heredity іп Oenothera, and may account for
some of the almost bafflingly difficult phenomena to be observed among these forms.

The following represents the history of one such О. Lamarckiana race. It was derived
from a single individual which appeared from an unknown source in the back-yard of a
citizen of St. Louis, Mo., where no Oenotheras had previously been seen, so that its
origin cannot even be conjectured. From pure seeds of this individual I grew in 1910 a
uniform offspring of 38 plants. Fig. 9 shows the range of variation of the seedlings,
which was not very great. The close similarity to DeVries’s race of O. Lamarckiana (at
a somewhat younger stage) is shown by comparing Pl. 1. fig. 1. Three plants of the F,
were self-pollinated as parents for the next generation, and yielded respectively 62, 29,
and 41 plants, making а total of 132 plants in the F, РІ. 1. figs. 10 and 11 show a
typical rosette and adult plant for comparison with Pl. 1. figs. 5 and 8. The
rosette-leaves of this race are crinkled similarly to the plants of DeVries’s race, but they
have decidedly broader leaves and the latter are conspicuously larger, so that the rosettes
reach a much greater diameter. The individual leaves of the mature rosettes also show
wider fluctuation than in DeVries’s race. The stem-leaves of this race are also much more
variable than in DeVries’s form. The same individual may bear leaves ranging all the
way from ovate, tapering at both ends, and petiolate (thus resembling О. grandiflora) to
sessile or nearly so, with broad and aurate base (as is usual in O. Lamarckiana). These
leaf types do not always form a definite transition series, but one type may be found
chiefly on the main stem and the other on a side branch. Both rosette- and stem-leaves
are, on the average, somewhat more crinkled than in DeVries's race, and the buds bear
fewer long hairs. On the whole, the fluctuating variability of this race is decidedly
greater than in the О. Lamarckiana of DeVriess cultures. This is, perhaps, due to the
fact that continued inbreeding has decreased the range of ordinary variation in the latter
race.

Among the 132 individuals of this race in the Е,, all belonged to the parent type but
two plants which must be classed as mutants. One of these was a dwarf, corresponding to
but not agreeing with О. nanella, the other a narrower-leaved one, corresponding to
О. rubrinervis. This race of О. Lamarckiana, therefore, is capable of producing aberrant
types analogous to, but not identieal with, those described by DeVries: a fact of
considerable interest. Тһе following table summarizes these facts :—

TABLE II.
Е. | К Mutants,
38 plants. | №. 3.5. 62 plants.
No. 1.1. 25 o 1 narrow-leaved plant,

No. 2.2. 4l. y; 1 dwarf plant.
182 ,

KNOWLEDGE OF THE MUTATING OENOTHERAS. 25

Another race of O. Lamarckiana, whose characters differ only quantitatively from those
of DeVries’s race, I have studied carefully and compared with the latter. This race was
derived from the coast near Liverpool and has already been referred to in connection with
the history of O. Lamarckiana (see p. 18). Bailey (1907) showed that this form was
O. Lamarckiana, and MacDougal (1907, p. 8) found from cultures that it produced several
mutants. I have since made more detailed comparisons of forms from this locality, and
have discovered a number of new forms, some of which have been isolated as races, and a
few of which will be referred to in this paper. Тһе O. Lamarckiana race from St. Anne's
is almost identical with that of DeVries, being much closer than the last race described,
yet there are a few quantitative differences in such features as width of rosette-leaves,
diameter of rosette, and number of long hairs on the sepals, which appear to be
constant.

I have seen one other race of О. Lamarckiana from this locality which differed
considerably in the shape of its stem-leaves, which were longer and tapering to a petiole
at the base. This race stood somewhat nearer to O. grandiflora in its characters. One
may therefore expect to find races of O. Lamarckiana bridging more or less completely
the gap between this species and O. grandiflora. A series of hybrids between certain
races of O. Lamarckiana and O. grandiflora, which I am now cultivating in Е,, may throw
some light on the origin of such intermediate races. For the F, contained two types
each of which approaches one parent although in certain features resembling the other.
The two hybrid races were therefore different from either parent, though intermediate.

Another very interesting race, which I received from the Bremen Botanical Garden
under the name О. Lamarckiana, Ser., var. cruciata, must be briefly described here. In this
culture 53 plants developed. Their rosette-leaves closely resembled О. Lamarckiana, being
deeply crinkled, but frequently broader in the middle of the blade than typical for that
form. Many leaves were also of rather larger size than O. Lamarckiana usually attains,
the maximum size being 9:5 х 36 ст. Of these rosettes, 11 showed a striking peculiarity
which I have not seen in any other form * (see РІ. 1. fig. 12), and several others showed
traces of it. The leaves were peculiarly curled and sickle-shaped, the blade being almost
wholly undeveloped on one side of the midrib. "There was no lack of chlorophyll and no
evidence of any parasite, although it is possible the condition may be pathological. Тһе
peculiarity appears when the seedlings are less than two months old. I have not yet
made the test, but it is probable that the peculiarity is at least partially inherited. "The
only сазе I have seen at all resembling this is a single rosette (see Gates, 1910, pl. 31)
in a culture of an О. Lamarckiana from Lancashire. In this rosette many of the leaves
were green on one side of the midrib and colourless (lacking chloroplasts) on the other.
The colourless portion continually died away from lack of nourishment, but there was no
curving around of the green half of the leaf. I interpreted that case as a probable
sectorial chimera, but it is not probable that that explanation could apply in the present
case. e
The flowers of this race also presented several remarkable peculiarities. "The style was

* Т have since observed similar aberrations in certain other races.
SECOND SERIES.— BOTANY, VOL. VIII. E

26 DR. В. В. GATES—CONTRIBUTION TO А

short, so that nearly all the flowers were self-pollinated, and the petals were partly
cruciate and partly normal on the same plant. Measurements showed ovary 11 mm.;
hypanthium 35 mm.; bud-cone 20 mm.; free sepal tips 6 mm.; petals 30 mm. long;
cruciate petals 7 mm. broad; normal petals 82 mm. broad; hypanthium very stout,
9:0 mm. in diameter.

In my eultures 737 plants of DeVries's race of O. Lamarckiana have been grown
from pedigree seed. Of the larger cultures, 170 plants in 1907 gave one mutant
О. rubrinervis, and 115 plants in 1909 gave two О. lata mutants. Тһе remaining plants
were in small eultures not exceeding 50 plants of each pedigree, and they produced no
mutants.

O. laevifolia.

This species has never appeared as a mutant from О. Lamarckiana, but was found by
DeVries in the locality at Hilversum from which his cultures of 0. Lamarckiana were
obtained. І have already referred to some of the differences from O. Lamarckiana.
The leaves both of the adult rosette and the stem are distinctly narrower than in
O. Lamarckiana, as seen from Pl. 2. figs. 16, 17, and 18; the latter, which shows the
range of variation in width and amount of crinkling in the leaves of the adult rosettes, is
instructive for comparison with О. Lamarckiana (Pl.1. fig. 6). Pl. 1. fig. 13 and
РІ. 2. fig. 14 again show clearly the three types of leaves which appear in the develop-
ment of a rosette. These are photographs of the same plant, taken with an interval
of 8 days. Fig. 16 also exhibits the three leaf types. It must be said that photographs
are frequently rather inadequate to represent the differences between related forms to
those unfamiliar with the characters, yet it is hoped that a careful comparison of figs. 5
and 17, of 6 and 18, of 8 and 19, will give a true conception of some of the main differ-
ences between О. Lamarckiana and О. laevifolia. The stem-leaves of the latter are more
or less furrow-shaped, and stand out stiflly from the stem. Fora systematic enumeration
of the characters, reference may be made to DeVries (1909, i. p. 309) or Gates (1909 c).

РІ. 2. fig. 20 shows the fruiting stage of a typical individual. Comparison with РІ. 3.
fig. 31 for О. rubricalyx shows that the bracts are much narrower and shorter than in that
form. Incidentally it may be observed in fig. 20 that late in the season the stigma
sometimes pushes out of the bud and expands before the latter opens. Ав is well-
known, the flowers of all the Oenotheras decrease rapidly in size towards the end of the
season. But the decrease is not always in the same proportion in all parts of the flower.
When the diminution in length of petals and sepals is more rapid than in length of
style, the stigma will protrude from the unopened bud. "This is not an uncommon con-
dition in all the large-flowered Oenotheras late in the flowering period, and it of course
increases the probability of crossing. But flowers which develop so late in the season
have a very poor chance of maturing seed, and it is not believed that such a phenomenon
has any adaptive significance in securing occasional cross-fertilization іп the race. Crosses
with other individuals are secured, in the large-flowered Oenotheras in America, through
the visits of moths in the evening when the flowers open. The fact that part of the species
composing the Onagra group are strictly self-pollinating while part are open-pollinated, |

KNOWLEDGE OF THE MUTATING OENOTHERAS. 27

and that, on the whole, the former species, such as О. biennis and О. muricata, ате more
successful and widespread than the latter, such as O. grandiflora and O. Hookeri, does
not favour the view that occasional crossing is a necessity or even an advantage in
this genus.

Schouten (1908) in several cultures of О. laevifolia has found that it yielded various
mutant forms, including О. laevifolia salicifolia, О. scintillans, О. laevifolia nanella, and
O. rubrinervis.

О. brevistylis.

This form also was discovered by DeVries in Hilversum, but has never appeared in
cultures of O. Lamarckiana. Its almost completely sterile ovaries render its propagation
possible only by using its pollen to cross with another form, so that it must clearly have
originated as a mutant, very probably in the locality where it was found. The charac-
ters of this form have not been accurately shown in previous accounts. РІ. 2. fig. 21
shows how distinct are the young seedlings from О. Lamarckiana (РІ. 1. fig. 1), О. rubri-
nervis (Pl. 2. fig. 25), or O. laevifolia (fig. 15), the blades being characteristically broader
than іп any except О. Lamarckiana, from which the leaves differ in having longer
petioles and (at this time) very little crinkling. РІ. 2. fig. 22 is a characteristic mature
rosette, showing some leaves broad-pointed, but many having obtuse, blunt tips such as
never occur in the mature rosette of О. Lamarckiana (cf. Pl. 1. fig. 5). There is always
this wide range of variation in the tips of the leaves of each rosette. "This is shown in
РІ. 2. fig. 25, with which РІ, 1. fig. 6 of O. Lamarckiana should be compared. It will
be seen that the ranges of variability overlap, though О. Lemarckiana never bears а
leaf with tip as blunt as the broadest O. brevistylis, and O. brevistylis never bears a leaf
with tip as narrow as the narrowest О. Lamarckiana. Hence, again as in О. laevifolia,
though their ranges of variation overlap there is no doubt whatever that these two forms
are perfectly distinct. If there were any doubt regarding the distinctness of the rosette
stage it could be set at rest when the plants mature, for not only has О. brevistylis
sterile ovaries but its style is extremely short so that the stigma barely projects above
the calyx-tube, and the free tips of the sepals are also very short, as shown by РІ, 2.
fig. 24. In addition, the terminal rosettes of the branches are more compact than in
О. Lamarckiana, the bracts being shorter, with broader and less pointed tips. I found
no difficulty in identifying the rosettes of O. brevistylis when mixed with O. Lamarck-
iana and was able to confirm my determination in each case when the plants bloomed,
by opening a bud and discovering the very short style. It would be impossible to obtain
more conclusive proof that a range of variation which is apparently continuous is in
reality not so, for in the ontogeny of one series of individuals we pass from the rosette
stage which yields a continuous series of leaf-shapes and in which there is some difficulty
in recognizing the two types, to the flowering stage in which all the flowers of each
individual are recognizable at a glance as belonging to one type or the other. This race
also furnishes a striking instance of what occurs in all these Oenothera forms, namely
the “ holding together” or correlated variability of characters affecting all parts of the
plant, the rosette and stem-leaves as well as the flowers and fruits. One must adopt the

E2

28 DR. В. В. GATES—CONTRIBUTION TO A

conception that some fundamental germinal change has occtirred in the origin of
O. brevistylis, presumably from O. Lamarckiana, that change exhibiting itself externally
in a different response as regards leaf-shape, but also in the very short style and badly
formed stigma, the short sepal tips, and the almost entire failure of the ovaries to
mature.

DeVries was able to show that O. brevistylis will breed true when seeds were obtained
by self-pollinating many flowers, but O. brevistylis has usually been propagated by using
its pollen to cross with O. Lamarckiana, the offspring being about 50 per cent. О. brevi-
stylis and 50 per cent. О. Lamarckiana. Му cultures of О. brevistylis were from seeds
of DeVries which had been grown in that way for five generations. The sixth
generation in my garden yielded 92 plants, of which 55 were O. Lamarckiana,
34 O. brevistylis and 3 O. lata mutants. The seventh generation, grown last year,
gave a similar result, showing that there is continued sharp alternative inheritance
when this mutant is crossed with its putative parent; and there is no tendency
for any of the leaf characters or flower characters to separate and be independently
inherited in the F, or later generations. The logical conclusion is, therefore, that all
these character-changes represent the external manifestations of one fundamental
germinal change. The sterility of the ovaries serves to emphasize the depths to which
the organism has been stirred. It appears that the changes have been so serious that
_ the development of the ovaries is rendered almost impossible. It is possible, however,
that the failure to produce seeds results, not from abortion of the ovules, but from the
fact that pollen will not germinate on the imperfectly formed stigma,

O. rubrinervis and O. rubricalyx.

The mutant O. rubrinervis probably bears a somewhat different relation to its parent
(O. Lamarckiana) from some of the other derivatives. DeVries (1907) has shown
that when O. Lamarckiana or one of its mutants is used to pollinate O. diennis, Linn., two
types are produced in the F,, which he calls respectively О. laeta and О. velutina. This
type of behaviour will be considered further in the section on hybridization, but it might
be pointed out here that, as Honing (1911) has shown, О. Lamarckiana and O. rubri-
nervis differ from each other in characters which are analogous to the differences
between О. laeta and О. velutina. In both cases the former has broader, the latter
narrower leaves, the former has darker green leaves and the latter more long hairs, ete.
I shall show later that O. muricata, Linn., from the wild condition also often produces a
broad-leaved and a narrow-leaved type, so that this “double” condition is of funda-
mental interest in the hereditary behaviour of the genus, though its full meaning is not
yet clear.

In my cultures of O. rubrinervis in 1907 a new form (О. rubricalyx) originated
suddenly in a single individual. Its origin and subsequent behaviour when bred pure
and in crossing have been described in a previous paper (Gates, 1911 b). It need only
be said that this form is morphologically identical with О. rubrinervis, but differs |

strikingly in a physiological character—pigment production; there being a large increase |

KNOW LEDGE OF THE MUTATING OENOTHERAS. 29

in the amount of anthocyanin, particularly in the sepals, hypanthium and rosette-leaves.
Photographs of О. rubrinervis and О, rubricalyx can, however, be used interchangeably,
for the morphological characters are exactly the same in both. РІ. 2. fig. 25 shows a
pan of seedlings of O. rubricalyx and О. rubrinervis, in which it will be seen that the
leaves are narrower, more pointed and less crinkled than in O. Lamarckiana (РІ. 1,
fig. 1. Fig. 26 represents an older rosette, in which the three types of rosette-leaves
(cf. the rosettes of О. Lamarckiana and О. levifolia) are present, though type 2 is
inconspicuous in O. rubrinervis, differing from type 3 only in the more even margin,
more rounded tip and less crinkling, but usually not having a broad blade sharply
narrowed to the petiole. The rosette in Pl. 2. fig. 26 shows three leaves of type 1, and
three of type 2.

There is considerable variation in the mature rosette of O. rubrinervis in width of leaf,
amount of red developed in the midribs, amount of crinkling and pubescence, etc., so
that certain individuals transgress the boundary line separating them from O. Lamarck-
iana in each of these characters or in several of them together. But the occasional
indeterminable rosettes which occur nearly always show decisive characters when the
plants come into bloom. Тһе leaves of the mature rosettes also vary in the width of
the leaf-tip, and Pl. 2. fig. 27 shows an individual with exceptionally broad tips. The
transgressive variability between О. Lamarckiana and О. rubrinervis becomes more
evident under varying or unfavourable conditions of growth. Thus in 1911, in a culture
of О. rubrinervis x О. Lamarckiana, F,, 64 rosettes developed, 12 of which were classed
as O. Lamarckiana and 52 as O. rubrinervis. These were transplanted too late in the
season to become annuals, so they continued to produce rosette-leaves. The leaves
appearing towards the end of the season (September 27) on the O. Lamarckiana rosettes
were much narrower and lighter green than normal, so that they were practically indis-
tinguishable from О. rubrinervis leaves. It is evident that local conditions of nutrition
have much to do with the fluctuating variability of the Oenotheras, often producing a
surprising range of variation. But this is clearly distinct from the germinal variations,
which are inherited in each generation.

Comparison of large numbers of rosettes of О. rubrinervis and О. Lamarckiana
shows the impossibility of discriminating, merely by inspection, between fluctuating and
germinal variations. Plants and animals frequently breed true to remarkably small
differences, as many recent breeding experiments, particularly the work of Johannsen
(1909) with “ pure lines” in beans and of Jennings (1911) with Paramecium, show. The
only way to determine whether the offspring of a variant individual will regress towards
the mean of the race, or will reproduce its peculiarity in all or a portion of its offspring,
is by growing its progeny and determining their characters.

Another example of the manner in which a plant may immediately react to a change
of environment by producing new characters, is furnished by a large plant of О. rubri-
nervis which was removed from the garden to a greenhouse in September, 1907, while
in flower. It continued to bloom, but the weight of the inflorescence soon bent the tall
stem into a nearly horizontal position. Growth in this position continued and a series

80 DR. В. В. GATES—CONTRIBUTION TO А

of rosettes soon developed along the stem, each apparently from a transformed bra
bud. One of these “lateral rosettes,’ which is evidently in part a response to a new
gravitational stimulus, is shown in Pl. 3. fig. 28. These rosettes were very uniform, all
having the same arrangement and shape of leaves. The latter were all of one type,
which is very distinct from any of the leaves in the ordinary ontogeny of O. rubrinervis,
or, for that matter, of any other form I know.

A flowering shoot of О. rubrinervis (this plant derived from О. rubrinervis X
О. Lamarckiana) is represented in fig. 29 for comparison with О. Lamarckiana (РІ. 1.
figs. 7 & 8). It will be seen that the stem-leaves of this individual are more crinkled
than those of the О. Lamarckiana plant represented in fig. 8, though in general the
leaves of О. Lamarckiana show this character more than those of О. rubrinervis. Honing
(1911) believes that the О. rubrinervis from О. rudbrinervisx О. Lamarckiana differs
from pure О. rubrinervis and agrees with the О. blanda of Schouten (1908). But I
have not been able to observe any constant difference between О. rubrinervis derived
from this cross and. O. rubrinervis of pure parentage.

РІ. 3. fig. 30 shows one of several tall plants of О. rubricalyx which were grown in
heavy clay soil and, apparently as a result, developed no side branches. In this typical
plant it will be seen that there is à gradual decrease in the amount of crinkling on the
upper stem-leaves. РІ. 3. fig. 31 represents the fruiting stage of О. rubricalya, showing
the characters of bracts and capsules.

One other strikingly aberrant individual of О. rubrinervis which appeared in my
cultures in 1911 must be briefly described. It occurred in a lot of 34 plants derived
from the original О. rubricalyx mutant іп the F,, which probably represented а pure
homozygous race of О. rubricalyx, for although all remained rosettes except the aberrant
individual, yet 82 showed clearly the О. rubricalyx pigment character of the rosette. The
aberrant plant, unlike the rest, scarcely developed any rosette but formed а tall stem. It
resembled O. rubrinervis most closely, from which it differed in the following characters.
The stem was smooth and terete, almost free from long hairs. The leaves were of
harder texture, considerably thicker, broader (42 mm.), and more pointed than in
O. rubrinervis. 'They were also stiffer, standing out from the stem, and some were
furrow-shaped, but they did not resemble О. laevifolia. The flowers, which all showed the
same peculiarities, had elongated squarish ovaries, hypanthium scarcely at all developed,
bud-cone somewhat distorted, sepals with red stripes as in O. rubrinervis. The stamens
were frequently more than 8 (one flower 10, another 13), crumpled and incompletely
coalesced with the petals, which were very much reduced and pale cream colour. The
stigma-lobes were variously distorted, and in several cases the pistil was composed of
three separate styles united at the base, each style bearing only 3 stigma-lobes. In
addition to these peculiarities, in six instances a leaf was formed in the axil of a
bract, two of these leaves being “ pitchers ” or ascidia. In one case а “ pitcher " and a
normal leaf were both found in the axil of a single bract.

_ This peenliar plant formed no seeds, but was preserved as an herbarium specimen.
( Its flowers. recalled in some respects the virescent flowers of a certain race of Oenothera

KNOWLEDGE OF THE MUTATING OENOTHERAS, 31

which I described several years ago (Gates, 1910 с), but were quite unlike them in
other features. This is another case of a mutation which is unable to perpetuate
itself, and its chief interest lies in showing how various are the germinal departures
from the parent form which may take place.

The data regarding my cultures of О. rubrinervis and О. rubricalyr may now be
summarized. In 1907 І grew a total of 1112 plants of О. rubrinervis, all but 82 of
which were the purely fertilized offspring of a culture of O. rubrinervis plants, num-
bering 45, grown in the previous year, 15 of the latter being used as parents for the
second generation. The О. rubricalyr mutant appeared in thislot. One О. rubrinervis
mutant also appeared in 1907, in a culture of 170 plants of О. Lamurckiana from
pedigree seed.

I shall not add anything to what is already published regarding О. rubricalyz
(Gates, 1911 5), except to state that in 1911 I obtained a pure (homozygous) race, in
which all the rosettes exhibited thered character on their ventral surfaces. Six other
cultures of О. rubricalyx in the F, gave altogether 41 О. rubricalyx and 51 О. rubri-
nervis.

Regarding the nature of the change which led to the appearance of the О. rabricalys
mutant, it seems clear that the original mutant individual was heterozygous, and
originated from a cross between a germ-cell in which the new dominant character appeared
and one in which it was lacking. This is shown (1) by the fact that all the rubricalyxr
plants of the F, tested were heterozygous, and (2) that the F, from this plant con-
tained in addition to the 11 О. rubricalyx, one О. rubrinervis. The latter must have
originated from two germ-cells of the original mutant, both of which contained only
the recessive O. rubrinervis character.

The germinal change in the germ-cell which united with a normal О. rubrinervis
germ-cell to form this heterozygous mutant was, as I pointed out elsewhere (Gates,
19100), probably fundamentally quantitative—a difference in capacity for anthocyanin-
production which manifests itself throughout the plant.

О. gigas.

As is now well-known, this mutant has 28 chromosomes, double the number present in
O. Lamarckiana and most of the mutants. The structural changes which were corre-
lated with this nuclear change have been investigated in previous papers (Gates, 1909 4,
1911с) and need not be further considered here, except to recall that the tetraploid
number of chromosomes is associated with a doubling in the volume of the nuclei and
a very large increase—in varying proportions according to the tissue—in the size of the
cells. It seems that the gigantic stature of О. gigas (which expresses itself in stoutness
of parts rather than tallness) results directly from the larger size of its cells, and it is
probable that the number of cells is approximately the same as іп О. Lamarckiana.
Several similar cases which confirm this result have since been discovered in other
plants. Compare Wikstroemia (Strasburger, 1910), Musa (Tischler, 1910), Моззез
(Е. & Е. Marchal, 1909 and 1911). In animals, however, at least in Echinoderm and

32 DR. В. В. GATES—CONTRIBUTION TO А

Crustacean larva, it appears from the results of Boveri (1905) апа Artom (1911) that,
although doubling in the number of chromosomes in the race is followed by corre-
sponding increase іп the volume of the nuclei and cells, yet a regulation in the number of
cells takes place, so that the resulting larve are not of gigantic size. This appears at
present as an interesting difference between plants and animals, there being apparently
' no marked diminution in the number of cells in certain at least of the tetraploid races
of plants. It seems not improbable that this lack of regulation in cell number in
plants may be due to the greater thickness and rigidity of the cell-walls in plant embryos
and tissues, so that the controlling action of the young organism as a whole is not able
to express itself in re-arranging the cells as formed, in such a way that fewer and larger
cells will compose the growing embryo.

O. gigas also exhibits a wider range of variability in leaf-shape than any other of the
mutants from О. Lamarckiana, РІ. 3. fig. 32 represents а typical rosette, but in a large
culture the leaf-blades range from practically orbicular to narrowly linear. In each
individual, however, all the leaves are usually of the same shape within narrow limits of
fluctuation, though occasional rosettes show considerable variation in this regard; nor
have I observed the same succession of ontogenetic stages of the rosette as in
О. Lamarckiana, and the other forms already described, in the O. gigas from DeVriess
cultures. So far as І have observed in this race of О. gigas, leaves corresponding to
type (1) and type (2) in О. Lamarckiana and О. laevifolia do not occur, at least under
ordinary conditions of culture. Іп 1909 I grew a culture of 434 plants of О. gigas,
constituting the sixth generation from the original mutation of DeVries from
О. Lamarckiana. РІ. 3. fig. 33 shows the range of variation in leaves taken from the
mature rosettes. In each rosette all the leaves belonged to one type. РІ. 3. figs. 34,
95,36 & 37 represent individual rosettes, the range of variation being astonishing.
Schouten (1908) attempted to isolate six types from the series of forms, but states that
the offspring of any one type showed the whole range of variation, so that this could
not be done. Оп account of the strongly biennial character of О. gigas only four plants
of my culture came into flower. From these I raised 48 plants in the following year,
but the numbers were not large enough to test the inheritance of this remarkable
range of variations. They indicated, however, that the offspring of each plant tended to
resemble its parent, though occasional rosettes departed from this rule *.

In РІ. 3. fig. 38 the buds of O. gigas are shown, two-thirds natural size. As will be
seen, they are very much larger than those of О. lata (Pl. 3. fig. 43, left), which they most
resemble. Cytological examination showed that the particular individual whose buds
are here represented possessed 28 chromosomes. This plant produced much pollen,
which was used in making several crosses.

An interesting form which frequently appears in families of О. gigas is a dwarf,
О. gigas nanella. It is photographed in РІ. 3. fig. 39 on the same scale as the previous
RN to show its comparative size. It has not the О. палеа leaf-shape, but is |

* It is probable that the meiotic chromosome distributions are порти concerned in the determination of
these various dod зі

KNOWLEDGE OF THE MUTATING OENOTHERAS. 33

merely a dwarf or miniature gigas of the broad-leaved type. Different dwarf rosettes
show much variability in leaf-shape, the dwarf character being the only thing they
have in common. Schouten states that his plants of O. gigas nanella showed the same
pathological conditions as his O. nanella, but I have found no indication of it in my
cultures. It is evident that the dwarf character is directly due to an inherited germinal
change in both cases.

The percentage of dwarfs in my cultures of O. gigas was large. Out of 339 plants,
one lot contained 4°28 per cent., another lot 8°76 per cent., while a third lot of only
fifteen plants contained 10:9 per cent. Schouten’s (1908) cultures of О. gigas yielded
‚ in 1906 2°01 per cent. О. gigas nanella, or 24 plants іп 1196; and іп 1907, among
about 1000 gigas plants, which were chiefly the offspring of selected individuals of the
previous year, appeared a total of 1:87 рег cent. dwarfs, one culture yielding as high as
14°81 per cent. The meaning of this behaviour is not yet clear.

All individuals of the O. gigas of DeVries’s cultures are descended from one indi-
vidual mutant. Nilsson (1909) reports obtaining another O. gigas mutant in a culture
of 80 O. Lamarckiana which were the offspring of one individual*. In the following
year this O. gigas individual yielded 8 plants, all having the characters of O. gigas.

In 1909 I obtained from the Botanical Garden at Palermo, Italy, a packet of seeds
under the name Oenothera cognata. A portion of these seeds were sown in 1911, and
yielded 56 rosettes. The young plants apparently belonged to a new and unknown race.
Their leaves had very long petioles, and the blades were ovate with broad tips, gradually
tapering to the petiole. Some of these leaves are shown in РІ. 3. fig. 40. They
resemble type (1) in the younger rosettes of O. Lamarckiana or O. laevifolia as already
described. As previously stated, I have never observed this type of leaf in cultures of
DeVries’s gigas. Ав the rosettes of О. cognata develop, however, there is a sudden
transition to the O. gigas type of leaf, and the mature rosettes—after more of the second
type have been produced and those of the first type have died away—are absolutely
indistinguishable from О. gigas. РІ. 3. fig. 40 represents the transition stage, when
both types of leaf are present. These plants were all biennials, and hence produced no
seeds, but a fresh sowing has yielded a few plants, from which the characters will be
more fully determined. The number of the chromosomes has not yet been examined {.

The mature rosettes of this culture exhibited the same variations as my cultures
of O. gigas, though the extremes of variation were not observed. ‘Three dwarf rosettes,
corresponding to O. gigas nanella, also occurred.

Enquiry at the Botanical Garden at Palermo elicited the information that the race
had been derived in cultivation, apparently being descended from a single individual
mutant. They have since lost the strain. It is evident, then, that this was an inde-
pendent mutation, and it remains to be seen whether the race differs constantly from
O. gigas in its earlier ontogenetic stages, for the mature rosettes are identical with that
form. From this and other data, some of which will be referred to in this paper, it is

* See also Nilsson (1912) for further data.
+ The chromosome number for typical plants has since been found to be 28.

SECOND SERIES.— BOTANY, VOL, VIII, Р

34 DR. В. В. GATES—CONTRIBUTION TO A

evident that the Oenotheras cultivated in Botanical Gardens have for years been exhibiting
phenomena of mutation closely similar to or identical with those studied by DeVries in
his garden. Certain of these new forms, such as О. cognata, have been noticed by
gardeners, given names, and made the basis of new local races. Their origin is just
as sudden and discontinuous in Botanical Gardens as in experimental plots, and is
therefore clearly not due to continued selection. Each race appears, at once, full-fledged,
the descendant of a single aberrant individual.

О. nanella.

І have grown hundreds of plants of О. nanella, and while they exhibited a wide
range of leaf-form, yet they have never been found to give rise to mutants. О. nanella
has never been known to throw any of the otlier mutants, and, in this respect, it differs
from several of the other types. But in very large cultures, dwarfs having the
characters of other mutants sometimes appear from О. nanella. Thus (Zeijlstra, 1911)
18,649 plants of O. nanell« derived from a previous generation were without exception
dwarfs, but three of them possessed oblonga and one elliptica characters. О. nanella
itself differs from О, Lumarckiana mot only in stature sara in leaf and flower
characters.

The rosettes of O. nanella in my cultures are more variable than those of any other
mutant except O. gigas. These variations will not be illustrated here, but the buds are
shown, two-thirds natural size, in Pl. 3. fig. 41, with a bract attached to one. The
bud-cone is usually shorter and slenderer than in the other forms.

Dwarf or nanella forms of nearly all the mutants are now known, including О. lata
nanella, О. gigas nanella, О. nanella-oblonga, О. nanella-albida, О. nanella-elliptica, and
О. nanella-scintillans.

Regarding the nature and cause of the dwarf character of О. nanella, Schouten
(1908) has suggested that it results from a pathological condition of the plant, due to
bacterial action, and invokes Eriksson's mycoplasma theory to explain its hereditary
transmission. Не states that the rosettes of O. nanella become sickly and die back ; but
careful observation has revealed nothing of this nature in my cultures, which shows
that the phenomenon he describes is due to some local condition. Zeijlstra (1911) has
since found that a Micrococcus inhabits the cells of the stem in О. nanella in his
cultures, which indicates its greater susceptibility than other forms. That different
forms differ widely in their susceptibilities to fungal attacks, I have found to be true of
other species of Oenothera. Zeijlstra believes that the peculiar leaf-characters and the
very short internodes of О. nanella are a result of the parasite, because occasional side
branches appear which are dwarf yet have О. Lamarckiana leaf-characters and longer
internodes than the main stem. Не therefore thinks that a dwarf О. Lamarckiana
appears as the result of a mutation, but that the parasite which attacks the dwarfs
distorts them and causes their development into the “ О. папеЦа ” form. Whether а
parasite really produees this effect can only be determined by infection experiments.
Even if the hypothesis were true, it would remain unexplained why all the nanellas

| which umet! are affected. I have obtained ктем dwarf races, amine quite unlike | |

KNOWLEDGE OF THE MUTATING OENOTHERAS. 85

characters, in cultures of 75 races related to О. grandiflora, so that evidently the
dwarfing of О. nanella is not primarily due to a parasite, although the presence of bacteria
may make it smaller than it otherwise would be.

In the case of О. gigas, I һауе shown (1909 4) that the larger size of the plant
is fundamentally concerned with the larger size of the cells and their nuclei, which in
turn is connected in a causal relation with the doubled number of chromosomes. Pre-
liminary measurements of cell-size іп О. nanella, by a student, indicate that they are
considerably smaller than the corresponding cells in О. Lamarckiana. I expect to
investigate the question of cell-size in О. nanella more fully, but in the meantime it is
very probable that the small stature of О. nanella may be at least partially explained on
the basis of the cell-size, the dwarfuess being a character of the cell. Тһе many other
changes in the shape of leaves, etc., are probably, like those of О. gigas, not so simply
explained (cf. Gates, 1909 d, 1911 с). It seems evident, in any case, that the dwarfness
of О. nanella is а cell problem (although the number of chromosomes remains 14) and
is not produced by any pathological condition induced by other organisms.

Gauchéry (1899) (and other papers might be cited) classes dwarfing as of two kinds,
(1) constitutional nanism, (2) induced nanism. From an anatomical examination of a
number of dwarfs, he concludes that the reduction in the number of elements in the stem,
andin their calibre, is very variable according to the tissue considered. "Thus, the dimen-
sions of the xylem vessels are always reduced, but this is not always true of the
dimensions of the other cells; and in one case there is even shown an increase in the
size of the cortical cells, though in all the other elements of the stem there is a decrease
in size. "The reduction in the dimensions of the epidermal cells was found to be much
less than in the vessels.

О. lata and О. semilata.
The mutant О. lata has occurred a number of times іп my cultures, as shown in the

following table :—
TABLE III.

N — Number
Culture. ғ

piki O. lata.

EN |
О; rubrinervis ое. 96 1
О. Lamarckiana x О. brevistylis .. 92 3
о АРВ 117 g
О. rubrinervis x О. nanella ...... 49 1
0. rubrinervis х O. Lamarckiana
(from О. lata x О. Lamarckiana). 64 1
ТИМИ ... 411 8 =1-95 рег cent.
| р

РІ. 3. fig. 42 shows а typical rosette. Тһе very characteristic buds are shown in
r2

36 DR. В. R. GATES—CONTRIBUTION TO A

Pl. 3. fig. 43, reduced. The three buds to the left of the figure are O. lata buds as
they appear in plants derived from seeds of DeVries. These flowers produce prac-
tically no pollen. An account of the pollen development in this form was published
in 1907. The phenomena of sterility are very similar to those occurring in hybrids,
as Tischler (1908) has also pointed out. The pollen mother-cells or young pollen-grains
break down in various stages of development. The tapetal cells disappear early, and
the wall-cells of the anther grow into the cavity of the loculus, more or less completely
filling it with tissue. This ingrowth of the surrounding cells is doubtless due to the
release of pressure on their surfaces caused by the disintegration of the tapetal cells and
the pollen mother-cells. In some cases it is evident that mitoses of the ingrowing cells
have also taken place. The sterility is, therefore, not due to the ingrowth of the tapetal
cells, but the pollen mother-cells themselves frequently begin disintegration, from some
unknown cause at an earlier stage of development.

Тһе O. iata derived from the Amsterdam cultures, wherever it is grown, is almost
or quite completely sterile in its anthers. This sterility is closely correlated with the
shape of the bud. Ав the illustration (РІ. 3. fig. 43, three buds to the left) shows, the
buds are thick and clumsy, rounded, not squarish, radially asy trical, exhibiting а
characteristic protrusion, swelling or hernia on one side of the cone near the top. This
is in turn due to the crumpling of the petals and their distorted disposition in the bud.
It is evident that this irregularity in the development of the petals is closely associated
with the failure of pollen development.

Whatever causes the sterility of the pollen is also concerned in the distortion of the
petals, and therefore of the buds, in this characteristic manner. "That this is the case is
clearly shown by the behaviour of a pollen-producing form of О. lata, which is one of
many types occurring in а culture of naturalized Oenotheras from Birkenhead, near
Liverpool an account of which occurs elsewhere in this paper. This form is probably
to be classed as О. lata, though it differs slightly from Н. DeVries’s mutant in the
characters of the rosette (see Pl. 4. fig. 44), the leaf-blades being somewhat less crinkled,
more elongated and with longer petioles than in typical О. lata, though the ranges of
variability of the two forms appear to overlap in these respects. "The individual repre-
sented in fig. 42 probably agrees most nearly with DeVriess O. semilata. But the
majority of these rosettes agreed more nearly with О. lata. Тһе flowers, however, were
fertile, producing considerable quantities of good pollen, and its buds have nearly or
quite the normal squarish shape of other pollen-producing mutants, as shown by the
four buds to the right in Pl. 3. fig. 43, although there remain some traces of the
crumpling of the petals seen in typical О. lata. Whether these plants should be
classified as O. semilata or as a polleu-producing race of O. /a/a is difficult to decide, on
account of the variability. MacDougal (1907, p. 15) first mentioned a successful self-
pollination of this Liverpool form, supposing it to be an exceptional condition. Му
1909 culture of 107 plants from one lot of seeds from Birkenhead contained 8 О. lata
individuals with the bud-characters just mentioned, all of which produced considerable
quantities of pollen, and two others whose place is somewhat doubtful,as they never

KNOWLEDGE OF THE MUTATING OENOTHERAS. 37

fully developed. One plant of this culture, however, corresponded in all respects with
the O. lata of DeVries, having the characteristically distorted buds and sterile anthers.

It is now clear that lata possesses 15 chromosomes in some at least, and probably in
all individuals. The presence of the extra chromosome is probably the cause of the
failure in the pollen development. The varying distributions of the odd chromosome in
the megaspores, and perhaps also partial apogamy, probably account for the inconstancy
of the offspring from this mutant. The chromosome-number of semilata is not yet
known, and the problem regarding the relation between the extra chromosome and the
lata characters therefore remains to be solved.

If one thinks of the lata characters as emerging from those of О. Lamarckiana, and
the semilata characters as only half emerged, one gets à picture of the relationships of
these three forms.

O. semilata is one of the rarest of the mutants, having appeared only three times
in DeVries's cultures, and every time from O. /aía. One such mutant self-fertilized
gave 358 plants, all of which were semilata except three which were nanella and four
lata. In a cross, O. lata х О. semilata gave 105 plants distributed as follows :—
61 Lamarckiana, 39 lata, 2 nanella, 2 oblonga, 1 albida, which is similar to the result
obtained when other forms are used to pollinate O. /ata.

I obtained four plants showing the characters of O. semilata in a culture of 75 plants
from О. lata x О. Lamarckiana. The offspring were as follows :—63 О. Lamarckiana,
4 О. lata, 2 О. lata nanella (?), 2 О. semilata, 1 О. laevifolia(?), and З aberrant
rosettes, 2 of which were probably О. semilata. Тһе kaleidoscopic fluctuations of
these plants during their ontogeny often renders their determination a matter of great
difficulty. Тһе F, from one of these self-fertilized O. semilata plants produced 40 very
variable rosettes. About 11 of them were О. вет а?а, but the rest formed a continuous
series running to О. Lamarckiana and could not be more accurately determined. Three
which came into bloom were all clearly О. semilata. The other О. semilata plant
produced in Е, only three plants, all of which were O. Lamarckiana.

О. semilata is only known as a mutant from О. lata, unlike most of the other mutants
which have appeared from various ancestors. As DeVries has stated, its external
characters suggest that it is an intermediate hybrid between О. ¿ata and О. Lamarckiana,
and the fact that its offspring breed true does not exclude this interpretation. "The fact
that O. semilata has only occurred from the single source, О. lata x О. Lamarckiana,
supports the view that it is such an intermediate hybrid, rather than a mutant, but
its chromosome number may explain why it originates only from О. lata.

My castration experiments with О. lata (1909 b) have furnished an indication that
some form of apogamy occasionally, though rarely, occurs in this mutant. It seems
possible that perhaps 1 per cent. of the ovules will develop seeds without fertilization,
though the remaining ovules require fertilization for their development.

The characters of the other mutants from О. Lamarckiana, which have been less
thoroughly studied, will not be discussed here. Тһе data presented in this section are

98 DR. К. Е. GATES—CONTRIBUTION TO A

meant to give a better understanding of the differences between the more common
mutants and their parent form, as well as their relationships to each other. It is clear
that the mutants differ from each other not in one but in many characters, and I am
in agreement with the conception of DeVries that all these characters are in each case
the expression of a single germinal change.

But the appearance of so many so-called ‘‘ combination types,” particularly in the
cultures of Schouten (1908), shows that all the characters do not necessarily remain
together under all circumstances. The “combination types" are supposed to originate
from the union of two germ-cells which have mutated in different directions, but the
significance and manner of origin of such types, and their relation to hybrid splitting,
will have to be determined by future investigations. If О. потеЦа, for instance, is
capable of producing occasional types in its offspring such as O. nanella-oblonga,
О. nanella-elliptica, О. n.-lata, and О. n.-scintillans, then each of these leaf-types must be
potential or “latent” in O. nanella just as they are in O. Lamarckiana, and the only
quality really absent from О. nanella is stature. But * latent" is a dangerous word to
use, for it is only a cloak for ignorance regarding the real cause. It is possible, though
not probable, that some of the mutants from O. Lamarckiana appear only as alternative
dichogenous conditions which are environmentally determined very early in the
ontogeny. On the other hand, certain of them may be present in some germinal
condition resembling the heterozygous condition in Mendelian hybrids. But this
question and others relating to the interpretation of the mutation process are discussed
in another section.

O. grandiflora, Solander.

The previous paragraphs dealt with O. Lamarckiana and its derivatives. A later
section will consider the derivatives from a naturalized colony of forms on the sand-
dunes of Lancashire which included both O. Lamarckiana and O. grandiflora. Тһе
latter species has already been carefully described (Vail, 1907) from plants derived from
the locality in Tensaw, Alabama, where it still survives. It is frequently grown in
Botanical Gardens, and, particularly in France, it often goes under the name О. swaveo-
lens, Desf. I have grown cultures from the O. grandiflora naturalized in Lancashire,
and also from plants collected in the original locality in Alabama by Prof. 8. М. Tracy.
I have already pointed out that this species formerly had a wide range in America,
extending eastwards to the Atlantic, where it long survived in Virginia and Carolina.
The eastern form probably differed in certain features from the Alabama form.

In July, 1907, I began a greenhouse culture from seeds collected wild on the
Lancashire coast, for which I was indebted to Dr. D. T. MacDougal. The rosettes were
transplanted to individual pots when quite young, and they developed slowly and
perfectly under these conditions. About half of them were proved by their striking
rosette characters, and afterwards by their flowers, to belong to O. grandiflora. РІ. 4.
fig. 45 shows one of these rosettes four months after germination, and РІ. 4. fig. 46 the
same type at the age of five and a half months. In the latter figure a new leaf-type has
appeared, with characteristic basal lobes, though the old type of leaf remains. This is

KNOWLEDGE OF THE MUTATING OENOTHERAS. 39

the most striking case I have observed, of change in leaf-characters during the ontogeny
of a rosette. Pl. 4. fig. 47 shows the adult rosette of the same type at the age of nine
months, the rosettes having been kept constantly under tropical conditions, so that they
developed slowly and displayed their characters to the full. In fig. 47 the characteristic >
type of leaf with pinnatifid basal lobes continues to be produced, and the first type has
disappeared. Іп the О. grandiflora type of rosette ontogeny there are, then, three
leaf-types as іп О. Lamarckiana and some of its derivatives, but only types (2) and (3)
are represented in figs. 45 and 46. Туре (2) in О. grandiflora is represented by a
number of leaves, while in О. Lamarckiana and its relatives types (1) and (2) are
only represented (at least in out-of-door cultures) by three leaves each. Тһе number
of leaves of each type is probably always variable, depending upon the conditions of
development.

When О. grandiflora is grown under ordinary cultural conditions—4. e., when the seeds
are sown in January and the young plants are planted out in May or June—the adult
type of rosette-leaf with pinnatifid lobes is usually omitted entirely ; and frequently, as I
shall show in other О. grandiflora races, there is no rosette at all, the young plant
forming а stem with internodes almost at the very beginning. This interesting
difference from О. Lamarckiana in physiological behaviour results in the plants
developing more rapidly, and frequently blooming earlier. РІ. 4, fig. 48 shows such а
plant of typical O. grandiflora m my cultures from Alabama. The absence of a rosette
gives the plant a shrub-like appearance. Such plants were top-heavy, and it was
necessary to have them staked in order to prevent their being broken down by the wind.
Fig. 48 also shows the ovate-lanceolate shape of the stem-leaves in O. grandiflora. They
taper at both ends, the lower end narrowing to a petiole, and (except the uppermost
bracts) are not sessile with aurate base, аз is usual in O. Lamarckiana. Such buds are
shown, reduced in size, in РІ. 4, fig. 49. Unlike those of O. Lamarckiana and all its
derivatives, the buds are free from long, muricated hairs, though they usually bear soft,
pubescent hairs, but are sometimes quite glabrous. Fig. 49 also shows the range of red
colour pattern in these buds.

O. biennis, Linn.

Yt is well-known that a large number of North American geographical races of
О. biennis exist. New species, such as O. strigosa and О. canovirens, are published as
segregates from time to time, but many others exist, and my experience in growing
numerous races from wild seeds collected in various parts of the continent has been
that usually each new locality yields a race which is recognizable as distinct by its
leaf-characters and habit. One of the most interesting questions in connection with the
evolution of the Oenotheras is, How have these races arisen and what relation do they
bear to the evolution of the group? Тһе fact that these forms are all self-pollinating
adds to their interest and makes the question of their manner of origin and evolution
somewhat different from that of the larger-flowered open-pollinated forms. For it is
clear that in the latter intercrossing of races must frequently take place, so that the

40 DR. В. В, GATES—CONTRIBUTION TO A

evolutionary unit is here not a series of pure, self-fertilizing biotypes, but a population of
interbreeding races whose constituency is changing within limits from generation to
generation.

The two series of forms—open-pollinating and self-pollinating—are, however, not
nearly so sharply separated as was formerly supposed. Later in the present paper I
shall cite several cases in my cultures, of the appearance of small-flowered, short-styled,
and hence self-pollinated plants from large-flowered, long-styled ones. I have even
found several cases of transition from long-styled to short-styled flowers or vice versa,
sufficient to cause the change from open-pollination to close-pollination, on the same
plant. All such cases probably occur in hybrids between long-styled and short-styled
forms. They are similar to the cases cited by Boulenger (1907). Since species are
known having large flowers and short styles (О. Simsiana as well as О. brevistylis, in
which the stigma is more or less rudimentary), and having small flowers with long
style (O. ornata), it is evident that in the further light of the experimental evidence of
transition referred to above, any sharp cleavage-line between open-pollinated and close-
pollinated forms, such as was formerly thought to exist, disappears. Added evidence is
furnished by the large-flowered, long-styled O. grandiflora, and its segregate, the small-
flowered, short-styled О. Zracyi, which otherwise clearly resembles О. grandiflora.
These grew together in Alabama and must have occasionally intercrossed.

I am growing the second generation of a number of distinct races of O. biennis from
various sources. They differ chiefly in leaf-characters and will be described in detail
at another time. The race of О. biennis which I have used іп my hybrid crosses has
already been described (MacDougal, 1905, p. 9), but I must add here certain photographs
which are necessary for a comparison of the parents with the corresponding parts of their
hybrid offspring. Pl. 4. fig. 50 represents a typical rosette of this race; Pl. 4. fig. 51,
a flowering shoot; and РІ. 4. fig. 52, the same in fruit, to show the capsules and the size
and other characters of the bracts. In РІ. 4. fig. 53 several buds are represented, reduced
in size. Three of them, taken from different heights of the stem, show the great range of
continuous variation in the size of bracts of an individual.

O. muricata, Linn.

This Linnean species has usually been distinguished from О. biennis by its smaller
flowers and narrower leaves. But, as mentioned elsewhere (Gates, 1911 а, p. 580),
I have found certain races having the flower-characters of O. muricata (i. e., smaller
than biennis), but with broad leaves. The narrower leaves of О. muricata, аз a distinction
from О. bienmis, therefore breaks down and there alone remains the difference in size of
flower.

The fact that O. muricata possesses both narrow-leaved and broad-leaved races, and
that both frequently occur in the same locality and are probably derived from seeds
of the same individual, is one of the most interesting discoveries in connection with my
cultures of this group. For the broad- and narrow-leaved individuals differ in the same

KNOWLEDGE OF THE MUTATING OENOTHERAS. 41

general way as do О. Lamarckiana and О. rubrinervis. And more particularly, they
differ in the same as do the “ twin” types first described by DeVries, which are obtained
in the F, when certain small-flowered forms, such as О. biennis and О. muricata, are
pollinated from O. Lamarckiana or its derivatives. It, therefore, seems probable that
this twin hybrid behaviour depends upon a double alternative (broad- or narrow-leaved)
condition in the wild state. My cultures of O. muricata from such widely separated
regions as Middleton, N.S., St. John, N.B., and Winnipeg, Manitoba, in Canada, have
revealed in each case two distinct races, the two being respectively identical from all
three localities. This shows not only the wide distribution of these particular types
East and West on the continent (though probably confined to a narrow range of
latitude), but since the two types were found in each case from a single collection of
seeds, and since they occur in strictly self-pollinating forms having the flower-characters
of O. muricata, it is evident that this twin behaviour can hardly be due to pollination
from large-flowered forms. This is especially true since no large-flowered forms are
known to occur in the regions where these seed collections were made, and they are
certainly absent at least from the Nova Scotia region.

In short, then, we have to account for the occurrence of broad-leaved and narrow-
leaved * twin" types in races which it is reasonably certain were purely self-pollinated.
And the same behaviour was repeated in collections of seeds from three widely separated
areas. All races of O. muricata do not show two types, however, several collections
of seeds having yielded uniform races with only one very constant type.

The broad-leaved and narrow-leaved types from my collection of seeds from Winnipeg,
are represented respectively in РІ. 4. figs. 54 and 55, while fig. 56 shows a plant of the
broad-leaved race from Nova Scotia, in flower. Its rosette is identical with that of
fig. 54. Ihave as yet been unable to test the behaviour of the twin races, except to
show that the broad-leaved type breeds true. Seeds of O. muricata sent to me from
the Bremen Botanical Garden were found also to contain two types which differed but
slightly if at all from the plants from the three localities in Canada already mentioned.
The total numbers of broad- and narrow-leaved individuals from these four sources were
respectively 30 and 27, indicating possibly a tendency to appear in equal numbers. It
must also be said that the two types were very distinct, as represented in the photographs,
without any tendency to be obscured by variation.

Another lot of seeds, received from the Kiel Botanical Garden under the name
O. biennis, yielded 37 plants belonging to two very distinct types: type (1) (26 plants)
had very broad, large, light-green rosette-leaves; type (2) (11 plants) had narrow,
dark-green leaves with long unmargined petioles. These two types were both unlike the
previously mentioned “twin” types, and were so unlike that it seems probable they may
have been thrown together by an accidental mixing of seeds.

On the other hand, I have collected seeds near Woods Hole, Mass., which yielded a
single uniform race of O. muricata, apparently corresponding with the var. canescens
of Robinson, and remaining true in the F, Seeds kindly sent by Mr. E. L. Rand,
from near the beach at Seal Harbour, Maine, also yielded a single, uniform race with

SECOND SERIES.— BOTANY, VOL. VIII. с

49 DR. В. В. GATES—CONTRIBUTION TO A

characteristic narrow, dark-green leaves often having red midribs. A young seedling
is shown in РІ. 5. fig. 57, and a nearly mature rosette in fig. 58. The young seedlings
are very characteristic, forming tiny, compact rosettes of dark-green leaves closely
hugging the ground. The number of leaves in the young seedlings seems to increase
much more rapidly than in O. biennis, O. grandiflora, or O. Lamarckiana forms, though
I have not determined the reason for this. The young plantlets also have a stouter,
harder (usually reddish) stem, and their roots are longer, making it more difficult to pull
them from the ground. АП these differences are perhaps adaptations to a more rigorous
environment. Another collection of wild (naturalized) seeds from near Berlin, sent
through the kindness of Professor E. Baur, yielded a uniform race apparently identical
with the last. In a culture of O. ammophila from the Bremen Botanical Garden
I was unable to find any constant distinction from certain races of O. muricata.

It is perhaps worth noting that in all cases where one uniform race was derived from
a collection, it was always a narrow-leaved type. Another interesting fact which shows
the extreme closeness between the organism and its environment, and the way in which
a slight environmental difference will induce a whole race of Oenothera to modify
markedly its development, is shown in my cultures, already mentioned, of O. muricata
var. canescens. Іп 1910 this race was grown іп a very wet, rich soil, with much humus,
and the whole race (69 plants) without exception formed very imperfect rosettes and
then shot up slender, unbranched stems. Тһе offspring from this race were grown the
following year in a much drier, poorer soil, without fertilizer, and with little humus.
These plants (22) again reacted uniformly in producing rather large rosettes, and,
finally, forming shoots with many long basal branches, giving them a pyramidal
appearance. The leaf and flower characters were, however, unchanged. Similar facts
are, of course, a matter of common knowledge to all who grow plants, but the numerous
and varied responses of which the plant is capable under varying environment are often
neglected or insufficiently considered, both in systematic descriptions and in experiments
on heredity.

My studies of O. muricata from collections of seeds thus far renders it probable
that the individual races are less numerous and more wide-ranging than is the case
with O. biennis. It also seems probable that the O. muricata forms are, on the whole,
distinctly more northern in distribution than О. biennis, there being possibly a relation
between smaller flower-size and higher latitude or altitude, though at present this is not
clear, for apparent exceptions occur. In any case, it could only be true in a very
general way.

The following table summarizes my cultures of O. muricata races. Several other
cultures were unfortunately killed. The detailed description of these races will be
considered at another time. |

KNOWLEDGE OF THE MUTATING OENOTHERAS. 43

TABLE ТУ.— Raees of О. muricata, Linn.

cata ШИШ ШИШ sss
О. muricata, Linn., Winnipeg, Man. ............ 11 4 T.i wa
А a Another sowing .. 2 0 2
The races from Winnipeg
BE Jo ND aes ue cepi 10 3 7 and Middleton are iden-
>tieal, and that from St
$ ^5 Another sowing .. 1 0 4 is at least very
closely similar.
я жы NOB. р 8 5 3
ji » Second sowing.... 5 2 3
š » Broad-leaved, F, .. 33 33 0 a рей of the F, with
broad leaves gave only
j; % й Жұ». 21 21 0 broad in F,
; Apparently identical with
k Bremen Botanical Garden .... 17 16 1 { iiie Wiwsipaq ceci
“0. biennis, Linn,” Kiel Botanical Garden........ 37 11 96 | vibe 5” {тош
О. muricata, Linn., Berlin (naturalized) .......... 148 = 148 | Uniform race.
2 Seal Harbour, Maine ........ 186 e 186 ще same as Berlin
var. canescens, Robinson, Woods 69 T 69 Тай fa.
Hole, Mass...
» » » Fj 29 .. 29 и

У. (А) CULTURES ОҒ OENOTHERAS FROM LANCASHIRE.

These cultures, which were begun іп July, 1907, have developed so many features of
interest that a separate account of them is necessary. We may recall that these
Oenotheras have been occupying the Lancashire sand-dunes for at least a century,
and probably date from a much earlier time. It is, therefore, of much importance to
know what has been happening to this complex of forms during the period of its
naturalization. An inquiry regarding the process of evolution in such a genus as
Oenothera must continually keep in mind a freely-intercrossing population as its
starting-point, and as the condition under which its evolution has gone forward. An
analysis of the changes which the plants may show as a result of freely commingling for
a long period is, therefore, of great importance, especially in its bearing on the relation
between crossing and the phenomena of mutation.

Bailey (1907) first called attention to the fact that O. Lamarckiana was growing on
the sand-dunes at St. Anne’s-on-Sea, and MacDougal (1907, p. 8) found that seeds from

G2

44 DR. В. В. GATES—CONTRIBUTION TO A

this locality produced О. Lamarckiana and several of its mutants, including О. lata,
О. rubrinervis, О. oblonga, and О. albida. Му cultures from near Birkenhead have
shown that О. grandiflora is an equally conspicuous member of the Oenothera population
of that locality. О. grandiflora and О. Lamarckiana have also intercrossed freely,
producing a complex of forms which I have since succeeded in part in analyzing. The
latter frequently display characters, particularly of the leaves, which are entirely different
from anything known in О. Lamarckiana or its mutants, though the bud and flower
characters usually correspond with О. Lamarckiana (as do the mutants of DeVries) or
are intermediate between О. Lamarckiana and О. grandiflora. This fact provided the
first clue to the relationships of these aberrant types. I have isolated а number of these
types, which are now being grown in the second or third generation. They will be
described at a later time. These cultures are, therefore, an attempted analysis of a
typical interbreeding colony of forms. Itis hoped that naturalized or indigenous colonies
of Oenotheras in other parts of the world will be subjected to a similar analysis, for it is
only by this means that any full knowledge of the factors at work in the evolution
of the génus can be gained. I shall be grateful for seeds sent to me from any locality.

Тһе question whieh must be asked regarding the Lancashire Oenotheras is, Are all
these new aberrant races the direct result of hybrid combinations, or must an additional
factor of mutation be invoked to account for the origin of some of them? We have
already seen that, in the case of the mutants from O. Lamarckiana, such a factor must
be invoked, for the origin of certain of the mutant characters cannot be explained on
the basis of hybrid splitting. In the case of the Lancashire Oenotheras, it is evident
that many of the new races at least are complex hybrids, even although they breed true.
Whether any of them have had a different origin, remains to be seen. In any case, the
mutational process (using the term in contradistinetion to hybrid splitting) is probably
a result of the germinal disturbance which the race has experienced through previous
crossing. Оп this view, mutation and hybridization are closely related processes, the
former being an indirect result of the latter. This does not exclude the possibility,
indeed probability, that mutational transformations may result also from environmental
action on the organism, or from other causes.

We may now consider some of the results of an experimental analysis of these forms—
an analysis which is not yet completed, but certain results of which need to be stated.
It will be remembered that these forms аге in part derivatives from О. Lamarchiana
and in part complex hybrids between О. Lamarckiana and О. grandiflora races. It is
possible that О. biennis may also have entered into their composition, though I have no
direct evidence of this.

My first culture of these forms consisted of 56 plants grown in a tropical hothouse
(see Gates, 1912). Half of these developed the characteristic O. grandiflora rosette (see
РІ. 4. fig. 47), while the remainder displayed a large number of curious rosette types,
scarcely two being alike, but all falling in the series of О. Lamarckiana and its mutants,
_ clearly rather widely divergent from О. grandiflora. The rosettes developed their
characters to the fullest extent, and seven о these (РІ. 5. figs. 59-65) аге selected to —

KNOWLEDGE OF THE MUTATING OENOTHERAS. 45

show the range of types represented *. The other rosettes approximated more or less
to these, though no two were quite identical, The variability, both of this and an
independent outdoor culture the following year, was far greater than DeVries’s race of
O. Lamarckiana exhibits. Fig. 59 represents a rosette (No. 5) belonging to the O. lata
race already described (p. 36) as originating from this locality. Its leaves differ
constantly from the О. lata of DeVries (cf. Pl. 3. fig. 42). Pl. 4. fig. 44 shows another
rosette of this O. lata race as grown in outdoor culture. The greater elongation of the
blades and petioles is the feature distinguishing both from the race of DeVries. The
buds of the two have already been contrasted (РІ. 3. fig. 43). The individual (No. 5)
represented in Pl. 5. fig. 59 produced a somewhat fasciated stem with straggling,
irregular branches, probably a result of the tropical environment. Its rosette closely
resembled No. 6, which was elsewhere described (Gates, 1912, pl. 2) as continuing to
form rosette-leaves without internodes. No. 5 was self-pollinated, and its F, grown
out-of-doors in the usual way yielded the following interesting series of types. Three
of the offspring produced rosettes like their parent, $. в. they were lata-like, but the
flowers were as small as in О. biennis and the anthers surrounded the stigma in the
bud. The buds bore numerous muricated hairs, and the bud-cones were somewhat
barrel-shaped (type 1). To the next type belonged 7 plants, a very distinct type
represented іп РІ. 5, fig. 66, having very narrow, almost elliptical, whitish leaves of
uniform width (type 2). This rosette agrees with DeVries’s О. elliptica, (af. ‘ Mutation
Theory,’ i. 394, fig. 83), but the flowers were small with the anthers surrounding the
stigma. Two other plants (type 3) resembled most those just mentioned, but were
dwarfs. Four plants belonged to another type (РІ. 5. fig. 67) having small flowers іп
which the anthers were almost or quite touching the base of the stigma-lobes (type 4).
This unexpected appearance of short-styled flowers would indicate an ancestral cross
with О. biennis. I have, however, never discovered О. biennis in my cultures, numbering
several hundred plants, from this locality. Of course, there remains the possibility that
the tropical conditions under which the fertilization and development of the seeds of the
mother plant took place may have served to call forth “latent” characters in the race,
which it had acquired through previous crossing.

However, I grew outdoors an independent culture from a second sowing of Lancashire
seeds, and these revealed (among many types) one constant type with flowers inter-
mediate between O. biennis and O. Lamarckiana, so this furnishes very good evidence
that O. biennis has, at a previous time at any rate, formed a part of this colony. The
behaviour of three races which belonged to O. grandiflora, though differing somewhat
among themselves, may be summarized here. One race (54х32) was typical rather
broad-leaved O. grandiflora, agreeing closely with the Alabama form. The Е, contained
11 plants and the F, 82 plants which bred perfectly true. Another race (No. 40), the
F, from one plant, contained 43 individuals. This race also belonged to O. grandiflora,
though differing in minor points from the last, and it was much less uniform. Several
plants omitted the rosette stage and 5 were dwarfs. One of the dwarfs produced in the

* It will be observed that in hothouse cultures the number of rosette-leaves is always greater than out-of-doors.

46 DR. В. В. GATES—CONTRIBUTION ТО А

F, а uniform race numbering 23 plants. They varied in height from one to two feet,
but otherwise agreed in having O. grandiflora flowers and general leaf characters.
This dwarf race of О. grandiflora is clearly different from О. nanella. Still another
dwarf race had the following history. It arose in another race (No. 49) which closely
resembled No. 40, the plant yielding in F, 80 offspring which were fairly uniform,
except one rosette which resembled an overgrown O. gigas, and 2 dwarfs. The next
generation from one of the latter yielded a uniform progeny of 20 plants. Unlike
the last dwarf races, they averaged only one foot high, and their leaves resembled
O. nanella, but they had O. grandiflora flowers and buds. It is evident, then, that
constitutional dwarf races may easily occur in these forms.

Numerous other types are now being grown in F, and will be described in detail
later. My cultures of these forms furnish abundant evidence that rather uniform
true-breeding races can be obtained from this evidently hybrid progeny. Indeed, my
experience has been that most of these races of hybrid ancestry will breed reasonably
true, at least, after the F,, though, as might be expected, there is usually more variability
in these races than in the mutants of DeVries which have been inbred for a number of
generations.

In all cases of which I am aware, of crosses between large-flowered and small-flowered
Oenotheras (see Section on hybridization, p. 52), the flowers of the F, hybrids are
intermediate, both as regards size of petal and length of style. As showing that an
analysis of a mixture of races such as I am describing is possible, I may refer to one
race (070.4) derived from a single individual іп my sowings of 1908. The first generation
of offspring from this plant numbered 60 uniform individuals, and four of these, self-
pollinated, yielded in all over 200 for the second generation. They constituted a very
uniform race, and a eareful comparison showed that their characters were identical in
every respect with those of the hybrid О. nanella x О. biennis, which also breeds true
(cf. Section on hybridization where the rosettes are described). Тһе stem-leaves of this
race (070.4) were found to be rather exceptionally variable, varying from pointed at
both ends and petiolate, to sessile with broad base, uncrinkled. Тһе dimensions of
the flowers were as follows:—length of ovary 8 mm., hypanthium 32 mm., cone
99 mm., вера! tips 5-8 mm., diameter of cone at base 5-7 mm. There is considerable
variability in the length of style, the stigma being above the anthers in some, but
touched by their tops in others, though it is apparently never so long as in
О. Lamarckiana or so short as in О. biennis. This race much resembles О. rubrinerris
in general appearance, on account of its usually red midribs in the rosette-leaves. But
the buds are shorter and more slender (base of cone 9 mm. in diameter), the rosette-
leaves are usually less crinkled, and the two forms cannot really be confused, the
smaller flowers, of course, making the race distinct.

The numerous other races I am now cultivating from this source will be described in
detail elsewhere. Abundant confirmation is, however, already furnished for the
interesting fact that in Oenothera hybrid races with characters intermediate between
their parents frequently breed true. The question of the gametic composition of such
races, and the reasons why they breed true, will be referred to in the next section.

KNOWLEDGE OF THE MUTATING OENOTHERAS. 47

These cultures also by showing that, in addition to O. grandiflora, O. Lamarckiana,
and its mutants as known from DeVries’s experiments, a host of other forms occur in a
freely interbreeding population, raise the question as to the origin of the latter forms,
and the limits and relations between hybridization and mutation. Тһе first step in
answering that question is taken by isolating the many races and learning their
characters. The next step is to recombine them in various ways and determine the
results of those recombinations or by independent syntheses to build up hybrid races
having the same characters. By such means we may expect ultimately to understand
the relation between hybridization and mutation phenomena.

Finally, it is important to note certain facts regarding the history of this colony of
Oenotheras. It must perhaps always remain uncertain whether the O. Lamarckiana in
this colony was transported as such from the other side of the Atlantic, or whether it
originated in the colony as a result of crossing. I have pointed out that in the latter
case it would have had an equal chance of originating in the region of Virginia, where
both О. biennis and О. grandiflora, the only forms which could have given rise to it,
were formerly wild. In any case, the important fact, from an evolutionary point of
view, is that О. Lamarckiana and its derivatives, such as О. rubrinervis, are now the
dominant feature of this Oenothera colony. Whether or not they have originated in
that colony, they are the forms which will provide most of its future population, and,
therefore, determine any future evolution it may undergo. If O. Lemarckiana has
originated there as a hybrid, then it has largely overcome its parent species in the
struggle for existence, and its descendants will be the survivors for the next
evolutionary step.

(B) CULTURES FROM BOTANICAL GARDENS.

I have obtained seeds of various Oenothera species from nearly all the Botanical
Gardens in Europe. А number of these cultures have contained distinct surprises.
I have already mentioned certain ones in other connections, but a few inore will be
added here. Of course, the names under which seeds are received cannot always be
relied upon. I have received О. Lamarckiana several times under the names О. biennis
or О. biennis grandiflora, For instance, seeds from the Botanical Garden at Kolosvar,
Hungary, under the name О. biennis grandiflora were found to be nearer О. Lamarckiana
than any other species, though not typical of its mean condition. Again, seeds from
the same garden under the name О. erythrosepala, Barb., developed О. Lamarckiana and
one typical О. lata mutant.

A lot of seeds of О. suaveolens, Desf. (= O. grandiflora), from Nantes Botanical Garden
yielded nearly 200 plants, which showed a great range of variability with every evidence
of previous crossing. Many of the rosettes closely resembled O. Lamarckiana, while
some were very similar to О. rubrinervis. Among other aberrant types, there were
one dwarf and two with short styles. The flower dimensions in one of the latter were
as follows: ovary 15-17 mm., hypanthium 47-51 mm., petals 40-42 mm. long. The
style, however, was very short so that the stigma was below the anthers. The sepals
were reddish, and the plant somewhat resembled a brevistylis form of О. rubrinervis, АП

48 DR. В. В. GATES—CONTRIBUTION TO А

the flowers on this plant were alike, with very little variation. These flowers would
not be self-fertilized in the bud because the style was too short. In the other short-
styled individual the styles were somewhat longer, so that in the first flowers the stigma
was just below the anthers, and in later flowers was touched by the anthers, as in
O. biennis. I have not yet determined whether these two aberrant plants resulted
from germinal or fluctuational variation, probably the former.

VI. HYBRIDIZATION IN OENOTHERA.

In this section will be taken up (1) various experiments in erossing the mutants with
each other and with О. Lamarckiana, and (2) the results of crossing various forms with
О. biennis. The race of О. biennis used in these crosses came from New York State,
and has already been described by MacDougal (1905, p. 9). Its main features are here
illustrated in РІ. 4. figs. 50-53.

The results of these erossing experiments partly confirm DeVries's experiences in
crossing Oenothera forms, though there are certain contradictory results. An effort has
been made—particularly in the description of the hybrids between О. biennis and other
forms, where it is most needed on account of the brevity of DeVries's descriptions
of the twin hybrids—to make the description and illustration of the various forms,
and their relation to each other and to their parents, as clear as possible. The
complete description and accurate comparison of these hybrids, not merely in taxonomic
language, but so as to bring out the relationship of their characters, is a difficult
task, but one which is essential to an elucidation of the laws of inheritance in these
forms.

After the hybrids have been described, and the results of crosses compared with those
of DeVries, an attempt will be made to givea résumé of the various types of inheritance
now known to occur in Oenothera. This seems particularly necessary because the results
of Oenothera crosses have received little attention in the recent literature of
hybridization. Іп the last section I showed that races of Oenothera which are of hybrid
origin frequently breed true, and this is in entire accord with the results of crossing-
experiments. From a survey of all the hereditary phenomena in the genus, it appears
that there are two general types of behaviour : (1) alternative inheritance; (2) blended
inheritance. How far these may represent fundamentally different processes will be
discussed after the data are presented.

(a) Crosses involving O. Lamarckiana and its Mutants.

DeVries made a number of crosses between O. Lamarckiana and its mutants, with the
result that both parental forms appeared in the F,, and both usually bred true in later
generations. І have made similar crosses, with results which corroborate those of
DeVries. The data from some of these crosses may now be stated.
_ О. lata X О. Lamarckiana yielded іп 1906 five О. lata and eight О. Lamarckiana.
Six of the latter, self-pollinated, produced a total of 104 plants іп 1907, all of which were
o. Lamarckiana, UN no difference whatever from О. aa a pure

KNOWLEDGE OF THE MUTATING OENOTHERAS. 49

parentage. Again, О. rubrinervis Х О. Lamarckiana (the latter derived from a cross
with О. lata) gave in one case (No. 115 x 77) 32 О. rubrinervis, 5 О. Lamarckiana,
and 1 which was uncertain. In another case (No. 121 x 78), with parents of the
same ancestry, the result was 49 O. rubrinervis, 16 O. Lamarckiana, and 1 O. lata. In
one case, from О. rubrinervis X О. Lamarckiana, the Е, from a rubrinervis plant of the
Е, gave 82 individuals exclusively О. rubrinervis, and a Lamarckiana plant from the Е,
of the same cross yielded 19 plants purely O. Lamarckiana, bearing out DeVries’s result
that splitting in the F, does not occur. The data from these and three other crosses
are brought together in the following table :—

TABLE V.
Number
Date. Cross. о OFFSPRING.
Plants.
1906. | O. lata x О. Lamarckiana, Е,........ 13 О. lata 5 + О. Lamarckiana 8,
0. Lamarckiana, Е, (6 plants) ..| 104 | 0. Lamarckiana 104.
1906. | О. rubrinervis х О. Lamarckiana, F, ..| 38 O. rubrinervis 32 + O. Lamarckiana 5 + 1 uncertain.
1906. | О. rubrinervis х О. Lamarckiana, Е, ..| 66 "e 49 4- » 16 + O. lata 1.
1907. О. rubrinervis, Е, (1 plant) 82 | О. rubrinervis 82.
1907. О. Lamarckiana, F, ,, 19 | O. Lamarckiana 19.
1909. | О. Lamarckiana х О. rubrinervis, Е, ..| 68 О. Lamarckiana 22 + 0. rubrinervis 45 + 1 mutant.
1910. | О, Lamarckiana х О. rubricalyz, Е, ..| 11 | О. Lamarckiana 2 + О. rubrinervis and О. rubricalyz 8 +
1 mutant.

1910. | О. rubricalyx x О. Lamarckiana, Е, ..| 4 | О. Lamarckiana 1 + О. rubricalyx З (rosettes only).
1909. | О. Lamarckiana x Q. gigas, Е, ...... 40 O. Lamarckiana 40.
1911. н Ж „ F,(2plants| 44 | О. Lamarckiana 44.
1907. | О. rubrinervis x O. nanella, Е poss] 42 | О. Lamarckiana 20 + О. rubrinervis 20 + 0. lata 1.
1907. | O. rubrinervis x О. папеПа, F, ...... 3 | O. Lamarckiana 1 + O. rubrinervis 2.
1907.| О. lata x О. rubrinervis, Е,.......... 1 | 0. oblonga mutant.
1907. pox s Fu. 4 | О. Lamarckiana 8 + O. lata 1,

The O. Lamarckiana remained true in F, and F,.
1910. | О. rubrinervis х О. папеЙа, Е, ...... 79 | О. Lamarckiana 95 (?) + О. rubrinervis 52 -- 2 aberrant.
1910. О. rubricalyz x О. папеЙа, F, ...... 42 | О. Lamarckiana 7 + О. rubricalyx 35.

Though the numbers involved were small in several of these cases, yet the results are
sufficient as confirmatory evidence.
SECOND SERIES.—BOTANY, VOL. VIII. H

50 DR. В. В. GATES—CONTRIBUTION TO А

In the cross О. Lamarckiana x О. rubricalyx, the Lomarckiana was trom the Е, of
О. Lamarckiana x О. gigas. Its characters differed in no respects from О. Lamarckiana
of pure ancestry, and its guarded offspring also bred true. Its behaviour was, therefore,
in every way the same as that of pure O. Lamarckiana.

Again, when two mutants from О. Lamarckiana are crossed, O. Lamarckiana in
addition to one or both of the parent types is usually obtained in the Е. Thus, the
cross О. rubrinervis x О. nanella was made by DeVries four times, and in each case
the F, showed only two types, О. Lamarckiana and О. rubrinervis, the percentage of
the latter varying from 92 per cent. to 25 per cent. And in the one case in which later
generations of the cross were grown, the О. rubrinervis of the F, gave five types:
(1) О. Lamarckiana, which proved constant ; (2) О. rubrinervis, which also bred true;
(3) O. nanella, whose constancy was not tested; (4) an intermediate type, O. nanella
rubrinervis, which bred true; and (5) O. rubrinervis, which in the F, split again in the
same manner into О. Lamarckiana, О. rubrinervis, and О. nanella rubrinervis (see
DeVries, Die Mutationstheorie, ii. p. 451). Тһе constant intermediate type, О. rubri-
nervis nanella, is virtually a much dwarfed О. rubrinervis. "Тһе foliage is that of
O. rubrinervis but the leaves are smaller. The amount of dwarfing is not so great
as in O. nanella, but it brings the plants much nearer the size of О. памейа than of
О. rubrinervis.

The crosses made between mutants also show that O. Lamarckiana as well as
one of the parent types appears in the F,. О. памейа, however, is apparently an
exception as it rarely or never appears in the F,, the dwarf characters being recessive,
These results are grouped together in the second part of Table V.

The sharp way in which the young seedling types are contrasted with each other
is clearly shown in РІ. 5. fig. 68. "This pan contains the Е, from О. lata x O.nanella.
Four of the seedlings (large and broad-leaved) are O. Lamarckiana, two O. rubrinervis,
one O. lata (with crowded, broad, somewhat crinkled leaves and short petioles), one
(small) uncertain, and one new mutant having narrow, elongated, whitish leaves with
very little chlorophyll. The latter soon died from lack of nourishment.

(b) Crosses with O. gigas.

The results derived from crosses between the mutant O. gigas and other forms
have been somewhat conflicting, but I think all can now be brought under one view-
point without involving a contradiction. In the first place, O. gigas, when crossed
with O. Lamarckiana, clearly differs from the other mutants in its behaviour, at
least in certain cases. DeVries first showed that О. gigas x О. Lamarckiana and its
reciprocal give an identical, uniform hybrid type, intermediate between the parents.
He has since (1908) brought together the results of his crosses with this mutant,

. which may be summarized in the following table :— x

KNOWLEDGE OF THE MUTATING OENOTHERAS. 51

Таві VI.
| 0. зл х О. Lamarekiana, В .....::-.. 62 plants. Е, 224. А constant, intermediate race.
| О. Lamarckiana x О. gigas, Fy .......... 069, Е, 180. Same as гесіргоса].
| O. gigas х О. brevistylis, Е QUSE EC pA UR i is Same as above.
| О. gigas х О. rubrinervia, Бу... D ^" »
| О. rubrinervis se О, дыша, КО. л, 288...
68 intermedi ee MES О. lata & О. gigas.

| Оша х О Колы eke Os 235 , Las 7 ОЕ е dii
| O. biennis X O pigat, Жұл ава... BÉ i Intermediate type.

О. gigas X @ боты, Ei. i ee ees 23 06 а " 9

О. muricata X О. gigas, Е.а. а... 66 , » »

Оаа X O. тия, E, sea a e cks Gb в 9; 39

То these may be added the results of Miss Lutz (1909), who from О. lata X О. gigas
obtained an T; of 40 plants, containing three types: (1) O. lata (two plants, each having
15 chromosomes); (2) O. gigas (six plants, each having about 30 chromosomes) ; (3) a
more or less intermediate and variable group of 32 plants, having also an intermediate
number of chromosomes—22, 23, or 21.

In similar crosses, Geerts (1911) reports the following results from counting the
chromosomes of О. gigas hybrids: О. lata X О. gigas gave an F, containing a /ata-like
type and a gigas-like type, both having 21 chromosomes. This is contrary to the results
of Miss Lutz, though he omits to quote the latter. O. gigas X О. Lamarckiana and its
reciprocal gave a constant intermediate . type, though the Е, contained 21 chromosomes
and the F, are said to contain only 14. That is, though the number of chromosomes
changes from 21 to 14, there is said to be no change whatever in the external
characters. In thelight of the results of Miss Lutz above quoted, this may well be
doubted. Further, I have shown (Gates, 19094) that О. gigas has conspicuously larger
cells in its various tissues, a fact which is undoubtedly closely connected with its
larger dimensions. It is, therefore, highly improbable that the nuclei and cells, and
hence various dimensional characters, of the F, are identical with those of the F.
Indeed, Geerts’s own figures, in the paper referred to, prove that the nuclei, and doubtless
also the cells, are smaller where the chromosome number is smaller, as I had shown
previously.

I made the cross О. Lamarckiana х O. gigas in 1909, and obtained ап Е, of 40
plants the following year. They were all identical with O. Lamarckiana in every
respect, and several which were examined contained only 14 chromosomes. Two of
these plants yielded an Е, of 44 plants which were indistinguishable from
O. Lamarckiana. Of course, there is the possibility that an error may have occurred,
in collecting the wrong capsule from the original mother plant, but all such errors are
guarded against as carefully as possible.

These facts are paralleled by a recent result of Miss Digby (1912) with Primula.
It was found that P. floribunda var. isabellina X Р. kewensis gave an F, resembling
the seed parent. The former parent has 9 chromosomes as z number, while the
latter has 18 as x number, yet the F, contained only 18 chromosomes as 22 number.
It seems probable that the extra 9 are extruded in the divisions of the embryo.

H2

52 DR. В. В. GATES—CONTRIBUTION ТО A

2 Inview of the results of Miss Lutz, іп which two О. lata plants having 15 chromosomes

occurred in the F, of a cross with O. gigas, and the result of Geerts’s that the chromosome
number in these crosses is 21 in F, and becomes 14 in Е,, we may assume that ап
extrusion of the extra chromosomes from the nuclei has occurred, in some cases during
the division of the fertilized egg in the original cross, in other cases during the reduction
divisions in the F, plants. The former case would be in harmony with certain results of
Baltzer, Herbst, and Tennant in connection with the behaviour of the chromosomes in
hybrid Echinoderm larvee, in which the extrusion of certain chromosomes takes place
during cleavage.

This leaves unexplained the result of Miss Lutz in obtaining six pure O. gigas plants
having the gigas chromosome number, in the Е, of О. lata x О. gigas. There is another
possible explanation of the two lata plants in the F, of this cross, namely, that they
originated apogamously. I obtained experimental results some time ago (Gates, 1909 5)
which showed with a certain degree of probability that a small percentage of the ovules
of O. lata will develop without fertilization.

It is clear that further experiments are required before these somewhat conflicting
results can be fully understood. It is at least evident, however, that O. gigas differs from
the other mutants in giving, in some cases, an Е, which is intermediate between the
parents, in other cases two or three different types; and it appears that the variation in
crossing-behaviour is closely connected with the distribution and fate of the chromosomes
in the various hybrids.

(c) Crosses between O. biennis and O. Lamarckiana and its derivatives.

These crosses between large-flowered and small-flowered forms display a number of
interesting and peculiar features. (1) The reciprocal crosses are not alike, the hybrids
being usually distinctly patroclinous. (2) These hybrid races breed true in later
generations, except in certain characters such as dwarfing, in which they split.
(3) Certain crosses (chiefly those in which О. Lamarckiana or one of its derivatives is the
pollen parent) exhibit two types in the F,, both of which are different from either parent.
These DeVries calls twin hybrids. They are respectively broad- and narrow-leaved, and
are called laeta and velutina. The same laeta and velutina types may be derived from
several different crosses, or they may differ in certain minor characters according to the
parents used. (4) The flowers of these hybrids appear to be in all cases intermediate in
character between those of the parents. The flowers are larger than O. biennis, but
smaller than O. Lamarckiana, and the styles are short enough so that the anthers touch
the base of the stigma and secure self-fertilization. There is, however, a considerable
range of variability in both these features of the hybrid flower.

DeVries has described this type of hybrid behaviour in several papers (1907, 1908,
1909), and Andrews (1910) has added anatomical notes on the structure of the hybrids,
but no figures of these hybrids have been published. In a number of crosses which
are summarized in Table VII. (p. 56), I have obtained results which are in general |
| 2. with those of DeVries, though certain probably important differences are to be 1

KNOWLEDGE ОҒ THE MUTATING OENOTHERAS. 53

noted. The numbers of plants are, unfortunately, often small on account of sterility,
the biennial habit, ete.

In regard to all these crosses, it must be said that fluctuating variability frequently
obscures the results and renders the formation of sharp categories very difficult. The
characters of these hybrids, such as broad and narrow leaves, are particularly responsive
to slight differences in nutrition, such differences producing in some cases great variation
in the width of leaves in one type, in other cases obscuring the line between two
different types. Apparently the best conditions for displaying the laeta and velutina
characters are, in a rather poor soil, with very little fertilizer but plenty of water. The
types then appear very sharply distinct, not only in width of leaf, but in colour,
pubescence, and crinkling. Owing to this obscuring fluctuation and to environmental
influences, I am still uncertain whether O. Lamarckiana х О. biennis and О. nanella x
O. biennis yield a single fluctuating type or two types which are distinguished only
by width of leaf. The two crosses with О. rubricalyx given in the table are merely
additional to those already published (Gates, 1911 0), in which it was shown that
O. rubricalyx behaved the same as its parent O. rubrinervis.

The two sharply distinct rosette types derived from О. biennis Х О. Lamarckiana are
shown in РІ. 5. figs. 69 & 70, which represent laeta and velutina respectively. That
O. biennis X O. laevifolia yield two types which are, in some crosses, identical with the
preceding, is shown by РІ. 6. figs. 71 & 72. The rosettes from this cross sometimes show
much variability, however, and I have obtained rosettes which were exactly intermediate
between O. biennis and O. laevifolia, as shown in РІ. 6. fig. 78 (cf. Pl. 4. fig. 50 and Pl. 2.
fig. 17). The basal jagged teeth of this hybrid are such as frequently occur in
O. biennis. Flowering shoots of the narrow- and broad-leaved hybrids derived from
O. biennis x O. Lamarckiana are shown in РІ. 6. figs. 74 & 75, where the difference in
width of stem-leaves is clearly seen. Other differences between these types, as regards
flowers, etc., have been described by Andrews (1910).

As the twin types mature, their characters gradually converge, so that in the fruiting
stage they are wholly indistinguish able. This is due to the rapid diminution in length and
width of the bracts of both types. In РІ. 6. fig. 76 are shown, two-thirds natural size,
several buds of the velutina type from О. biennis x О. Lamarckiana, in which the bracts
pass from relatively broad and as long as the flower-bud, to very narrow and less than half
the leneth of the bud. The characteristic hairiness of this type of bud is also seen in the
photograph (cf. РІ. 3. fig. 41 for О. nanella and РІ. 4. fig. 53 for O. biennis). A fruiting
shoot of О. biennis x O. laevifolia is shown in РІ. 6. fig. 77. The twin types are at this
time wholly indistinguishable from each other and from the two types in О. biennis X
О. Lamarckiana. The measurements of the buds іп О. biennis X О. laevifolia were as
follows: ovary 11 mm., hypanthium 28 шш, bud-cone 20 mm., вера! tips 9 mm.,
petals 30 mm. maximum, usually less.

The results derived from the reciprocal crosses are not quite so clear on account of the
obscuring effect of fluctuation, but in a general way there appears to be one rather widely
fluctuating type. Its broader-leaved varieties (w hich are the most frequent) are very

54 DR. В. В. GATES— CONTRIBUTION TO A

similar to the broad-leaved type derived from the reciprocal cross, as will be seen by
comparing Pl. 6. fig. 78 with РІ. 5. fig. 69. But a narrower-leaved type frequently
appears and breeds fairly true in later generations. This type has identical characters іп
О. Lamarckiana X О. biennis and іп О. nanella X О. biennis. А typical rosette is shown

in Pl. 6. fig. 79. It is markedly narrower-leaved than fig. 78, and comes true to its,

difference, yet it does not resemble the velutina or narrow-leaved type derived from the
reciprocal cross (cf. Pl. 6. fig. 79 with figs. 70 & 72). Whether, therefore, with О. biennis
as the pollen-parent we derive two types differing only in leaf-width (and not in other
characters, as do laeta and velutina), or whether this represents a single type with widely
fluctuating variability, I am not prepared to decide. Another cross, О. nanella X
O. biennis, yielded a large offspring with a continuous range of variation in leaf-width,
but agreement in other characters. If there are two types in this cross they differ from
each other, then, only in leaf-width, and not, like /aeta and velutina, in a series of other
foliage characters.

Another point with regard to these hybrids, О. nanelia x О. biennis, О. Lamarckiana
x О. biennis, and (О. nanella x О. biennis) x (О. Lamarckiana x О. biennis), which are
all identical, is that some of the young seedlings are frequently lacking somewhat in
chlorophyll, giving them a whitish appearance. "These individuals form smaller rosettes
with yellowish centre, but otherwise agree with the type. These rosettes finally,
however, “ catch up" to the others, and the appearance of two distinct types in the
seedlings is seen to be due to the presence of weaker individuals which contain less
chlorophyll and hence grow more slowly. І have not determined whether this
constitutional weakness is inherited. А fruiting shoot of О. Lamarckiana x О. biennis
is shown in РІ. 6. fig. 80, for comparison with fig. 77 and РІ. 4. tig. 52.

That the reciprocal hybrids differ from each other, even at an early age, is evident from
а comparison of Р]. 6. figs. 81 & 82, which represent respectively О. biennis х О. nanella
and its reciprocal. Comparison with fig. 83, which represents рше О. пале а seedlings,
shows that both hybrid types differ from О. nanella even at this stage. It also shows
that O. biennis X О. nanella is at this time patroclinous, most resembling О. nanelia, while
the reciprocal, with its long leaves and petioles, is nearer О. biennis.

As shown by the table (VII.), in О. палеа X О. biennis, and also in its reciprocal, I
have obtained dwarfs in the F,. This is contrary to the results of DeVries, who found
the dwarf character recessive in the F,, reappearing іп the F,. Thus (DeVries, 1908) in
О. muricata x O. nanella, the F, gave laeta and velutina, about 50 per cent. of each, and the
laeta type continued to breed true while the velutina split in the Е, and later generations,
giving about 60 per cent. dwarf velutina (murinella) and the remainder normal tall
relutina. In the paper above referred to, DeVries concludes that laeta breeds true because
its pollen carries only tallness, which is dominant over its own heterozygous egg-cells
(tall and dwarf) but recessive to pure “dwarf” cells. The latter, however, is contra-
_ dietory to the fact that in velutina tallness is dominant to dwarfness, and also leaves

3 unexplained why laeta does not split while velutina does split. Further experiments will ` 4

be required before this contradiction can be cleared up.

A

KNOWLEDGE OF THE MUTATING OENOTHERAS. | 55

Another result which has not previously been recorded is as follows: The velutina from
the Е, of O. biennis x О. Lamarckiana was pollinated by О. Lamarckiana, giving the three
types laeta, velutina, and О. Lamarckiana. This cross when repeated gave the same
result, the proportionate numbers being, moreover, about the same in both cases—
i. e., the velutina were five or six times as numerous as the other types, though the
numbers were rather small for determining ratios (see Table VII.) It was hoped that
a number of other crosses which I had made would permit a solution of this interesting
'situation, but many of them failed to develop, so the results are still incomplete.

With the above should be compared the two cases (Table VII., p. 56) in which the
hybrid О. Lamarckiana X О. biennis was pollinated by О. Lamarckiana and yielded a
series of rosettes which apparently contained the two types О. /aeta and О. Lamarckiana,
though there is much variability. А further series of crosses will be necessary before this
behaviour can be fully understood. According to DeVries (1911) (О. Lamarckiana x
biennis) x Lamarckiana gives pure Lamarckiana at least as regards rosette characters,
for he does not speak specifically of the flowers. It seems questionable whether the
latter will not remain intermediate in all these hybrids. If the rosettes from this cross
all belong to О. Lamarckiana they show, at least in my cultures, a much greater range
of fluctuation than pure cultures of that form yield.

In the important paper above referred to, DeVries describes double reciprocal hybrids
between O. biennis and O. muricata, the results being in agreement with the predictions,
on theoretical grounds, of Giglio-Tos (1908, 1911). Thus (0. biennis x muricata) х
(muricata x biennis) gives О. biennis which remains true, and (О. muricata X biennis)
х (biennis X muricata) gives О. muricata.

(d) Results.

It would at first appear, from a survey of all the results of crossing in the genus
Oenothera, that we are dealing with four distinct types of hybrid behaviour: (1) crosses
of O. Lamarckiana with its mutants, except O. gigas; (2) crosses of O. gigas with various
forms; (3) crosses between the large-flowered and small-flowered species of Oenothera ;
and (4) Mendelian behaviour in occasional characters such as in О. brevistylis and O. rubri-
calyx when crossed with their parents. But I believe it can be shown that types (1) and
(3) are essentially the same, and the peculiarities of type (2) in which it differs from the
behaviour of the other mutants are, as I have already indicated, probably to be explained
in connection with the presence of a double set or tetraploid number of chromosomes,
the members of the extra set of chromosomes varying in their distribution in different
hybrid crosses.

In the first place, it is possible that the appearance, for example, of O. rubrinervis
and O. Lamarckiana in the F, of crosses between these forms is only an extreme case
corresponding to the ¿aeta and velutina appearing іп, е. g., О. biennis X О. Lamarckiana.
In both cases, later generations of both types breed true (usually), the difference being
that in the former case the types derived from the cross are at least closely similar to,
if not identical with, their parents, while ¿aeta and velutina are quite as different from

56

DR. В. В. GATES—CONTRIBUTION ТО A

TABLE УП.
Date. Cross. peus OFFSPRING.
1908. | О. Lamarckiana х О. biennis, Е,.............. 9 Uniform intermediate type.
Туре І., 37 plants, same аз О. nanella x О. biennis.
1909. » ж ” F.H 40 | » Il., 3 plants, broad-leaved.
1909. Ж x E ТОИ 1 0. oblonga rosette.
1 Lamarckiana rosette.
1910. я x 5 Балл 8 | 1 velutina rosette.
1 small rosette Lamarckiana (?).
1911. % х " Базы wanuy s 2 Small rosettes, apparently oblonga.
1911. | (О. Lamarckiana x O. biennis, Е.) x Lamarekiana| 9 G клр тену us ea its u a
4 6 Lamarckiana.
1911.|( i; x Е,) x Lamarckiana 13 Er or lacia.
px 1 а smaller flows.
1911. O. rubricalyz x e biennis, Ho E UE, 20 Uniform type.
' 24 velutina.
1909. | 0. Бітік x 0. Lamarckiana, F Қалы RE тА 39 | 14 broader-leaved (laeta).
1 very close О, biennis.
2 narrow-leaved (velutina).
"apb. T а p itte кызу, ^ | 2 broad-leaved (laeta).
1910. ” x T latte, E, l rios 21 All laeta, flowers intermediate.
1910. au » velutina, ES 2552 51 АП velutina, one dwarf rosette.
30 velutina rosettes.
1910. (0. biennis x 0. Lamarckiana) velutina x 0. 4 laeta rosettes.
marckiana| | 40 5 Lamarckiana rosettes.
1 mutant rosette.
1911. (0. biennis x О, Lamarckiana) velutina, Е,...... 2 2 velutina.
; 10 velutina rosettes.
1911. | ( » X » ) velutina x О. La- | 14 | 9 laeta rosettes.
marckiana 2 Lamarckiana rosettes.
1909. | O, biennis x О. levifolia, Ё,.................. 4 Б [и
: ES 2 laeta, ” » » ”
1910. | ( pee и учета, И 12 19 velutina.
но s Mette; ҚОСА 18 F laeta.
1 velutina, flowers intermediate.
1910. | O. biennis x О. rubricalyz, Е,................ 34 laeta and velutina, much variation. :
1 1908. | О. nanella x О. biennis, Е, .................. 7 One type, same as О. Lamarckiana х О. biennis. |
1910. = x % нра 9 > > ваше as last.
| 1909. $5 x % Е, Аа енто curtas 5 ” ”
1910. ii x . № а 13 „ ваше аз ‘last generation.
60 laeta, broad-leaved, somewhat variable type.
| 1910. i x a Foa ‚г... 69 | 1 velutina (?), narrow-leaved rosette.
| u 1 dwarf, with O. biennis flowers.
чы, ы. UM EE ame) 16 | One type, samo as О. nanlla х 0. bien
быш 0. nanella x 0. omen. EQ басымы арық; 18 „ breeding true.
he 98 velutina.
1910. 0. biennis x O. amd, Е, же ашлыгын dim s. ui 7 laeta.
ch ы C 6 dwarfs.
= | С 2 aberrants.

KNOWLEDGE OF THE MUTATING OENOTHERAS. 57

their parents as the latter are from each other, and the flowers remain intermediate.
However, there is the difference that (О. rubrinervis x Lamarchiana) x (Lamarckiana
Х rubrinervis) presumably gives both О. rubrinervis and О. Lamarckiana, while
(О. Lamarckiana X biennis) х (biennis x Lamarckiana) gives, if the result of DeVries
is invariable (1911, р. 100), only О. Lamarckiana. But, as already stated, there is some
reason to doubt the constancy of the latter result.

I have already expressed my agreement with Honing that O. Lamarckiana and
О. rubrinervis differ from each other in the same way that the laeta and velutina hybrids
differ—4. e., in width of leaves, crinkling, hairiness, ete. This point has now been
adequately proved by the careful studies of Honing. But, I have further shown in the
present paper (p. 40) that О. muricata in the wild condition also frequently contains
two types whieh differ from each other in similar ways. Тһе appearance of both (twin)
types in the F, of many crosses is probably a result of this dimorphic condition, which
is therefore a fundamental one in this section of the genus. It remains for further
studies to diselose the nature of the relation between the two types as regards origin,
but it seems not impossible that they represent alternative conditions, either of which
may come into expression under certain circumstances.

Certain results of Miss Saunders with Stocks (1908) and Petunias (1910) also
indieate that races may breed true and yet be heterozygous, and it is probable that this
is the explanation of the behaviour in such crosses as О. rubrinervis x О. Lamarckiana,
although that is not yet proven.

From one point of view, especially regarding the origin of new characters, one of the
most important features of heredity in Oenothera is that both blended and sharply
alternative inheritance occur. The cases in Oenothera of blends which breed true іп
specific crosses are now well established. Му cultures of Lancashire Oenotheras have
also yielded several true-breeding races whose characters are evidently a blend of
previously existing races. Certain crosses between these and other forms give the
impression of an absolute blending of characters between races in the F,. Тһе behaviour
of later generations of these hybrids is not yet known. It remains to determine whether,
with repeated crossing back, blending continues to occur until the differences are
obliterated. Тһе reason for intermediate or blended races breeding true is not yet
explained, but from the behaviour of certain of the меш and velutina types (DeVries,
1908) it is possible that the failure of segregation may here also be due to a difference
between the characters borne by the male and female germ-cells. Оп the other hand, it
is not improbable that there is in such cases а real blend without segregation of the
materials representing the germinal characters. It seems probable that when real
blends occur, they will be found chiefly in characters which have a quantitative basis,
yet certain apparent cases of this kind, such аз О, rubricalyz, show sharp segregation
phenomena.

Through the maze of these complicated types of behaviour, it is beginning to appear
evident that there are certain connections between the systematic relationships of а
form and its hereditary behaviour, and crossing therefore is one of the most fundamental

SECOND SERIES.—BOTAN Y, VOL. VIII. ї

58 DR. В. В. GATES—CONTRIBUTION TO А

methods of analyzing those relationships. It is clear that forms which originate from
the same parent do not all have the same hereditary behaviour towards that parent.
Hence the relationship of their characters to the parent form is different in different
eases, and, therefore, the change implied in their method of origin has been different.
The hereditary behaviour of the mutants of Oenothera is also usually different from that
of the related wild species. It seems justifiable to conclude that not only in Mendelian,
but in other types of hereditary behaviour, the method or type of inheritance is deter-
mined in large measure by the nature of the characters themselves.

Our knowledge of heredity іп Oenothera, therefore, shows that several types of
inheritance exist, having an evident connection with the systematic relationships of the
forms, or, in other words, with the nature of the differentiating characters of those forms.
Blending and segregating inheritance hoth occur in Oenothera, and it is not yet clear
in how far the former may be explained in terms of the latter.

УП. GENERAL CONSIDERATIONS AND SUMMARY.

From the foregoing pages it will appear that the explanation of the mutation pheno-
mena in Oenothera Lamarckiana is by no means simple, and a full understanding of
the nature and cause of this behaviour has not yet been reached. A number of facts,
however, lead to the belief that the previous crossing which О. Lamarckiana has under-
gone, either in the wild condition or in Botanical Gardens, or both, has been the chief

cause of the germinal disturbances which make themselves manifest in the appearance |

of mutants. The exact nature of these disturbances is not at present clear. Cytological
studies, as well as the hereditary behaviour, show that the various mutational changes
are of fundamentally different kinds, involving germinal alterations which are not only
diverse among themselves but even appear in some cases to originate at different points
in the life cycle. Thus, іп O. gigas the essential change appears to have taken place іп
the megaspore mother-cell or in the early divisions of the fertilized egg, and possibly also
during the reduction divisions in the pollen mother-cell; while the hereditary behaviour
of the original О. rubricalyx mutant indicates that it was the result of a cross between
a mutated and a non- mutated germ-cell. On the other hand, it is not easy to explain the
behaviour of such mutants as O. rubrinervis and O. nanella on the theory of their hybrid
or heterozygous origin, without involving a contradiction. О. rubricalyz gives in each
generation a certain percentage of reversions to the parent form, O. rubrinervis, until à
homozygous race is obtained, but in most of the other mutants no reversions occur.

It seems entirely probable, as DeVries believes, that the cause of the mutations is

ап internal one; and yet it cannot at present be decided whether this is due, for instance, |

to irregularities in chromosome distributions or to the action of slight releasing stimuli.

In any ease, the capacity for producing aberrant types is а result of germinal instability |
probably occasioned by previous crossing, the phenomena accompanying meiosis in |
|. the mutants resembling those found in hybrids in various particulars. It is evident |
_ that any explanation of mutation must take into account the peculiar chromosome |

Has.

KNOWLEDGE ОЕ THE MUTATING OENOTHERAS. 59

It by no means follows, however, that mutations appear only after a mixture of germ-
plasms, and it is necessary to distinguish two types of mutations: (1) those following
an ancestral mixture of germ-plasms, and (2) those oceurring in pure lines.

The numerous and rapidly-accumulating cases of mutations, not only in Oenothera
but in many other genera of plants and animals, make it impossible to ignore this
factor as a method of species-formation, whatever its evolutionary significance may be.
But series of chance-wise mutations could scarcely result in the orderly phylogenics we
often see in the paleontological record ; nor are they adequate to account for the more
complex cases of adaptation, especially those involving inter-relationships between
various organisms, though they may be capable of explaining cases of “ climatic
adaptation " in various species of a genus.

Darwin, in his theory of natural selection, also usually assumed a change in the
climatic or biological environment of a species, which led ultimately to its modification.
But the effects of climatic and distributional vicissitudes in the earth’s history are again
scarcely competent to explain the occurrence of long, orderly sequences of forms. Тһе
only effect of natural selection which is experimentally proven is its conservative effect
in maintaining the species in its present equilibrium and at the maximum of its
efficiency in the struggle for existence. That natural selection can produce a modifica-
tional effect is, like the inheritance of acquired characters, at present experimentally
unproven.

A conspectus is given, showing roughly the relationships of the Oenothera species
belonging to the group which contains the mutating forms.

In Section III. are summarized my studies on the present distribution of the mutating
Oenotheras, and the history of their introduction from North America into European
Botanic Gardens. They have now become naturalized in nearly all regions of the world.
The early introductions into Europe can be identified in many cases by means of figures,
descriptions, and specimens. Certain of these races appear to have resembled closely
O. Lamarckiana, but no early record can be positively identified with this form to the
exclusion of О. grandiflora от (in one case) О. biennis.

In Section IV. the mutation phenomena іп О. Lamarckiana are considered in detail,
the fact being emphasized and proven by illustrations that, although the ranges of
variation of the various mutants frequently overlap in various characters, yet their
modal conditions are different for each mutant, and the mutants are therefore clearly
discontinuous in their origin from the parent form. When the mutants from О. La-
marckiana appear, their characters are fully developed in the first individual, and there
is no evidence that subsequent selection from the offspring of a mutant has any effect in
intensifying its characters.

Several of the mutant types, notably О. gigas and О. nanella, show a surprising
range of variability, particularly in leaf-characters. The Oenotheras are also
very sensitive to environmental changes, and the environment is a very important
element in interpreting the wide range of fluctuation often observed in physiological

: | 12

60 DR. В. В. GATES—CONTRIBUTION ТО A

as well as structural characters. Particular races of Oenothera grown under different
conditions of culture may be so unlike as scarcely to be recognizable, or indeed they
may be wholly unrecognizable if the conditions of their cultivation are very diverse.
These facts serve to emphasize the point of view that heredity and variation represent
the two sides of a single process, namely the interaction of the organism with its
environment in every stage of ontogeny.

Several races of О. Lamarckiana from various sources show constant differences
from the form of DeVries’s cultures, though some of them at least produce mutants
analogous to those of DeVries. The differences between these O. Lamarckiana
races are largely in the foliage, and they appear to be chiefly of a quantitative sort.
The race of DeVries itself shows a considerable range of fluctuation, particularly in
the crinkling of the leaves and the method of branching. O. Lamarckiana is thus a
polymorphic species, containing several elementary species.

It is shown that in О. Lamarckiana and nearly all the mutants, three distinct leaf-types
succeed each other in the ontogeny of the rosettes. This necessitates great care in
comparing only corresponding stages of the various forms, a precaution which is
especially urgent in comparing hybrids with their parent types. The mutant O. gigas
usually does not show this succession of leaf-types, the young rosettes merely showing a
succession of leaves of the same type but of increasing size. One race, however,
received from the Botanic Garden at Palermo under the name O. cognata, whose adult
rosette was identical with O. gigas, in my cultures of 1911 passed through stages
apparently analogous to those of the other mutants.

One remarkable feature of the Oenothera mutants is the way in which the differing
characters of leaf, stem, and flower remain together and refuse to be separated and
redistributed by crossing as in Mendelian hybrids. The differences are not in single
characters but affect every part of the plant, most of the former showing recognizable
peculiarities from a very early stage right through their ontogeny. But many of the
forms are more easily recognizable in certain stages than in others. Thus О. brevistylis
shows marked peculiarities in the young seedlings, the mature rosettes are difficult to
distinguish from O. Lamarckiana, but the adult plants are easily recognized by their
bracts, sepals, and styles.

O. grandiflora, like the O. Lamarckiana forms, passes through several distinct
stages in its rosette ontogeny, but under ordinary conditions of culture usually omits
entirely the very characteristic adult stage of its rosette, and not infrequently the
rosette stage is omitted altogether under the same conditions of culture in which it is
produced in the O. Lamarckiana forms.

A large number of geographic races of O. biennis from North America have been grown. .
They differ chiefly in leaf-characters, but their behaviour in crossing is not yet known.

My cultures of O. muricata from many sources show that, especially in Canadian
forms, a dimorphism exists which is comparable with the twin hybrids first observed by
_ DeVries in crosses between О. biennis and O. Lamarckiana, These twin forms are |
_ respectively broad- and narrow-leaved, and they were observed in cultures from wild `
seeds collected in Nova Scotia, New Brunswick, and Manitoba. This indicates that the |

KNOWLEDGE OF THE MUTATING OENOTHERAS. 61

explanation of the twin hybrids of DeVries is to be found in a fundamental alterna-
tive dimorphic condition present in many of the wild forms. It is further probable, as
Honing thinks, that O. Lamarckiana and O. rubrinervis are twin types corresponding
to the broad- and narrow-leaved forms,

Several years’ cultures from a colony of Oenotheras naturalized on the coast of
Lancashire near Birkenhead has shown that not only O. Lamarckiana and several of
its mutants (as previously known), but also O. grandiflora and a host of other races
previously unknown, occur in this locality. Many of the latter races have been isolated
and several of them are found to breed true. Their bud characters are usually more or
less intermediate between О. Lamarckiana and О. grandiflora, which indicates their
hybrid origin. But they show the important fact that in Oenothera, hybrid races
frequently form blends which breed true. Some of these races exhibit quite unex-
pected characters, and it remains to be determined whether a mutation factor in addition
to hybridization is necessary to account for some of them—~. e., whether, as in the case
of O. Lamarckiana and its mutants, the characters of some of the races can not have
originated merely through hybrid blending and splitting.

Numerous cultures of Oenotheras from Botanic Gardens, as well as TN from other
sources, have shown that the line of distinction between the large-flowered, long-styled
and the small-flowered, short-styled forms is by no means so sharp as was formerly
supposed. In certain large-flowered races which have undergone crossing the style may
range from long to short in different flowers of the same plant. Wild species are also
known having large flowers and short style, and others having small flowers with long
style. There are in reality not merely two but a number of different lengths of style,
ranging from long, as in О. grandiflora or О. Lamarckiana, to very short, as in
О. brevistylis, And in each species or race there is usually a narrow range of fluctua-
tion in length, which occasionally (apparently always in forms that have been crossed)
becomes а remarkably wide range.

Oenothera hybrids show a variety of types of behaviour, depending upon the relation-
ships of the forms concerned. Both blending and segregating types of inheritance occur,
the former usually in interspecific crosses (though О. gigas shows this type of behaviour
in some cases), and the latter usually in crosses between the mutants. Occasional
mutants—e. g., О. brevistylis and О. rubricalyz—behave in typical Mendelian fashion.
The twin hybrids, first described by DeVries and here illustrated for the first time,
are produced when О. biennis and certain other species are pollinated by О. La-
marckiana or one of its derivatives. They are respectively broad- and narrow-leaved, and
are called ¿aeta and velutina. Тһе velutina type crossed back with Lamarckiana gives
velutina, laeta, and Lamarckiana in about the proportion 6: 1:1. The twin types are,
so far as known, intermediate between the parents in flower-characters, but divergent
from either parent in leaf-characters. "When selt-pollinated they breed true, though it
has been shown by DeVries that they produce two kinds of pollen-grains or egg-cells.

This type of inheritance, in which new, blended characters arise and breed true, occurs
also in numerous interspecific and interracial crosses in Oenothera, and is probably of
fundamental significance in connection with the origin of new races.

62 DR. В. В. GATES—CONTRIBUTION TO A

It is held that there is no fundamental distinction between mutations and fluctuations,
except that one is inherited and the other is not—or, in other words, one is due to
a germinal alteration, the other to a somatic variation.

VIII. BIBLIOGRAPHY.
(This list includes only papers actually cited in the text.)

Anpxrws, Е. M.—1910. Twin Hybrids (laeta and velutina) and their Anatomical Distinctions. Bot.
Gazette, І. 193-201.

Автом, С.--1911. Та sistematica del genere Artemia in relazione col numero dei cromosomi delle
cellule sessuali e in relazione col numero e colla grandezza delle cellule somatiche. Biolog.
Centralbl. xxxi. 104—108.

BAILEY, CuanLEs.—1907. De Lamarck's Evening Primrose (Oenothera Lamarckiana) on the sandhills
of St. Anne’s-on-Sea, North Lancashire. Address at Annual Meeting, Manchester Field-Club,
pp. 28, pls. 6.

BanrLETT, Н. H.—1907. Some new Washington Plants. Bot. Gazette, xliv. 302-304.

1911. Systematic Studies on Oenothera. I, Oenothera Tracy, sp.nov. Rhodora, xiii. 209-211,

pl. 93.

Bateson, W., Е. В. Saunpers, & В. C. Рохметт.—1908. Experimental Studies in the Physiology of
Heredity. Report IV. to the Evolution Committee of the Royal Society, pp. 40. London:
Harrison & Sons.

BourENGER, G. A.—1907. On the Variations of the Evening Primrose (Oenothera biennis, Linn.).
Journ. Botany, xlv. 853-363.

Boveri, Тн.—1905. Zellenstudien: V. Ueber die Abhängigkeit der Kerngrüsse und Zellenzahl der
Seeigel-Larven von der Chromosomenzahl der Ausgangszellen. Jena: Gustav Fischer; pp. 80,
pls. 2, text-figs. 7.

Davis, B. M.—1910. Notes on the Behaviour of certain Hybrids of Oenothera in the First Generation.
Amer. Nat. xliv. 108-115. ;

—— 1911. Some Hybrids of Oenothera biennis and O. grandiflora that resemble O. Lamarckiana.
Amer. Nat. xlv. 193-233. E

DeVries, Hvoo.—1901-3. Die Mutationstheorie. Leipzig. 2 Vols.

1907. On Twin Hybrids. Bot. Gazette, xliv. 401—407.

—— 1908. Ueber die Zwillingsbastarde von Oenothera nanella. Ber. deut. Bot. Ges. ххуі. а
667-676.

—— 1909. Bastarde von Oenothera gigas. Ber. deut. Bot. Ges. l. c. рр. 754-762.

—— 1909. On Triple Hybrids. Bot. Gazette, xlvii. 1-8.

—— 1909. The Mutation Theory. Translated by Prof. J. B. Farmer and A. D. Darbishire. Chicago :
Open Court Pub. Co.

--- 1911. Ueber doppeltreziproke Bastarde von Oenothera biennis, Linn., und О. muricata, Linn.
Biolog. Centralbl. xxxi. 97-104.
Diesy, L.—1912. The Cytology of Primula kewensis and of other related Primula Hybrids. Annals

... of Botany, xxvi. 357-388, pls. 41-44, figs. 2.
GarEs, В. R.—1907. Pollen Development in Hybrids of Oenothera lata x O. Lamarckiana, and its

x relation to Mutation. Bot. Gazette, xliii. 81-115, pls. 2-4.

_— 1908. А Study of Reduction in Oenothera rubrinervis. Bot. Gazette, xlvi. 1-34, pls. 1-3. p.

—— 1909 а. The Behaviour of the Chromosomes in Oenothera lata x О. gigas. Bot. Gazette, xlviii. |

E 179-199, pls. 12-14, 0

KNOWLEDGE OF THE MUTATING OENOTHERAS. 63

Gares, В. R.—1909 b. Apogamy in Oenothera. Science, n. s. xxx. 691-694.

1909c. An Analytical Key to some of the Segregates of Oenothera. 20th Rep. Missouri

Botanical Garden, 123-137.

--- 19094. The Stature and Chromosomes of Oenothera gigas, DeVries. Arch. f. Zellforsch. iii.
525-552.

--- 1910а. The Chromosomes of Oenothera Mutants and Hybrids. Reprint, Proc. 7th Internat.
Zool. Congress, 1907, pp. 4.

— 19106. The material basis of Mendelian Phenomena. Amer. Nat. xliv. 203-218.

—— 1910c. Abnormalities in Oenothera. 21st Rep. Mo. Bot. Garden, xxi. 175-184, pls. 29-81.

--- 1911 а. Early Historico-botanical Records of the Oenotheras. Proc. Iowa Acad. Sci. 1910, 85-
124, pls. 6.

--- 19110. Studies on the Variability and Heritability of Pigmentation in Oenothera. Zeitschr. f.
ind. Abst- und Vererbungslehre, iv. 337-372, pl. 6.

— 1911 с. Pollen Formation in Oenothera gigas. Annals of Botany, xxv. 909-940, pls. 67-70.

--- 19114. Mutation in Oenothera. Amer. Nat. xlv. 577-606.

1919. Ап Onagraceous Stem without Internodes. New Phytol. xi. 50-53, pls. 2-3.

СлоснЕву, P.—1899. Recherches sur le nanisme végétal. Ann. des Sci. nat., Bot. 8° Série, ix. 61-
156, pl. 4, figs. 32,

GEERTS, J. M.—1909. Beiträge zur Kenntnis der Cytologie und der partiellen Sterilitiit von Oenothera
Lamarckiana. Recueil des Travaux Bot. Néerlandais, v. 93-208, pls. 5-22. |

--- 1911. Cytologische Untersuchungen einiger Bastarde von Oenothera gigas. Ber. deutsch. Bot.
Ges. xxix. 160-166, pl. 8.

Спемо-Тоз, Екмаххо.--1908. L/eredita e le leggi razionali dell’ ibridismo. Biologica, vol. ii. no. 10,
pp. 1-36.

| ——- 1911. Les dernières expériences du Prof. DeVries et léclatante confirmation de mes lois
rationelles de l'hybridisme. Biol. Centralbl. xxxi. 417-425.

Hitt, A. W.—1912. The History of Primula obconica, Hance, under Cultivation, with some remarks on
the History of Primula sinensis, Sab. Journ. of Genetics, ii. 1-20, pls. 1-2.

Номімв, J. A.—1911. Die Doppelnatur der Oenothera Lamarckiana. Zeitschr. f. ind. Abst- und
Vererbungslehre, iv. 227-978, figs. 10. ;

Jennines, Н. S.—1911. Assortative Mating, Variability, and Inheritance of Size, in the Conjugation
of Paramecium. Journ. Exp. Zoology, xi. 1-134.

JoHANNSEN, W.—1909. Elemente der exakten Erblichkeitslehre. Jena: G. Fischer ; pp. 517.

Песіккс pu SABLoN, Maruigv.—1910. Sur la théorie des mutations périodiques. Comptes Rendus,
cli. 830-332.

Lutz, ANNE M.—1909. Notes on the First Generation Hybrid of Oenothera lata 9 x O. gigas d.

Science, n. s. xxix. 263-207.

1912. Triploid Mutants in Oenothera. Biol. Centralbl. xxxii. 385—435, figs. 7.

MacDoveat, D. T:—1903. Mutations in Plants. Amer. Nat. xxxvii. 737—770, figs. 10.

MacDovear, Ули. ӛнгіл, & Эмлил.— 1905. Mutants and Hybrids of the Oenotheras. Carnegie
Institution Pub. no. 24, pp. 57, pls. 22.

MacDoveat, Ули, & ӛнгіл,-1907. Mutations, Variations, and Relationships of the Oenotheras.
Carnegie Institution Pub. no. 81, pp. 92, pls. 22, figs. 73.

Macovun, Jonn.—1883. Catalogue of Canadian Plants. Montreal.

MARCHAL, Ex. & Ём.—1909. Aposporie et sexualité chez les Mousses. II. Bull. de PAcad. roy. de
Belgique, 1909: 1249-1288.

— 1911. Aposporie et sexualité chez les Mousses. ПІ. Bull. de Acad, roy. de Belgique, 1911:
750—778, pl. 1.

64 DR. R. R. GATES—CONTRIBUTION TO A

NELsON, Aven.—1911. Contributions from the Rocky Mountain Herbarium. IX. New Plants from
Idaho. Bot. Gazette, lii. 261-274.

NirssoN, М. НЕківЕвт-.-1909. Oenothera gigas, framgången som mutation i Sverige. Bot. Notiser,
1909, 97-99.

---- 1912. Die Variabilitit der Oenothera Lamarckiana und das Problem der Mutation. Zeitschr.
f. Abst- und Vererbungslehre, 8 : 89-231, pls. 3-5, figs. 36.

SavxDERS, Ерітн R.—1910. Studies in the Inheritance of Doubleness in Flowers. I. Petunia. Journ.
of Genetics, 1. 57-69, figs. 5.

ScHourEN, А. R.—1908. Mutabiliteit en Variabiliteit. Dissertation: Groningen ; pp. 196.

Этомрз, Turo. J.—1912. Die Entstehung von Oenothera gigas, DeVries. Ber. deutsch. Bot. Ges. ххх.
406-416.

ScRAsBURGER, E.—1910. Chromosomenzahl. Flora, с. 398-446, pl. 6.

Тізсні,ек, G.—1908. Zellstudien an sterilen Bastardpflanzen. Arch. f. Zellforsch. i. 33-151.

1910. Untersuchungen über die Entwicklung des Bananen-Pollens. I. Arch. f. Zellforsch. v.

622-670, pls. 80—81, figs. 4.

Varl, Anna M.—1905. Onagra grandiflora (Ait.), a Species to be included in the North American
Flora. Torreya, у. 9-10

ZEILSTRA, H. Н. Fzw.—1911. Oenothera nanella, DeVries, eine krankhafte Pflanzenart. Biol.
Centralbl. xxxi. 129-138, figs. 5.

--- 1911. On the Cause of Dimorphism in Oenothera nanella. Kon. Akad. Wetens. Amsterdam,
1911: 680—685, figs. 3.

IX. EXPLANATION OF THE PLATES.

Thé figures representing whole mature plants, or shoots in flower or fruit, were reproduced without
reduction from the size of the original photographs. АП others were reduced one-third. Тһе
buds were all originally photographed natural size, and are therefore reduced to two-thirds
natural size in reproduction.

PLATE 1.

Fig. 1. O. Lamarckiana, young seedlings.
2, 8. О. Lamarckiana, young rosettes.
4. O. Lamarckiana, older rosette, showing three leaf-types (see text).
5. O. Lamarckiana, mature rosette.
6. O. Lamarckiana, selected leaves from mature rosettes, showing range of fluctuation.
7. O. Lamarckiana, adult plant. An extreme variant having very much crinkled leaves.
8. The same, another individual. Terminal shoot in bloom, upper stem-leaves practically without
crinkles.
9. Young seedlings of another race of O. Lamarckiana found in St. Louis, Mo.
10. Mature rosette of St. Louis race of- O. Lamarckiana. Cf. figs. 4 and 5 showing the
O. Lamarckiana of H. DeVries.
11. Adult plant of same race as last. Cf. figs. 7 and 8.
12. A rosette from a culture of “ O. Lamarckiana var. cruciata," from the Bremen Botanic
Garden. The mesophyll of the leaf develops on only one side of the midrib, the other
half curving around to form a sickle-shaped blade. E
13, 14. Developing rosettes of О. laevifolia. The figures represent the same plant taken at an | t
interval of 8 days, to show how new leaves appear. They show the three types of |
rosette-leat described in text. — — U

KNOWLEDGE OF THE MUTATING OENOTHERAS. 65

PLATER 2.

Fig. 14. See lost Plate, No. 13.

15. 0. laevifolia, young seedlings. СУ. figs. 1, 21, 25, 83.

16-17. Rosettes of O./aevifolia. The latter is typical, while the former is much crinkled aud
scarcely distinguishable from О. Lamarckiana in a photograph. Cf. figs.4 and 5. Fig. 16
also shows the three ontogenetic leaf-types.

18. О. laevifolia, selected leaves from mature rosettes, showing range of variation in leaf-width
and стіп пе. Cf. figs. 6, 23, and 33.

19. Typical flowering-shoot of О. laevifolia. Leaves narrower, smoother, and more furrow-shaped
than O. Lamarckiana. Cf. figs. 7 and 8.

20. О. laevifolia, shoot in fruit. Cf. figs. 31, 52, 77, and 80.

21. О. brevistylis, young seedling. Cf. fig. 15, О. laevifolia, same age; also figs. 1 and 25.

22. Mature rosette, О. brevistylis. Each such rosette contains some leaves with broad, rounded
tips. Cf. fig. 5 (O. Lamarckiana).

23. Selected leaves from mature О. brevistylis rosettes, showing range of variation. Cf. fig. 6
(O. Lamarckiana) and fig. 18 (O. laevifolia).

24. Buds of O. brevistylis, two-thirds natural size, sepal tips very short.

95-96. О. rubricalya.

25. Young seedlings. Cf. fig. 1 (О. Lamarckiana).

26. Typical rosette of О. rubricalyx. Су. fig. 4.

27. O. rubrinervis rosette, leaf-tips broader than typical.

PLATE 3.

. 28. “ Lateral rosette” from a plant of О. rubrinervis whose stem was bent horizontally and

afterwards developed a number of rosettes in place of lateral branches.

29. Flowering shoot of О. rubrinervis, from О. rubrinervis x О. Lamarckiana.

30. О. rubricalyx, adult plant, grown in heavy clay soil and devoid of lateral branches.

31. Fruiting shoot of same, showing capsules and bracts.

32. O. gigas, typical young rosette, showing nearly orbicular thick leaf-blades.

33. O. gigas, leaves from mature rosettes, showing range of variation in a culture grown in 1909.

34. Rosette of O. gigas, broad-leaved type.

35. Rosette of O. gigas, an extreme variant with very narrow, almost linear leaves.

36-37. O. gigas, narrow-leaved rosettes.

38. 0. gigas, buds, two-thirds natural size, much larger than those of other mutants. Cf. figs. 24
41, 43, 49, and 53.

39. O. gigas nanella, a dwarf form of O. gigas.

40. О. cognata, Palermo Botanic Garden. Earlier rosette-leaves with long petioles, later ones
identical with O. gigas. An independent mutation.

41. О. папеПа, buds, two-thirds natural size, with bract attached to one.

42. 0. lata, typical rosette from the race of DeVries.

43. Buds, two-thirds natural size.

Buds 1-8 (on the left), O. lata of DeVries, which produces almost no good pollen.
Buds 4-7 (on the right), from a race of O. lata or О. semilata from Lancashire, producing

pollen. The sterility of the pollen is о by crumpling of the petals and
consequent distortion of the buds.

SECOND SERIES.—BOTANY, VOL. VIII. K

66

Fig. 44.

Fig. 57.

DR. В. В. GATES— CONTRIBUTION ТО A

PLATE 4.
Rosette of О. зет аға от О. lata from Lancashire.

. Lancashire Oenotheras, а race of О. grandiflora from Lancashire. Rosette in tropical hot-

house, age 4 months.

. Same as last, 6 weeks later. A new type of rosette-leaf has appeared, which is characteristic

of O. grandiflora, but is usually omitted altogether in ordinary cultures of this species.

. O. grandiflora race from Lancashire. Mature rosette in hothouse, age 9 months.

Cf. fig. 46.

Adult plant of O. grandiflora from Alabama, under ordinary conditions of culture at Mo. Bot.
Garden. It produces no rosette but branches freely, giving the plant a shrubby
appearance.

. Buds, two-thirds natural size, from a Lancashire race of O. grandiflora. They are smooth and

free from hairs, rounded, with thin papery sepals. Cf. fig. 24.

Rosette of a race of O. biennis from New York State; photographed from a culture at Woods
Hole, Mass. This pure race was used іп my hybrids with О. Lamarckiana and other
forms.

. Flowering shoot of the same.

Shoot of the same in fruit, showing capsules and bracts.
Buds of the same, two-thirds natural size, showing diminution in size of bracts.
Broad-leaved rosette of O. muricata from Nova Scotia.

. Narrow-leaved rosette of О. muricata from Winnipeg, Manitoba. Both these races occur

together with identical characters in both these localities. See text, p. 40.

. Flowering stage of broad-leaved type of O. muricata from Winnipeg. It will be seen that the

rosette is identical with fig. 54, but reduced in size in reproduction.

PLATE 5.
Very young rosette of О. ammophila, which, in this stage at least, is identical with certain
narrow-leaved races of О. muricata, such as represented in the next figure.

. Rosette of a uniform race of О, muricata from Seal Harbour, Maine.

Lancashire Oenotheras. Rosette No. 5 from culture in tropical hothouse; between O. lata
and О. semilata, age 9 months. This plant was fertile. See text.

. Lancashire Oenotheras, No. 8. Rosette in tropical hothouse, age 91 months.
. Same culture as last, Хо. 18, age 4 months.
. Same culture as last, Хо. 22, age 5 months, a dwarf rosette with characteristic type

of leaf.
Lancashire Oenotheras, No. 23. Rosette in tropical hothouse, age 4 months. Cf. fig. 45,
from which it differs in having longer petioles and narrower blades.

. Lancashire Oenotheras, No. 25. Rosette in tropical hothouse, age 91 months. This has

since given rise to a characteristic race. (Cf. Gates, 19112, p. 350.)

. Same culture as last, No. 33, age 92 months.
. Type 2 in Ё, from No. 5 of Lancashire Oenotheras. The rosette closely resembles DeVries’s

О. elliptica. See р. 45.

. Туре 4 in Е, from No. 5 of Lancashire Oenotheras. Adult plant, showing small flowers with

short style. Sce р. 45. x

A pan of seed'ings showing the F, from О. lata x O.nanella. The forms represented are
О. Lamarckiana, О. rubrinervis, О. nanella, and an aberrant (colourless) type. See text,
р. 50. |

69. Rosette of O. biennis x O. Lamarckiana, F,, broad-leaved or laeta form.

70.

Rosette of O. biennis х O. Lamarckiana, F,, narrow-leaved ог velutina form.

KNOWLEDGE ОҒ THE MUTATING OENOTHERAS. 67

PLATE 6.

Fig. 71. Rosette of O. biennis x О. laevifolia, Fi, broad-leaved type. Cf. fig. 69.

72. Rosette of О. biennis x О. laevifolia, F,, narrow-leaved type. СХ. fig. 70.

79. Intermediate type from О. biennis x О. laevifolia, Е,, another cross.

74. Flowering shoot of О. biennis х О. Lamarckiana, F,, velutina type.

75. Flowering shoot of О. biennis x О. Lamarckiana, Fi, laeta type.

76. Buds, two-thirds natural size, from О. biennis x О. Lamarckiana, F, (velutina). Cf. bracts
and buds with fig. 53 (O. biennis). The buds are larger and more hairy, the bracts more
nearly entire.

77. Shoot showing flowers, capsules, and bracts from О, biennis x О. laevifolia.

78. Rosette of О. Lamarckiana x О. biennis, Еу, broad-leaved type. Cf. fig. 69.

79. Rosette of О. nanella x О. biennis, Fy.

80. О. Lamarckiana x О. biennis, shoot in fruit showing capsules and bracts. Cf. figs. 77 and 31.

81. Pan of seedlings of О. biennis X О. nanella, Е. Cf. figs. 82 and 83, same age.

82. Pan of seedlings of О. nanella x O. biennis, F,.

. 88. О, nanella, pan of young seedlings. СУ. figs. 81 and 82, same age.

Note added January 2nd, 1913.

THE recent important paper of N. Heribert-Nilsson (1912) has shown that in an
independent race of O. Lamarckiana obtained from a garden in Southern Sweden, which
differs in some respects from the race of DeVries, mutations are produced which are
parallel to, though differing from, the DeVriesian mutants. Thus it is further shown
that О. Lamarckiana, like ordinary wild species, contains various elementary species,
and new difficulties arise before those (e. g., B. M. Davis) who would derive all the races
of О. Lamarckiana from a single (hypothetical) source.

Nilsson also expresses the view, with which I am in agreement, that from an
evolutionary standpoint it is immaterial just what forms may have taken part in the
complex ancestry of O. Lomarckiana.

Тһе valuable data of Nilsson's paper, and particularly his description of the behaviour
of a new giant type, are of great interest. His explanation of the mutation phenomena
without regard to the cytological facts has, however, led him into theoretical errors
which might otherwise have been avoided, but these in no way detract from the useful-
ness of the mass of breeding data contributed. I shall refer in greater detail to these
giant forms in a paper now in the press in the ‘ Biologisches Centralblatt.

| В. В. С.

TRANS. Linn. Soc., SER 2. Bor. Vo . VIII. Pr. 1.

В. R. Gates, phot.

| Grout, sc. et imp.
MUTANTS OENOTHERA.

Pr 2

VoL. УШ

Trans. Linn. Soc., SER 2. Вот

GATES.

MUTANTS Í OENOTHERA

5
а
а
LÀ
v
Е.
о
x
x

TRANS. Linn. Soc., SER. 2. Вот. Vor. VIII. PL. 3.
GATES.

Grout, sc. et imp
R. R. Gates, phot.

E Trans. LINN. Soc., SER. 2. Bor. VoL. VIII. PL. 4.
GATES. ;

Grout, sc. et imp
R. R. Gates, phot

Trans. Linn. Soc., SER. 2. Вот. Vor. VIII. PL. 5.
GATES.

68

| Grout, вс. et imp
MUTANTS 97 NOTHERA.,
R. R. Gates, phot. |

<
САТЕЗ.

TRANs. Linn. Soc., SER 2. Bor. Vor. VIII. PL. 6.

T x Сгош, өс. et imp.
NTS Š ;
R. R. Gates, phot. M U + A 1

LINNEAN SOCIETY OF LONDON.

ae MEMORANDA CONCERNING TRANSACTIONS

- The First Series of pere Transactions, containing both Botanical and Zoological ee es been n complet f

in 30 Vols. and a few entire sets are still for sale. Only certain single volumes, or part omplete sets, m:

EM obtained at the o riginal priees. The к of the Index to Vols. 1-25 is 8s. to the publie, oe Bs. to Fellows;

22. Vols. 26-30, 4s. to the public, and 3s. to Fel р

— Тһе Second Series of the Transactions із divided into Zoological and Botanical sections, The prices es of
Botanical parts of these which have been paci: аге as undermentioned. (For the Zoological parts see Zoolo

Не)

Secoxp Surtes.—Borany (continued).

Secoxp SrRIES.— Borat,

x МАСА - When Price to the:
b Volume. Published. Publie.
Co.» 5

бы ee 2 VL Pet Lo 100L... D10 0
Part IL 19001..... 0 6 0

"Pat Hi 1002... 018 0...
"Part IV. 1908. .... зо S
“Баб У. 1908..... 0 8
Part ҮІ. 1903. о

2nd Ser. ВОТАМУ.] (VOL. VIII PART 2.
THE

TRANSACTIONS ©

1. ON SOME INDIAN JURASSIC GYMNOSPE

AND

9. RHEXOXYLON AFRICANUM, À

EW MEDULLOS

[ 69 ]

II. On some Indian Jurassic Gymnosperms. Ву NELLIE BANCROFT, B.Sc.(Lond.),
F.L.S., 1851 Research Scholar, Newnham College, Cambridge.

(Plates 7-9.)
Read 21st November, 1912.
I. INTRODUCTION.

THE fossil plants under consideration are contained in four siliceous blocks sent by
Mr. Hayden, Director of the Geological Survey of India, at the request of Professor
Seward %, The strata from which the blocks were obtained belong to the Rajmahal
Series, a division of the Upper Gondwana Group, and are probably of approximately
the same age as the Lower Jurassic (Liassic) rocks of Europe (4, 18, 15). Тһе Rajmahal
Series of strata is best developed in Bengal, north of Calcutta, and takes its name from
the Rajmahal Hills, where it was first recognised. It also occurs in the Southern Godavri
district and in the south and south-west of Madras. The fossil plants now to be described
came from the Rajmahal Hills, the exact locality being Amrapara, on the Bansloi
River, Puchwara Pass. The geology of the district has been described by Oldham and
Morris (19); more fully, with an accompanying map, by Ball (1); by Medlicott
and Blanford (15); and by R. D. Oldham (18).

The blocks (referred to in the sequel as I., IL, III., IV.) are closely packed with
cycadean fronds; they contain also cycadean stem-fragments, a fructification, a probable
fructification-axis, and numerous coniferous branches. Similar Jurassic plants have
been figured and described by Messrs. Oldham and Morris and by Dr. Feistmantel in
the Memoirs of the Geological Survey of India. Indeed, several of the outstanding
features of the four blocks have been already noted by these writers.

Oldham and Morris (10) figure the frond on the under surface of block I. (cf. their
pl. 21. fig. 8 with РІ. 9. fig. 8 а); also the stem-structures on blocks I. and II. (cf. their
pl. 84. figs. 2 & 3 with Pl. 9. figs. 4 & 1). Тһе impression on block ТУ. (РІ. 7. fig. 1)
is similar to that shown by Oldham and Morris, pl. 34. fig. 1. Feistmantel (7, p. 78)
describes the block II. stem, noting the similarity to Carruthers's figures of Bucklandia
Milleriana and some species of Yatesia, which show alternating zones of large and small
leaf-scars. He cannot decide, however, whether this character is accidental or constant
in the Rajmahal stems, which, he points out, have generally only the larger scars and
are somewhat like those drawn by Williamson (40, pl. 53. fig. 5). The same author (6,
Taf. 3, fig. 1a & b) gives figures of the stem and fructification contained in block II.
(РІ. 9. fig. 1; Pl. 7. fig. 2). He refers the stem to Wiiliamsonia, and considers the

* [I take this opportunity of thanking my friend Mr. Hayden for his ready response to my request for
permission to borrow the specimens, and to have sections cut from the blocks.—4A. C. Szwaxp. |

SECOND SERIES.—BOTANY, VOL. VIII. L

70 MISS NELLIE BANCROFT ON SOME

fructification to be probably that of W. gigas, on account of the large number of bracts
which surround it.

Professor Seward (27, pp. 198, 194; text-fig. 30) describes a British Museum specimen
of an Indian eycadean stem from the Upper Gondwana Series. This agrees very closely
with the stem-fragment in block I. (Pl. 9. fig. 4), both externally and in anatomical struc-
ture (cf. 27, text-fig. 30 B and РІ. 8. figs. 9 & 10). The fronds of Ptilophyllum cutchense
(27, p. 195; text-fig. 32) in connection with this stem are identical with many of those
in the Amrapara blocks (cf. Pl. 9. fig. 3).

IT. METHODS оғ EXAMINATION OF THE BLOCKS.

'The blocks consist of a very hard and compact porcellanous substance, the result of the
metamorphism by igneous rocks, of very fine close shales (19, p. 4). "Tests with acid
showed that this * porcellanite " is very largely, if not entirely, composed of silica, for it
dissolved readily in dilute hydrofluoric acid, while dilute hydrochloric acid failed to reveal
any trace of limy substance.

The very complete silicification of the plant-remains enclosed in the blocks renders
them unsuitable for satisfactory microscopic examination in thin sections, these being
extremely transparent. Various attempts were made to overcome this difficulty, such as
staining the sections with fuchsin; the use of coloured sereens in the condenser of the
microscope; examination by polarised light and by dark ground illumination; апа the
examination of sections cleared of Canada balsam, and mounted in media of different
refraction, e. g., cedar-wood oil and water.

None of these methods were, however, wholly satisfactory, the transparent and homo-
geneous nature of the material rendering impossible, in most sections, an investigation
of the finer anatomical details. Іп some cases, smooth cut surfaces were etched with
dilute hydrofluorie acid in the hope that this would reveal differences in the degree of
silicification, and thus facilitate examination by reflected light. This method also was
unsuccessful, for silicification and a certain amount of metamorphism due to contact with
igneous material have evidently obliterated much of the structure, and hence little
differentiation was apparent on the etched surfaces.

Finally, surfaces from which sections had been cut were treated with immersion oil
without previous etching. Such surfaces, examined by reflected light and considered
in conjunction with their corresponding sections, provided the means by which all
the evidence with regard to anatomical structure was obtained.

Professor Nathorst's collodion-film method (17) was used to elucidate the surface-
features of some of the leaves present in the blocks. It was determined that the cells
above the course of the vascular bundles are elongated and that those between are shorter -
ànd more isodiametric (Pl. 7. Во. 10). No certain evidence with regard to form and
distribution of stomata can be obtained from a surface-film of this kind, especially if
the stomata are at all sunk. Іп this case it is possible that some apparent cell-groups |

_ represent stomata-openings surrounded by epidermal celis (Pl. 7. = 11; df. stoma- |

ж of "Macrozamia атысы fig. 183 2

INDIAN JURASSIC GYMNOSPERMS. 71

ІІІ. DESCRIPTIONS OF THE SPECIMENS.
A. Coniferous Remains,
BRACHYPHYLLUM, Brongn.
BRACHYPHYLLUM MAMILLARE, Brongn.

All the blocks contain fragmentary remains of foliage-shoots. The leaves of these
shoots are small and fleshy, triangular in shape, with a median dorsal keel, crowded, and
more or less spirally arranged; in some cases a small dot-like protuberance may be
seen towards the apex of the leaf (Pl. 7. fig. 6). The name Brachyphyllum mamillare,
Brongn. (2, p. 109), has been applied by various writers to coniferous remains of this type,
Thuites expansus, Phill. (20; see also 13), and Echinostrobus expansus, Feistm. (6), being
apparently synonymous (see Seward, 27, pp. 297 and 298 for further references). As
Seward (30, p. 148) remarks, it is not always easy to decide whether coniferous twigs of
the above type should be referred to Brachyphyllum or Thuites, and though a distinction
has been made between the genera, in that Brachyphyllwm should have spiral phyllo-
{аху and Тш ез whorled, this seems to be a variable character, at least in Thuites,
a departure from the whorled arrangement being frequently noted (30, p. 140). The
generic name Sphenolepidium has also been applied to coniferous twigs similar in habit
to the Indian specimens, though possessing, perhaps, less markedly adpressed leaves.
Oldham and Morris have referred similar examples from Amrapara to Arthrotaxites
indicus (19, pl. 82. fig. 8).

In the present case, however, the generic name Brachyphyllwm has been adopted, as
it does not infer, like Thuites or Arthrotazxites, any relationships with recent genera.
Seward (23, p. 214) points out that the genus “has been compared with several recent
genera; but we cannot regard Brachyphyllum, with its numerous species from various
geological horizons, as more than a purely provisional genus, the actual botanical nature
of which is very uncertain, probably more than one family of Conifers being represented
by the forms referred to under this name.”

Further, on account of their similarity to В. ола аге, these Indian forms may be
referred to that species, which also is regarded as provisional.

With regard to systematic position, the similarity of habit of Brachyphyllum to that
of certain Cupressinese has suggested comparison with that group, while Hollick and
Jeffrey (11) have definitely referred B. macrocarpum, Newberry, to the Araucarineze on
anatomical grounds, making Brachyphylloidese a sub-tribe of the wider group (11, p. 75).
They do not, however, suggest that all the species of Brachyphyllum should be included
under that head (p. 34). For the present, however, the Indian coniferous remains
referred to Brachyphyllum mamillare may be considered as of doubtful affinity, too little
being known of their anatomical structure and organs of fructification to render more
definite classification advisable. | x

It may be noted that sections and cut surfaces of the blocks show transverse and

15 L2

72 MISS NELLIE BANCROFT ON SOME

longitudinal sections of the leafy twigs (РІ. 7. figs. 3f & 7). Ко structure can be
definitely determined, though the impression is given of leaves surrounding a central
stem, the leaf-sections having a vascular bundle, a fair amount of palisade mesophyll,
and a definite epidermis. In general outline and arrangement of parts there is agreement
with sections of the leafy twigs of some recent Cupressinez, but this, of course, is to be
expected from the form of the twigs.

Coniferous Wood.

In blocks II. and ПТ. there are several examples of coniferous wood (РІ, 8. fig. 17 0),
which in section shows, more or less indefinitely, narrow medullary rays and bordered
pits, such as Hollick and Jeffrey figure for Brachyoxylon notabile (qf. 11, p. 56, pl. 13. -
fig. 8, with Pl. 7. fig. S. No examples of flattened or alternating pits were seen. There
is the possibility that the wood originally belonged to the same plant as the leafy twigs,
though there is no evidence of this beyond the association of the plant-remains in the
blocks. Тһе wood is fragmentary and too ill-preserved to allow of reference to any
genus or species.

CONIFEROCAULON, Fliche.
CONIFEROCAULON sp.

Block III. shows a portion of a large stem, 14 ems. long by 10 ems. wide (Pl. 9. fig. 2).
The stem has externally the appearance of a Zamia without leaf-bases, e. g., Z. Skinneri
ог Z. Loddegesió (Seward, 24, pl. 14, figs. 1 & 2). Тһе surface is irregularly grooved
in а transverse direction, with occasional elliptical protuberances. Хо leaf-scars сап be
distinguished. The interior of the stem is occupied by a much-shattered woody cylinder.
What appears to be the reverse of the external layers of this stem is represented on the
surface of the block (Pl. 7. fig. 19).

Seward refers to а Lower Greensand type showing external characters somewhat
similar to those of the Indian stem, as Benstedtia (24), substituting this name as not
implying any particular affinity, for the more definite name Dracena Benstedtii given
by König, following Mackie's suggestion that the stem was very like that of a Dracena
(14). Fliche (ro) figures a stem-fragment from the Lower Cretaceous of Saint-Dizier,
similar externally to the stem described above. Не notices the resemblance, in the
external features at least, to the Araucarias of the Colymbea section, and definitely
refers the stem to the Araucarian Coniferze (р. 7), naming it Coniferocaulon colymbee-
forme.

A similar stem is recorded from the Uitenhage Series of South Africa (Upper Jurassic
or Wealden Age). It is compared with Fliche’s stem, but the name Benstedtia is adopted
as the more non-committal term (29, p. 84; pl. v. fig. 2; text-fig. 5). In his description
of the Jurassic Flora of Sutherland, however, Seward (31, p. 690, pl. ix. fig. 41) refers
a stem showing externally irregular transverse wrinklings to Coniferocaulon, on account
of the narrowness of the pith, a character which suggests coniferous affinities rather than
сусайеап. In a footnote he mentions that Dr. Stopes, in the course of her examination
of the British Museum Benstedtias, has found traces of xylem elements with pits of

INDIAN JURASSIC GYMNOSPERMS. 73

coniferous type. Dr. Stopes herself (32) in a recent paper concludes that the so-called

Benstedtias are merely the decayed remains of gymnospermous types of higher coniferous
affinities, and includes them under the name Coniferocaulon Benstedti. Still more
recently Knowlton (12) proposes to substitute the name Benstedtia Benstedti for these
types, since, according to him, the rules of nomenclature will not permit Dr. Stopes's
combination of generic and specific terms. Dr. Stopes has replied to this criticism
(33) by pointing out that the types variously referred to Benstedtia and Coniferocaulon
do not constitute а genus at all, as they are merely fragments referable to coniferous
stems, and that her use of the term Coniferocaulon Benstedtii is merely for the sake of
convenience and historic interest.

In the present case, on account of external similarity to these plant-fragments, the
block III. “stem” is provisionally referred to the Coniferse under the designation
Coniferocaulon sp., that genus being used in Dr. Stopes’s sense of the term (32, p. 59)
as a convenient and non-committal means of referring to fragments of stems of higher
coniferous affinities.

B. Cycadean Remains.
i. Fronds.
PTILOPHYLLUM, Morris.

PriILOPHYLLUM CUTCHENSE, Feistm.
PrrILOPHYLLUM ACUTIFOLIUM (Oldh. & Morris).

Block I. (РІ. 9. fig. 3a) shows the remains of a pinnate frond; the pinnæ appear as
impressions on either side of the rachis, upon which their scars of attachment are to be
seen (РІ. 8. fig. 1; cf. 19, pl. 21. fig. ба). This impression was about 9 ems. long
before the block was cut and represented probably almost the whole length of the
original frond, since the pinne decrease in size towards either end of the fragment
(4, р. 18). This frond is figured by Oldham and Morris (19, pl. 21. fig. 8) as
Ptilophyllum cutchense (see Feistmantel, 4, р. 13, and 5, pp. 42 & 43, for synonyms).
Other examples are distinctly referable to Ptilophyllum, but differ somewhat from the
typical P. cutchense. Feistmantel (4 and 5) unites these varieties as the “Group of
Ptilophyllum cutchense,” characterized by short oblique ршиг, а little curved, and
obtusely acuminate, slightly auriculate at the base, imbricate or contiguous, sub-
alternately and obliquely inserted on the rachis. At on the photograph (РІ. 9. fig. 3)
is the impression of a frond-apex, while d is apparently the true upper surface of a
frond (Pl. 9. fig. 3; Pl. 8. fig. 2). The apex of a frond is also shown on РІ. 8. fig. 3.

Тһе frond с seems to be of а different type, and agrees with Feistmantel’s diagnosis of
Ptilophyllum acutifolium, Morris (4, p. 11), being distinguished from P. cutchense, Morris,
by its longer, narrower pinn:e forming a larger frond. The ріш are more acuminate
and eurved than in the allied species, and are clearly subauriculate at the upper basal
angle, and decurrent at the lower (Pl. 8. fig. 5; qf. 4, Taf. 1 and Taf. 2. figs. 1-4).

74 MISS NELLIE BANCROFT ОМ SOME

There are several examples of this type in block Т., it being, according to Feistmantel
(5, p. 45), of frequent occurrence in the Rajmahal Hills.

The venation of the Ptilophyllwm pinnee is of an ordinary cycadean type, there being
no midrib (РІ. 8. fig. 4a). The nerves are parallel and apparently end blindly.
Dichotomies occur even in the upper parts of the ріппе. Іп both species the pinne
are entire, subalternate and closely-set on the upper side of the terete rachis, which
thus appears almost entirely on the under side of the frond (Pl. 8. fig. 18). In most
cases compression has been too great to allow of the subauriculate nature of the upper
basal angle being distinguished, though the example figured shows it well (РІ. 8. fig. 5;
cf. 4, Taf. 1, figs. 2а & 8 a).

The abundance of P/ilophyllum fronds, particularly P. eutchense, in block I., indicates
their possible connection with the stem represented (cf. Seward, 27, pp. 192-194, text-
figs. 30A & 32, a reference to a similar stem with P. сшісһепве leaves attached).
There are, however, no indications of actual attachment in the material now under
consideration.

DiIcTYOZAMITES, Oldh.

DicTYOZAMITES FALCATUS (Morris) nob.

Block II. shows many fragments of detached fronds referable to the genus Dietyo-
zamites, Oldh. (Pl. 9, fig. la; РІ. 8. fig. 6; cf. 8, pl. 2. figs. 5 & 6; 9, pls. 3, 4, 5;
4. Taf. 4. figs. 7 © 8; Tafs. 5 & 6; 19, pl. 24. figs. 1 & 2, Dictyopteris falcata).
Тһе pinnz are alternate, varying somewhat with regard to size, shape, and form of
apex, and are inserted by the middle only of the base, being sessile or slightly stalked.
The basal angles are apparently subauriculate. Тһе pinns are specially characterized
by their numerous veins radiating from the base, dichotomising and anastomosing to
form a reticulum, of which the central meshes are long and nearly parallel, while those
towards the apex and margin are shorter. Morris terms leaves with similar pinnæ
Dictyopteris falcata, believing them to be of filicinean affinities, but Oldham (19, p. 39)
prefers to regard them as cycadean fronds. He accordingly substitutes the generic
name Dictyozamites, retaining Morris's species. Feistmantel (4, pp. 18 % 19), con-
sidering Dictyozamites to be characteristic of the Rajmahal Group and purely Indian,
distinguishes the type as D. indicus, but recent discoveries of exactly similar fronds in
other parts of the world render it advisable to retain the former name Dictyozamites
falcatus (вее 16, 28, 41, 42).

Sections of rachises and ріппе.-Іп the majority of cases, the rachises in section show
no clear structure, though in some examples the vascular system is seen to be composed
of separate strands arranged in a double series in the form of a y (РІ. 9. fig. 7; РІ. 7.
fig. 18; cf. 34, pl. 62. fig. 1; pl. 68, fig. 1; and 37, p. 99, fig. 52), a disposition probably
derived by the depression of the айқай members of an original circular or elliptical
system having the xylem directed inwards. Such a system is seen in the leaf-bases
— ы: the сша іш block 1. т» fig. 8; PL 7. figs. 17 & 18; oe

INDIAN JURASSIC GYMNOSPERMS. 75

p- 59, figs. 80 & 88). In vascular bundle arrangement, the present examples agree
with the majority of Wieland’s Cycadeoidean rachises, which, he points out, show a fern-
like V-shaped or closed grouping of the bundles (37, p. 62), characteristic of “а pinnular
bundle supply arising from the upper sides of the system” producing closely-set pinnules
as in Ptilophyllum and Otozamites, while the omega type of bundle arrangement
of most recent Cycads * results in a heavy continuous ridge between the two rows of
|» pinnules” (p. 63).

Nothing can be determined from longitudinal sections with regard to the nature of
the xylem elements; indeed, in most sections, longitudinal or transverse, it is difficult
to distinguish any vascular tissue at all. Nor can the rachis ‘sections be referred with
certainty to either Ptilophyliwm ог Dictyozamiles, except іп one ease where a transverse
section showing the U-shaped vascular arrangement is definitely correlated with an
impression of a Ptilophylium leat.

Where the attachment of the pinns with the rachis is seen in section, it is clear
that in some cases the pinne are much more closely-set than in others; this condition
probably corresponds with greater approximation of the arms of the vascular system, as
some of the sections seem to suggest (РІ. 8. fig. 18a © b). It appears possible from
examination of the fronds on the surfaces of the blocks that the less closely set
examples are Dictyozamites, but no essential difference of bundle arrangement can be
detected. |

In none of these cases of pinna-attachment is it possible to obtain a clear idea of the
pinna structure. In an example of Péilophyllum cutchense lying at the surface of
block II. however, the structure, as far as can be determined, shows an upper small-
celled epidermis, with possibly an upper hypoderm, palisade and spongy mesophyll with
vaseular bundles not showing structure, and a lower epidermis, well-marked but not
sufficiently clear to show stomata (Pl 8. fig. 4а & b; see also Pl 9. fig. 6) Тһе
pinnee decrease in thickness somewhat abruptly towards the margin and appear pointed
in transverse section. It is possible, however, that this is a variable character.

Other pinnze present in the sections, though of essentially the same structure as those
of Ptilophyllum cutchense, were evidently originally thinner in texture. Further,
they possess relatively a greater number of vascular bundles, while the pinna-margins
are decidedly rounded (РІ. 9. fig. 5). In the pinna of this type figured, each vascular
bundle is seen to possess a small mass of xylem presumably centrifugal in nature; for
adjaeent to this xylem, above and below, are eavities representing, no doubt, centripetal
xylem and protoxylem towards the upper surface of the leaf, and phloem towards the
lower surface (РІ. 8. fig. 7). Pinnz of this second type may be tentatively referred to
Dictyozamites falcatus, since the external characters of this species suggest a closer
correspondence with the structure as seen in section than do those of Ptilophylium
acutifolium.

There can be little doubt with regard to the cycadean affinities of these fronds. Both
types are cycadean in general plan, agreeing with such recent Cycads as some of the
Zamias, which show palisade mesophyll. In the possession of palisade tissue there is
à тее LE nt with Cycadeoidea ingens, Ward (35, pl. 7. fig. 4).

76 MISS NELLIE BANCROFT ON SOME

The distribution of sclerenchyma in connection with the vascular bundles in
Ptilophyllum and Dictyozamites is not clear, though there is evidently a much
smaller amount than in the cycadeoidean type figured by Wieland. (With regard

to the surface-features of the leaves, see Section II. “ Methods.")

1. Fructifications.
WILLIAMSONIAN TYPE.

Block II. shows a Williamsonia-like fructification referred by Feistmantel, as already
mentioned, to Williamsonia gigas, Carruth., on account of the large number of bracts
surrounding it (Pl 7. fig. 2; cf. 4, Taf. 2, figs. 1, 2, & 3; Taf. 3, figs. 10 & 4).
The specimen, although apparently poorly preserved, was sectioned longitudinally and
transversely. The transverse section reveals a central mass of petrified tissue surrounded
by the transverse sections of the bracts, between which and the central structure is
matrix material (Pl. 7. fig. 3). Тһе bracts show seven vascular bundles as in the case
of the leaf-bases surrounding the stem to be described in the next sub-section. Тһе
bundles are embedded in a ground-mass, having a broad peripheral band of sub-
epidermal sclerenchyma (see under ** Stems " for further description). In places there
are indications of hairs between the bracts enclosing the central structure ; in this
respect, and also in the structure of the bracts, there is agreement with a Williamsonian
female fructification from the Upper Jurassic of Scotland, described by Prof. Seward at
a meeting of the British Association, 1911. No indications of male organs can be
detected in the Indian specimen. The central mass consists of cellular tissue, the
arrangement of which is suggestive, in spite of its indefiniteness, of transverse sections
of seed-stalks and interseminal scales, as shown by Wieland for Cycadeoidea (РІ. 9.
fig. 10; cf. 36, pl. 8. fig. 6; pl. 10. figs. 14 & 16; and 37, р. 119, fig. 61). Owing to
lack of structure at the edge of the central portion it is impossible to determine whether
the transverse section was taken in or below the ovule-bearing region.

The longitudinal section of the fructification is useless from a structural point of
view. In the extreme upper part of the central mass, however, the oiled cut surface
corresponding to the longitudinal section shows the ends of its vertically-running
constituents, the appearance being suggestive of the presence of interseminal scales and

ovules (РІ. 7. figs. 4 & 5).

In spite of the meagre evidence afforded by this petrifaction, its similarity to
Seward's Williamsonia is sufficient to support its reference to the “ Williamsonian
type” of fructification, this being regarded as of Bennettitalean affinities on account
of the close correspondence to such forms as Bennettites Gibsonianus, Carruth.

Block II. shows the moulds of bract-like structures and the reverse of the bract-
bases. Below these, and slightly to one side, is a raised mass, apparently of silicified
wood, probably belonging to the axis which bore the bracts (Pl. 7. fig. 1) Тһе
arrangement of the whole structure suggests that it represents part of a bract-bearing |
fructification axis, referable to the Williamsonian type. Тһе James Yates collection of

‹

INDIAN JURASSIC GYMNOSPERMS. 77

Yorkshire Williamsonias, in the Paris Museum *, contains several such axes, certain
of which have been mentioned by Saporta (21, pl. 81. fig. 1), Seward (26, p. 276), and
Wieland (38, p. 98; 39, p. 448, fig. 6). (For further references, see Seward, 27, p. 182;
Saporta, 22, pl. 240 to pl. 243.)

11. Stems.
WILLIAMSONIAN ТҮРЕ,

Block I. shows a portion of а cycadean stem (РІ. 9. fig. 4) 10:1 ems. in length and
about 5 or 6 cms. in diameter. Part of the fragment shows the armour of leaf-bases
with indications of ramental packing between them (РІ. 7. fig. 16), while part has the
outer cortical parts broken away, exposing the stele. At the widest part of the stem
is a scar (РІ. 9. fig. 4, a; РІ. 7. fig. 14), evidently of a lateral structure which has
completely fallen away. Below it the wood is badly shattered, and it is impossible to
discover how the branch was supplied with vascular elements. There are no indications
as to the nature of this lateral branch which may have been vegetative or reproductive.
The presence of bract-like structures surrounding the region of the scar (РІ. 7. figs. 14
& 15) suggests the possibility that it was left by a bract-bearing floral axis.

Owing to the absence of a large part of the armour, it is not possible to determine
with certainty whether there are zones of large and small leaf-scars, though towards
the lower end of the stem-fragment the few remaining scars are large, and longer than
they are broad; while in the region of the scar they are small, and broader than long
(РІ. 9. fig. 4), All the scars are, however, more or less rhomboidal to lozenge-shaped
(19, p. 36).

In the lower part of the stem there is a large weathered leaf-base, showing seven
slight depressions, representing the vascular bundles of the leaf-supply. The vascular
bundles correspond in position to those of the leaf-base sections, there being two adaxial,
three abaxial, and two more or less lateral (Pl. 7. fig. 17а &6; cf. fig. 17 e, an example
of a transversely broken leaf-base; also Pl. 9. fig. 8, a section of a leaf-base). The
reverse of several leaf-bases, below the actual stem, shows the same occurrence of seven
vascular bundles passing out to the leaf.

Cut surfaces and thin sections reveal the general structure of the stem and the leaf-
bases (Pl. 8. figs. 8-15; Pl. 9. fig. 8) Іп transverse section, two or three series of
leaf-bases are seen surrounding the stem. These possess an outer epidermis, a fairly broad
peripheral layer of sclerenchyma, and an inner eround-mass of parenchyma with many
secretory sacs. There are constantly seven vascular bundles arranged as in the diagram,
the xylem being directed inward (РІ. 7. fig. 18; cf. arrangement in the weathered leaf-
base). There isno indication of the expected mesarch structure of the vascular bundles,
but it is probable that the protoxylem and centripetal xylem are not preserved, and
that the mass of radially arranged elements seen in the most favourable cases is all

ж Part of this collection has recently been acquired by Yale University.

SECOND SERIES.—BOTANY, VOL, VIII. M

78 MISS NELLIE BANCROFT ON SOME

centrifugal xylem; the phloem also is indistinguishable (Pl. 9. fig. 9). Between the
leaf-bases are indications of ramental hairs (РІ. 7. fig. 18).

The cortex of the stem is fairly wide, and has many secretory sacs (РІ. 8. fig. 14). А
tissue suggesting periderm is developed more or less irregularly in its inner layers
(Pl. 8. figs. 10 & 13). Traces of leaf-supply strands are seen here and there (fig. 10),
but it is impossible to obtain any idea as to their course. Phloem and cambium are not
preserved. The wood is fairly compact, with numerous uniseriate medullary rays, and
is of almost coniferous appearance (РІ. 8. figs. 9-12; ef. 27, р. 193; 37, р. 75, fig. 41).
The xylem elements themselves are small and radially arranged. The radial longitudinal
section of the upper part of the stem-fragment shows very occasional indications of
tracheides with bordered pits arranged similarly to those of recent Cycads, as noted by
Oldham and Morris (Pl. 8. fig. 15; see 19, р. 35). Тһе wood, although much broken
up, with a certain amount of separation into rings, appears to have been developed
originally in a single continuous zone, the primary cambium having persisted. The
breaking up of the xylem zone (РІ. 8. fig. 10) may have been occasioned by the conditions
of preservation, or, as Oldham and Morris suggest (19, p. 35), by * the partial decomposi-
tion of the wood previously to its becoming mineralised." A similar appearance in an
Indian cycadean stem is figured by Seward in his ‘Catalogue of Mesozoic Plants’
(27, p. 194, text-fig. 30 B).

The structure of the woody zone is undoubtedly endarch (РІ. 8. fig. 12; of. 37, p. 75,
fig. 41d). The pith is parenchymatous with numerous secretory saes, such as those in
the medulla of Cycadeoidea (37, p. 77); there are no medullary vascular strands.

This stem possesses characters some of which are cycadeoidean, while others point to
contact with the recent Cyeads. Тһе endarch nature of the xylem is common to all
Cycadophyte stems, but the single compact woody zone with its narrow medullary rays
is a characteristic feature of the cycadeoidean stems, wholly different from the looser
structure of recent cycadean wood, formed in certain cases from successive cambiums.
The Indian stem has multiseriately pitted tracheides similar to those of Cycads, as con-
trasted with the scalariform type usually occurring in the Cycadeoidee (37, p. 75). Like
these, however, it has numerous secretory sacs in its parenchymous ground-tissues,
rather than the gum-canals of recent Cycads. бо far as external characters are
concerned, the stem described resembles that of a recent “ armoured ” Cycad in general
appearance, for the leaf-bases, unlike those of the Cycadeoidee, do not possess axillary
structures; while in place of the ramental scales figured by Wieland (37) for his
American Cycadeoidez, and by Seward (25) for the English type Cycadeoidea gigantea,
the leaf-bases are separated by a packing of ramental hairs.

In the structure of the xylem and medullary rays, and of the leaf-bases with their
ramental hairs, the Indian stem agrees with the axis of Prof. Seward’s Scottish
Williamsonia fructification *. The leaf- and bract-bases in the two forms show exact
correspondence in shape and size, in the amount of sclerenchyma and character of the
ground-mass, and in the number and nature of the vascular bundles.

* Prof. Seward has now published a full account of this specimen (see Phil. Trans. Roy. Soc. ser. B, |
vol. SEU p. 101). ы

INDIAN JURASSIC GYMNOSPERMS. 79

This agreement in structural details in the two specimens lends support to the
reference of the stem to the Williamsonian type, a reference which is further
strengthened by its association with fronds of Ptilophyllum cutchense, or, as Seward
prefers to regard them, Williamsonia pecten. It will be remembered that fronds of this
type are described in actual connection with a similar stem in the ‘ British Museum
Catalogue of Mesozoic Plants’ (27, p. 193).

The stem in block II. (Pl. 9. fig. 1; Pl. 8. figs. 16 & 17) is about 18:6 ems. long and
2:5 ems. wide, showing two complete zones of large leaf-scars and part of a third, and
three zones of narrow scars. In the alternation of zones of large and small leaf-scars,
its external characters agree with those of the Bucklandian type of stem (3). Some of
the larger scars in their upper parts show five slight depressions marking the position
of the vascular bundles which supplied the leaves (see Feistmantel, 7, p. 78).
Lateral compression has occasioned the breaking up and displacement of much of the
harder tissue, so that in section it is only possible to determine general characters. The
pith and cortex are apparently of the usual type for this group, both having secretory
sacs. In the cortex are several vascular structures, presumably leaf-trace bundles and
irregularly developed periderm. Тһе xylem, though now much broken up owing to
conditions of preservation, shows signs of having formed one continuous zone, as appears
to be usual in the fossil Cycadophyta (37, p. 74). It is compact with narrow medullary
rays, and is clearly endarch (РІ. 8. fig. 16). Хо evidence with regard to the pitting of
the elements is obtainable. In general characters this stem agrees with the previously
described example, and is thus referred to the same comprehensive type.

C. Іпсегію Sedis.

On the surface of Block IV. was noted a small rounded structure, which was carefully
removed by chipping away a little of the surrounding matrix. The structure is
evidently a small bilateral seed 3 mm. long, 3 mm. wide, and 2 mm. in thickness just
above the base (Pl. 7. fig. 9). At the base there is an indication of a pedicel. It is
impossible to determine whether the seed is Cycadean or Coniferous. Oldham and Morris
(19, pl. 35. fig. 9) figure similar examples and incline (p. 36) to refer them to cycadean
affinities, on grounds of association, although, as they point out, “ во far as the seeds
themselves are concerned, they might be coniferous."

IV. CONCLUSION.

The above investigation of certain gymnospermous types of Liassic age from the
Rajmahal Hills in Bengal adds some facts as regards structural details to the already
fairly extensive literature concerning Indian Mesozoic Plants (see references to the
works of Feistmantel, 4-9; Oldham and Morris, 19). This literature, although dealing
mainly with the external morphology of the types plainly indicates the similarity of the
Mesozoic Flora of India to floras of the same age in other parts of the world. The
geographical and geological distribution of the genera described above are too well
known from the writings of various authors to render a discussion upon oe subject

: ` M

80 MISS NELLIE BANCROFT ON SOME

necessary or desirable. In this connection, reference may be made especially to the
works of Prof. Seward on the Jurassic Flora (27, Introd. pp. 1-42; see also 28); of
Dr. Wieland, whose ‘ American Fossil Cycads’ includes an account of the geographical
distribution of Mesozoic Cycadophyta (37, chap. i.); and of Feistmantel (6, pp. 47, 48),
who remarks on the similarity of the Liassic and Inferior Oolite floras to be found along
a tract running in a north-westerley direction from India to England. It is sufficient
to note here that the anatomical evidence obtained from the present study of Indian
fossils is in accordance with the accepted view of uniformity of the Mesozoic Floras.

V. SuMMARY.

1. Certain gymnospermous types of Liassic age from the Rajmahal Hills in Bengal
are described.

2. The cycadophyte remains present amongst the material show the association of
stems and fructifications of Williamsonian type with foliage which may be considered
as identical with the English fronds Williamsonia pecten. Portions of Dictyozamites
falcatus fronds are also present. There are, however, no indications of actual attachment
of leaves or fructifications with the stems.

3. The vegetative organs show a combination of recent and fossil cycadean characters.
Externally the stems resemble those of recent “armoured " Cycads, the leaf-bases being
separated by ramental hairs, instead of the scales usually found in the fossil forms. The
single compact woody zone with its narrow medullary rays is a characteristic feature
of the fossil stems, differing considerably from the looser structure of recent cycadean
wood, formed in certain cases from successive cambiums. Тһе Indian wood, however,
shows multiseriately-pitted tracheides like those of recent Cycads, instead of the
scalariform type usually occurring in the fossil stems. Like these, the parenchymatous
ground-tissue has numerous secretory sacs rather than the gum-canals of the living
forms. Тһе “ closed" arrangement of the vascular bundles of the leat-bases and rachises
is similar to that seen in the American fossil Cycads, as contrasted with the open or
omega type of arrangement in most recent genera. Both types of frond are cycadean
in general plan, agreeing with such living forms as some of the Zamias, which show
palisade = in the possession of which there is also agreement with Cycadeoidea
ingens.

4. In the structure of the xylem and medullary rays, and of the leaf-bases with their
ramental hairs, there is exact similarity between the Indian stems and the axis of
eos Seward’s Scottish Williamsonia fructification. |

. The Williamsonian fructification amongst the Indian types seems to agree, so far
as pisi is available, in general characters and arrangement of parts with other
Williamsonian and Bennettitean forms.

6. The structural evidence obtained, in conjunction with the external morphology of
the specimens, supports the already accepted idea of a uniform Mesozoic Flora.

INDIAN JURASSIC GYMNOSPERMS. 81

The writer desires to express gratitude to Prof. Seward for help and advice throughout
the above investigations; also to Mr. H. H. Thomas for many useful hints with regard
to methods of examination of the material.

о»

REFERENCES.

. Barr, V.—The Geology of the Rajmahal Hills. Mem, Geol. Survey, vol. xiii. 1877.
. Broneniart, А. Prodrome d'une histoire des végétaux fossiles. 1828.
. Слввотневз, W.—On Fossil Cycadean Stems from the Secondary Rocks of Britain. Trans. Linn.

Soc. vol. xxvi. p. 675. 1870.

. Евтвтмахтет, O.—Ueber die Indischen Cycadeen Gattungen Ptilophyllum, Morr., und Dictyo-

zamites, Oldh. Palzontologische Beiträge, i. 1876.

. —— Fossil Flora of the Gondwana System. Jurassic (Oolitic) Flora of Kach. Mem. Geol.

Survey of India, vol. ii. ser. xi. pt. i. 1870.
— Ueber die Gattung Williamsonia, Carr., in Indien. Paláontologische Beiträge, ii. 1877.

‚ —— Jurassic (Liassic) Flora of the Rajmahal Group in the Rajmahal Hills. Paleeontologia

Indica, vol. i. ser. 2, pt. ii. 1877.
— Jurassic (Liassic) Flora of the Rajmahal Group from Golapili (near Ellore), S. Godavri
District. Ibid, pt. и. 1877.
Outliers of the Madras Coast. 104. pt. iv. 1879.

. Елене, P.— Contribution à la Flore Fossile de la Haute-Marne. Bull. de la Soc. des Sciences de

Nancy. 1900.

. Ношиск, A., and Jzrraey, E. C.—Cretaceous Coniferous Remains. Memoirs of the New York

Botanical Garden, vol. 11. 1909.

. Kxowrros, Е. H.—The correct Technical Name for the Dragon Tree of the Kentish Rag.

Geological Magazine, decade v. vol. viii. Oct. 1911.

. LrxprzY, J., and Ноттох, W. Тһе Fossil Flora of Great Britain, vol. iii. 1835.

Млскть, 8. J.— The Dragon Tree of the Kentish Rag. Geologist, vol. v. p. 401. 1862.

. Мерілсотт, H. B., and Buanrorp, W. T.—A Manual of the Geology of India. 1879.
. Хатноввт, А. б. Sur la Présence du Genre Dictyozamites, Oldham, dans les Couches jurassiques

de Bornholm. Overs. kongl. dansk. Vidensk. Selsk. Forhandl. p. 96. 1889.
— _ Ueber die Anwendung von Kollodiumabdrücken bei der Untersuchung fossiler Pflanzen.

Ark. för Botanik, үй. 1908.
Orpmaw, В. D. А Manual of the Geology of India. 1893. (A Revision of the Manual by

Medlicott and Blanford, 1879.)
Oxpuam, T. and Моквтз, J. Fossil Flora of the Rajmahal Hills. Palseontologia Indica, vol. i.

ser. 2, pt. i. 1862.

‚ Puurs, J. Illustrations of the Geology of Yorkshire. 1829.

20

21. ЗАРОВТА, Le Marquis de. Plantes Jurassiques. Paléontologie Frangaise; T. ii. Cycadées. 1875.

22. Ibid. Т. iv. Types Proangiospermiques, etc. 1891.

23. Sewarv, А.С. The Wealden Flora. Part II. Gymnosperme. (Catalogue of Mesozoic Plants in
the British Museum.) 1895.

24. —— Notes on the Geological History of Monocotyledons. Ann. of Bot. vol. x. p. 205. 1896.

25. —— On Cycadeoidea gigantea, a new Cycadean Stem from the Purbeck Beds of Portland. Quart.

27.

Journ. Geol. Soc. London, vol. liii. Feb. 1897.
— — On the Leaves of Bennettites. Proc. Camb. Phil. Soc. vol. ix. pt. v. Mar. 8, 1897.
—— Тһе Jurassic Flora. Part I. Yorkshire. (Catalogue of Mesozoie Plants in the British

Museum.) 1900.

MISS NELLIE BANCROFT ON SOME

. SEWARD, А. C.—On the Occurrence of Dictyozamites in England, with Remarks on European and

Eastern Mesozoic Floras. Quart. Journ. Geol. Soc. vol. lix. p. 217. Мау 1903.

. —— Fossil Floras of Cape Colony. Annals of the South African Museum, vol. iv. pt. i. 1903.
. — The Jurassic Flora. Part II. (Catalogue of Mesozoic Plants in the British Museum.)

1904.
The Jurassic Flora of Sutherland. Trans. Roy. Soc. Edin. vol. xlvii. pt. iv. no. 98. 1911.

. Storrs, М. C.— The Kentish Rag Dragon Tree. Geological Magazine, п. s., decade v. vol. viii.

р. 55. Feb. 1911.
The Name of the Dragon Tree. 104. p. 468.

. Warp, L. F.—Status of the Mesozoic Floras of the United States. United States Geological

Survey, monograph 48, pt. II. (plates). 1905.

. WizLAND, С. R.—The Leaf Structure of Cycadeoider. А Study of some American Fossil Cycads,

Pt. П. Amer. Journ. of Science, vol. vii. Apr. 1899.
--- Тһе Female Fructifieation of Cycadeoidea. А Study of some American Fossil Cycads,
Part ПІ. Ibid., May 1899.

. —— American Fossil Cycads. Publication of the Carnegie Institute of Washington. 1906.

--- Historic Fossil Cycads. Amer. Journ. of Science, vol. xxv. 1908.
On the Williamsonian Tribe. Ibid. vol. xxxii. Dec. 1911.

. WILLIAMsoN, W. C.— Contributions towards the History of Zamia gigas. "Trans. Linn. Soc.

vol. xxvi. p. 663. 1870.

. YanE, H.—Mesozoic Plants from Korea. Journal of the College of Science, Imperial University

of Tokyo, Japan, vol. xx. article 8. 1905.

. Үокотлма, M.—Jurassic Plants from Kaya, Hida, and Echizen. 104. vol. iii. p.1. 1890.

EXPLANATION OF THE PLATES.

PLATE 7.

Fig. 1. Probable axis of fructification of Williamsonian type, shown on the surface of block IV.

a, moulds of bracts ; 6, reverse of bract-bases ; с, axis. Nat. size.

2. Williamsonian fructification of block II. Nat. size.

3. The same, in transverse section, showing also the matrix of the block. а, central part of
fructification ; 5, surrounding bracts; c, a petiole in transverse section ; d, a petiole in
longitudinal section; e, pinnz in transverse section; f, sections of coniferous twigs.
Nat. size,

4. The same, in longitudinal section. a, central part of fructification ; 0, bracts ; c, matrix ;
d, portion of the fructification shown in fig. 5. Nat. size.

5. d, of fig. 4, showing appearance suggestive of seeds and interseminal scales. x 6.

6. Brachyphyllum mamillare, portion of a leafy twig. x 3.

7. The same, in transverse section, showing the leaves surrounding the compressed central stem.
x 20.

8. Coniferous wood; longitudinal section showing pitted tracheides. x 180.

9. Seeds of block IV. a, nat. size; b and с, two views of a seed showing bilateral symmetry ;
mote indication of pedicel. x 4.

кеі
сл

=
>

=
“3

м
о

o р қом н

эы,

09

9.

10.

INDIAN JURASSIC GYMNOSPERMS. 83

Collodion-film taken from the under surface of а pinna of Ptilophyllum cutchense, indicating
а, the elongated epidermal cells marking the course of a vein; and 4, the shorter cells
between the veins. х 110.

. The same, indicating a probable stoma-opening, a, with its surrounding cells. x 350.

(Cf. fig. 12.)

. Macrozamia corallipes, a portion of the lower cuticle, showing a stoma-opening, а, and the

surrounding cells. х 350. (For comparison with fig. 11.)

. Diagrammatic representation of the petiole shown on РІ. 9. fig. 7. Note the arrangement of

the bundles, outlining a U, with the xylem (shaded) directed inwards, x 20
Scar of the stem shown on РІ. 9. fig. 4 (see а). а, moulds of bracts surrounding the scar
(cf. fig. 15 а). Nat. size.

. Longitudinal section through the block, immediately to the left of the scarred stem shown in

fig. 14, showing a, the moulds of the bracts surrounding the scar; 0, the transverse
sections of the bracts; c, matrix with petiole and pinna sections. Nat. size.

Williamsonian stem of Block I., leaf-bases. Note the indications of a “ packing” between the
leaf-bases. Nat. size.

. The same. а and 0, two views of a large weathered leaf-base; с, a leaf-base in the matrix,

broken transversely. Note in each case indications of the presence of seven vascular
bundles, and cf. fig. 18 and Pl. 9. fig. 8. a and b, nat. size ; с, X 6

. The same, showing, somewhat diagrammatically, transverse sections of leaf-bases with

indications of ramental hairs separating them. x 4
(9) Reverse of the outer layers of the Coniferocaulon stem shown on Pl. 9. fig. 2. Note the
indications of wrinkling and of elliptical markings. Nat. size.

PLATE 8.

Ptilophyllum cutchense, a portion of the frond shown on РІ. 9. fig. 3 а. Note the impressions
of the under surfaces of pinne and their scars of attachment on the rachis, х 2.

P. cutchense (see Pl. 9. fig. 8 d). Upper surface of pinne. x 2.

P. cutchense, а Írond-apex. Nat. size.

P. cutchense. a, portion of a frond shown on the surface of a block, x 2; 5, section of one of
the pinnz of a. x 50. (Cf. Pl. 9. fig. 6.)

P. acutifolium. Мое indications of sub-auriculate nature of upper basal angle of pinne.
x 2.

Dictyozamites falcatus, portion of frond, showing the attachment of the pinne and the
reticulate venation. х 2.

Part of transverse section of the pinna also shown оп РІ. 9. fig. б. а and 5, cavities left
presumably by the non-preservation of the centripetal xylem and phloem respectively ;
с, supposed centrifugal xylem. х 100.

Part of transverse section of a leaf-base, showing a, epidermis ; b, peripheral sclerenchymatous
layer; с, ground-mass with secretory sacs; d, vascular bundle. x 50. (See РІ. 9.
fig. 8.)

Transverse section of the Block I. Williamsonian stem taken about 1 cm. below scar (see
Pl. 9. fig. 4, 2). а, the leaf-bases surrounding the stem. Хай. size. Note that much of
the cortex and xylem have fallen away from the stem.

Transverse section of the same stem taken at the level of a’, Pl. 9. fig. 4. Note apparent
separation of the xylem into rings. а, cortex with secretory sacs ; b, periderm ; с, xylem ;
d, probable leaf-trace; е, pith; f, bending-out of woody cylinder in the region of the
scar. Nat. size. |

84

Fig.

lik

m
оо

=

E-

MISS NELLIE BANCROFT ON SOME

Transverse section of the xylem of the Block I. stem. Note the uniseriate medullary rays.
x 275

. Transverse section of the xylem, showing its endarch nature. а, the passing-out of a leaf-

trace; b, the pith with secretory sacs. х 50. |

. Transverse section of the periderm-like tissue shown in position in fig. 10,0. х 50.
. Transverse section of the ground-tissue of the Williamsonian stems, showing secretory sacs,

x 50

. А tracheide of the Block I. stem, showing multiseriate fitting. х 350.
. Transverse section of the Block II. Williamsonian stem (see Pl. 9. fig. 1). « and 5, probable

leaf-traces ; c, leaf-base. х 3.

. Section of part of Block IL, showing a, longitudinal section of Williamsonian stem ;

b, oblique sections of coniferous wood ; c, the petiole shown on РІ. 9. fig. 7. Nat. size.

. Diagrammatie representation of two conditions of the petiolar vascular system ; in 0 the

arms of the U are slightly closer than in a, seeming to correspond to the closer setting of
the ропе. х 20.

PLATE 9.

. Williamsonian stem of Block II. Note the zones of large and small scars. а, Dictyozamites

falcatus. Nat. size.

Coniferocaulon stem of Block III. Slightly reduced.

Fronds of Block I. а, Ptilophyllum cutchense; b, frond-apex ; c, P. acutifolium ; d, upper
surface of P. cutchense. Nat. size.

Williamsonian stem of Block I. а, scar; д and 2’, levels of sections figured on РІ. 8. figs. 9
& 10. Note slight variation in size of leaf-bases, seen at the right-hand side of the
stem. Nat. size.

Transverse section of the pinna, a portion of which is shown on РІ. 8. fig. 7. x 20.

P. cutchense, transverse section of pinna (cf. Pl. 8. fig. 4, 5). х 20.

Transverse section of a petiole showing arrangement of bundles (cf. Pl. 7. fig. 13). х 90.

Transverse section of a leaf-base. а, two vascular bundles shown in fig. 9 (cf. РІ. 7.
figs. 17 & 18; РІ.8. fig. 8). х 20.

Two vascular bundles shown at g, fig. 8. х 100.

Portion of transverse section of Williamsonian fructification. х 100.

INDIAN JURASSIC GYMNOSPERMS. 85

APPENDIX.
(Nore on THE INTERNAL STRUCTURE or Coniferocaulon sp.)

Since the reading of the paper on * Some Indian Jurassic Gymnosperms,” two sections, a longitudinal
and a transverse, have been obtained from the block III. stem, Coniferocaulon sp. These shed very
little light on the general structure of the specimen; the few facts that they contribute, however,
serve to confirm the view that it represents the remains of a coniferous stem.

'The transverse section passes through the stem below a lateral structure, evidently a branch-base *.
It shows that the stem is much compressed and imperfectly preserved, the breaking up of the tissues being
probably due in part to decay before petrifaction. "The greater part of the preserved tissues consists of
the woody axis of the plant (text-fig. 1, 2). Partly surrounding this із a layered tissue, с. The thickness
of this layer is much greater on the side of the stem protected by the matrix of the block than on the exposed
surface, where it is very thin (text-fig. 1, с). This indicates that the wrinkled surface shown in РІ. 9.
fig. 2 does not represent the true external layer of the stem, so that in the case of this specimen it is

Text-fig. 1.

Exposed surface шарды

Coniferocaulon sp.—Diagram of the transverse surface of the stem, showing the incomplete,
imperfectly preserved woody axis 2, with the surrounding layered tissues с. As the
diagram indicates, the woody structure is not equally clear throughout the axis.

useless to compare the exposed surface with the external surface-features of Zamia Skinneri (according
to Seward, 24, pl. xiv. figs. 1, 2, & 3) or with those of Araucarias of the Colymbeea section (according
to Fliche, 10, p. 7). Dr. Stopes reaches the same conclusion in the case of a British Museum
specimen (32, p. 56).

Neither the transverse nor the longitudinal section give any information concerning the nature of
the protuberances present on the wrinkled exposed surfaces; it cannot be determined whether they
have any connection with the vascular system or are merely cortical in origin. Пт. Stopes interprets
similar protuberances in the case of some Cretaceous specimens of Coniferocaulon as being due to the
presence of teredo-borings in the original structures (32, p. 55); but this explanation does not hold,
apparently, in the present case. U

On the cut transverse surface of the stem there are indications of zoning of the wood, though it
cannot be determined whether the zones are seasonal ora result of petrifaction. In the transverse
section the wood presents a “ coniferous » appearance, the xylem elements being small, polygonal, and
compactly arranged. The medullary rays seem to be uniseriate. Іп parts of the axis, structure 1з

el

ж This lateral structure is not included in РІ. 9. fig. 2. Its position before sectioning was i diately above the
shattered portion of the axis shown in the top left-hand corner of the photograph.

SECOND SERIES.—BOTANY, VOL. VIII. N

86 ON SOME INDIAN JURASSIC GYMNOSPERMS,

almost entirely obliterated, as seen in transverse section (those areas not marked by curved lines in the
diagram) ; the woody nature of these areas is indicated by a comparison with the fractured vertical
surfaces corresponding to them, these surfaces being regularly striated. The outer layered tissues show |
no definite structure; it is possible that they may originally have been corky in nature.

The longitudinal section passes through the lateral structure mentioned above, showing part of its -
axis in transverse section and part of the main axis in longitudinal section. Тһе connection between
this lateral structure and the main axis is obscure, owing to imperfect preservation and compression
of the tissues. The longitudinal section of the wood shows a compact structure, but actual details are
not clear. Nothing сап be determined as to the appearance of the medullary rays in longitudinal
section of the wood; and although occasionally there are indications that the tracheides may possess
a single series of bordered pits, such as are shown in Dr. Stopes's figure (32, p. 57), preservation is not
sufficiently good to allow of a definite statement being made on this point.

Although the general anatomical features of the stem cannot be determined, the coniferous appear-
ance and texture of the wood is clear, and hence the reference of the specimen to Coniferocaulon is
justified. N. B.

14th February, 1913.

Bancroft Trans ух. Soc. SER. 2. Bor. Vou. УШ.Рь.

ME cono

wee wr
3 2

т”. °
IND

vue 5 PAN

s ары. -

é

Te

ооо

4 o s

неке pss =й

—

^ oocoocccoóc
ree € ET

Gecocooecceoo

INDIAN JURASSIC GYMNOSPERMS.

Bancroft

Trans. Law. Soc. SER. 2. Bor. Vor. VIII Pr 8.

г P -
%

„т

BRK OR

р

ecee

/

172 8««f80

95,

h

$

М.В. del. .

INDIAN JURASSIC GYMNOSPERMS.

Bancroft Trans. Шин. Soc. SER. 2 Bor. Vou. VIII. Pr. 9.

W Tams. Phot. — ?
INDIAN JURASSIC GYMNOSPERMS.

[87

ТП. Rhexoxylon africanum, a new Medullosean Stem. By NELLIE BANCROFT,
B.Sc.(Lond.), F.L.S., 1851 Research Scholar, Newnham College, Cambridge.

(Plates 10 & 11.)

Read 6th February, 1913.

I. INTRODUCTION.

THE specimen which forms the subject of this paper was sent to Prof. Seward by
Dr. A. W. Rogers, the Director of the South African Geological Survey, as an unusual
type of petrified stem-structure. The block was sent to Dr. Rogers without any infor-
mation as to its locality or associations, but in his letter in reference to the fossil he
writes :— I have no doubt that the specimen came from the Cape.” Не adds: “It
probably came from the Karroo rocks.”

Since the Karroo Series ranges from the Dwyka boulder-beds and shales to the
Stormberg Series—i. e., in terms of European geological chronology, from the Upper
Carboniferous (approximately) to the Rheetic or Lower Jurassic (5, pp. 233-243)—the
exact age of the specimen must be left in doubt, so far as concerns evidence other than
that supplied by the structural features themselves.

ІІ. DESCRIPTION OF THE SPECIMEN.

(a) External Characters.

'The specimen is silicified, and represents what is evidently a portion of a stem, the
outer layers of which are absent. Тһе length of the specimen before cutting was about
6:5 em. and the diameters 7:9 5:7 cm., the structure being somewhat elliptical in
transverse section (Pl. 10. fig. 1; РІ. 11. fig. 8).

Externally the fossil has an irregularly ribbed appearance, with vertical and alter-
nating bands of varying width and of two different textures. The more outstanding
bands consist apparently of a fairly porous and loose ground-tissue, while the alternate
bands show fine vertical striations and are very compact (Pl. 11. fig. 1, а & st). These
correspond to the irregular masses of compact texture which are seen on the transverse
surface of the stem (РІ. 10. fig. 12), and which are evidently vascular structures.

Тһе eut transverse surface shows general ground-tissue, more compact towards the
interior of the stem than at the outer parts, where here and there it is very porous
(PL 11. fig. 8). Embedded in this ground-tissue, mainly at the periphery of the block,
are the masses of compact tissue mentioned above. Each of these is presumably a stele

SECOND SERIES.—BOTANY, VOL. VIII. о

58 MISS NELLIE BANCROFT ON

or part of a stele, and each of the complete steles is more or less heavily outlined by a
dark brown coloration of the adjacent tissues (РІ. 11. figs. 3, 7, & 8).

Since it is impossible to know how much of the outer layers of the stem is lacking,
the precise arrangement of the “steles” * must remain undetermined, although appa-
rently one or more series were originally arranged peripherally. In addition to these,
the central ground-tissue contains irregular isolated patches of tissue of the same
compact structure and of varying size and shape. Situated near two steles belonging
to what is considered as the inner series is an irregularly arched mass of the same tissue,
consisting of two distinct parts (Pl. 10. figs. 1, 8, 9, & 11; Pl. 11. figs. 7 & 8).

A comparison of the cut and rough transverse sections of the block considered in
connection with the outer surface-features shows that the masses of compact tissue run
fairly regularly through the short length of stem without changing their position to
any great extent, hence apparently without much branching or anastomosing. The cut
vertical surface figured (Pl. 10. fig. 18) shows the straight and independent course of
two of the steles. Тһе irregular arched mass mentioned above slightly changes its
relative position with regard to the members of the inner series of steles.

(b) Internal Structure.

Several transverse sections were cut from one end of the block, one of them being
practically complete. This will be referred to as Section A. Three incomplete sections
were taken from the same original thick section as this, and are therefore in approxi-
mately the same plane. These, with another incomplete transverse section, cut about
a centimetre higher, do not reveal any facts as to change of position, branching, or
anastomosing of the steles. An almost complete section (referred to as Section B) cut
from the other end of the block, its position being about 4 cm. below the first complete
section, shows no change in general arrangement of the steles, and very little change in
their individual outlines. The central arched mass is more irregular and broken up ;
while one of the steles of the inner series (stele 5; see Pl. 10. figs. 8 & 10; Pl. 11. figs. 7
& 8) assumes a different form, possibly in relation to the separation of а trace. It also
moves slightly away from stele 4 towards stele 6 in passing from the level of Section A
to that of Section В. АП the sections give the same details of structure.

Section A, corresponding approximately to the cut surface of the block shown in
diagram (cf. РІ. 10. fig. 1 and РІ. 11. fig. 8), shows eight complete steles (1-8) embedded
in the ground-mass, one almost entire (9), and parts of four others (10-13), all belonging
to what is termed the “inner series.” In addition to these there are seven groups of
vascular tissues (a-g), evidently parts of steles which are not completely represented.
These seem to belong to an outer series. Here and there amongst the steles, and more
or less in connection with them, particularly in the case of the inner series, are groups
of xylem elements, possibly representing a leaf-supply system (PL. 10: fig. 1; PL H:
figs. 3,7, & 8). Finally, in the central ground-tissue are patches of vascular elements

+ 1 ‹ 4 A š
The term “ stele” is used provisionally for convenience іп describing the structure of the stem. See footnote
to p. 48 of Dr. Scott's monograph on Sutcliffia insignis,

А NEW MEDULLOSEAN STEM. 89

varying in size and form. In this central system is included the irregularly arched
group of vascular elements mentioned above, the group as a whole consisting of two
distinct parts separated by a thin line of ground-tissue. In the description of the
individual masses of vascular tissue this will be referred to as the “ central stele.”

Pl. 11. fig. 7, a photograph of Section B from the opposite end of the block, shows
the steles lettered and numbered as in РІ. 11. fig. 8 and РІ. 10. fig. 1.

Structure of the Ground-tissue.

Silicification has rendered the material so homogeneous and. transparent that it is
impossible to determine definitely the structure of the ground-tissue in which the steles
are embedded. It seems, however, to have been composed of moderately large paren-
chymatous cells, not generally very compactly arranged, though the apparent. loose
structure is probably to some extent due to the shrinking together of the cells and the
consequent production of large spaces. A similar process is described by Weber and
Sterzel for JMedullosa (13, p. 74). As before mentioned, the ground-tissue at the
periphery of the block appears more porous than at the centre. This may be due to
more marked shrinking and destruction of the cells in the outer tissues.

In the curve of the central “stele” is an irregularly oval mass of fairly compact
ground-tissue surrounded by a band of varying width of what appears to be periderm
(Pl.10.figs. 1, 8, 9, & 11; Pl. 11. figs. 7 & 8). Неге and there the cells of the band
are seen to be tabular in transverse section and radially arranged (Pl. 11. fig. 6). The
mass of tissue enclosed by the band of periderm varies in area at different levels,
according apparently to the variation of the curve of the central stele (cf. Pl. 11. figs. 7
& 8; also Pl. 10. figs. 1, 8, 9, & 11). Other more or less circular or irregular bands
occur here and there, as shown in Pl. 11. figs. 7 & 8, pd.

In Section B the band of periderm encloses a mass of ground-tissue containing
scattered and erratic xylem elements ; some of these are cut obliquely, some transversely,
but the majority are seen in longitudinal section (Pl. 10. fig. 11, zy. Тһе band of
periderm itself encloses here and there patches of xylem elements transversely cut.
There is no trace of accompanying phloem (РІ. 11. fig. 6, гу”, А similar enclosure of
xylem elements by periderm is figured by Dr. de Fraine (1, p. 1050, text-fig. 16) in the
case of two leaf-trace bundles in a stem of Swicliffia. The author suggests that the
abnormal condition may be due to injury perhaps by fungal growth. It is difficult to
account for the formation of xylem enclosed by periderm in the present case. In the
parenchymatous ground-mass are occasional groups of sclerotic cells, these occurring
particularly in the neighbourhood of the steles (РІ. 10. figs. 3, 7, 10, & 16; РІ. 11.
fig. 4). There is no certain evidence as to the presence of secretory cells or gum-canals,
although here and there are cells with dark contents which may possibly represent
secretory elements.

Briefly, then, the ground-tissue consists, as far as may be determined, of large-celled
parenchyma, with sclerotic nests and bands of periderm in the neighbourhood of some

of the steles.
| 02

90 MISS NELLIE BANCROFT ON

Structure of the Steles.

These are irregular in nature, in some respects recalling the plate-rings of certain
Medullosezm. The stele 3 of the inner series, as seen in Section А, is a fairly typical
example (РІ. 10. fig. 2; Pl. 11. fig. 3). At first sight, the impression is given of a
somewhat elliptical mass or bundle, apparently a single complete structure. On closer
examination, however, this is seen to consist of two closely-developed though distinct
parts, the inner being much larger than the outer. These are presumably the product
of two cambiums, one giving rise to normally, the other to inversely, orientated elements.
The fact that the outer and inner, or normal and inverse, parts are independent
structures is demonstrated by a slight space between them, and by the lateral dis-
continuity of the brownish elements, which are considered to represent crushed cambium
and phloem (РІ. 10. fig. 2). The crushed tissue between the two masses shows here and
there traces of vascular elements, and may represent a much reduced “ partial pith”
such as shown definitely іп Medullosa stellata, Cotta, or М. porosa, Cotta (13, Taf. үп.
figs. 1 & 4; p. 64, fig. 9), or it may represent ordinary ground-tissue with transfusion-
cells, such as occurs іп Cycas, between the normal and inverse bundles of the stem. Оп
to this tissue abut what appear to be the protoxylems of the two masses (РІ. 10. figs. 2
& 15) The inverse group shows a lateral projection, probably a leaf-trace about to
separate. Тһе direction of the xylem elements is here slightly oblique (РІ. 10. fig. 2, 05;
РІ. 11. fig. 3). In the diagram the dark line marking the external borders of the xylem
represents the brown crushed cells, probably the remains of cambium and phloem. In
the case of stele 3 nothing could be determined with regard to the form and arrange-
ment of these cells. In Section B stele 3 shows very little change in form and structure ;
a slight interruption in the continuity of the radial series of elements appears in the
inverse part, which is larger in proportion to the whole stele than in Section А (PL if.
fig. 7, cf. fig. 8).

А structure such as stele 3 шау be compared with the plate-rings of Medullosa
stellata, Cotta, which, it will be remembered, consist of a “ partial pith ” surrounded by
secondary vascular tissues. In stele 3 the partial pith may be regarded as either much
reduced in amount or crushed by subsequent secondary growth; while the lateral
portions of the complete ring may have remained undeveloped by discontinuity of the
cambium at these points. In this way a stele, consisting apparently of two distinct
parts, may be derived from a complete ring of the Меди Поза stellata type. Worsdell (14)
considers that the so-called anomalous vascular structures in Cycas may be derived from
Medullosean steles by local suppression of cambiums*, and the type of structure
possessed by this South African stem may also have been produced in the same manner.
Weber and Sterzel's figures (13, Taf. viii. fig. 3; Taf. Ш. fig. 5) of М. Solmsii, Schenk,
indicate a similar condition in some of the steles of that form, partial pith being reduced

or practically absent, while there is also a smaller lateral development of the secondary
tissues of the rings.

* See also Chodat (“ Les Ptéridopsides des temps paléozoiques; étude critique,” Arch. des Sci. phys. et nat. t. xxvi.
1908) and Matte (* Recherches sur l'appareil libéro-ligneux des Cycadacées’: Caen, 1904) оп this point.

А NEW MEDULLOSEAN STEM. 91

Stele 4 (Pl. 10. fig. 3; Pl. 11. fig. 8) in Section A shows a very irregular type of
development. Smaller normal and larger inverse parts (а and В) are present as before,
but at one side these are so closely in contact as to appear almost continuous, the lateral
break being less evident than in the case of stele 8. At the opposite side, however, the
break is marked (РІ. 10. fig. 3). The position and arrangement of the xylem elements
marked y suggest that they belong to @, the inverse xylem; the apparent separation
may be due to a break in cambial activity, such as mentioned by Dr. Scott for Medullosa
anglica (7, p. 90). At the side of the inversely developed xylem is a small group
of obliquely running elements, possibly a leaf-trace. The crushed tissue external to
stele 4 shows no trace of structure; while between the normal and inverse parts,
which, as the diagram shows (Pl. 10. fig. 3), are very closely in contact, the tissue
is obliterated. |

In Section В (РІ. 11. fig. 7) stele 4 still shows the break in the development of the
inverse xylem. The inverse and normal parts show no distinct separation laterally,
the elements swinging round and appearing oblique and confused.

Stele 5 (РІ. 10. fig. 3; Pl. 11. fig. 8) presents in Section A the same arrangement of
normally and inversely developed elements, the product of two independent cambiums.
The inverse group apparently consists of two parts, 6 and у; the continuity of the
medullary rays at one point, however, suggests that |3 and у are not really independent,
although the arrangement of the xylem elements shows evidences of a slight disturbance.
At the other side, however, there is a distinct break in the continuity of the medullary
rays and of the radial series of elements, owing perhaps to a change of direction of the
cambium at that point, or to a temporary cessation of cambial activity. At the left-
hand side, the inverse group shows a separating trace ; and at the right, further towards
the periphery of the stem, is a separate trace, of interest on account of a few well-
preserved cells in the crushed tissue accompanying it on its external border. These
cells have delicate walls, and show evidences of having been arranged originally in a
radiating series. Their nature and position suggest that they are phloem-cells. Traces
of similar cells occur in the crushed tissue in connection with the outer normal part of
stele 5. No partial pith is present between the closely-developed normal and inverse
parts of the stele. In section B, stele 5 (Pl. 11. fig.7; Pl. 10. fig. 10) shows a considerable
amount of change in form and structure. The normal part is much reduced and, like
the inverse, shows discontinuity of development. The elements of both groups are more
or less oblique in section. The inverse part shows a large limb, possibly a trace, bending
round towards the partial stele d, the direction of the elements being usually oblique.
On the inner edge of the limb, closely approximating partial stele d, the elements
run horizontally and show spiral thickenings on their walls. They are evidently the
protoxylem of the limb.

Stele 6, which in Section A shows a structure very similar to that of stele 3 (РІ. 10.
fig. 4; Pl. 11. fig. 8), is considerably changed in form at the level of Section В (РІ. 10.
fig. 10; Pl. 11. fig. 7). The inverse part shows the break in development as before, and
at the left-hand side (towards stele 5) is a large, much broken-up limb or trace with
scattered sclerotic nests between the groups of xylem elements. This trace approaches

92 MISS NELLIE BANCROFT ON

the normal part of stele 5 very closely. It is impossible to say whether the two steles
5 and 6 are really distinct, for the appearances almost suggest an anastomosis between
them. In stele 6, as in 5, the elements are mostly oblique in direction.

Stele 7 is more suggestive of a complete stele than any of the others shown in
Section A. The break in continuity of development at the left-hand side is very slight,
the chief evidence of it being a confusion of direction of the xylem elements here
(Pl. 10. fig. 5). It seems as if the outer and inner cambiums had almost joined at
this point. At the right-hand side, however, where a trace is evidently about to separate
from the inner mass, separation of the normal and inverse parts is very distinct.
As usual, the structure of the tissues external to the xylem and between the two
groups is obliterated. The transverse section from the opposite end of the stem-
fragment includes only a small portion of this stele; its straight course, indicating
but slight change of position or form, is shown by the vertical cut surface figured
in РІ. 10. fig. 13. 3

The other steles of the inner series, as shown in Section A, do not present any new
points of structure (see diagrams of steles 8 and 9, Pl. 10. figs. 6 & 7). In Section В,
stele 9 shows the same outlines as in the first, 7 and 8 are incompletely represented, but
their straight course and unbroken contours are shown in the diagram (Pl. 10. fig. 13).
The position of steles 6, 7, 8, 9, and 10 iu Section B appears slightly closed in towards the
interior of tlie stem as compared with their position in Section A (cf. Pl. 11. figs. 7 & 8).
Stele 10 is complete in Section B. It is much smaller than the others, consisting of a
comparatively large, fan-shaped, inversely orientated part and a very small normal part.
There is a slight indication of a partial pith, and both parts of the stele possess much of
the brown phloem-like tissue (Pl. 11. figs, 7 & 4).

Immediately external to the inner steles, at the periphery of the stem-fragment, are
several portions of large masses of normally orientated xylem (Pl. 10. fig. 1; Pl. 11.
figs. 7 & 8, a-g), representing presumably a second series of steles. They are evidently
the product of independent cambiums, which, so far as can be judged from the form of
the incomplete masses, must have had a more or less arched course, as in the case of the
cambiums of the inner series of steles. The first-formed xylem elements of these masses
are sometimes almost in contact with the inner steles; in every case the ground-tissue
between the two series of vascular structures is crushed.

The members of the second series of xylem groups correspond to the small parts of
the inner ring, having no inversely orientated groups pairing with them. They are
thus, in terms of the inner series, “ partial steles.” Partial steles с, d, and 6, as seen in
Section A, show breaks in the xylem, the medullary rays and radial series of elements
being discontinuous along certain lines, although the general direction of development
remains the same after the break as before (Pl. 10. fig. 3). This indicates a cessation of
cambial activity for a time; or a new cambium may actually have arisen.

In association with the majority of these partial steles are patches and strands of
xylem elements. In the case of b a large strand is seen in actual connection; but their
nature, and in most cases their origin also, is extremely problematical. They may be
developed by abnormal extensions of the cambium, or they may be ultimately connected

A NEW MEDULLOSEAN STEM. 93

with the leaf-supply, although nothing could be decided on this point owing to the
absence of the external layers of stem.

It is interesting to note that the direction of the xylem elements in these strands is
very varied, for though the main masses in transverse section usually show transverse
sections of the individual elements, the elements of the strands often run obliquely or
horizontally, owing possibly to the extreme abnormality of the cambial growth (РІ. 10.
figs. З & 10). Crushed tissue occurs with some of the lateral strands, showing occasionally
thin-walled radially arranged cells, as in the case of some of the inner steles. Portions
of partial steles а, 0, c, and d are shown in Section B (РІ. 11. fig. 7). а, 6, and с present
practically the same appearance as in Section A. Partial stele 4, however, shows great
irregularity in structure (Pl. 10. fig. 10). Тһе xylem is much broken up and the
direction of the individual elements frequently varies from vertical to horizontal within
а smallarea. The appearances suggest that the stem has been locally injured, causing
abnormality in the development of the neighbouring tissue. Іп one place a band
of periderm-like tissue borders the outer margin of the main mass of xylem (Pl. 10.
fig. 10, pd.). Preservation is not sufficiently good to give any idea as to the formation of
this tissue. It may be of the same nature as that described by Holden for Myelorylon
(3, p. 253). External to this is а band of badly preserved ground-tissue with sclerotic
nests (s) and an isolated group of inversely orientated vacular elements (vb). Beyond
this is a group of irregular xylem elements, apparently inversely developed, for a band of
brown crushed tissue showing here and there traces of radial arragement appears on the
inner edge of the group. The rest of the tissues of the partial stele are mostly oblique
and show occasional breaks in development. А large trace appears at the left-hand side
separated from the “trace " of stele 5 by а band of sclerotic nests. In the ground-tissue
between the normal part of stele 5 and partial stele d are sclerotic cells and groups of
variously orientated xylem elements accompanied by crushed tissue.

Other Vascular Structures of the Stem.

Тһе central arched stele in Section A consists of two parts separated by rather more
of the crushed and structureless tissue than occurs in the case of the inner steles (Pl. 10.
fig. 8). The larger part of the stele shows a change from vertically running elements at
one extremity to horizontal elements in the thin elongated limb at the other. This limb
swings round in a hook-like manner, the elements maintaining their horizontal course.
In the smaller part of the stele (that towards the centre of the stem) the elements are
more or less oblique in direction throughout. The crushed tissue considered as repre-
senting phloem outlines the external margins of the two parts. In the thin hooked limb
of the larger, where the elements run horizontally, this tissue becomes twisted from the
inner side of the hook to the outer side (Pl. 10. fig. 8).

In one of the incomplete sections, slightly above the level of the complete Section A,
only a trace of this hook is to be seen (Pl. 10. fig. 9); while no trace of it appears on the
cut surface of the block, which is slightly below the level of the section (РІ. 10. fig. 1).
Nor does the hook oceur at the level of the section taken from the opposite end of the

94 MISS NELLIE BANCROFT ON

stem-fragment, although there is a slight swinging round of the elements (РІ. 10. fig. 11).
The break indicated in the smaller part of the stele in the complete Seetion A becomes
more marked in passing through the stem to the opposite end of the block (cf. РІ. 10.
figs. 8, 1, & 11). А break also appears in the larger portion of the stele. At the level
of Section B the elongated limb beyond the break is fairly strongly marked, the xylem
elements running more or less horizontally throughout its length. In the main part of
the stele the xylem elements are oblique (Pl. 10. fig. 11).

The appearance of the central stele at the different levels indicates a probable sinuous
and irregular course in a vertical direction.

It is interesting to note that opposite the ends of the central stele there are larger
spaces than usual between the steles of the inner series; this is seen particularly in
Section A (Pl. 11. fig. 8; Pl. 10. fig. 1). Whether or not it has апу significance
connected with the separation of traces from the central stele cannot be determined.
Besides this structure the central ground-mass, particularly at the level of Section В,
contains small isolated patches of xylem and crushed phloem arranged collaterally
(РІ. 11. figs. 7 & 8) *. The course of the constituent elements varies from vertical to
horizontal, the variation owing, no doubt, to the abnormality of cambial formation and
growth indicated throughout the whole vascular system of the stem. It is difficult to
estimate the significance of these bundles; they may be the remnants of a cauline
vascular system, representing the “ star-rings ” of some Medullosez.

Structural Details.

In all the transverse sections the wood is seen to be compact and coniferous in type,
resembling that of Cordaitean and Araucarian stems (Pl. 11. figs. 3 & 4). The steles
and partial steles consist of radiating rows of tracheides, with uniseriate medullary rays,
both primary and secondary, occurring at intervals. The number of rows of tracheides
between the medullary rays varies from three to eight or nine. The tracheides are
closely arranged and polygonal in shape; the lumen of each element is filled with
layered siliceous material, this being less compact towards the centre (РІ. 10. figs. 14
& 15) In one or two cases the middle lamella may be seen (РІ. 10. fig. 14, ml).

The xylem masses are apparently endarch in structure; stele 3, for example, shows
probable protoxylems in both normal and inverse parts at the extreme inner margins.
The other steles of the inner series do not show the first-formed tracheides clearly. ‘The
partial steles of the second series, however, clearly show endarch structure (Pl. 11. fig. 2)
and at their inner margins the medullary rays seem to widen slightly, although they
apparently still consist of one row of cells.

In Section А the horizontally running elements of the limb of the central stele are
spirally thickened in the majority of cases; a few have bordered pits. In Section В,

where the limb is much larger, most of the elements show a double series of bordered
pits (Pl. 11. fig. 5).

* See page 89, reference to enclosure of xylem in periderm.

Bancroft.

$ 2
N 22
mS \ Z

M
NSQNE
ү

б \\ RUN 3
TAS

OPS
алау

Y
ок

9.

%

Bancroft

Trans. Low Soc. Ser. 2. Bor. Vo

ўа 9 999
449934 МЕНЕ РЕЗ 0»
ior (Огде (041 во -

at? %
ga* One 24% 167

eec
ç ` ch
ч {Сун — "et
x 9 °

еі, деб

авто ©
py АМЕ

26
a? Pisce
3%

London Steseoscopic Co. imp.

А NEW MEDULLOSEAN STEM. 95

The lack of structure in the tissue immediately external to the xylem has already been
commented upon in dealing with the separate steles. Only here and there are indi-
cations of radial rows of thin-walled cells, presumably representing the phloem.

The longitudinal sections add very little to the facts already obtained from the trans-
verse sections. "The evidence of the latter with regard to the loose parenchymatous
nature of the ground-tissue is confirmed; gum-canals or secretory sacs cannot be
distinguished, but patches of sclerotic cells are fairly abundant, especially near the steles.
'The masses of xylem in longitudinal section show little evidence of branching or anasto-
mosing (cf. steles 7 and 8 on cut surface, Pl. 10. fig. 13). In some cases small branches
or traces, such as occur in the transverse sections, are seen separating. Тһе brown
tissue outlining the xylem is crushed and structureless, as seen in longitudinal
sections. |

Тһе structure of the xylem may be determined only occasionally in the longitudinal
sections. Тһе tracheides have large bordered pits on their radial walls, typically in two
alternating series (Pl. 10. fig. 17а; Pl. 11. fig. 5), although here and there only one
series, and in one or two cases three series, are present (РІ. 10. fig. 17 b). The pits are
always in contact and are usually flattened, resembling those of the Cordaitez and
Araucariese. Where the wood is cut tangentially sections of the bordered pits are seen,
but the middle lamella is not definitely distinguishable in any of the cases noted. The
medullary rays, as seen in tangential section, are fairly high, consisting of three or four
to fifteen cells in a vertical series. Their uniseriate nature is clearly shown (РІ. 10.
fig. 18).

In the main masses of vascular tissue there are no indications of protoxylem in
longitudinal section, but in the ground-mass between two of the steles is a strand of
xylem consisting almost entirely of spiral elements.

(c) Summary of Structure.

(1) The stem-fragment under consideration shows parts of two series of vascular
structures embedded in a parenchymatous ground-mass, which also contains a central
arched stele and many isolated patches of vascular elements; sclerotic nests and bands of
periderm occur irregularly in the ground-tissue. |

(2) Each of the “‘steles” of the inner series consists of two parts, the outer small and
normally orientated, the inner larger and inversely orientated. The two parts are
almost in contact, the tissue between being more or less obliterated.

(3) The members of the second series consist of normally orientated xylem only ;
they are therefore termed “ partial steles.”

(4) The steles pursue a fairly straight course through the short length of stem,
apparently with little branching or anastomosing. Here and there traces are seen
separating from the inverse portions of the inner series ; while the lateral portions of the
partial steles are much broken and irregular.

(5) The xylem is compact and coniferous in type, having the protoxylem elements at
the inner margin of the masses.

SECOND SERIES.—BOTANY, VOL. VIII. Р

96 MISS NELLIE BANCROFT ON

(6) The xylem elements have bordered pits arranged typically in two alternating
series, The pits are flattened and in contact, as in the Cordaitez and Araucarian
conifers.

(7) The medullary rays, both primary and secondary, are uniseriate.

(8) The tissue accompanying the xylem-masses externally is much crushed, but
occasionally shows traces of thin-walled cells, radially arranged. It is considered as
representing phloem.

III. AFFFINITIES AND REFERENCE OF THE SPECIMEN.

In many respects the stem-fragment shows an approach to certain members of the
fossil group Medulloseze, the structure of which is well known from the researches of
Góppert and Stenzel (2), Weber and Sterzel (13), Solms-Laubach (тт), Scott (7, 8, 9), and
other authors. The forms included under the genus //е4и оза have been arranged by
Weber and Sterzel in form-cycles, each form-cycle being represented by a typical species
with varieties which may or may not be quite distinct. The principal form-cycles are
those of M. stellata, Cotta, М. porosa, Cotta, М. Solmsii, Schenk, and M. Leuckarti,
Gopp. & Stenz. Evidently belonging to the same Medullosean family or plexus are
forms from different horizons and showing certain differences in structure and histology
from Medullosa itself. Such genera are Steloxylon, Sutcliffia, and Colpoxylon. Тһе
Cladoxylez are а distinct group showing certain points of resemblance to the Medul-
loseæ. The South African stem agrees most closely in general appearance, at least, with
М. Solmsii. The characters of this form-cycle may be briefly summarised as follows
(see 13, p. 73; Taf. iii. figs. 1-5; Taf. viii. fig. 3):--М. Solmsii shows typically two
series of plate-rings, each ring representing a stele, and appearing to consist of a
normally orientated and an inversely orientated mass of secondary tissue, the lateral
portions of the ring being often poorly developed (see 13, Taf. iii. fig. 5; Taf. үш.
fig. 3). In variety ¢ypica the normal and inverse parts of the external rings are equally
strongly developed ; var. incrassata has the normal part most strongly developed; and,
lastly, in var. lignosa there are several zones of secondary wood and bast outside the
two zones of plate-rings. In many cases the “partial pith” containing primary
tracheides, and typically occupying the central part of a ring, is practically absent (13,
Taf. ñi. fig. 5; Taf. viii. fig. 3). It may have been originally insignificant in amount,
or it may have become crushed during the life of the plant or during fossilisation (13,
р. 74). It seems to have entirely escaped preservation in the cases where a distinct
space is seen іп the interior of the ring. Мейш(ова Solmsii, according to these
authors, has the closest xylem-structure known іп the genus Medullosa (13, р. 114);
nothing seems to be actually known of the type of pitting of the tracheides in
M. Solmsii, though in the genus, as a whole, multiseriate bordered pits are the rule.
Phloem is not recognisable; the dark-coloured tissue surrounding the steles is
almost structureless. In the interior of the stem JM. Solmsii shows very small
star-rings, and in the cortex also are similar structures arising from the inner series
of plate-rings.

A NEW MEDULLOSEAN STEM. 97

The age of Medullosa Solmsii and of other Continental species is Permian (13, р. 48);
the English species, M. anglica, however, is of Lower Coal Measure age (7, p. 84).

In the case of the South African stem, the general appearance and arrangement of the
innermost series of steles recalls that of the plate-rings of М. Solmsi, except that in
the African type the inverse part is much more strongly developed than the normal
part, possibly because towards the centre of the stem there is more space for such
development to proceed after the second zone of xylem-masses has begun to develop,
cramping the normal parts of the first zone. It has been shown, however, that there
is variation in the relative degrees of development of the normal and inverse parts of
the rings in M. Solmsii; while in M. porosa, in the case of the outer bundles of the
pith (13, p. 70) and sometimes in М. anglica (7, p. 89), there is greater development of
xylem towards the central parts of the stem. Further, in the South African stem the
outer plate-rings are not represented, unless the portions of normally orientated xylem
at the periphery of the fragment correspond to these, the inverse parts being absent.
Or, these masses may belong to an outer zone of normally orientated tissue, such as
occurs in М. Solmsii, var. lignosa, although in this form the zone is more or less con-
tinuous. Тһе outer plate-rings in this case would be entirely unrepresented.

The new stem shows agreement with M. Solmsii, and differs from M. stellata and
М. porosa, in the extreme reduction or at least non-preservation of partial pith and
primary tracheides. There is a marked difference from JM. anglica, the steles of which
show a large primary mass of tracheides and parenchyma (7, p. 88). The isolated
irregularly orientated’ xylem-groups which occur in the South African stem may be
vestigial star-rings, comparable with those very generally present in the central ground-
mass of Меди ова stems. :

The bundles arising from the inner ring of steles in the new type may be of the
same nature as the leaf-traces, which in Medullosa typically arise from the inner plate-
rings. Their ultimate fate, however, cannot be decided, as no cortex is present.

Secretory canals, which are present іп the central ground-tissue of M. Leuckarti (13,
p. 133) and in the ground-tissue and pericycle of the steles in M. anglica (7, p. 96),
have not been distinguished in the South African stem, nor can they with certainty be
shown to occur in species of Wedullosa other than the two mentioned.

Sclerotie nests are present in the ground-tissue of the African stem; according to
Weber and Sterzel, however, their presence in the species of Medullosa is uncertain
(13, р. 115). Sclerenchymatous bands, isolated in the ground-tissue, are mentioned by
Scott for М. anglica (7, р. 96; pl. x. fig. 2); while for the same species internal
periderm is described (7, р. 96; pl. v. figs. 1 & 3; pl. vii. fig. 18; pl. x. figs. 6 & 7),
occurring in a continuous though irregular zone. In the African stem isolated bands
of periderm occur apparently without definite purpose.

Owing to the variability and wide range of structure shown by Meduliosa stems, the
only distinct points of difference between them and the South African stem lie in the
compact wood of the latter, with its uniseriate medullary rays, and in the pitting of
the tracheides. In the new specimen the biseriate pitting of the tracheides resembles
that exhibited by the Cordaitee and Araucariez rather than the multiseriate type

Р2

98 MISS NELLIE BANCROFT ON

occurring in the species of Meduilosa and in Pieridosperms generally. It is unfortunate
that, as mentioned above, nothing is known of the pitting in M. Solmsii, which shows
an approach to the South African stem in the more compact nature of the xylem and
also іп the apparent reduction of “ partial pith.”

Steloxylon Ludwigii, a stem of doubtful Permian age, is another member of the
Medullosean group. It was first somewhat superficially described by Góppert апа
Stenzel as Medullosa Ludwigii (2, Taf. xvii. figs. 18-20; see also 6, Taf.1. figs. 1-16;
and 12, Taf. iii. figs. 6 & 9); it exhibits a homogeneous ground-tissue with usually
roundish or elliptical steles. Some are, however, considerably elongated and arranged
somewhat radially, instead of tangentially as in the case of the plate-rings of Medullosa.
The structure of the ground-tissue is not clear, but appears to have been paren-
chymatous. The medullary rays of the wood are of varying heights, and are uni- or
biseriate, separating groups of two or three series of tracheides. The tracheides bear
roundish oval bordered pits in one or more rows. Phloem is not preserved; while the
steles have a partial pith, the structure of which is not clear. Steloxylon Ludwigii
differs from the species of Medullosa and from the South African stem in the arrange-
ment of the steles. It approaches the new type in the pitting of the tracheides, but
differs from it in the less constantly uniseriate nature of the medullary rays and in the
definite presence of a partial pith (12, p. 552). |

Sutcliffia insignis, from the Lower Coal Measures of Shore, is considered by Scott (8)
to be a probable primitive type of Medullosean stem. Тһе recent work of Dr. Е.
de Fraine indicates its affinities with M. anglica. It possesses a large central stele, very
similar in structure to a single stele of M. anglica. From the main stele arise subsidiary
steles at intervals. Phloem is well preserved, and the tracheides usually possess multi-
seriate bordered pits. 8шей а is not comparable with the South African stem, except
through other members of the Medullosean group.

Colpoxylon, represented by the species C. eduense, Brongn., has been described by
Renault as coming from the Permian of Autun (4, pl 67, figs. 1 & 2). Itis in agree-
ment with regard to structure with some of the simpler Medullosez, but itis monostelie
for some length of the stem. It is perhaps an aberrant Medullosa, and like Swtcliffia is
only indirectly comparable with the South African stem.

The Cladoxylez are а group of forms (Cladoxylon and Vélkelia) approaching the
Medullosez in the possession of numerous steles. Solms described the structure of
Cladoxylon in 1896 (то) and of Vélkelia in 1910 (12). The group differs from the
Medullosez in the arrangement of the steles, which are usually more or less radially
disposed (cf. Stelozylon), and also in the minute structure of the wood, the tracheides
usually having scalariform pits. The age of the group is Devonian or at least early
Paleozoic. Scott concludes that the Cladoxylez and Medullosez have no intimate
relation, but they probably present a case of parallel development, showing superficial
resemblances to each other in the possession of many steles (8, p. 64; 9, p. 497).

Cladoxylon Kidstoni, however, deserves mention on account of an approach in certain
characters to the South African fossil. It has been described by Solms-Laubach as
coming from the Lower Carboniferous of Berwickshire (12, p. 537). Ав the stem-

А NEW MEDULLOSEAN STEM. 99

fragment is very small (2:4 em. x 2 em. diam. and 1:3 ст. long), only four steles and part
of two others being represented, it is impossible to determine the original arrangement
of vascular structures. If anything may be judged from the position of those present,
however, it would seem that they were more or less radially arranged—that is, the
narrow partial pith lay in a radial direction. The steles are almost complete rings; only
at one pole of the partial pith is the continuity of the secondary growth broken, and at
this point, in the case of two of the steles, is attached a peculiar wing-like mass of
secondary elements, probably representing a trace. In one instance, the wing is less
closely in contact with the main stele and has divided into two parts (12, Taf. iii.
fig. 5). At the other pole of the stele, secondary growth shows its greatest development.
The partial pith is narrow and, like the phloem, usually not preserved: in one case its
cellular nature may be determined (12, Taf. iii. fip. 11). The steles pursue a fairly
straight course, so far as can be determined from the fragment. The wood is compact,
with uniseriate rays separating groups of from two or three to eight or nine rows of
tracheides (12, Taf. iii. fig. 11). In this there is agreement with the South African
fossil. The chief feature of interest in C. Kidstoni, from the point of view of comparison
of the two types, lies in the pitting of the tracheides. The radial walls possess typically
a single series of elongate-elliptical bordered pits, giving a scalariform effect
(cf. character of the group Cladoxyleze). Неге and there, however, often in the same
tracheide with the single series, a double row of bordered pits occurs (12, Taf. iii.
fig. 7) *, the pits being either alternate or opposite—they are always flattened and in
eontact. In the case of alternate pitting the appearance is very similar to that presented
by the tracheides of the South African fossil. 'This cannot, however, be regarded as
indicating any relationship between the two types, which show points of dissimilarity in
other respects, such as stelar organisation. C. Kidstoni is interesting, in that the pitting
of its tracheides points to a transition from a simpler more fern-like type to that
exhibited by the South African fossil, Cordaiteze, and Araucariez. Тһе Cladoxylex
need not be considered so far as the closer relationships of the new type are concerned,
but must be mentioned with the Medullosez as representative of the plexus of polystelie
types to which the stem under consideration is apparently referable.

With regard to the wider affinities of the new specimen, it is possible that the biseriate
Araucarian type of pitting exhibited by its xylem elements indicates a point of contact
more or less remote with the ancestral forms of the Атапсагіее. Тһе combination of
Araucarian pitting of the xylem elements with the Medullosean arrangement of steles is
interesting, and the possibility is suggested that there may have existed other simpler
or even monostelic forms, as yet unknown, also exhibiting Araucarian pitting, and
that from these or their ancestral stock the Araucariee may have been derived either
directly or indirectly. It is extremely unsafe to base any conclusions upon the pitting
of the xylem elements, for how far such a character may serve as a criterion in the
consideration of relationship between types is not as yet thoroughly understood, and
the above suggestion is made with due regard to this fact. It is, however, interesting

* [n Solms-Laubach’s figures the bordering of the pits is not shown.

100 ; MISS NELLIE BANCROFT ON

in that it provides a parallel case to that put forward by Miss de Fraine (1) for the
Сусадасеге, which she considers may have originated from a monostelic Medullosean
form typified by Sutcliffia *, which shows a more Cycadean type of pitting.

The age of the South African specimen, as has been shown, is doubtful, but its structure
justifies its reference to the Paleozoic era, while its agreement in general anatomy with
the Continental Medulloseæ rather than with the earlier English types is consistent
with reference to an Upper Permian horizon.

From early Palseozoic to Permian times there seems to have been a plexus of types
showing polystelie structure, and the South African stem may be included as a member
of this plexus. То recapitulate its probable affinities :— The stem seems to agree most
closely with the genus Medullosa, which is recognised as showing considerable variation
of stelar structure. It differs from the typical species +, however, in the extreme
reduetion of the partial pith, in the general coniferous texture of the xylem with its
compact structure and uniseriate rays, as wellas in the pitting of the elements, which
in this respect resemble those of the Cordaiteze or the Araucarian conifers. Steloxylon
is approached chiefly in the type of pitting of the tracheides, but it differs in the less
definite arrangement of the steles and in the presence of a partial pith—this being much
reduced in the South African stem.

These differences, united to the facts that the horizon and locality of the stem-
fragment are indefinite and that its associations are indeterminable, sufficiently isolate
it from all hitherto-described types to render its reference to any one of them undesirable.
Further, this fossil appears to be the first, and so far the only, specimen of a stem
of this nature known from the Southern Hemisphere, so that it seems justifiable to
institute а provisional genus for its reception and to include the genus in the family
Medulloseze.

The generic name Rhexoxylont is suggested as being descriptive of the broken-up
nature of the vascular structures of the stem.

ТУ. SUMMARY AND DIAGNOSIS.

(1) Rhexozylon africanum, a stem-fragment of doubtful later Paleeozoic age, from an
unknown South African locality, is described and diagnosed as follows :—

RHEXOXYLON, nov. gen.

Vascular system of the stem consisting of an inner ring of elliptical steles, each com-
posed of a large inversely orientated part, from which traces are given off, and a small
normal part; external to these a series of “ partial steles ” consisting only of normally
orientated elements; the lateral margins of these partial steles often much broken by
the separation of traces. Xylem compact, having spiral elements on the inner margin;

ж Cf. Worsdell’s views on the origin of Cycadean structures.
T М. Solmsii must be excepted. Note remarks on p. 96.

A NEW MEDULLOSEAN STEM. 101

metaxylem elements typically with biseriate bordered pits. Medullary rays uniseriate.
Ground-tissue parenchymatous, with scattered and irregular vascular strands, bands of
periderm, and sclerotic nests.

В. AFRICANUM, nov. sp.
Characters of the genus.

(2) The general structure of Rhexoxylon africanum suggests relationship with the
genera Medullosa and Steloxylon, and the new genus is included in the family
Medullosez.

In conclusion, the writer’s expressions of gratitude are due to Professor Seward
for help and advice during the above investigation; and also to Dr. R. Kidston for
slides of Oladoxylon Kidstoni which he kindly lent for comparison with the South
African stem.

Botany School, Cambridge.

REFERENCES.

(т) De Fearne, E.—On the Structure and Affinities of Sutclifia, in the light of a newly discovered
Specimen. Ann. of Bot. vol. xxvi. 1912.

(2) бӧррект, H. R., and Этемивь, G.—Die Medullosez, eine neue Gruppe der fossilen Cycadeen.
Palzontographica, 1881.

(3) Hornen, H. S.—Note оп a wounded Myelozylon. New Phyt. vol.ix. 1910.

(4) Вкмасіл, B.— Bassin Houiller et Permien d'Autun et d'Epinac. Flore Fossile, iv. pt. II. 1896.

(5) Rogers, А. W., and Du Toir, A. L.—An Introduction to the Geology of Cape Colony. 1909.

(6) Зснемк, A.—Ueber Medullosa, Cotta, und Tubicaulis, Cotta. Abhandl. der math. phys. СІ. der k.

. Sšchs. Gesellschaft der Wissenschaften, vol. xv. no. vi. 1889.

(7) Scorr, D. H.—Structure and Affinities of Fossil Plants from the Palzozoic Rocks. III. On
Medullosa anglica, a new Representative of the Cycadofilices. Phil, Trans. Roy. Soc. B,
vol. exci. 1899.

(8) — On Sutclifia insignis, а new Type of Medullosez from the Lower Coal Measures. Trans.
Linn. Soe. vol. vii. p. 45. 1906.

(9) Studies in Fossil Botany. 1909.

(то) Sorms-Lavsacu, H. Grar zv.—Ueber Ше Seinerzeit von Unger beschriebenen structurbietenden
Pflanzenreste des Unterculm von Saalfeld іп Thüringen. Abhandl. der К. pr. geologischen
Landesanstalt, Neue Folge, Heft 23. 1896.

(11) —— Ueber Medullosa Leuckarti. Bot. Zeit. vol. lv. p. 175. 1897.

(12) —— Ueber die in den Kalksteinen des Culm von Glatzisch-Falkenberg in Schlesien erhaltenen
structurbietenden Pflanzenreste. IV.—Vólkelia refracta, Stelozylon Ludwigii. Zeitschrift
für Botanik, 1910.

(13) Weser, O., and SrznzzL, J. T.— Beitrüge zur Kenntniss der Medullosex. Ber. der Naturw. Ges.
zu Chemnitz, vol. xiii. 1896.

(14) Мовѕретл,, W. C.—The Structure and Origin of the Cycadacem. Annals of Botany, vol. xx.
p.129. 1906.

102

Гір. 1.

MISS NELLIE BANCROFT ON

EXPLANATION OF THE PLATES.

PLATE 10.
Figures 1-19 diagrammatic, figures 14-18 drawn with the aid of a camera lucida.

Transverse surface of the stem, after cutting, showing the main features (cf. РІ. 11. fig. 8,
a photograph of the complete Section А, taken from this surface, and representing a level
slightly higher in the stem) : 1-18, steles of the inner series ; a-g, partial steles of the outer
series. Nat. size.

Figs.2-8. Transverse sections of the vascular structures of the stem, drawn from the complete

Fig. 9.

Section A, shown in Pl. 11. fig. 8. Comparison should be made with this and with fig. 1
for the relative positions of the vascular masses of the stem. £r signifies transverse section
of the xylem elements ; ob, oblique section ; and lg, longitudinal section. The shaded out-
lining of the steles indicates the brown crushed tissue.

Fig. 2, stele 3 of the inner series : pm, partial pith ; pz, protoxylem ; at ob a trace is seen
separating from the larger inverse part of the stele. Fig. 3, steles 4 and 5 of the inner
series, and partial steles с and d of the outer series: а and В, normal and inverse xylem,
respectively, of the inner steles; гу, xylem evidently belonging to В, but more or less
separated from it by breaks in the continuity of the radial series of elements : similar
breaks are seen in partial stele с; 8, sclerotic nests. Figs. 4, 5, and 6, steles 6, 7, and 8,
respectively. Fig. 7, stele 9, showing sclerotic nests in the crushed tissue ats. Fig. 8,
central stele ; pd, bands of periderm. Figs. 2-7, x3; fig. 8, x 5.

Central stele, at a slightly higher level than fig. 8. (From an incomplete Section) x5.

Figs. 10 & 11. Vascular structures at the level of Section B, shown in РІ. 11. fig. 7. Fig. 10, steles 5

and 6 of the inner series, and partial steles d and e of the outer series : v5, isolated and
irregular vascular strands ; other lettering as before. x3. (Cf.figs. 8 & 4.) Fig. 11, central
stele: губ, xylem elements enclosed in periderm ; vy’, xylem elements scattered in ground-
mass. х5. (Cf.figs.8 & 9.) For the position of the central stele as shown in figs. 8, 9,
& 11, relative to that of the other vascular structures of the stem, refer to fig. 1 and РІ. 11.
figs. 7 & 8.

. Diagram showing the relation between the vertically striated bands of the outer surface of the

stem (cf. Pl. 11. fig. 1) and the vascular masses of the cut transverse surface. Nat. size.

. Portion of a tangential longitudinal surface, showing the straight course of two inner steles

(7 and 8): а, ground-mass. Маф. size.

. Xylem elements, showing the layered siliceous filling of the cell-cavities (see also figs. 15

& 18). Note indication of middle lamella, ml. х 450.
Protoxylem group of stele 3 (cf. fig. 2). x 450.

. Sclerotic nests in the crushed tissue of stele 9. x90. (Cf. fig. 7, s, and РІ. 11, fig. 4.)
. Pitting of the tracheides : a, typical ; 0, from the limb of the central stele. х 280. (Cf. Р1. 11.

fig. 5.)
Tangential section of xylem, showing the uniseriate medullary rays. Note the layering of
siliceous material in the cells of the rays. x 280.

Fig. 2.

Fig. 3.

Fig. 4.

Fig. 5.
Fig. 6.

Figs. 7&8. Sections B and А respectively.

A NEW MEDULLOSEAN STEM. 103

PLATE 11.

A photograph of the decorticated stem-fragment, showing the external app a, ground-
tissue ; st, stelar tissues. Nat. size.

The inner margin of partial stele 0, showing protoxylems (pz) and uniseriate medullary
rays (mr). x70.

Stele 3, from Section A (cf. fig. 8), a fairly typical example of an inner stele: ¢, trace ;
pm, partial pith. х5. Note the arrangement of the xylem elements and the outlining of
dark crushed tissue.

Brown tissue of stele 10 (from Section B), showing radial arrangement of elements :
s, sclerotic nests. x70. Note the structure of the xylem, with uniseriate medullary rays.

Xylem elements, showing biseriate pitting. X325.

Portion of periderm band in arch of central stele (Section В). The periderm (pd) encloses a
strand of xylem elements (#y*). x70. (Cf. Pl. 10. fig. 11.)

x2. Steles and partial steles are numbered and lettered

as in Pl. 10. fig 1: cs, central stele ; pd, epiderm bands ; vb, isolated vascular strands.

SECOND SERIES.— BOTANY, VOL. VIII. Q

“MEMORANDA CONCERNING TRANSACTIONS.

Transactions, containing both ы. та Zoological contributions, has been completed
t vod bet are sn for sale. Only се single volumes, or parts to complete sets, may
The priee of the Index to. Vols, 125i is 8s. to the public, and 6s. to Fellows; to
ellows Ws.

| уре: Series of the "Tissus is divided into Zoological and Botanical sections. The prices of the
inical parts of these whieh hav nei reen аге аз undermentioned. (For the Zoological parts see Zoological

_ Весов — (continued).
сс s he? to the Price to
a v Published. Public. Fellows.
EAA Ë gs.

о. тов £
W 0.6 05-29%
ПЕСНЕ: 2018 0 — 019
‘Part 1V.1908..... 0 8 0...
Esa 9 š 0.

1905 .... 0 12
1904. .... 012
ceres

uses 42 I. 4%

2nd Ser, BOTANY.)

THE

TRANSACTIO?

THE LINNEAN SOCIETY

A Mn.
: vu ed
у dco

MARINE ALGE, RHODOPHY

ы : (Communicated by Prof. J. SrANLEY Gaxprven, М.А, F B.S.

ta 1067-3

ТУ, Marine Alge, Rhodophycese, of the ‘Sealark’ Expedition, collected by Mr. J. Stanley
Gardiner, M.A. By Mrs. A. WEBER-VAN ВоззЕ, Ph.D. (Communicated by
Professor J. STANLEY GARDINER, M.A., F.R.S., F.L.S.)

(Plates 12-14 and 1 text-figure.)

Read 5th June, 1913.

| I. Introduction.

THE alge treated of in the following paper were collected by Mr. J. Stanley Gardiner
during the Percy Sladen Trust Expedition to the Indian Ocean in 1905. It was at first
intended that Mrs. A. Gepp should work out the collection, but she handed it over
to me, and illness prevented her joining in the work, which we would only have been too
glad to carry out together. . She had, however, sorted the material before sending it
to me.

The collection does not contain all the Rhodophycee collected by Mr. Stanley Gardiner ;
the Lithothamnia have been worked out by the late lamented Mr. F oslie. The literature
on the algz of this part of the Indian Ocean is scarce, for besides the paper by Mr. Foslie
and the Report on the СШогорвусее and Pheophyce of the ‘Sealark’ Expedition by
А. Gepp, M.A., F.L.S., and Mrs. E. Gepp, I only know of one paper in which more than
a few algæ of this region are mentioned. Mr. Reinbold, in “ Meeresalgen der Deutschen
Tiefsee-Expedition," mentions 26 species of Rhodophyceze from Diego Garcia and Mahé,
10 of which have been also collected by Mr. Stanley Gardiner. Besides this paper of.
Mr. Reinbold, I have found two other short references to alge from Diego Garcia and
the Seychelles. Mr. W. B. Hemsley adds a note on Algs to his “ Report on the
Vegetation of Diego Garcia.” Of the alge named only Dasya indica, J. Ag., has any
interest for us at present; it is the only Red alga mentioned. A note by “J. B." in
‘Knowledge’ treats only of Мухорһусеге, and may therefore be passed over in silence.

The “белігі” collection had special interest for me because I am just publishing,
with the aid of Mr. Reinbold, the first part of a list of alge collected during the
* Siboga' Expedition in the Malay Archipelago. This part contains the Myxophycem,
Chlorophycee, and the Phseophyces; the second part, containing the Rhodophycee,
will I hope, follow soon. Тһе study of a region so near to the Malay Archipelago
afforded me great interest from a phyto-geographical standpoint. It had, however,
one drawback ; some species found by me in the ‘Siboga’ collection and recognized as
new are also found in the Stanley Gardiner collection, and the faet that the paper on
the algze of the Indian Ocean will appear before the Siboga paper, will diminish
the number of novelties in the latter. But this could not be avoided, and I have
described the new species as fully as I could; only in one case, when the barren,

small fragment could not possibly have been recognized apart from the fertile
specimens of the ‘Siboga’ Expedition, have I referred to the forthcoming paper.

SECOND SERIES,— BOTANY, VOL. VIII. E

106 MRS, А. WEBER-VAN BOSSE ON THE

Amongst the alge of the Indian Ocean I found species of two new genera, Tapeino-
dasya and. Oligocladus, described for the first time from the Malay Archipelago in * Les
Annales du Jardin botanique de Buitenzorg,’ уо]. ix. 1910. It afforded me great pleasure
to describe these species, which confirm the validity of my new genera. One genus,
Pseudendosiphonia, is new in the present collection; it is most closely allied to Endo-
siphonia, and yet is so different that it can be easily distinguished ; one species, Dasya
indica, had to be described ав a new genus—Amphisbetema.

Many of the alge are barren, and can therefore not be specifically named; others are
only fragments. I have given the generic names of all, because I think that, for our
knowledge of the geographical distribution, even a generic name may have some value.
The collection is rich in novelties; I often hesitated before describing a given alga as
new, but when, after careful investigation, I could not identify it with any existing
diagnosis, I felt I had no alternative. Of the novelties, which include one genus,
17 species, one variety and one form, 11 are from deep water, 2 from reef and deep
water, 7 from reef alone. It may be well to recall here that Mr. and Mrs. Gepp
found Avrainvillea amadelpha in material from reef and deep water. Мапу alge appear.
to accommodate themselves to both modes of living. The ‘Sealark’ Expedition had,
like the * Siboga’ Expedition, the advantage of the use of a dredge, and several hauls
from deep water, where many Rhodophycez flourish, have given a rich harvest. The
nature of the bottom surrounding the reefs and islands, whence these algze come, has
been treated in detail in the above-mentioned paper by Mr. Foslie. I think it therefore
unnecessary to do this again.

The geographical distribution shows that a great resemblance exists between the
algological flora of the Indian Ocean and that of the Malay Archipelago as well as that
of the east coast of Africa. Of the 26 species found at Mahé and Diego Garcia by the
“ Deutsche Tiefsee-Expedition," 7 occur also at Dar-es-Salaam, and of the 8 other
species collected at Dar-es-Salaam by that expedition 5 are found in the Stanley
Gardiner collection *. But it is a puzzle to know how CladAymenia oblongifolia and
Calliblepharis prolifera, only known hitherto from New Zealand, and Chondria pusilla,
only known from the West Indies, came to be inhabitants of the present region. 1%
emphasizes the fact that the geographical distribution of alge is a factor that should
always be very carefully considered. The finding of Peyssonnelia Harveyana, known
only from the Atlantic Ocean at Brest, puzzled me less, because it may be possible that
the determination of the barren plant is erroneous. I have therefore queried it.

I have given a list of the localities with the algze found at each, and a general list of
all the Rhodophycez collected by Mr. Stanley Gardiner. The systematic observations
on the different genera and species conclude the paper.

I wish here to express my sincere thanks to Mr. Stanley Gardiner for entrusting me
with his collection, and to Messrs. Gepp and Cotton for the kind help they have given
me; and I may perhaps add a word of deep-felt regret that Mrs. Gepp was unable to do
her share of the work,

* The number of species common to both regions is probably far greater, for in the recently published paper by

Br. Schröder, “ Zellpflanzen yon Ost-Afrika,” Hedwigia, Bd. lii., four more species are mentioned, also found in the
present collection, 3 : |

MARINE ALGÆ OF THE ‘SEALARK’ EXPEDITION. 107

II. List of Localities and Names of the Alg found at each of them.
CHAGOS ARCHIPELAGO: DIEGO GARCIA, EGMONT, SALOMON, PEROS (ÎLE DU Соту).

1. Gelidium crinale (Turn.), Lamx.; 2. Gelidiopsis rigida (Vahl), Web. v. В.;
3. Hypnea pannosa, J. Ag.; 4. Laurencia obtusa (Huds.), Lamx.; 5. L. spinulifera,
Kütz.; 6. L. pygmea, n. зр.; 7. Chondria pumila, Vick.; 8. Herposiphonia prorepens
(Harv.), Schm. ; 9. Dasyopsis Ферри, n. вр.; 10. -D. palmatifida, n. sp.; 11. Dictyurus
purpurascens; Bory ; 12. Antithamnion adnatum, J. Ag.; 13. Ceramium cinnabarinum
(Gratel.), Hauck; 14. Reinboldiella Schmitziana (Reinb.), de Toni; 15. Peyssonnelia
calcea, Неудт.; 16. Melobesia callithamnioides, Falk.; 17. Amphiroa Sragilissima
(Linn.), Lamx.; 18. Jania tenella, Kitz.

COETIVY, REEF.

1. Chantransia Liagore,n.sp.; 2. Liagora orientalis, J. Ag.; З. L. hawatiana, Butt. ;
4. Gelidium pannosum, Grun.; 5. Champia compressa, Harv.; 6. Laurencia obtusa
(Huds.), Lamx.; 7. L. papillosa (Forsk.), Grev., f. australica in herb. Kütz.; 8. L. cey-
lanica, J. Ag.; 9. Oligocladus Prainii, n. sp.; 10. O. pusillus, n. sp.; 11. Dasyopsis
aperta, п. зр.; 12. D. беррй, п. sp.; 13. D. palmatifida, n. sp.; 14. Heterosiphonia
spec., fragment ; 15. Dictyurus purpurascens, Bory; 16. Jania tenella, Kitz.

SEYCHELLES ARCHIPELAGO (various localities).

1. Sarconema furcellatum, Zan.; 2. Gelidiopsis variabilis (Grev.), Schm. ; 3. Coral-
lopsis Cacalia, J. Ag.; 4. Hypnea Valentie, Turn. ; 5. H. pannosa, J. Ag.; 6. Champia
spec. ; 7. Laurencia papillosa (Forsk.), Grev., f. australica in herb. Kütz.; 8. Acanthophora
orientalis, J. Ag.; 9. Roschera (Tolypiocladia) glomerulata (Ag.), n. comb., fragment ;
10. Peyssonnelia coccinea, J. Ag.; 11. P. biradiata, n. sp.; 19. Jania tenella, Kitz.

Манв, Cap TERNE, REEF.

1. Dermonema dichotomum, Harv.; 2. Galaxaura fastigiata, Decne. ; 3. Gelidiopsis
rigida (Vahl), Web. v. B.; 4. Corallopsis Cacalia, J. Ag.; 5. Hypnea pannosa, J. Ag.;
6. Acanthophora orientalis, J. Ag.; 7. Amphiroa fragilissima (Linn.), Lamx.; 8. Jania
tenella, Kitz.

PRASLIN, REEF.
1. Galaxaura Liebmanni (Aresch.), Kjellm. ; 2. Gelidiopsis rigida (Vahl), Web. v. B. ;
3. Hypnea pannosa, J. Ag.; 4. Polysiphonia spec., fragment; 5. Amphiroa foliacea,
Lamx.; 6. Jania tenella, Kütz.

AMIRANTE.

1. Galaxaura veprecula, Kjellm.; 2. Iridea зрес.; З. Polycelia van Hoevellii, n. sp.,
fragment; 4. Gracilaria compressa (Ag.), Grev.; 5. Calliblepharis prolifera (Harv.),
J. Ag.; 6. Hypnea spec.; 7. Champia compressa, Нату.; 8. Celarthrum | Albertisii
(Picc.), Bórg.; 9. Hypoglossum spec. (spathulatum, Kiitz.?); 10. Nitophyllum spec.,

R2

108 MRS. А. WEBER-VAN BOSSE ON THE

fragment; 11. Laurencia ceylanica, J. Ag.; 12. Pseudendosiphonia Gardineri,
п. g., п. зр.; 13. Cladhymenia oblongifolia, Hook. et Harv.; 14. Dasyopsis Geppii,
п. sp.; 15. D. palmatifida, n. зр.; 16. Tapeinodasya Ethele, п. sp. ; 17. Dictyurus pur-
purascens, Bory; 18. Griffithsia spec., fragment; 19. Cryptonemia seminervis, J. Ag.;
20. Peyssonnelia polymorpha (Zan.), Schm., var. Gardimeri, п. var.

ALDABRA (collected by J. С. Е. Fryer).

1. Galaxaura hawatiana, Butt.; 2. Hypnea divaricata (1), Grev.; 3. Н. pannosa,
J. Ag.; 4. Laurencia nidifica, J. Ag.; 5. L. obtusa (Huds.), Lamx.; 6. L. papillosa,
(Forsk.), Grev., f. australica in herb. Kütz.; 7. Chondria simpliciuscula, n. sp.;
8. Amphisbetema indica (J. Ag.), n. n.; 9. Spyridia filamentosa (Wulf.), Harv., frag-
ment; 10. Halymenia polyclada, А. & E. S. Gepp, var. aldabradensis, n. var.

SAYA DE MALHA.

1. Galaxaura fastigiata, Decne., fragment; 2. G. obtusata (Sol.), Lamx., fragment ;
3. Eucheuma Cottonii, n. sp. ; 4. Fauchea microspora, Born.; 5. Gloiodermo (!) expansa,
n. sp.; 6. Chylocladia perpusilla, n. sp.; Т. Dictyurus purpurascens, Bory; 8. Crypto-
nemia seminervis, J. Ag.; 9. Cryptonemia spec.; 10. Peyssonnelia coccinea, J. Ag.;
11. P. Harveyana (1), Crouan; 12. Cruoriopsis cruciata, Duf.

CARGADOS CARAJOS.

1. Gloiophlæa articulata, n. sp.; 2. Iridea spec., fragment; 3. Eucheuma Cottonii,
n. Sp.; 4. Dasyopsis Stanleyi, n. sp.; 5. D. Geppii, n. sp.; 6. Heterosiphonia Rendlei,
n. sp. ; 7. Dictyurus purpurascens, Bory ; 8. Haloplegma Preissii, Sond. ; 9. Ceramium
spec.

ПІ. Systematic Survey of _ the Algz collected by Mr. J. Stanley Gardiner

the Indian Ocean.
ко | У d. салын я
ЕЕЕ ОЕ ;
FAMILY. | LOCALITY. Эсей: DEPTH DISTRIBUTION,
|
HELMINTHOCLADIACEZ,
1. Chantransia Iiagoræ, п. sp. ............ Coetivy Reef.
1. Liagora orientalis, J. Ав. ............... Coetivy ӨБ. Indian Ocean, Ceylon.
2. Liagora hawatiana, Butt. ............... Coetivy Ro ты с Pacific Ocean, Laie Point, Koolauloa,
Oahu,
1. Dermonema, dichotomum, Harv. ......... Mahé, Terné НӨ 2222. Ceylon, Malay Archipelago, Formosa.
CHATANGIACES.
E oe (2) articulata, n. вр. ......| Cargados Carajos. | ....... 30-45 fms
ura Liebmanni (Ar.), Kjellm....| Praslin. Bes — — | |» Gulf of Mexico.
2. Pete watiana, Butt. ............ Aldabra Passe Но | = Pacific Ocean, French Frigate Shoal,
reau, outside Bird Island,
3. Galazaura fastigiata, Decne. ............ Mahé, Terné. Reef. Е. Pacific Ocean, Philippines ; Malay Archi-
: Saya tie Malba. D] з р. Over 26 :
4. Galaxaura veprecula, Kjellm. ......... WAR Е 20-25 fms Indian Ocean, Madagascar.
5. байағамға obtusata (Sol.), Lamx. Paya de Mala. с 1 26 fms. Warm Atlantic, West Indies; Pacific
Ocean, Choo Isl., Japan.

MARINE А10 Ж OF THE ‘SEALARK’ EXPEDITION.

109

REEF OR
FAMILY. LOCALITY. Ladoo. DEPTH. DISTRIBUTION.
GELIDIACEZ,
1. Gelidiwm crinale (Turn.), Lamx .| Diego Garcia. Roe | 8 Mediterranean ; Red Sea.
2, Gelidium pannoswm, Grun. ............ Coetivy. Boos Up oa Ceylon, Friendly Islands, Fiji, Tonga,
and Samoa Islands
GIGARTINACEZ.
А Cargados Carajos. | ......... 2 fms.
5; HMM ID. т 1 Азаа, ус | eed -25
1. Polycelia van Hoevellit, n. вр. ......... СО ада 30-100 fms. | Malay Archipelago.
RHODOPHYLLIDACEZX.
1. Sarconema furcellatum, Zan. ............ Seychelles, Long 181/ Reef., | eee Red Sea; Indian Ocean
1. Eucheuma, Cottoni, n. sp. ............... Cargados Carajos. | ..... 26 fms. Many species of Buchouma occur on
Saya de Malha, | ..... 25 fms. the east coast of
SPHJSROCOCCACEX.
1. Gelidiopsis variabilis (Grev.), Schm. ...| Seychelles, Long Isl. REGN та Indian Ocean, east coast of Africa.
Seychelles, Praslin
2. Gelidiopsis rigida (Vahl), Web. v. B.4 | Diego Garcia. Reef. Tropical seas.
Mahé, Terné.
1. Gracilaria compressa (Ag.), Grev. ...... дайы us lo а 25 fms ise Ocean; Mediterranean; Gulf of
1. Corallopsis Cacalia, J. Ag. ............ Mahé, Terné. Нег osa š Bed ih. east coast of Africa.
Беус chelles, Long Isl.| Reef.
: Calliblepharis prolifera (Harv.), J. Ag. Amirante. | оси 25 fms. New Zealand.
1. Hypnea divaricata (2), Grev. ............ та. Legoon. | . wi» Gulf of Mexico; Pacific, Tongatabu.
2. Hypnea Valentiz im y Won. o ç. Seychelles, Long Isl.) Вее, |... Red Sea; Indian Ocean; Port n.
3. Hypnea раптова, J. Ас................... Chagos Are о — To Fee Coast of Traed, Mauritius; Ceylon ;
dabra Reef. Tonga Archipelago; New Ca ledonia.
Seychelles Reef.
slin. Reef.
Mahé, Terné Reef.
4. Hypnea spec. Bol — [| 9 25 fms.
RHODYMENIACEX.
1. Fauchea microspora, Born. ............... Saya de Malha. | ......... 53 fms. Mediterranean.
1. Gloioderma (P) expansa, n. Sp. ......... Saya de Malha. | 24... 53-55 fms à
1. Champia ri do Harv. кй n Amiante. А 20-25 fms. | Capeof Good Hope; Malay Archipelago ;
Coe Reef. Pacific Ocean.
2. Champia spec. ы Long Isl. Reef.
1. Chylocladia perpusilla, n. Sp. ......... ya de Mala. | .... 29 fms. T
1. Celarthrum Albertisii (Pice ва ЯТА оо 20 fms. Canary Islands; Bermudas; Malay
Archipelago.
DELESSERIACEZ.
1. Hypoglossum spec. ........................ Ame. 5 re 25 fms.
1. Nitophyllum spec. АЕ 15 ry nn 25 fms.
RHODOMELACEX.
1. Laurencia nidifica, J. Ag. ............... Aldabra Bw. i Sandwich Islands; Malay Archipelago.
Aldabra Mediterranean; Fiji Islands; Malay
2. Laurencia obtusa (Huds.), Lamx. ...4 | Coetivy. Reef. { Archipelago.
Egmont.
3. Laurencia papillosa f. australica ...... Aldabra. к тн New Caledonia.
ivy. Reef.
Seychelles, Long Isl.| Reef. :
4. Laurencia spinulifera, Kütz. ............ s Arch., м 1 Indian Ocean.
1%.
5. Laurencia рудтга, n. Sp. ............. Chagos Arch., Diego Reef.
ia, ;
6. Laurencia ceulanica, J. Ag. ............ Loo uw no м 20 fms. Ceylon; Malay Archipelago
Reef. Zanzibar Archipelago; №
1. Acanthophora orientalis, J. Ag. .........| Mahé, Tow Ж |] o š ,
chelles, Long Isl. Reef Australia ; Tonga Archipelago ; |
сы чш 7 Marianne |
1. Pseudend һ Gardineri. xsp|Amirnte | 1 --- 25 fms.
1. Cladh аа ане алай ат Ааа |... 25 fms. New Zealand. |
1. Chondria pumila, Viek. .................. Chagos Arch., Diego ро bados. |
2. Chondria simpliciuscula, n. вр .| Aldabra. Reef.

110

MRS.

REEF OR

A. WEBER-VAN BOSSE ON THE t

FAMILY. LocALITY. Liao: DEPTH. DISTRIBUTION.
RHODOMELACE® (con.).
1. Polysiphonia Praslin. Reef. [pelago.
1. — gimerulata (Ag. ), п. қарыны .| Seychelles, Long Isl. Reef. |0222. Zanzibar; Indian Ocean; Malay Archi-
Е т сов h. uy sŠ goa Bay; Western Austra lia.
" Oligétladws Рида. п, sp. м Coetivy Reef,
2. Oligocladus pusillus, n. вр. ............ Coetivy. Reef.
1. Dasyopsis Stanleyi,n.sp. ............... Cargados Carajos. | ......... 30 and 47 fms.
3 Coetivy. Reef.
2. Dasyopsis aperta, n. sp. ............... { с Зов Carajoas | ...— 47 fms.
АШАР. с l Ж и ты шл. 20-25 fms.
; s: Cargados Carajos. | 2.2... 25 fms.
3. Dasyopsis Geppii, п. sp. ............... Coetivy. Reef.
Chagos Arch
: АРИВ r. he ho, 20-25 fms. | Malay Archipelago.
4. Dasyopsis palmatifida, n. sp. .........4 | Coetivy. Reef.
Chagos Arch
т Tapeinodasya ЕЙ ах, п. зр. ............ Ани: 1. ee ae 20-25 fms.
1. Heterosiphonia Rendlei, п. зр. ......... Cargados Carajos. | ......... 47 fms.
2, Heterosiphonia spec. ivy. Reef.
1. Dictyurus purpurascens, Bory ...... Saya de Malha. 55 fms. Cape Comorin ; Indian Ocean, Mauritius,
Coetivy. Reef. Ceylon; Malay Archipelago; Diego
Amirante. 30 fms. Garcia.
Cargados Carajos. 30 and 47fms.
Chagos Arch., 22 fms.
Salomon.
1. Amphisbetema indica (J. Ag.) Aldabra Reef. Diego Garcia.
CERAMIACEX
BONN QNEM. ы... Amano р. 20-95 fms
1. Haloplegma Preissii, Sond. Cargados Carajos. | |2... 20, 30, and | Western Australia, Tasmania; Malay
fms Archipe
1. Antithamnion adnatwm, J. Ag. ......... omon В New Zealand; Malay Archipelago
1. Spyridia flamentosa (W (Wulf.), Harv. ...| Aldab: Nw — ] —. Mediterranean; Red Sea; Dar-es-Salaam;
Malay pelago ; West Indies
1. Ceramium cinnabarinum(Gratel),Hck.| Salomon. || ..... |. Mediterranean; Malay Archipelago.
2. Ceramium Cargados Carajos. | 22... 47 fms.
1. Reinboldiella Bhetstona bees ) |Chagos Arch, Eg-| Ref. 1 -... Japan; Malay Archipelago.
Toni| mont, Salom
GRATELOUPIACEJX.
1. Halymenia prete А. & E. S. Gepp,| Aldabra. Outer reef,
aldabradensis, n. v.
Алаа ы т 20-25 fms. | Atlantic; Mediterranean; Red Sea.
1. Cryptonemia seminervis, J. Ag. . =| Saya йе Маша. [| _..... 55 bas.
2. Cryptonemia spec. ........................ Saya de Malha: |. 55 fms,
SQUAMARIACEX.
1. Peyssonnelia coccinea, J. Ag. ............ Seychelles. ухо an 31 fms. West coast of Australia; Malay
Saya de Malis. | ........, 55 fms. Archi ; :
2. Peyssonnelia Harveyana (2), Crouan .. | Saya de Malba. | ...... 55 fms. Coast of France near Brest.
3. Peyssonnelia calcea, Heydr. ............ gmont Re о I oa x Tami, near German New Guinea; Malay
У Регов, Coin. Archipelago,
4. Peyssonnelia polymorphaf.Gardineri, Amirante. § | ........ 30, 45-60
n. f. fms.
5. Peyssonnelia biradiata, n. sp............. уема з |] aan 31 fms. Malay Archipelago,
6. P. spec. ya de Malha.
1. Cruoriopsis cruciata ©, Dut HERD Saya де Mala. | ........ ç 29 fms. Adriatic ; Malay Archipelago.
CORALLINACER.
E Melobesia Wii ides, Falk. ...... Egmont. [ ВЕ Atlantic (Р); Mediterranean.
1. Amphiroa fragilissima, Талак. ........ Chagos Arch, AE C ы ср West Indies; Pacific ; Malay Archi- |
Es pe ; Mahé, Сар Terné. | Reef. pelago ;-
2. Amphiroa foliacea, Lamx. Pac | Praslin. Reef. Marianne Islands ; у
1. Jania tenella, Иш. |... Coetivy. EM |. D Malay Archipelago ; Gulf of Mexico,
Seychelles, Reef. We
Mahé, Cap Terné. | Reef.

MARINE ALGJE OF THE ‘SEALARK’ EXPEDITION. 111

ТҮ. Systematic Account.
Fam. HELMINTHOCLA DIACE.&.

CHANTRANSIA (DC.), Schmitz.
1. CHANTRANSIA LIAGORA, n. Sp.

Thallo endophytico et epiphytico, nano, czespitibus globosis sparsis Liagorze insidente,
constante e filis hospitem intrantibus et filis erectis. Filis erectis parce ramosis,
ramis post monosporangia terminalia delapsa egredientibus; pilis non visis,
verisimiliter raris. Monosporangiis aut terminalibus aut lateralibus, pedicello
ramoso suffultis, ramis denuo monosporangia ferentibus. Filis latis 6-8-12 и,
articulis 16-20 и longis. Monosporangiis non maturis; antheridiis et carpogoniis
non visis.

Coetivy, reef, on Liagora hawaiiana ; in alcohol.

This little Chantransia forms a small tuft, which penetrates between the horizontal
filaments of the Liagora and attaches itself at their base. Т could detect no creeping
filaments in the host plant giving rise to new tufts. It has а height of 1 mm., the axes
are sparingly and irregularly branched and bear monospores both at the apex and
laterally. After the monospore has fallen off, or perhaps while still in situ, the filament
or branch that carries it grows out sideways. I also noted clusters of carpospores ;
these are not arranged in series as in C. efflorescens. Тһе filaments have a breadth of
6-8, rarely of 12 и, and their cells a length of up to 20,4. On account of the alge
having been first preserved in formalin and afterwards in alcohol, it is difficult to judge
the shape of the chromatophore.

I believe that this alga is a new species; there are, so far as I know, only two other
species of Chantransia on Liagora, namely, Ch. Barbadensis and Ch. Nemalionis, which
were found on Liagora elongata by Miss Vickers. Our alga differs from both species
in the smaller size of its cells and filaments.

Liacora, Lamouroux.
1. LIAGORA ORIENTALIS, J. Ag.
J. Agardh, Analecta algologica, 1896, p. 99.
Coetivy, reef; in alcohol.
Distribution. Indian Ocean, Ceylon.

2. LIAGORA HAWAIIANA, Butt.
Butters, “ Liagora and Galazaura," Minnesota Bot. Studies, vol. iv. part 2, 1911, p. 169.
Coetivy, reef; in alcohol.
Distribution. Laie Point, Koolauloa, Oahu, Pacifie Ocean.
Liagora hawaiiana belongs to the group of Liagore with dichotomously branched
fronds. It differs from all known species of this group in having an articulate frond ;
at the top of the thallus the articulations are inconspicuous, but lower down they are

112 MRS. А. WEBER-VAN BOSSE ON THE

very distinct. The articulations are at the base of each dichotomy and the internodes
are as a rule long.

DERMONEMA, Greville.
1. DERMONEMA DICHOTOMUM, Harv.
Harvey, Ceylon Alga, n. 93. | i
Schmitz u. Hauptfleisch in Engler u. Prantl, Natiirl. Pflanzenfam., Teil 1, Abt. 2, 1896, p. 335.
De Toni, Syll. Alg. vol. iv. sect. i. 1897, p. 102.
Terné Mahé; dry specimens.
Distribution. Ceylon; Formosa; Malay Archipelago.

Fam. CHATANGIACES. |

GroroPHLGA, J. Agardh.
1. GroIOPHLGAA (P) ARTICULATA, n. sp. (Plate 12. fig. 1; Plate 14. figs. 26, 27.)

Frondibus dichotomis, articulatis, usque ad 18 ст. altis, ramulos adventitios ex nodis
orientes gerentibus; internodiis cylindraceis, subcompressis, ad basin contractis;
internodio terminali apice rotundato. Frondibus constantibus e cylindro centrali
filis hyalinis, tenuibus, peripheriam versus excurrentibus contexto. Filis per
dichotomiam divisis, stratum periphericum cellularum minimarum moniliformium
arcte conjunctarum et cuticula crassa coopertarum formantibus. Organis
generationis ignotis. | |

Cargados Carajos, dry specimen, 30 fms.; in alcohol, 45 fms.

The algz from Cargados Carajos are barren, the determination is therefore not beyond
doubt, but the anatomical structure of these plants is so typical and so like that of the
family Cheetangiacee (РІ. 14. figs. 26, 27) that I should be very much surprised if the
organs of fructification should prove these algæ to belong to another family. Amongst
the Chzetangiacez they stand nearest to Gloiophlea. They are distinguished, however,
from Gloiophlea scinaioides—the only known species of the genus—by their articulate
frond ; the internodes are usually—though not always—contracted into a short pedicel-
like node, and from this node adventitious branches may spring, but as a rule the
branching is dichotomous. The peripheral layer consists of small cortical cells, covered
by a rather thick cuticle; these small tightly adhering cortical cells distinguish the
genus Gloiophlea from the genus Scinaia.

GALAXAURA, Lamouroux.
1. GALAXAURA LIEBMANNI (Aresch.), Kjellm. :
Areschoug, Phyce nove, in Act. Reg. Soc. Scient. Upsal. ser. 3, vol. i. 1854, p. 356.
Kjellman, “ Floridé Slagtet Galazaura,” in Kongl. Sv. Vet.-Akad. Handl. Bd. xxxiii. 1900, p. 46.
Praslin, reef; in alcohol. |.
Distribution. Mexico; Vera Cruz. _. г.
_ Though Galazaura Liebmanni is only known from the West Indian tropical seas, the
peculiar structure of the short cortical filaments and the compact outward appearance

MARINE ALG OF THE ‘SEALARK’ EXPEDITION. 113

of the alge from Praslin seem to indicate that they belong to G. Tiebmanni rather
than to С. соЙабепв from Australia, described by Kjellman as “solitaria, erecta et
robusta." Тһе plants from Praslin are short, intricate, and grow close together; the
short cortical filaments consist, as a rule, of four cells only, and of these the lower
one is by far the largest.

2. GALAXAURA HAWAIIANA, Butt. (Plate 18. fig. 25.)
Butters, “ Liagora and Galaxaura,” Minnesota Bot. Studies, vol. iv. part 2, 1911, p. 176.

Aldabra, reef, outside Passi Honakan.

Distribution. Pacific Ocean, French Frigate Shoal, Bird Island.

The alge from Aldabra are very much like Galaxaura hawaiiana, Butt., and the
differences between the description of С. hawaiiana (I have seen по type specimen) and
the alge from Aldabra are not, I think, sufficiently great to warrant the description of à
new species. Тһе alge from Aldabra are a little longer (they have a length of 7 cm.,
a breadth of 2 mm.), and the long internodes are not tapering, but rounded or
truncate. Тһе frond is slightly but distinctly annulate, and in the dried specimens of a
remarkable dark brown-red colour; the anatomical structure is like that of С. hawaiiana,
Butt. Butters records б. hawaiiana from depths of 14-16 and 26 fathoms; the
specimens from Aldabra were found at the outside of Passi Honakan. This difference
of locality may perhaps account for the small differences observable in the two plants.
I have given a figure of the Aldabra plant, which shows very well the long terminal

internodes.

3. GALAXAURA FASTIGIATA, Decne.
Decaisne, Sur les Corallines, 1842, р. 16.
Kjellman, ** Floridé Slágtet Galaxaura,” in Kongl. Sv. Vet.-Akad. Handl. Bd. xxxiii. 1900, p. 64.
Terné Mahé ; dry specimen.
Saya de Malha; fragment only. |
Distribution. Pacific Ocean; Philippines; Malay Archipelago.

4. GALAXAURA VEPRECULA, Kjellm.
Kjellman, “ Floridé Sligtet Galazaura,

Amirante, 20-25 fms.; in alcohol.
Distribution. Indian Ocean, Madagascar.

» in Kongl. Sv. Vet.-Akad. Handl. Bd. xxxiii, 1900, p. 80.

5. GALAXAURA oBTUSATA (Sol.), Lamx.
Solander in Ellis, Zoophytes, 1786, p. 113.
Lamouroux, Polypiers coralligénes, 1816, p. 262.
Agardh, Epicrisis systematis Floridearum, 1876, p. 525.
Saya de Malha, 26 fms.; only a small specimen.
Distribution. Warm Atlantic, West Indies; Pacifie Ocean, Loo Choo Islands,
Japan.
SECOND SERIES.— BOTANY, VOL. VIII.

114 МЕЗ. А. WEBER-VAN BOSSE ON THE

Fam. GELIDIACE.E.
GELIDIUM, Lamouroux.

1. GELIDIUM CRINALE (Turn.), Lamx.
De Toni, Syll. Alg. vol. iv. sect. t. 1897, p. 146.
Fucus crinalis, Turner, Hist. Fuc. tab. 198. .
Chagos Archipelago, Diego Garcia, in alcohol; no indication of depth.
Distribution. Mediterranean ; Red Sea.

2. GELIDIUM PANNOSUM, Grun.
Grunow, Algen der Fiji, Tonga u. Samoa Inseln, 1874, p. 40.
Coetivy, reef, on a piece of coral; in alcohol.
Distribution. Upolu, Samoa Archipelago ; Ceylon; Friendly Islands.

Fam. GIGARTINACELE.

Iripma, Bory.
1. IRIDZA spec.

Cargados Carajos, fragment with cystocarps, іп alcohol, 22 fms.

Amirante, fragment with tetraspores, in alcohol, 20-25 fms.

The specimens from both localities are too fragmentary to allow of a specific
determination, though the tetraspores being scattered among the peripheral cells
indicate that the plant belongs to the subgenus Rhodoglossum, J. Ag. Both specimens
are so alike in anatomical structure and outward appearance that, as far as can be
judged from the small fragments, I feel almost certain that they belong to the same
species.

Рогхсалда, J. Agardh.

1. Рогхсатла VAN H@VELLII, n. sp.

Amirante, 30-100 fms. ; in alcohol.

Distribution. Malay Archipelago.

The species is represented in the collection by such a poor and barren fragment that
I should have been unable to name it, had it not been for my studies on the rich
material collected during the ‘Siboga’ Expedition in the Malay Archipelago. This alga
will therefore be described and figured in the forthcoming paper on the * Siboga" algze.

Fam. RHODOPHYLLIDACE.
SARCONEMA, Zanardini.

l. SARCONEMA FURCELLATUM, Zan.
Zanardini, Plante Mar. Rubro Enum. р. 264.
De Toni, Syll. Alg. vol. iv. sect. 1. 1897, p. 367.

MARINE ALGÆ OF THE ‘SEALARK’ EXPEDITION. 115

Seychelles, Long Island.
Distribution. Red Sea; Indian Ocean.

EvcHEUMA, J. Agardh.

1. EUCHEUMA Соттохп, п. sp. (Plate 12. fig. 2.)

Thallo crasso, carnoso, depresso, decumbente, lateraliter et subdichotomice ramoso,
latere superiore verrucis et tuberculis aggregatis, latere inferiore tuberculis remoti-
oribus obsito. Verrueis nonnunquam forma irregulari; tuberculis vulgo parvis et
rotundis, nonnunquam apiculatis. Regione mediana thalli ad apices e cellulis satis
magnis hyphis circumdatis constante, hyphis infra item auctis et cellulas satis
magnas formantibus; peripheriam versus cellulis diminutis et ibidem series breves
formantibus. Organis fructificationis ignotis.

Saya de Malha, 25 fms. ; in alcohol.

Cargados Carajos, 26 fms.; dry specimen.

E. Cottonii is unfortunately a barren plant, and shows, in a dry state, a great likeness
to species of Gracilaria. I moistened it, to restore as much as possible its natural
appearance, and was then struck by its great resemblance to Zucheuwma. Тһе figures
are made from these moistened specimens, and show well the horizontally spread,
laterally branched thallus, covered with warts and more or less pointed tubercles.
The anatomical structure at the apex is also entirely different from that of Gracilaria,
though, in a cross-section at some distance from the apex, only round, rather large, and
somewhat thick-walled cells are to be seen. But these cells are of different origin to
the primary large cells, for at the apex the latter are seen to be surrounded by narrow
hyphee, which grow out lower down into large cells and are then no longer discernible
as hyphe.

Тһе genus Zucheuma is described by Schmitz in Engler u. Prantl, ‘ Pflanzen-Familien,
as having a central axis of slender elongated hyphe. This axis is found, for instance,
іп Ё. spinosum. Agardh has, however, called attention to the fact that Е. Schrammi
has a structure in which “fila strati interioris sparsim inter cellulas strati intermedii
quasi introducta obveniunt.” Schmitz, again, supposes that E. Schrammi, Crouan, is
not an Eucheuma but an Luryomma,

My studies on the Eucheumata of the Malay Archipelago seem to indicate that there
are species of Eucheuma without a central axis of longitudinal hyphe. Тһе investi-
gations are not complete, but it appears to me not improbable that the cylindrical erect
species have an axis of narrow cells, and that in the flat prostrate species this is absent.
I did not find an axis іп any prostrate specimens of Eucheuma of which I was able to
study the cystocarp. The whole question needs to be further investigated, but for
the present it seems highly probable that the algze from Saya de Malha and Cargados
Carajos belong to the genus Eucheuma, and the name 1 propose for it is in kind
remembrance of Mr. Cotton's friendly help.

116 MRS. А. WEBER-VAN BOSSE ON THE

Fam. SPH HROCOCCACE.

GELIDIOPSIS, Schmitz.

1. GELIDIOPSIS VARIABILIS (Grev.), Schmitz.
Schmitz, Marine Florideen von Deutsch Ost-Afrika, in Engler’s Bot. Jahrb. Bd. xxi. H. 1, 2,
1895, p. 148.
Seychelles, Long Island, reef.
Distribution. Indian Ocean; East coast of Africa.

2. GELIDIOPSIS RIGIDA (Vahl), Web. v. B.
Recueil des Travaux botaniques Neérlandais, no. 1, 1905, p. 9.
Fucus rigidus, Vahl, in Naturh. Sellskabets Skrifter, v. p. 46.

Seychelles, reef ; dry specimen.

Praslin, reef ; dry specimen and in alcohol.

Diego Garcia, reef; dry specimen.

Terné Mahé, reef; dry specimen.

Distribution. Common in tropical seas.

In 1904 I wrote in the above-mentioned paper that it appeared to me that the well-
known Gelidium rigidum of the tropical seas was not a Gelidium, but belonged to the
genus Gelidiopsis. In 1909 Prof. Okamura gave, in his ‘Icones of Japanese Algee,’
pl. lix., a figure of this alga and also of its large apical cell, and named it Gelidium
rigidum. Т had failed to detect this large apical cell, and its absence was one of
the reasons why I thought that the alga could not be a Gelidium; I had seen no
cystocarps. Quite recently, in ‘Icones of Japanese Algz, vol. ii. No. 10, р. 188,
Prof. Okamura has recalled his statement, made before he had seen my paper, and
agrees with me in naming Gelidium rigidum, Gelidiopsis rigida. This alga varies in
its ramification, the lateral branches being short or long, or sometimes unilaterally
inserted. І believe, however, that these variations are only a form of growth.

Gelidiopsis being a feminine substantive in the Greek language, the adjective should
have the feminine ending, and I made an error when I called the alga Gelidiopsis
rigidum in 1904.

GRACILARIA, Greville.
l. GRACILARIA COMPRESSA (Ag.), Grev.
Agardh, Species Algarum, 1823, p. 308.
Greville, Algæ Britann. 1830, p. 125.
De Toni, Syll. Alg. vol. iv. sect. п. 1900, p. 438.
Amirante, 25 fms.
Distribution. Atlantic Ocean ; Mediterranean; Gulf of Mexico.
The specimen is rather slender for б. compressa and scantily branched, but I assume

MARINE АГОЖ OF THE ‘SEALARK’ EXPEDITION, | 117

that the great depth at which it was gathered may in some measure account for
this. It bears cystocarps, and these have a little funnel at the top, exactly as in
G. compressa.

CoRALLOPsIS, Greville.

1. CoRALLOPSIS САСАТЛА, J. Ag.
J. Agardh, Epicrisis systematis Floridearum, 1876, p. 409.
De Toni, Syll. Alg. vol. iv. sect. 11. 1900, р. 459.
Terné Mahé; dry specimens.
Seychelles, Long Island; dry specimens.
Distribution. Red Sea. Г
Тһе specimens of the Seychelles аге more branched than those of Terné Mahé,
still I believe that they belong to the same species; some branches recall those of
С. minor.

CALLIBLEPHARIS, Kützing.

1. CALLIBLEPHARIS PROLIFERA (Harv.), J. Ag. (Plate 12. fig. 3.)
Epicrisis systematis Floridearum, 1876, p. 432.
Rhodymenia prolifera, Harvey, in Flora Nova-Zeelandia, p. 249.

Amirante, 25 fms.; dry specimens.

Distribution. New Zealand.

Harvey described this alga under the name of Rhodymenia prolifera, but J. Agardh
pointed out that the structure of the cystocarp was unlike that of Rhodymenia and
similar to that of Calliblepharis. The alga differs, however, from the other species of
this genus in bearing the cystocarps on the broad part of the thallus and not on special
leaflets; Agardh called it therefore Calliblepharis ? prolifera. Ав the specimens from
Amirante were fertile, I studied the ripe cystocarps; they have the same structure as
those of Calliblepharis, and I do not think that the absence of special leaflets should be
а reason to maintain J. Agardh's note of interrogation.

It is remarkable to find Calliblepharis prolifera in Mr. Stanley Gardiner's collection,
since it had previously only been found in New Zealand.

HYPNEA, Lamouroux.

1. HYPNEA (DIVARICATA P), Grev.
Greville, Alg. Brit. р. lix. `
De Toni, Syll. Algarum, vol. iv. sect. 11. 1900, p. 478.
Aldabra, fragment only, lagoon; dry specimen.
Distribution. Gulf of Mexico ; Mascarenes (?); Tongatabu.
As far as I could judge by thé small fragment, the specimen belonged to H. divaricata,
Grev. It differed from H. Valentie in the absence of small stellate spines.

118 MRS. А. WEBER-VAN BOSSE ON THE

2. НурмЕА VALENTLE (Turn.), Mont.
Cryptog. Сапат. р. 161.
De Toni, Syll. Algarum, vol. iv. sect. 11. 1900, p. 479.
Fucus Valentie, Turner, Historia Fucorum, 1809, tab. 78.
Seychelles, Long Island ; dry specimen.
Distribution. Red Sea, Indian Ocean; Port Denison.

3. HYPNEA PANNOSA, J. Ag.
J. Agardh, Alg. Liebman, p. 14.
Grunow, Algen der Fiji, Samoa u. Tonga Inseln, 1874, p. 39.

Aldabra; dry specimen.

Seychelles, reef; dry specimen.

Praslin; in alcohol.

Chagos Archipelago, Coin Peros, lagoon ; in alcohol.

Terné Mahé, reef; dry specimen.

Distribution. Coast of Mexico; Mauritius; Ceylon; Tonga Archipelago; New
Caledonia.

Among the specimens from Praslin were some overgrown by a sponge; the branches
thus hemmed in were obviously less branched and more cylindrical than the free-growing
ones. The specimens from Terné Mahé are exactly like those from Mexico in Kiitzing’s
herbarium and figured in Tab. Phyc., Band xviii. tab. 27. The specimens from the
other localities are like those gathered by Harvey at the Friendly Islands, No. 44, and
by Schmarda on the west coast of Ceylon. There is a difference between Harvey’s and
Schmarda’s specimens, but both being barren I abstained from further examination.

4. HYPNEA spec.

Amirante, 25 fms.

An indeterminable specimen, altered perhaps a little in consequence of the great
depth at which it grew.

Fam. RHODYMENIACE.E.

FAUCHEA, Montagne et Bory.

1. FAUCHEA MICROSPORA, Born.
Bornet in Rodriguez, Algas de las Baleares, 1889, p. 258, et in Bull. Soc. Bot. t. xxxvii. 1890,
р. 142.

Saya de Malha, 53 fms.; dry specimens.

Distribution. Mediterranean.

The plants are all barren, but the whole outward appearance of these algæ corresponds
so well with the description given by Borret that I feel no doubt that they belong to
this species.

MARINE АТС Ж OF THE *SEALARK' EXPEDITION. 119

GLOIODERMA, J. Agardh.

1. GLOIODERMA (P) EXPANSA, n. sp. (Plate 14. figs. 28, 29.)
Thallo membranaceo, plano, verisimiliter horizontaliter expanso, margine prolifera-
tionibus in thallos juveniles excrescentibus. Thallo seepe anastomosante, constante
e strato unico cellularum maximarum centralium, utroque latere duobus stratis
cellularum parvarum et uno strata cellularum minimarum periphericarum cincto.
Cystocarpiis et antheridiis ignotis ; tetrasporangiis in strato corticali parum mutato
dispersis, cruciatis divisis.

Saya de Malha, 53-55 fms. ; dry specimens.

This alga seems to be new to science, and I have referred it to the family of the
HRhodymeniacee on account of its large central cells placed in one series only. Its
outward appearance recalls Rhodophyllis peltata, Grun., of the Tonga Archipelago, but
it differs from that alga in having only one series of large central cells, whilst all species
of Rhodophyllis have two series. Amongst the Rhodymeniacez it belongs, I think, to
the genus G'loioderma, because it has very small peripheral cells and cruciate
tetrasporangia scattered in the cortical layer, which is, nevertheless, very little
changed.

Unfortunately I have seen neither procarps, cystocarps, nor antheridia, and therefore
the generic position of the alga remains uncertain. It needs to be further investigated,
for it was gathered at a great depth (53-55 fms.), from which few Rhodophycee аге
known.

By its broad, membranous, unbranched frond, which bears excrescences along its
margin, and sometimes, though rarely, on the frond itself, this alga is easily distinguished
from all other species of Gloioderma, as these have mostly a dichotomous or pinnate
ramification. The entirely different mode of growth of the alga from Saya de Malha,
and the great depth at which it was gathered, almost suggest that it may be the type of
а new genus.

CHAMPIA, Desvaux.

1. CHAMPIA СОМРВЕЗЗА, Нагу.
Harvey, Genera South African Plants, ed. 1, p. 402.
De Toni, Syll. Aig. vol. iv. sect. 11. p. 561.
Amirante, 20-25 fms. ; small fragment, in alcohol.
Coetivy, reef; dry specimen.
Distribution. Cape of Good Hope; Ceylon; Malay Archipelago; Friendly Islands;
Australia (?) ; New Caledonia.

2. CHAMPIA spec.
Seychelles, Long Island, reef ; dry specimen. : :
A small specimen intermingled with Gelidiopsis variabilis. The simple main axis
bears a few short branches, irregularly placed. Several main branches have grown

120 MRS. А. WEBER-VAN BOSSE ON THE

together, and cannot be separated without damage. The plant has distinct inflated
articulations blunt at both sides, and is certainly different from Champia compressa. It
strongly recalls Ch. parvula, especially the figure given by Kützing in his Tab. Phyc.,
Band ху. tab. 89, where Ch. parvula is figured under the name of Lomentaria intertexta.
The plant from the Seychelles is, however, very small, having a height of only 1 cm., and
it is barren. 7

CHYLOCLADIA, Greville.

1. CHYLOCLADIA PERPUSILLA, n. sp. (Plate 12. figs. 4, 5.)

ТһаПо minusculo, decumbente, ramuloso, cavo, sine diaphragmate; ramo. pri-
mario interne multis cellulis magnis ; ramulis constantibus e strato unico peripherico
cellularum magnarum et parvarum secundum quarum latus interne fila ramosa
decurrunt, nonnunquam glandulas ferentia. Sporangiis in ramulorum dilatatorum,
cavitatem prominentibus, nec in cavernis propriis sitis. Cystocarpiis et antheridiis
ignotis. Thallo lato 0:5-0:6 m.

Saya de Malha, creeping on Udotea, 29 fms.

This tiny plant was found creeping on Udotea, though it did not appear to possess
rhizoids. It has a main axis, from which spring slender ramuli, sometimes opposite, but
mostly irregular; the specimen is, however, so small that it is impossible to say much
as to its ramification. The main axis has a length of 1 em. and a breadth of 600 u, the
branches have a length of 2-5 mm. The branches are entirely hollow, but the down-
growing filaments that are proper to the genus bear glands (?); the main axis is filled
with large cells, loosely attached to one another. I could see no diaphragms, neither in
the branches nor at the base of the branches; the tetraspores are scattered over the
inflated top of some branches ; they protrude into the cavity of the branch, but they are
not situated in a special hollow, as they are in Lomentaria.

On account of the position of the tetraspores and the absence of diaphragms, I believe
that this little plant belongs to the genus Chylocladia. Its small size characterizes it as
а new species.

C@LARTHRUM, Borgesen.

1. C@LARTHRUM (ALBERTISII P) (Рісе.), Borg.
* Some new or little-known West Indian Florides, 11.,” in Bot. Tidsskr. Bd. xxx. 1910, р. 189.
Lomentaria Albertisii, Piccone, Alghe in “ Crociera del Corsaro,” 1884, р. 37.
Amirante, 20 fms.; dry specimen.
Distribution. Canary Islands; Guadeloupe and Bermudas; West Indies ; Malay
Archipelago.
I think it very probable that the alga from Amirante is Celarthrum Albertisii, because
I found that species amongst the ‘ Siboga’ material I cannot deny, however, that
the specimen has smaller articulations than the type and that I did not see in my

MARINE ALGZ OF THE ‘SEALARK’ EXPEDITION. | 121

preparations the peculiar glands described by Mr. Bórgesen. In how far, however,
these differences may depend upon the depth at which the plant was gathered, I do not
venture to say. Тһе specimen—and it is only a fragment—gives the impression of being
still very young and therefore I wish to be doubly cautious in my opinion as to it.

Fam. DELESSERIACE.

Hypoeiossum, Kützing.
1. HYPOGLOSSUM spec.
Amirante, 25 fms.; dry specimen.
A fragment; probably Hypoglossum spathulatum (Kütz.) J. Ag. (Delesseria hypo-
glossoides, Harvey, Austral. Alg.), but the specimen is too badly preserved to be
identified with certainty.

NITOPHYLLUM, Greville.
1. NITOPHYLLUM spec.

Amirante, 25 fms.
Only a poor fragment, attached to other algæ and indeterminable, but with a

tetraspore-bearing sorus typical for this genus.

Fam. RHODOMELACEA.
Subfam. LAURENCIE..
LAURENCIA, Lamouroux.

1. LAURENCIA NIDIFICA, J. Ag.
J. Agardh, Spec. Alg. ii. 1863, p. 749.
De Toni, Syll. Alg. vol. iv. sect. 111. 1909, р. 785.
Aldabra, reef.
Distribution. Sandwich Islands; Malay Archipelago.

2. LAURENCIA OBTUSA (Huds.), Lamx.
Fssai de classification, p. 42.
De Toni, Syll. Alg. vol. iv. sect. тїт. 1903, р. 791.
Fucus obtusus, Hudson, Flora Anglica, ed. iii. p. 586.
Aldabra, reef; dry specimens.
Coetivy, reef; dry specimen.
Egmont, reef and shoal; in alcohol.
Distribution. Mediterranean ; Fiji Islands; Malay Archipelago.
Laurencia obtusa is represented by two forms in the Stanley Gardiner collection ; from
Aldabra there are three packets (поз. 4, 5, 8). No. 4 is the stoutest plant, no. 5 and
SECOND SERIES.——BOTANY, VOL. VIII. T

122 MRS. А. WEBER-VAN BOSSE ON THE

no. 8 are more slender, while the alga from Coetivy Reef is very much like specimens in
Kützing's herbarium and named L. obtusa уат. racemosa. The specimens from Egmont
Reef are rather short, but they probably represent young plants. They show the charac-
teristic branching of the species, and all the branches have blunt truncated apices.

3. LAURENCIA PAPILLOSA (Forsk.), Grev., f. AUSTRALICA, in herb. Kützing.

Aldabra, reef; dry specimen.

Seychelles, Long Island; dry specimen.

Distribution. New Caledonia.

I have been іп doubt about these alge for they differ from typical specimens of
L. papillosa. They come near to two barren specimens from New Caledonia in
Kützing's Herbarium. Мг. Stanley Gardiner's specimens being also barren, I abstained
from giving а diagnosis.

4. LAURENCIA SPINULIFERA, Kütz.
Kützing, Tab. Phyc. Bd. xv. 1865, tab. 61, p. 22.
Chagos Archipelago, Diego Garcia; in alcohol.
Distribution. Indian Ocean.
The specimens come near to some forms of L. obtusa, but they are distinguished by
their small size. I think that until all the forms of L. obtusa from the East Indies are
better known, we may retain Kützing's species.

5. LAURENCIA РҮСМ ЖА, п. sp. (Plate 12. fig. 6.)

Fronde tereti, filiformi, tenui, cirea 1 cm. alta, pulvinatim esespitosa, intricata ; ramis
primariis decumbentibus, secundariis erectiusculis; ramulis clavatis, truncatis.
Ramis primariis usque ad 250 д crassis, ramis secundariis et ramulis tetrasporiferis -
100-150 u crassis. Cystocarpiis et antheridiis non visis.

Chagos Archipelago, Diego Garcia, no indication of depth ; in alcohol.

Laurencia pygmea is distinguished by its size, which is indeed dwarf for a Lawrencia.
It grows in tufts, and the alcohol material looks as if it had grown on stones or shells
and had been cut off with a knife. Тһе thallus consists of decumbent branches—I have
seen no rhizoids—which give off suberect ascending ones; these are branched irregularly.
Sometimes the ramuli are subopposite, sometimes subverticillate or even single. They
are always slender, and not particularly short as are the ramuli of L. pannosa, Zan.
From this plant, L. pygmaea may also be known by its thalli, which, though intricate,
are easily teased out under a pocket-lens, whereas in the case of L. pannosa this cannot
be done, according to Zanardini, without damaging the frond. The mode of branching
reminds me of L. obtusa, but the thallus is very slender, the main branches having a
diameter of 200-250 » and the side-branches and tetrasporiferous ramuli from 100-150 p.
Characteristic for this small alga are the swollen light-reflecting lateral membranes of
the central cells. From 2. indica, Hauck, it is easily known by its small size.

MARINE ALGZ OF THE 'SEALARK ' EXPEDITION. 123

6. LAURENCIA CEYLANICA, J. Ag.

J. Agardh, Epicrisis Syst. 1876, p. 660.

De Toni, Syll. Alg. vol. iv. sect. 111. 1903, p. 805.
Amirante, at 20 fms.; dried specimen and in alcohol.
Coetivy, reef; in alcohol.

Distribution. Ceylon.

Ë Subfam. СНОХОВТЕХ.
ACANTHOPHORA, Lamouroux.

1. ACANTHOPHORA ORIENTALIS, J. Ag.
J. Agardh, Species Algarum, ii. 1863, p. 820.
De Toni, Syll. Alg. vol. iv. sect. 111. 1908, p. 822.
Bérgesen, “Some new or little-known West Indian Floridee, 1," in Bot. Tidsskr. Bd. xxx.
1910, p. 201.

Terné Mahé; dry specimen.

Seychelles, Long Island; dry specimen.

Distribution. Zanzibar; Malay Archipelago; North Australia; Tonga Archipelago ;
Marianne Islands.

Both specimens are full grown and large, but both are barren, and thus a study of
the stichidia is out of the question. Ina recent paper Mr. Bórgesen has called atten-
tion to the fact that the name of 4. Thierii ought to be changed to A. spicifera and that
the antheridia of 4. orientalis are very much like those of A. spicifera.

PSEUDENDOSIPHONIA, n. 5.

Thallus radicibus parvis substrato affixus et verisimiliter horizontaliter expansus,
constat ex axi principali, ramis lateralibus incremento infinito et ramis dorsalibus
ventralibusque incremento finito. Ad apicem vegetationis axis principalis radiatim
se findit; axis principalis postea etiam inferius bilateralis. Axes laterales eodem
modo quo axis principalis nascuntur. Thallus formatus e cellula centrali et quatuor
cellulis pericentralibus pluribus stratis cellularum corticalium peripheriam versus

Cystocarpia et antheridia ignota. Stichidia in apice

diminutarum cinctis.
Pedicellus

pedicelli monosiphonii, sporam unam in singulis segmentis continentia.
monosiphonius ramificatus ; stichidia nuda.

1. PsEUDENDOSIPHONIA GARDINERI, n. sp. (Plate 13. figs. 13-15; Plate 14. fig. 30.)
Stichidiis sparsis extra axillaribus; ceteroque confer diagnosin generis.
Amirante, 20-25 fms.; in alcohol.
Pseudendosiphonia Gardineri recalls, by its short thorny branches and its stichidia

borne on a monosiphonous stalk, the genus Endosiphonia. It has, however, а different
| т2

124 MRS. А. WEBER-VAN BOSSE ON THE

habit, having a well defined stem and regularly placed lateral branches, whereas Епао-
siphonia has no definite stem nor regular ramification, Anatomically the genus is
characterized by its four big pericentral cells (fig. 30, р.с.), surrounded by several layers
of cortical cells which are always much smaller. In Zndosiphonia the four pericentral
cells are surrounded by cells of the same size.

The stem of Pseudendosiphonia is attached to small flat stones or particles of rock
by means of small rhizoids and around it stand four branches with a divergence of í;
they take their origin from the central cell and are therefore endogenous; the lateral
ones are of unlimited growth (Langtriebe), the dorsal and ventral ones remain short
(Kurztriebe). The lateral branches have the same development as the principal stem,
but their side-branches remain at first short and give the plant a thorny appearance
that strongly reminds one of Endosiphonia. At the top of some branches there are
luxuriously developed trichoblasts (“ leaves,” Falkenberg), but on most they had fallen
off, probably because they have so many side-branches and such large cells.

The stichidia bear a great resemblance to those of Zndosiphonia, particularly to
E. clavigera, for they are placed irregularly on the stem and branches (РІ. 13. fig. 14)
and not in the axils of the latter. They spring from peripheral cells and have a mono-
siphonous pedicel, but this bears branches, which is not the case with Z. clavigera. The
stichidium itself was bare in all the specimens examined, but I think it likely that the
branches of the pedicel may develop further stichidia. I noted at the top of these
branches some divided cells that I thought would develop into stichidia, but no later
stages were observed. The cells covering the tetraspores in the stichidium are not
horizontally disposed as in Endosiphonia but are arranged more or less obliquely. IL is
not certain, however, how far this may change as the sporangium attains maturity.
This alga differs from the genus JEndosiphonia by its habit and anatomical structure,
and though nearly related to it, I believe I am justified, on account of these differences,
in describing it as a new genus.

I have great pleasure in dedicating this interesting species to Mr. J. Stanley
Gardiner.

CLADHYMENIA, Hooker et Harvey.

1. CLADHYMENIA OBLONGIFOLIA, Hook. et Harv. (Plate 12. fig. 7.)
Hooker et Harvey, Alg. Nov. Zel. no. 70.
De Toni, Syll. Alg. vol. iv. sect. 111. 1903, p. 850.

Amirante, 25 fms. ; dry specimens and in alcohol.

Distribution. New Zealand.

The specimens from Amirante are small (Pl. 12. fig. 7) but identical with some that I
received from Prof. J. Agardh from New Zealand under the name of Cladhymenia
oblongifolia, var. І think that the small size of the Amirante specimens may be
accounted for by the great depth from which they come.

MARINE ALGZ OF THE SEALARK’ EXPEDITION. 125

CHONDRIA, C. Agardh.

1. CHONDRIA PUMILA, Vick. (Plate 12. fig. 8.).
Vickers, “Liste des algues marines de la Barbade," in Ann. d. Sc. Nat. 9"* série, Bot. vol. i.
p. 45.

Chagos Archipelago, Diego Garcia, no indication of depth ; in alcohol.

Distribution. Barbados.

On comparing the Chondria from the Chagos Archipelago with Ch. pumila, Vick,
from Barbados, I could detect no character whereby to distinguish it. I conclude,
therefore, that they belong to one species even though the distance between the two
localities is very great; but perhaps this minute plant will be detected later in inter-
vening localities. Miss Vickers intended to give a figure, but her untimely death
prevented her from carrying out this intention. I have therefore given an illustration
of the Chagos plant.

9. CHONDRIA SIMPLICIUSCULA, п. sp. (Plate 12. figs. 9; 10.)

Thallo filiformi, repente, intricato, pauci ramoso ; ramis interdum oppositis, plerumque
singulis; apice filorum truncato, cellulis apicalibus in fovea crateriformi
immersis. Cellula centrali et quinque cellulis pericentralibus uno aut duobus
stratis cellularum corticalium circumdatis. Cellulis periphericis desuper visis ad
apicem hexagonalibus, in parte inferiore cum pariete undulata et ea re moniliformi.
Organis fructificationis non visis.

Aldabra, reef, on Laurencia papillosa.

This small species is characterized by its filaments being mostly horizontally disposed,
sparingly branched, and bearing the tetrasporangia not in special stichidia but at the
apex of ordinary branches, which ultimately may grow out into long vegetative shoots
when the formation of tetraspores has ceased.

After having been dried Ch. simpliciuscula adhered firmly to the alga on which it had
spread itself, but I detected no union between the two plants. It belongs to the section
Celochondria, Falk., as is clearly shown by fig. 10 on Pl. 12.

PorySsIPHoNIA, Greville.

1. POLYSIPHONIA spec.
Praslin, reef, on Hypnea pannosa.
A fragmentary specimen that 1 was unable to name.

ROSCHERA, Sonder.

1. ROSCHERA GLOMERULATA (Ag.), n. comb.
Tolypiocladia glomerulata, Schmitz, Marine Florid. von Deutsch Ost-Afrika, 1895, p. 160.
Seychelles, Long Island ; dry specimen. |
Distribution. Zanzibar; Indian Ocean; Malay Archipelago.

126 MRS. А. WEBER-VAN BOSSE ОХ THE

A small fragment, attached to Gelidiopsis variabilis. In the ‘Siboga’ paper I hope to
point out that Schmitz was perfectly right, when he supposed that Roschera africana,
Sonder, and Tolypiocladia glomerulata belonged to the same genus. Т have therefore
made use in this paper of Sonder's older generic name.

HERPOSIPHONIA, Naegeli.

1. HERPOSIPHONIA PROREPENS (Harv.), Schm.
Harvey, Phycol. Australica, tab. 185 (B).

Chagos Archipelago.

Distribution. Algoa Bay ; Western Australia.

After describing the present plant, Harvey adds: “ This alga was first described from
Algoa Bay, where it occurs on corallines. The Australian plant is more slender, with
fewer tubes, but otherwise the same." Тһе Herposiphonia on Dasyopsis Geppii has
8 pericentral tubes; it stands therefore nearest to Harvey's Australian form. My
material is too scanty to make out whether these forms are specifically distinct from
the Algoa plant, as some authors have thought probable. |

OLIGOCLADUS, Weber-van Bosse.

1. OLIGOCLADUS PRAINII *, n. sp. (Plate 14. fig. 31.)

Thallo filiformi repente, ramis normaliter dorsaliter exeuntibus, erectis, endogenis,
ramulos unilaterales ferentibus. Filo constante ex uno ordine cellularum centrali
et ex quattuor ordinibus cellularum pericentralibus. Trichoblasti desunt. Filis
latis 190-900 u, ramulis 12-80 и. Organis fructificationis ignotis. Rhizoidis
unicellularibus.

Coetivy, reef, on Dasyopsis aperta; in alcohol.

Тһе alga that I have named Oligocladus Prainii belongs to the subfamily of the
Herposiphones, and is very nearly allied to the new genus Oligocladus, provisionally
described in ‘Annales du Jardin Botanique de Buitenzorg, 1910, p. 31. Тһе specimens
were found growing among the penicilli of Dasyopsis aperta and are barren; the
determination is therefore not quite certain, but in anatomical structure these algæ so
closely resemble 0. Boldinghii, Web. v. B., that 1 feel almost sure that they belong
to this genus. I may also add, that іп a few long branches I have seen a con
densation of protoplasm in two cells of each segment. This reminded me of the
two sporangia in each segment in Oligocladus Boldinghu, but in the present case I
could not detect the slightest trace of division in the protoplasm. They were, however,
too young, and I must leave it to future investigators to settle the question.

О. Prainii is а creeping filamentous alga with four pericentral cells, dorsally placed,
endogenous branches, and unicellular exogenous ventral rhizoids. Thus far it is like
O. Boldinghii. It differs from this alga in having no trichoblasts, but a naked, straight

* Named in honour of Sir David Prain, in remembrance of a lovely walk in the Botanical Gardens at
Calutts — — zo ТҰЗ ы

MARINE ALGZ OF THE ‘SEALARK’ EXPEDITION. 127

apical cell, and a much denser ramification. The absence of trichoblasts is no reason
to separate O. Prainii from the genus Oligocladus, for in the genus Polysiphonia we
find species with and without these organs.

The main axis of the plant was attached by means of its long rhizoids to the
branches of Dasyopsis, and, as may be seen in fig. 31, the youngest part raises itself
from its support, grows vertically upwards, though its apex shows a slight curve. The
apical cell is rather long and cuts off the segments at its base by a transverse wall.
These segments remain undivided for a long time—I have counted eight of them under
the apical cell before the pericentral cells are begun to be cut off.

The dorsally placed branches arise irregularly with a distance of 3, 4, 6, 7, 8, 9, and
even more segments between them; they bear branches of the second order on the side
turned towards the main axis. In some cases I observed near the apex of the main axis
long ramified branches alternating with short simple ones, but in later stages I could
detect no further difference. Perhaps the long ones bear the tetrasporangia and the
short ones remain sterile, but this is a mere suggestion. The first segment of the
branches issuing from the main axis has only short pericentral cells; the second
segment has pericentral cells of the normal size; all the branches of the first and
second orders arise from the central tube. The size of the branches and of the main
axis differs considerably according to the place where the measurement is taken. Тһе
main axis has a breadth of 120-200; the branches 40-80 in the lower parts, but at
the top only 12-20 и; the segments are 13 times as long as the diameter, and at the
top as long as broad.

At the top of the main axis all the branches are arranged on the dorsal side of the
filament, but at some distance from the apex their position is sometimes altered ; this
is the result of a slight twisting of the cells of the main axis. -

2. OLIGOCLADUS PUSILLUS, n. Sp.

Thallo filiformi, repente; ramis erectis, endogenis, simplicibus, normaliter dorsaliter
exeuntibus. Filisconstantibus ex uno ordine cellularum centrali et quattuor ordinibus
cellularum pericentralibus. Trichoblastis numerosis, filorum apices cingentibus.
Filis latis 40-100 и. Organis fructificationis ignotis. Rhizoidis unicellularibus.

Coetivy, reef, on Dasyopsis aperta ; in alcohol.

Oligocladus pusillus, оп account of the numerous trichoblasts that surround the
slightly curved top, stands nearer to O. Boldinghii than О. Prainii. It is а smaller
plant, though its filaments are broader. Its creeping filaments give off side-branches
on the dorsal side at a distance of 5, 7, 12, 20 segments and even more. These
ascending side-branches remain simple; when, however, they develop into creeping
filaments, as occasionally happens, they likewise give off branches on their dorsal
side. The creeping branches have a diameter of 100 4, the side-branches a little less,
and at the top of both the diameter sinks to 404. The segments have a height of
80—100 u, but this sinks at the top to 20 ш and even less.

Organs of fructification have not been observed.

128 MRS. A. WEBER-VAN BOSSE ON THE

Dasyopsis, Zanardini.

1. Dasyopsts STANLEYI, n. sp. (Plate 13. fig. 16.)

Thallo elongato, subdichotomo, pauce ramoso, penicillis longis, gracilibus cincto;
ramificatione radiata, constante e penicillis basim versus deciduis, spinas non
relinquentibus. Stichidiis oblongis et cystocarpiis ex penicillis orientibus. An-
theridiis non visis. Thallo alto 16 cm. Habitu Dasye elegantis.

Cargados Carajos, 30 fms.; in alcohol and dry specimens.

To rightly understand the new species of Dasyopsis, it may be worth while to give

а short account of the characters of the genus. Dasyopsis has a sympodial mode of
growth; the stem or axis being composed of short branches with definite growth ;
these are pushed aside by their own first branch, and the distance between the base of
the mother-branch and the point of insertion of its first branch varies in different
species. |

But this is a feature Dasyopsis has in common with all Dasyez; its special character
consists in the absence of true pericentral cells. The cells that surround the central
tube are not originally segments of this tube, as in the other Rhodomelacez, but they `
are downward-growing hyphe, the mother-cell of which is cut off from the branch that
is pushed aside when sympodial growth sets in (Pl. 14. fig. 32, т). Тһе hyphz may be
of the same diameter as the central cells, in which case they resemble true pericentral
cells, but they have quite a different origin. The displaced branches continue to
develop, they are shorter or longer, naked or corticated, according to the species, and
carry several monosiphonous filaments, the so-called penicillus, at the apex.

Dasyopsis Stanleyi consists of a subdichotomously divided principal axis or stem,
carrying at indefinite intervals branches of unlimited growth, which may ramify again
and are surrounded at the top by monosiphonous penicilli. This pretty alga so much
resembles Dasya elegans that I felt at first inclined to take it for that species; after
careful examination, however, I discovered that the plant has the characteristic
structure of Dasyopsis; the branches of definite growth are separated by only one
short segment from each other, and grow out in any direction from the main axis. Аз
every branch or so-called penicillus gives off a downward-growing hypha, the central
axis soon increases in thickness; true pericentral cells are of course absent.

The penicilli are composed of cells having a breadth of 6-13 и; their length varies
from being almost square at the base to 72 и at the top. They are very delicate and
soon fall off, unless they happen to bear cystocarps or stichidia at their base. Тһе
stichidia were very rare іп my preparations; I observed only one but unfortunately
lost it; I saw, however, that it was long and cuspidate. The cystocarps were numerous
in fertile specimens. '

2. DASYOPSIS APERTA, n. sp. (Plate 18. fig. 17; Plate 14. fig. 32.)
Thallo ex axibus principalibus primo surgentibus, postea decumbentibus constante, unde
denuo rami adscendentes nascuntur, Apice axis principalis et ramorum penicillis
ramulosis, monosiphoniis cincto, penicillorum pilis postea decidentibus, basi

MARINE АТ@ Ж OF THE *SEALARK ' EXPEDITION. 129

monosiphonio brevi tantummodo restante. Thallo sympodiali, radiato; ramis
cum divergentia i. Primo segmento libero ramorum ad basim satis magnam
cellulam incremento basipetali disjungente. Cellulis centralibus magnis, hyphis
numerosis angustis cinctis. Stichidiis ramulis penicillorum sessilibus; 8-10 sporas
in unoquoque segmento ferentibus. Cystocarpiis et antheridiis ignotis.

Coetivy, reef ; in alcohol. ;

This new Dasyopsis attains a height of 2:5 cm., and consists of a small tuft of primary
ascending cylindrical axes or stems, that finally become horizontal and give rise to new
ascending ones. Тһе tufts, on growing older, become more or less intricate, but I
never saw the fronds anastomose. Dasyopsis aperta is a good species for the study of
the peculiar Dasyopsis structure, because the cells are large, even at the apex of the
shoot. It is sufficient to spread the alga on the slide and to tear off carefully with
a needle the surrounding penicilli to see plainly the row of central cells, each carrying
a pushed aside apex and its first branch whose lowest cell builds up the sympodium.
These branches stand spirally around the main axis, as far as I could ascertain, with
a divergence of 1, and carry long, incurved filaments at their top, which form together
the penicillus. The first cell, after the branch has been pushed aside, cuts off a cell at
its base (РІ. 14. fig. 32, m); this is the mother-cell of the downward growing hyphae,
which are at first almost as large as the cells of the central axis or tube. At the top ofa
stem or axis, where the mutual position of the different cells is still unaltered, the central
tube appears to consist of a central and four pericentral cells, but on following the
pseudo-pericentral cells to their origin, it will be clearly seen that they are downward
growing hyphe, arising from the first free cell of the pushed aside branch. ‘The sym-
podial axis is afterwards covered by a thick layer of smaller hyphe, which by growing
between them alter the original position of the cells.

The stem or main axis is cylindrical, though tapering towards its apex, which is
densely covered with penicilli; these, however, soon fall off, and the base of older stems
is almost denuded, though not entirely, for remnants of the basal portion of the penicilli
often remain.

The sessile stichidia are borne on the side-branches of the penicilli; they are
cylindrical and, when the tetrasporangia have acquired their full size, a little torulose.
I counted from 6-8 sporangia in different segments; and the whole is crowned by one
or two barren cells. Neither cystocarps nor antheridia were seen. 224,

There are, according to Falkenberg *, two other species of Dasyopsis, which, like
D. Stanleyi and D. aperta, have penicilli facing every way, namely Dasyopsis spinella
and D. cervicornis. From the first both D. Stanley; and D. aperta differ in the absence
of the little spines that have given D. spinella its specific name, and from D. cervicornis
in the shape of their stem, cylindrical throughout, not angular at the insertion-point

of the penicilli.
- PR PA pes ‘ Die Rhodomelaceen des Golfes von Neapel; 1901, p. 667.

SECOND SERIES.—BOTANY, VOL. VIII.

180 MRS. A. WEBER-VAN BOSSE ON THE

3. Dasyopsis GEPPII, п. sp. (Plate 13. figs. 18-20; Plate 14. fig. 33.)

Thallo carnoso, plano, palmato, inciso, horizontaliter expanso, sympodialiter con-
structo, ad apicem radioso ; divergentia ignota. Penicillis lateralibus tantum bene
evolutis, dorsalibus et ventralibus aliquamdiu conspicuis, postea evanescentibus.
Ramis lateralibus pro parte cum axi principali coalitis. Strato corticali cellularum
rotundarum aut angulosarum contiguarum. Antheridiis cylindricis in turmis in
facie inferiori thalli sitis. Cystocarpiis et stichidiis ignotis.

Amirante, 20-95 fms. ; in alcohol.

Cargados Carajos, 25 fms.

Chagos Archipelago, Solomon Island, no indication of depth; in alcohol.

Coetivy, reef.

Dasyopsis Geppii has a broad, flat, fleshy frond, slightly palmatifid and attached to
the substratum by a thick pedicel (РІ. 13. figs. 18, 19). Its outward appearance
differs entirely from any known species of Dasyopsis, and it is only by the study of
the youngest segments of the frond that an insight into its anatomical structure is
obtained. This is in principle like that of Dasyopsis, but differs in so far as the
displaced branches, instead of turning in every direction, as they do in D. Stanleyi,
spread themselves principally in a bilateral direction and grow congenitally into a
leaf-like frond. The penicilli are situated on the margin of this frond, but also on its
upper and lower surface; the cells of the filaments have a breadth of 16-28 p, and
a length of 32-88, 100, and 140 р.

At the top of the different segments the penicilli are so closely packed together, in
consequence of the smallness of the cells, that it is impossible to trace the central
tube down into the frond or to observe the divergence of the penicilli around its apex
(Pl. 14. fig. 33).

Dorsi-ventral organization does exist, though I could not trace it at the apex of the
frond. The ventral side, that turned towards the substratum, bears antheridia at
the base of old penicilli. The antheridia stand in groups of five and six together
(РІ. 13. figs. 19, 20), and are borne оп a monosiphonous pedicel. The dorsal side of the
frond is barren in all my specimens. The position of the antheridia is interesting,
for D. Geppii is both a reef and a deep-water plant. If it were only collected on the
reef, one might infer that the antheridia kept to the ventral side to ensure protection
from the sun, but since these alge have been found with ventrally placed antheridia,
not only on the reef but also in deep water, this explanation does not hold good.
Dorsi-ventral organization of any kind was hitherto unknown in Dasyopsis.

I have great pleasure in dedicating this alga to my kind friend, Mr. Gepp, of the
British Museum.

4. DASYOPSIS PALMATIFIDA, n. sp. (Plate 13. fig. 21.)
ТһаПо carnoso, plano, profunde palmatifido, segmentis aut laciniis ѕере terminan-
tibus in filamento cylindrico, substrato bulbo affixo. Segmentis ad apicem radiatis,

postea horizontaliter expansis. Peniciliis circum apicem thalli bene evolutis,
infimis evanescentibus. Penicillis constantibus e filis multo crassioribus quam

MARINE АТ@Ж OF THE ‘SEALARK’ EXPEDITION. 131

penicillis Dasyopsis Geppii; filis stichidia in apicibus pedicellorum brevium
monosiphoniorum circumdantibus. Stichidiis sporis numerosis. COystocarpiis et
antheridiis ignotis.

Amirante, 20—25 fms.; in alcohol.

Chagos Archipelago, Solomon Island; in aleohol.

Coetivy, reef; in alcohol.

Dasyopsis palmatifida resembles in essential points D. Geppii, but its frond is far
more deeply palmatifid and the different segments often terminate in a cylindrical
branch. These fasten themselves to any hard object and produce at that spot a little
tubercle, which much resembles the thick pedicel of D. Geppit. І have considered
the question whether D. palmatifida might not be the stichidia-bearing form of
D. Geppii, but have been unable to solve it. In addition to its deeply palmatifid frond,
D. palmatifida differs also from D. Geppi? in the size of the filaments forming the
penicilli; these cells having a length of 104—106 р and a breadth of 12-20, 36-40 p.
Stichidia with tetrasporangia are borne at the base of the репісіШ on monosiphonous
pedicels ; they have a blunt apex.

TAPEINODASYA, Weber-van Bosse.

1. TAPEINODASYA ETHEL, п. sp. (Plate 13. figs. 22, 28.)

Thallo procumbente, constante e sympodio bilaterali; tertio quoque segmento aut
ramulum lateralem spinosum incremento finito ferente, aut ramulum lateralem
spinosum incremento sympodiali infinito. Ramulis adventitiis przeterea ex axillis
ramulorum spinosorum emergentibus sed parvis etiam adultis. Axi constante
e cellula centrali eum quattuor cellulis pericentralibus ; quo numero verumtamen
cellularum pericentralium in segmentis sæpe majore ob concrescentiam ramulorum.
Axi centrali cellulis cortiealibus, peripheriam versus diminutis, cincto. Hyphis
numerosis in parietibus inter cellulas pericentrales nascentibus. Organis fructi-
ficationis ignotis.

Amirante, 20-25 fms.; dry and in alcohol.

Tapeinodasya Ethele is the second known species of this genus; the type, T. Bornetii,
being described in the ‘Recueil des Travaux botaniques Néerlandais, Хо. 1, 1904. The
plant from Amirante differs in having a less dense ramification, which is, as a rule,
bilateral. The branches have the outward appearance of little spines. At the top of
the main axis they point outwards in three directions, but immediately below the
growing-point the branches stand in two rows. In my preparations the distance
between the succeeding branches in the sympodium is three cells, but I should not be
surprised if in parts of the frond the distance amounted to only two cells, as in
the other species, 7. Bornetii, the number is known to vary. The branch that is
displaced ends in a spine; it is either of definite growth and develops no further, or of
indefinite growth, in which case it may give rise to а lateral sympodial branch. The
branches grow congenitally with the main axis for one or two segments, and in a
cross-section through such a segment a greater number of pericentral cells is visible.

v2

139 MBS. A. WEBER-VAN BOSSE ON THE

The central сей is surrounded normally by four pericentral cells, but between these
other large cells soonvappear and numerous hyphe develop in the membranes. I have
not seen them enter the pericentral cells as they do іп 7. Bornetit, but it may be that
this takes place in the basal part of the frond. А layer of small cortical cells surrounds
the main axis. The latter has a winged appearance on account of the pericentral cells
of the spine-like lateral branches growing sideways together with it for a few segments.

Adventitious branches appear occasionally on the dorsal side of the frond in the axil
of lateral branches, but they remain short; in 7. Bornetii these branches attain full
development. :

The characteristic dorsi-ventral organization of T. Bornetii is far less conspicuous in
T. Ethele, and shows itself in barren specimens only in the rudimentary axillary
branches. No organs of fructification were present, and it was only because l was
familiar with the genus that I was able to name the specimens from Amirante.

HzTEROSIPHONIA, Montagne.

1. HETEROSIPHONIA RENDLEI, п. sp. (Plate 12. figs. 11, 12.)

Thallo fasciculato, pluribus ramis principalibus e callo basali surgentibus; ramis
principalibus totis ecorticatis, pauce ramosis, penicillos ad segmentum quodque
secundum, distichos et alternantes gerentibus. Cellulis pericentralibus 7-8-10.
Penicillis ramulos unilaterales et stichidia potius obtusa quam acuta ferentibus
tetrasporis instructis. Antheridiis et cystocarpiis ignotis. Thallo alto 8 cm.

Cargados Carajos, 47 fms.; dry specimen.

I believe Heterosiphonia Rendlei* to be a new species because none exists amongst
the Heterosiphonie ecorticate that is exactly like it. It belongs to the small group with
more than four pericentral cells, the branches or penicilli of which arise at a distance
of two or three segments from each other and are turned regularly to the right and
left. H. Hendlei stands nearest to H. subsecundata (Suhr), Falk., but differs in the
pericentral cells not being transversely divided.

The principal axis has a height of about 8 cm. ; it is erect and carries small horizontally
spread penicilli which have unilateral monosiphonous branchlets incurved at the top.
Тһе number of pericentral cells in the main axis varies from 6-10, and the segments
are three times as broad as long. In the penicilli the few basal polysiphonous segments
are almost as long as broad or a little broader.

I have seen neither cystocarps nor antheridia, only stichidia, which are borne on the
side-branches of the penicilli ; they are elongated. with a blunt apex.

2. HETEROSIPHONIA spec.

_ Coetivy, reef; dry specimen.

_ This is a small specimen, 1 em. high; it grows on Codtum. I failed to recognize it
n the described species and yet do not dare to describe it as new.

* Named in honour of. Dr. Rendle, Keeper оҒ the Botanical Department of the British Museum, renowned for
the kind e: that it offers to every visitor.

MARINE ALGZ OF THE ‘ SEALARK ' EXPEDITION. 129

AMPHISBETEMA, п. ©.

Thallus constat e pede crasso rhizomatoideo, unde nascuntur axes erecti, corticati,
regulariter sympodii modo compositi e ramis incremento finito conspicue mono-
podialibus dorso-ventraliter crescentibus. Axis constat ex una cellula centrali et
sex cellulis pericentralibus, tribus aut quatuor ordinibus cellularum longitu-
dine cellule primariz pericentralium cinctis; quos ordines stratum corticale
crassum e cellulis parenchymaticis fissis compositum cingit. Cystocarpia et
antheridia ignota. Stichidia sporis verticillatis, eque ac in Dasya non omnino
cellulis corticalibus obtectis.

1. AMPHISBETEMA INDICA (J. Ag.). (Plate 13. fig. 24; Plate 14. fig. 34, and text-fig.)
Dasya indica, J. Agardh, Till Alg. Syst. xi. 1890, p. 111.

Diagnosis the same as that of genus.

Aldabra, reef, Passi Honakan, outside; dry specimens.

Distribution, Diego Garcia.

The alga for which I propose the generic name of Amphisbelema was described in
1890 by J. Agardh under the name of Dasya indica; the structure of the stichidia,
almost identical with that of Dasya, probably induced this eminent algologist to
describe the plant under that name, for he was quite cognisant of the differences in
ramification between his alga and other Dasye. In De Toni's ‘Sylloge Algarum,'
iv. p. 1146, 1903, Dasya indica has received the name of Wilsonea (?) indica.

— I found this alga in the collection of Mr. Stanley Gardiner, and through the kindness
of Sir David Prain I was able to compare it with authentic material from Diego Garcia,
now preserved in the Herbarium at Kew, and I wish to express шу best thanks to him
for so liberally placing this material at my disposal. The study of the type-specimens
enabled me to identify my barren material from Aldabra. This differs only from the
type in being a little stronger; the alge from Diego Garcia bear stichidia with
tetrasporangia, a fact which may account for their being more slender. |

Amphisbetema has a strong rhizome-like base from which arise erect main axes,
attaining a height of 8-9 cm., with distichous bilateral branches. These main axes
are composed of branches or shoots of definite growth, succeeding each other at a
distance of six segments to be then displaced by their own first branch. Тһе branches
have, as far as Г could make out, monopodial growth ; sympodial growth sets in only
when each in turn is pushed aside. All the branches have a dorsi-ventral organization ;
their first branch of the second order is always inserted at a distance of six segments
from the base, and all the succeeding ones at a distance of two segments; they stand
in two rows on the ventral side of the branch, and only by subsequent growth do they
ultimately take their bilateral position. Тһе branches of the second order have the
same ramifieation as the primary ones and are always turned to one side; each
principal branch carries as a rule six pairs of branches of the second order; but the
nearer to the apex the less do they develop.

After the formation of these six pairs of branches, the vigour of each mother-
braneh seems to be exhausted. At a distance of three or four cells from the apex,

154 MRS. А. WEBER-VAN BOSSE ON THE

a segment remains undivided : its contents, balled together, look like a large drop of oil
and have a reddish-brown colour. The first segment succeeding this undivided cell may
show a division into a central and pericentral cells, the next is undivided, and the
ultimate one has usually the shape of a spine. Sometimes the top of a branch is
crowned by a simple or ramified monosiphonous filament. I could find no explanation
for this difference in development; the monosiphonous filaments are not trichoblasts,
nor could any trace of these organs be detected. The large undivided cell, with the cells
above it, fall off after a shorter or longer period, and the branch that carried them is
blunt in consequence.
Fig. 1.

Schematic figure of apex of Amphis-
betema indica, showing the sym-
podial growth of main axis and the
monopodial growth of the displaced
shoots. After 6 segments shoot I
is pushed aside by its first branch,
shoot II is also displaced after
6 segments by its first branch, and
the other shoots follow in like
manner. The displaced shoots
carry side-branches at every second

segment.

`

The axes have in each segment a central and six pericentral cells ; in the last-formed
side-branches this number may sink to four, but as a rule it is six. The pericentral cells,
in common with those of Hndosiphonia and Melanothamnus, cut off to the outside cells
quite as long as themselves; and this may be repeated several times. In the branches
the cells on the dorsal side are a little longer than those on the ventral, but this difference
disappears with subsequent growth. "The whole central strand is surrounded by a rather
thick layer of parenchymatouscells. Тһе peripheral cells are of equal size and elongated
longitudinally, not radially as in Melanothamnus. Though the cortical layer is pretty -

MARINE ALGÆ OF THE ‘SEALARK’ EXPEDITION. 135

thick, the transverse membranes of the succeeding segments within are plainly visible,
owing to the fact that the central strand consists of a large number of cells of equal
length. The segments are short, and this fact gives, whenever the apex of a shoot is
pushed aside, the impression that the main axis is divided into three branches. Careful
examination will, however, show that this impression is erroneous, being brought about
by the displaced shoot, its first branch—the new shoot—and its second branch standing
very close together, owing to the short segments and the congenital growth of the first
two segments. -

My plants are barren: J. Agardh, however, described the stichidia, which he stated
were like those of Dasya and bore verticillately developed tetraspores partially uncovered
on the outside. |

Amphisbetema has received its name * on account of the difficulty experienced in
assigning to it its true systematie position. It has characters in common with several
genera of Rhodomelacew and also with Melanothamnus, Born. et Falk., a genus of
uncertain position so long as the organs of fructification remain unknown.

With Streblocladia, Amphisbetema agrees in the regular alternating mode of mono-
podialand sympodial growth ; with Melanothamnus and Endosiphonia in the numerous
secondary pericentral cells of equal length; with Heterosiphonia in the dorsi-ventral
organization of its branches; and with Dasya in the structure of the stichidia.

The association of all these characters in one genus makes me inclined to consider
Amphisbetema as an old form of Rhodomelace which existed before the genera that
have now developed one or more of the characters—in our new genus still blended
together—in their own special way and to the exclusion of the other characters.

Dictyurts, Bory. `
1. DICTYURUS PURPURASCENS, Bory.
Bory de St. Vincent, in Bélanger, Voy. Indes orient. p. 170.
Falkenberg, Die Rhodomelaceen, 1901, p. 675.
De Toni, Syll. Alg. vol. iv. sect. 111. 1903, p. 1173.
Saya de Malha, 55 fms.; dry specimen.
Coetivy, reef ; dry specimen and in alcohol (fragment).
Cargados Carajos, 30 and 47 fms.; dry specimen and in alcohol.
Amirante, 30 fms. ; in alcohol.
Chagos Archipelago, Solomon Isl., 22 fms.; in alcohol.
Distribution. Cape Comorin; Coast of “India; Mauritius; Ceylon; Malay Archi-
pelago.
Fam. СЕВАМТАСЕМ.

GRIFFITHSIA, C. Agardh.
1. GRIFFITHSIA spec.
Amirante, 20-25 fms.; in alcohol.
Only a small fragment, but with the characteristic tetraspores of the genus.
* GpgeoBirqpa, controversy, a point in dispute.

136 MRS, A. WEBER-VAN BOSSE ON THE

HaroPLEGMA, Montagne.
1. HALOPLEGMA PREISSII, Sond.
Sonder, Alg. Preiss, p. 24.
Harvey, Phyc. Austr. 1859, tab. 79.
De Toni, Syll. Alg. vol. iv. sect. 111. 1903, p. 1366.

Cargados Carajos, at 22, 30, and 47 fms. ; both dry and in alcohol.

Distribution. Western Australia; Tasmania, in the River Tamar above Georgetown ;
Malay Archipelago. !

The specimens of И; Preissii from Cargados Carajos are shorter, less branched, and
more flabellate than most plants found in the herbaria under that name. l have not
ventured, however, to describe these specimens as new, for I have found no microscopic
character to distinguish them from the type, the size of the main branches and the
peripheral filaments being essentially the same in the different plants. The specimens
from Cargados Carajos bear tetraspores; I have not seen cystocarps or antheridia, and
it is possible that these may show differences in structure. Dr. Harvey, in his * Phye.
Australica,’ however, calls attention to the fact that the external form of H. Preissvi is
very variable. Не found it growing in the River Tamar, a considerable distance above
Georgetown, where it appeared at first to be a distinct species; but he was able to trace
it down to the Heads of Port Dalrymple, and found that it blended into the usual variety.
I feel very much inclined to think that the specimens from Cargados Carajos owe their
small size in great measure to the depth at which they were gathered.

ANTITHAMNION, Naegeli.
1. ANTITHAMNION ADNATUM, J. Ag.
J. Agardh, Analecta algologica, 1892, p. 12.
Solomon Island.
Distribution. New Zealand; Malay Archipelago.
А small specimen with tetraspores, attached to Herposiphonia.

SPYKRIDIA, Harvey.
1. SPYRIDIA FILAMENTOSA (Wulfen), Harv.
Harvey in Hooker, Brit. Flora, vol. ii. р. 887; Phycol. Brit. 1871, pl. xlvi.
Aldabra reef, outside Passi Honakan.
Distribution. West Indies; Mediterranean; Red Sea; Indian Ocean; Malay
Archipelago. | °
Only a small fragment attached to .4mphisbetema indica,

CERAMIUM (Roth), Lyngbye.
1. CERAMIUM CINNABARINUM (Gratel.), Hauck.
Hauck, Die Meeresalgen von Deutschland, 1885, p. 112.
Solomon Island.
Distribution. Mediterranean; Malay Archipelago.
A small specimen amongst other Algz. ES

-

MARINE ALG OF THE *SEALARK' EXPEDITION. 137

2. CERAMIUM spec.
Cargados Carajos, 47 fms.
On Heterosiphonia Rendlei; barren specimens.

REINBOLDIELLA, De Toni.
1. REINBOLDIELLA SCHMITZIANA (Reinb.), De Toni.

Рһусеге japon. nove, 1895, p. 35.

Gloiothamnion Schmitzianum, Reinbold, in Hedwigia, 1895, p. 205.
Chagos Archipelago, Egmont Reef; in alcohol. On Gelidiopsis.
Solomon Island. On Gelidium.

Distribution. Japan; Malay Archipelago.

Fam. СВАТЕГООРТАСЕ Ж.

HALYMENIA (C. Ag.), J. Agardh.

1. HALYMENIA POLYCLADA, А. & E. В. Gepp.

Var. ALDABRADENSIS, n. var.—Thallo e callo basali subito cuneatim expanso 10 cm.
alto, 4-5 cm. lato deinde palmati-partito constante; ramis usque ad 6 cm. longis,
05-1 em. latis, ramulis lateralibus iterum iterumque subdichotomicis pinnatis,
angustis.

Aldabra, outer reef; dry specimens.

Distribution. Christmas Island.

Halymenia polyclada belongs to Agardh's Acanthymenia section of Halymenia.
According to Schmitz, it is highly probable that this group is identical with the
genus Gelinaria, Sond. Schmitz is further of opinion that the differences existing
in anatomical structure are sufficient to separate Acanthymenia from the genus Haly-
menia. So long, however, as we do not know the cystocarps of the genus Gelinaria
I think it more expedient to follow Agardh, although ultimately we will probably have
to follow Schmitz and call the species of the Acanthymenia group by the generic name
of Gelinaria.

The specimens from Aldabra are very similar to H. polyclada, which is nearly allied
to H. formosa. It has а subdichotomous ramification, but lateral ramuli are plentiful
along the margins of the upper branches. In the lower part of the fronds the margins
are almost entire. It is a stronger plant than И. polyclada, but has the same short stalk
expanding suddenly into a broad frond. In how far H. formosa, Durvillai, and polyclada
are really good species, is a question that future investigations must decide.

SECOND SERIES.—BOTANY, VOL. VIII. X

188 MRS. А. WEBER-VAN BOSSE ON THE

CRYPTONEMIA, J. Ag.
1. CRYPTONEMIA SEMINERVIS (Ag.), J. Ag.
J. Agardh, Algæ Liebman. р. 11 in not.
De Toni, Syll. Alg. vol. iv. sect. ту. 1905, р. 1610.
Saya de Malha, 55 fms.; dry specimens.
Amirante ; dry specimen.
Distribution. Mediterranean; Atlantic ; Red Sea.
In a dry condition the plants appear to have no midrib, but on moistening a delicate
midrib became distinctly visible, and this extended upwards to about the middle of the
leaf. The plants are barren.

2. CRYPTONEMIA spec.

Saya de Malha, 55 fms.; dry specimen.

A poor fragment, but differing from Cryptonemia seminervis by its much thicker frond
and dark red colour.

Fam. SQUAMARIACEA.

It is much to be regretted that so many of the specimens of Squamariacez collected
by Mr. J. Stanley Gardiner are sterile, for without fruit it is almost impossible to name
the genera and species of this puzzling family. I happened to have studied a large
collection of Indian Peyssonnelie before I undertook to name the present collection, and
this has helped me to recognize some of the species mentioned hereafter.

In my “ List of the * Siboga' Algze" I will treat in detail of the Peyssonnelia, and only
mention here that I have found it of great help to keep up a distinction already hinted at
by the late lamented Schmitz*, namely, a distinction based on the differences in arrange-
ment of the filaments of the hypothallus. These horizontal filaments, which run over
the substratum, give off the ascending vertical filaments of the perithallus, and Schmitz
observed that some hypothalli consisted of straight filaments running close to one
another ( Peyssonnelia), and that others consisted of curved fan-shaped groups of filaments,
as in the basal layer of Cruoriella. To these two subgenera, if I may distinguish them
provisionally by such a name, I wish to add a third, differing in having no hypothallus
proper, but a mesothallus, i. e., a layer of cells occupying the middle of the thallus (as is
the case, for instance, in Ralfsia expansa), which gives off branches both downward and
upward. My investigations do not yet allow me to decide whether this division of the
genus Peyssonnelia, based as it is upon anatomical characters, is supported by differences
in the structure of the fruit. Sometimes this is the case, but my specimens are mostly
barren.

For this third subgenus I should like to propose the name of Ethelia, in honour of my
dear friend, Mrs. Е. S. Gepp. |

ж Schmitz, “ Marine Florideen von Deutsch Ost-Afrika," in Engler's Bot. Jahrbuch, Bd. xxi. Heft 1, 2, 1895,
p. 173. | : |

MARINE ALGJE OF THE ‘SEALARK’ EXPEDITION. 139

PrYSSONNELIA, Decaisne.
Subgen. EUPEYSSONNELIA, n. subg.

Hypothallus filis arcte iunctis, flabelli modo fusis, maxime erectis.

1. PEYSSONNELIA COCCINEA, J. Ag.
J. Agardh, Epicrisis Floridearum, 1896, p. 385.
De Toni, Syll. Alg. vol. iv. sect. ту. 1905, р. 1695.
Seychelles, 31 fms. ; dry specimen.
Distribution. West coast of Australia; Malay Archipelago.
The specimen is small and sterile; it was only by comparing it with ‘Siboga ' material
that I was able to name it.

9. PEYSSONNELIA HARVEYANA (2), Crouan. (Plate 14. fig. 35.)
Crouan іп 7. Agardh, Spec. Algarum, ii. 1852, p. 501.
De Toni, Syll. Alg. vol. iv. sect. 1v. 1905, p. 1694.

Saya de Malha, 55 fms., on several dry, calcareous, little pebbles.

Distribution. Coast of France, near Brest.

The specimens are of a bright red colour and adhere firmly to the substratum; they
approach both P. conchicola and P. Harveyana, but on account of the large cells of the
hypothallus I feel inclined to regard it as Р. Harveyana. With regard to distribution,
it is more likely that they belong to P. conchicola, an inhabitant of the Red Sea.

The specimens are all barren, and it is impossible to name them with certainty.

3. PEYSSONNELIA CALCEA, Heydr.
Heydrich, “ Neue Kalkalgen von Deutsch Neu-Guinea " in Biblioth. Bot. von Frank u. Luerssen,
1897, p. 10.
Egmont, reef ; in alcohol.
Coin Peros; in alcohol.
Distribution. Tami, near German New Guinea; Malay Archipelago.
The specimens bear nemathecia with tetraspores.

4. PEYSSONNELIA POLYMORPHA (Zan.), Schmitz.

Form GARDINERI, п. f.—Thallo primo adherente, deinde a substrato soluto, tubera
forma diversa, sed superficie levi formante, colore parum rubro cum siccatus est.
Hypothallo cellulis longis 96-44-60 u, latis 8-10-16 a. Perithallo constante e
seriebus cellularum sursum directarum, cellulis infimis majoribus, cellulis superiori-
bus minoribus, longioribus aut brevioribus quam latis ; latis 16-20-24 и. Organis
fructificationis ignotis.

Amirante, 30, 45-60 fms. ; dry specimens. |

The specimens from Amirante are nearly allied to P. polymorpha, but the loose thalli

are smaller and the crusts are thinner than in the typical specimens of that species. The
anatomical structure of the two alge is much alike, as far as the vegetative thallus is

140 MRS. А. WEBER-VAN BOSSE ON THE

concerned, and reproductive organs were not to be found. The cells of the f. Gardineri
are perhaps, on the whole, somewhat smaller than those of the type, and seen from above
f. Gardineri is covered by a smooth peripheral layer, which Р. polymorpha had lost in
all the specimens I have seen.

I conclude that f. Gardineri is a deep-water form of P. polymorpha.

Subgen. ETHELIA, п. subg.

Mesothallus filis rectis aut ramosis, sinuatis, utroque latere fila perithalli emittens.

5. PEYSSONNELIA BIRADIATA, п. sp.

Thallo per substratum extento, tantum hic illic adherente, coriaceo, rubro, rotundo,
diametro 2 ad 3 centimetrum, constante e mesothallo et perithallo. Mesothallo
constante е filis ramosis, sinuatis, parva flabella efficientibus. Perithallo constante
e filis sursum et deorsum directis. Perithallo summo diviso in partem inferiorem
obliquam, cellulis magnis, et in partem superiorem erectam, hic illic zonatam,
cellulis parvis. Organis fructificationis ignotis.

Seychelles, 31 fms. ; dry specimen.

Distribution. Malay Archipelago.

A small but characteristic specimen.

Among the algze from Saya de Malha there are several specimens that I believe belong
to the genus Peyssonnelia, but they are unfortunately barren. They have much in
common with the Peyssonnelie of the Ethelia-group, for they have a mesothallus that
gives off branches both in a downward and upward direction (РІ. 14. fig. 36). They are
furthermore much calcified, hard as stone, and attached to bits of coral.

But besides these features found among the Hthelia-group, the Peyssonnelie from
Saya de Malha are characterized by long filaments (Pl. 14. fig. 37) almost free from each
other and covered by a layer of pectic substance, which colours red with ruthenium.
The filaments are like those we find in the sori of other Peyssonnelie, but in the new
plants they are not confined to sori alone, but cover the whole upper surface. Although
I made sections through all parts of the thallus, I could find no trace of fructification.

We know that the fronds of Stragularia species amongst the Ralfsiaceze are covered
by filaments, and it appears probable that we have here something analogous amongst
the Реуззоппейе; but until the reproductive organs are discovered it is difficult to
speculate as to these structures. А doubt still lurks in my mind whether, after all,
these filaments may not belong to an encrusting parasitic alga; but it is well worth
while to call attention to these specimens. |

Cruoriopsis, Dufour.
1. CRUORIOPSIS CRUCIATA, Dufour.

Dufour, Elenc. Alg. Lig. p. 35, after De Toni.
De Toni, Syll. Alg. vol. iv. sect. rv. 1905, p. 1689.

Saya de Malha; dry specimen on Udotea.

MARINE ALG OF THE *SEALARK ' EXPEDITION, 141

Distribution. Adriatic; Malay Archipelago.
The alga is barren, but its resemblance to specimens from Nias, Sumatra, bearing
cystocarps and antheridia, is so great that I believe they belong to the same species.

Fam. CORALLINACE.
MELOBESIA, Lamouroux.

1. MELOBESIA CALLITHAMNIOIDES, Falk.
Falkenberg, Die Meeresalg. d. Golfes v. Neapel, 1878, p. 265.
Egmont Reef, on different algze ; in alcohol.
Distribution. Atlantic, Brest? ; Mediterranean.

AMPHIROA, Lamouroux.
1. AMPHIROA FRAGILISSIMA (Linn.), Lamx.
Lamouroux, Pol. flexibles, p. 300.
A. Weber-van Bosse and M. Foslie, “Тһе Corallinacez of the ‘ Siboga’ Exped.," Monogr. lxi.
1904, p. 89.
Chagos Archipelago, Solomon Island ; in alcohol.
Distribution. West Indies; Pacific; Malay Archipelago ; Dar-es-Salaam.

2. AMPHIROA FOLIACEA, Lamx.
Lamouroux, Voyage de ’Оташе par Freycinet, Zoologie par Quoy et Gaimard, p. 628.
A. Weber-van Bosse and M. Foslie, * The Corallinacez of the ‘ Siboga" Exped.," Monogr. lxi.
1904, p. 99.
Praslin; in alcohol.
Distribution. Marianne Islands ; Malay Archipelago.

JANIA, Lamouroux.
l. JANIA TENELLA, Kütz.
Kützing, Tab. Phyc. Bd. viii., 1858, tab. 85, p. 41.

Coetivy, reef ; in alcohol.

Coin Peros ; in alcohol.

Praslin ; in alcohol.

Seychelles; dry specimens.

Distribution. Malay Archipelago; Gulf of Mexico.

SECOND SERIES.— BOTANY, VOL. VIII.

Fig.

Fig.

ON THE MARINE ALG OF THE ‘SEALARK’ EXPEDITION.
EXPLANATION OF THE PLATES.

PLATE 12.

1. Gloiophlea articulata, n. sp.. х $.
9. Eucheuma Cottonii, п. sp. Nat. size. Two forms.
8. Calliblepharis prolifera, J. Agardh. x 3.
4. Chylocladia perpusilla, n. sp. x 2.
5. Fertile portion of a ection ramulus of Chylocladia perpusilla. х 63.
6. Laurencia pygmea, n. sp. x 2.
7. Cladhymenia oblongifolia. X 5.
8. Chondria pumila, Vick. x 4.
9. Chondria simpliciuscula, n. sp. x 2.
10. Top of frond, with regular rows of hexagonal cells. х 80.
21. калда Rendle, n. ap. Xi
12. Penicillus of Heterosiphonia Rendlei, with stichidia. x 24.

PLATE 18.

. 13. Pseudendosiphonia Gardineri, п. gen. et sp. х $.

14. Top of frond of Ps. Gardineri. х 19. st.=stichidia.
15. Stichidium of Ps. Gardineri. x 100.

16. Dasyopsis Stanleyi,n. sp. х $.

17. Dasyopsis aperta, п. sp. х 4.

18. Dasyopsis Серри, n. sp. х $.

19. h » Underside of thallus, with clusters of antheridia. х 4.
20. Antheridium.

Хы Бат pore п. вр. x $.

22. Tapeinodasya Ethele, n. sp. x 5.

23. Part “of Trond. x II
24. даа эш (J. Ag.). Nat. size.

25. Galaxaura hawaiiana, Butt. Nat. size.

PLATE 14.
26. Gloiophlea articulata, n. sp. Section through apex of frond. х 32.
27. » » Peripheral filament, showing small cortical cells. x 340.
28. Gloioderma ? expansa. Nat. size.
29. b Transverse section, showing large central cells. x 40.

80. ошар Gardineri. Section through thallus. х 16. c.=central tube; p.c.=peri-
central cells.

31. Oligocladus Prainii, п. sp. х 32.

32. Dasyopsis арегіа, п. sp. Part of penicillus. x 46. т. = mother-cell of downward-growing
hypha.

33. Dasyopsis Серри, п. sp. Young displaced branch. x 130.

34. Amphisbetema indica. Apex of frond. х 24.

35. Peyssonnelia Harveyana. Transverse section. х 128.

ке a ? subgen. Ethelia. Section through thallus; highly magnified.

№ Жол plant, section through Super Mss of thallus ; highly magnified.

WEBER-VAN Воввк.

TRANS. LINN. SOC., SER. 2. BOT. VOL. VIII. Pl. 12.

a

4
ғ
Lot
a
T 1

J. Obbes del.

MARINE ALGZE FROM THE INDIAN OCEAN.

тии Жи TRANS. LINN. БОО, SER. 2. BOT. VOL. VIII. ees

4. Obbes del.

MARINE ALGZE FROM THE INDIAN OCEAN.

Wzszn-vAN Bosse. TRANS. LINN. SOC., SER. 2. BOT. VOL. VIII. РІ. 14.

КУ ҮЛҮ?
| Vf Й / /
í í ff

ү

|

\\

===
~

n A

Wh

W | /)

==

— =
Е

—
—
—

S ss
A

==
MI

(17H
7 72
Y ИЛА

co

=

x NI
ЕГЕ
EX

09990009909

= ES zecegees 20085 295222 ә 2 -

asa
== = 22545255:

Е Е
тис sm

29
А, * `
З қон
o \
e Ou s
x т n ©
aoe В e E ° is
CA o. DH АН % ү»
S ыт %, ЕН < в
\ у A | S X
VV HI G H và H
X S H m A UA IH Q
N Н KH © T & H АХО
<Я ала W WU SY
A 5 D АХ
i ARONA
АХО
RAY
1
үүнү N А
HLH LE EH SY
NOIRES 3
АЯ (УХ)
Пеано
644642 4 p

rU шан
Ep

а
Ame.
as
am

<}

J. Obbes del,

MARINE ALGÆ FROM THE INDIAN OCEAN.

LINNEAN SOCIETY OF LONDON,

MEMORANDA CONCERNING TRANSACTIONS.

сМ First Series of the Transactions, containing both Botanical and Zoological contributions, has been compel
de 30 eres and a few кте ‘sets are still for sale. Only certain single volumes, or parts to complete sets, =
‘be obt: original prices. The price of the Index to Vols, 1-25 is 8s. to the publie, and 6s. to Fellows;
Vola, 26-30, 4s, to the tebe, and 3s. to Fellows.

Second Series of the Transactions is divided into Zoological and Botanical sections. The prices of the

X 8,
ee сы те ° which have been са ieee ыы are as пп Mu oaa Up (For the ды parts зее Zoologi

SECOND vong et (continued). ie
Volume. кыма сыз Fellows:
fc
үт. Part f 9. Ее
Pat IL I90L ... : 0g. G 9
Part ПЕ 1909... 018-0
| Te. .— 0 8 O
18025-26 5 8
1908: ... 012 6
dee. uo
: 9
0
0
0
g

- P

E y
904 .. 014
1905.....0 +

duc ie

im mw "wo

2nd Ser. BOTANY J | Т

[ 148° j

У. Observations on the Morphology and Anatomy of the Genus Mystropetalon, Harv.
By В. J. Harvey-Greson, M.A., F.L.S., Professor of Botany in the University of
Liverpool.

(Plates 15 & 16.)

Read 5th June, 1913.

IN the * Annals of Natural History, vol. ii. no. 12, p. 385 (1839), Sir W. J. Hooker
communicated a short paper by Dr. W. H. Harvey, * On two species of a new South
African genus of the Natural Order Rhizanthez of Blume," in which Harvey furnishes
brief diagnoses of two species of what is now recognised as a genus of Balanophoracese, viz.
Муз торе от. Thomii and M. Polemanni. ln a memoir on “Тһе root parasites referred
by authors to Rhizanthee, &c.", Griffith provides a more detailed diagnosis of M. Thomii
with critical remarks, based on specimens obtained from Harvey (Trans. Linn. Soc.
vol. xix. 1845, p. 303), and fourteen years later Sir J. D. Hooker, in his paper “ On the
structure and affinities of the Balanophorace: ” (Trans. Linn. Soc. vol. xxii. 1856, р. 1),
contributes some additional observations on the subject. Briefer references are made to
the genus in Harvey and Sonder’s * Flora Capensis’ (vol. ii. p. 573, 1861-2), in Bentham
and Hooker's ‘Genera Plantarum’ (vol iii. p. 234, 1880), and in Baillon’s ‘ Natural
History of Plants’ (Engl. Ed. vol. vi. p. 505, 1880). А short account of the genus is
given also by Engler (Pflanzenfamilien, III. r. 1889, p. 252), but Marloth in his * Das
Kapland, insonderheit das Reich der Kapflora’ (Wissenschaftliche Ergebnisse der
deutschen Tiefsee-expedition auf dem Dampfer ‘Valdivia, 1898-99, vol. ii. pt. 3,
p. 901) merely mentions the occurrence of the two species in the neighbourhood of
Caledon, Cape Colony.

While on a visit to South Africa in 1905, my friend, Mrs. Solly, of Knor Hoeck, Sir
Lowry's Pass, Cape Colony, told me that both species grew in or near the pass leading
over the Hottentot Holland Mountains, from Sir Lowry's Pass to Caledon. I had no
opportunity, during my visit, of searching for them, but Mrs. Solly kindly undertook, if
possible, to obtain specimens of M. Thomii for me. This she was able to do, two years
ago, when I received from her several specimens not only in full flower but with the
rhizomes (hitherto undescribed) and the host's roots intact. Mrs. Solly described the
specimens as those of М. Thomit, but after examination, as will be seen later, I found
that they could not be referred to that species and, as 1 think, not even to the rarer
species, M. Polemanni, the only description of which, so far asI am aware, is іп the
short paper by Harvey in the ‘ Annals of Natural History,’ quoted above. Mrs. Solly
informs me that the host roots are those of a Protea, although she says she had
previously found M. Thomii parasitic on grass roots. The first time the rhizome of that

SECOND SERIES.—BOTANY, VOL. VIII. 2

144 PROF. В. J. HARVEY-GIBSON ON THE MORPHOLOGY

species was seen was in 1907, but it was again found іп 1909. Тһе specimens, the
description of which forms the subject of the present paper, are, so far as I am aware,
the first plants of this remarkable genus that have been found entire. It seemed worth
while, under the circumstances, to attempt as complete a description as possible of
the plant, and to present the results of the investigation to the Linnean Society, in
whose Transactions the earlier accounts by Griffith and Hooker were published.

Harvey gives a brief diagnosis of the two species included by him under the genus,
and distinguishes them as follows :—

“М. Thomii, Harv. Anterior bractez broadly oblong (in anthesis), one-third longer
than the lateral. Limbs of the segments of the male perianth lanceolate; female
perianth subglobose, obtusely tridentate.” |

* M. Polemanni, Harv. Anterior bractea spathulate, with a narrow claw. Limbs of
the segments of the male perianth elliptical, very concave ; female perianth tubular,
trifid.”

In addition, Harvey’s figure of M. Polemanni shows the scales on the upper part of
the flowering axis as being ciliate along their margins, while those of M. Thomii ате
smooth.

Harvey gives as the habitat of M. Polemanni, “at Hoouw Hoch Pass, Mrs. Denys,
who communicated the only specimen yet seen to Polemann, from whom I received it.”

The specimens sent to me were collected on March 25th, 1910, near Caledon, growing
on rough rocky ground, in a very exposed situation on the side of a hill. Mrs. Solly
informs me that on the day after her visit to the locality “a veldt fire reduced the
hill-side toa blackened waste.” It would seem, therefore, that the chances of obtaining
many more specimens of the plant at present under consideration are very remote.

HISTORICAL SUMMARY.

The inflorescence of Mystropetalon is described by Harvey as consisting of а thick
fleshy axis covered by linear scales and ending in a spike 3—4 inches in length, densely
covered with male flowers above and female flowers below. The male flower is sur-
rounded by three bracts, the anterior of which is oblong in М. Tomi and spathulate in
M. Polemanni, while the two postero-lateral bracts are more or less united and con-
siderably smaller. All the bracts are ciliate along their margins and keels. The perianth
in both species is tripartite, the two posterior parts being more or less united above,
while the anterior part is free almost to its base, where it fuses with the two posterior
lobes. АП the perianth parts are spathulate. Harvey figures two stamens “ opposite
to and inserted on the posterior segments of the perianth," each with versatile anthers
and cubical pollen grains “with fluted angles.” Іп the centre of the male flower
lies a “ minute abortive ovary.”

The female flower is described also as having three bracts, but here the postero-lateral
segments are larger than the anterior іп M. Thomii, but considerably smaller than
іп И. Polemanni. The perianth is described as trilobed and epigynous and from the

AND ANATOMY OF THE GENUS MYSTROPETALON. 145

centre arises a filiform style. The ovary in both species is “ ellipsoidal, attenuated at
the base," and contains, in М. Thomii, “a fleshy mass (without distinction of ovules)”
and is surrounded by an arilloid “ torus” Harvey was thus unable to distinguish either
ovule, seed, or embryo, although he figures what is, apparently, intended to represent an
embryo in M. Polemanni.

Griffith (Trans. Linn. Soc. xix. 1845, p. 337) gives a rather fuller description of
JM. Thomii from specimens sent him by Harvey. Не states that there are three stamens
in the male flower, one (when present) rudimentary attached to the anterior *sepal" He
also mentions the occurrence of two abortive stamens in the female flower. Тһе style is
described as trisulcate and the stigma as trilobed. Griffith refers to a doubtful embryo
(the word is followed by a note of interrogation) centrally placed and separated from tbe
“epicarpium " by a spongy cellular tissue. Probably Griffith did not see the embryo
at all, for its position is apical, not central, and it is so small that it might well escape
notice unless investigated by means of serial microtome sections. Indeed, in his
appended observations he speaks of the “central tissue of the fruit" as “ not separable
as one body, breaking up rather on the slightest pressure occasioned by attempts at
dissection." “Та this respect," he adds, “1% differs considerably from the embryo of
Balanophora and Pheocordylis.” Later on he speaks of it as a *sporuliferous mass."
He does not mention M. Polemanni at all.

Hooker (Trans. Linn. Soc. xxii. 1859, p. 31) draws attention to Harvey and Griffith's
failure to find the real embryo and describes it in M. Thomii * at the upper part of the
seed, lying quite loose in its axis, with the radicle pointed to the hilum and close to it."
The “ sporuliferous mass” referred to by Griffith is, according to Hooker, “albumen.”
In the * Flora Capensis' (vol. ii. p. 573), Harvey and Sonder apparently accept Hooker's
interpretation, for they describe the fruit as one-seeded with a “thin үшеу epicarp and
a crustaceous endocarp," while the embryo is stated to be “very minute in the base of
easily friable albumen." Bentham and Hooker (Gen. Pl. iii. p. 235) state with greater
accuracy that the embryo is apical, Engler (Pflanzenf. ПІ. т. p. 252) adds one item of
interest with reference to the structure of the ovary (under generic characters). He
describes it as containing three ovules arising from a free placenta, each being reduced
to an embryo-sac. i

Beyond brief references to the genus in a few other systematic works, I know of no
further published accounts dealing with the subject of the present memoir.

Tur RHIZOME AND ITS RELATION TO THE Host PLANT.

No description of the morphological and anatomical features of the subterranean parts
of either species of Mystropetalon has, so far as I am aware, ever been published ; i
the authorities quoted, when they mention its rhizome at all, describe it as *' unknown. :
The first time it was found was, as above stated, in 1907, but whether these specimens
were really those of М. Thomii or of the present plant, I am unable to say, although,

probably, as they were seen by Dr. Marloth, they were of M. Thom If that be so,
Z2

440 7 PROF. В. J. HARVEY-GIBSON ON THE MORPHOLOGY

then the first rhizomes of the plant, whichever it is, under investigation, to be gathered
are those on which the present observations have been made.

The rhizome is usually, as І am informed, parasitic on the roots of a Protea, and the
material in my possession shows rhizomes in several stages of development.

A fully adult rhizome has the appearance of a very nodular, irregular mass, about
40 mm. in diameter, closely investing the root of the host. Such a rhizome is figured in
Pl. 15. fig. 1, while fig. 2 shows younger rhizomes in various stages of development, one
of them beginning to give origin to two erect flowering shoots.

Anatomically, the rhizome is covered with a layer of irregular thickness composed of
eutinised cells without intercellular spaces, forming a superficial cork. The main mass
of the rhizome is composed of fairly large parenchymatous cells, the majority of which
contain a substance which I at first presumed to be related in chemical composition to
the resinous material described as “ balanophorin” by Goeppert (Verh. K. Leop.-Car.
Akad. d. Nat. vol. xviii.) in Balanophora. Cells containing this deposit are scattered all
through the rhizome, but are specially abundant in the cortex and mixed with the
vaseular tissue (figs. 5 & 6). An attempt was made to determine the chemical nature
of the brown deposit, which amounts to about 50 per cent. by weight of the dried
rhizome, but only negative results were obtained. Since the substance is insoluble in
ether, aleohol, and benzene, and does not burn readily, it may be concluded that it is
neither a resin nor a wax. It is unaffected by boiling in water and the solution gives no
reducing-sugar tests. The possibility of its being a glucoside is excluded by the fact
that itis insoluble after boiling for three minutes in dilute sulphurie acid. It gives
none of the reactions of protein. It is not affected by strong sulphuric acid in the cold,
but the powdered rhizome gives, on standing, a vigorous reaction to Molisch's test. One
is thus forced to the conclusion that 16 has relations with the carbohydrates. It gives no
reaction to iodine nor to iodine and sulphuric acid. It cannot be a simple carbohydrate,
but is probably a peculiar carbohydrate of the cellulose group. The Molisch test when
repeated under the microscope was not entirely satisfactory, for the substance took on
not a violet but a deep red-brown colour. Unfortunately, the amount of tissue at my
disposal was too limited to permit of a detailed quantitative analysis. For convenience
I shall speak of the substance in the course of this paper as “ mystrin," for if Goeppert
and Poleck’s determination of the corresponding substance in Balanophora as а wax ог
resin be correct it is certainly not “ balanophorin."

It was found to be quite impossible to determine any definite arrangement of vascular
bundles in the rhizome owing to its extremely wrinkled and irregular form. Bundles of
short tracheids with reticulate thickenings, almost verging on spiral, wander irregularly
through the ground parenchyma. These strands are accompanied, usually, but not
always, by elongated elements obviously with sieve-tube characters. Many of the
tracheids may be described as scalariform, although Solereder (* Systematic Anatomy of
Dicotyledons,’ vol. ii. р. 739) doubts the presence of such elements in the Balanophoracee,
although this is asserted by Eichler. Fig. 5 shows such * scalariform ” tracheids mingled

AND ANATOMY OF THE GENUS MYSTROPETALON. 147

with “ mystrin " parenchyma taken from the rhizome just above the point of origin of
the haustorium, while fig. 6 shows the short reticulate xylem elements accompanied by
“mystrin” cells and elongated nucleate phloem parenchyma. Accompanying these
elements were chains of cells with large nuclei resembling those described by Pierce in
Brugmansia, Rafflesia, and Balanophora (Ann. Bot. vol. vii. 1893), but the mode of
preservation of the material did not permit of the determination of the ultimate
connection of the cells of the haustorium with those of the host.

THE INFLORESCENCE.

The inflorescence is a clavate spike from 80 to 100 mm. in length by about 20 mm. in
diameter near its apex (РІ. 15. fig. 1). The axis tapers to its point of origin from the
rhizome, and is covered for two-thirds of its length with closely-set linear leaves, while
the upper third bears female and male flowers in two distinct zones, the former
occupying the lower, the latter the upper portion of the floral area.

The leaves are rather pointed in the basal region of the axis, but become more
truncated as the floral region is approached. They are arranged in a dense spiral and
are of a bright red-brown colour due to the abundant deposits of **mystrin." The leaves
have broad bases, are relatively thick centrally, and bordered by thin wings. Тһе
abaxial surfaces, as well as the margins, are densely covered with long unicellular
hairs, but the median region of the adaxial side is quite glabrous. There is one median
vascular cord consisting of a variable number of short, more or less irregular, reticulate
and spiral tracheids, which lie immediately below the adaxial epidermis, accompanied
by а very small-celled phloem. As in the case of the rhizome, “mystrin” cells are
interspersed among the vascular elements. The main mass of the leaf consists of
spherical and polygonal parenchyma with thick pitted walls and with small intercellular
spaces; most of these cells contain “ mystrin." The epidermis, which consists of cubical
or polygonal cells, is destitute of stomata (fig. 11).

Fig. 7 (Pl. 15) shows a transverse section of the axis taken at the level of the origin
of vegetative leaves. The superficial cells (fig. 8) are completely filled with ** mystrin ”
and are comparable in size and general appearance with those of the body of the axis as
a whole. Many of these fundamental cells are also more or less filled with the same
brown deposit. The vascular bundles are twenty to thirty in number and are distributed
quite irregularly through the ground-tissue. These bundles have apparently no cambium,
and in all cases show cells filled with “ mystrin " interspersed among the vascular tissue
elements proper. At fig. 9 (Pl. 16) one of the smaller bundles is figured, and it will be
seen that it consists of a number of tracheids, some of which have spiral, others reticulate
thickenings, while the phloem consists of thin-walled, elongated, nucleated elements
resembling phloem parenchyma. Хо sieve-tubes, exhibiting definite sieve-plates, could
be distinguished. The larger bundles have several patches of tracheids and several
patches of phloem cells more or less separated by larger, thin-walled parenchyma
containing “ mystrin” (fig. 10).

148 PROF. R. J. HARVEY-GIBSON ON THE MORPHOLOGY

TuE MALE FLOWERS.

The male flowers, when young, are arranged in closely-packed alternating whorls,
which gradually open out as maturity is reached. The central axis cells are loaded with
** mystrin,” as are also the cells of the bracts and perianths. Within the perianth-lobes
lie the anthers of the two stamens, whose connectives are also surrounded by cells
containing the same brown deposit. The male flower (fig. 13) is surrounded at its base
by three bracts. One of these, anterior in position, is stalked, with a spathulate lamina,
and covered with long unicellular hairs on the abaxial side. There is a single central
vascular bundle as in the axis-leaves, and all the cells of the ground-tissue are filled
with * mystrin." Тһе latero-posterior bracts are very much smaller and more scale-like,
sessile, and fused for a certain distance upwards from their point of origin. These
bracts are also furnished with unicellular hairs along their margins and on their posterior
surfaces. Within the bracts and arising opposite to them are three perianth segments
which are fused at their bases to form a very short cup. The perianth segments are
nearly twice the length of the anterior bract. The anterior segment is slightly shorter
than the two posterior segments and stands apart from them. It has a single median
vascular bundle running through the slightly convexo-concave claw and is continued
unbranched into the concave spoon-shaped lamina. The perianth segments are only
slightly concave, and resemble exactly the figures given by Harvey (Ann. Nat. Hist.
ii. tt. 19, 20) for those of М. Thomii. All the cells, both of claw and limb, are filled
with “ mystrin." Griffith (Trans. Linn. Soc. xix. р. 337) in his description of М. Thom
says, *stamen tertium, dum adest, rudimentarium, adnatum ungui sepali tertii antici.”
I have examined over a hundred male flowers of the plants in my possession, but I
have never found the slightest trace of any such third stamen. The posterior perianth
segments have free claws, but the limbs are united above. They are similar in form and
histological structure to the anterior segment, and have also concave limbs forming
a covering to the two functional stamens. The stamens arise just where the claw
broadens out into the limb, and the filament is inserted on the anther about half-way
up its length. Harvey (2. с. р. 386) describes the anthers as versatile, but Griffith and
Hooker speak of them as extrorse. As a matter of fact, in the unripe flower the
anthers are extrorse, but as the flower opens the anther becomes versatile and ends by
being introrse, owing to a complete inversion of position of the anther with reference to
the axis of the flower.

The anther is markedly bilobed and the lobes open by longitudinal dehiscence. The
cells of the filament, including the vascular bundle, are full of “ mystrin,” and several
layers of cells surrounding the connective also contain the same deposit. The wall when
ripe is composed of one layer of cells, and the walls of these cells show the usual rib-like
thickenings common in anther-walls, save in the region of longitudinal dehiscence (fig. 14).
The pollen grains, as described by Harvey, are cubical as a rule, though some are
pentagonal or, rarely, hexagonal (fig. 15). Harvey figures the angles of the grains as

AND ANATOMY ОЕ THE GENUS MYSTROPETALON. 149

fluted, but I have not been able to distinguish any such indentations. In most grains one
prominent nucleus is visible, although two can be made out in some of the larger grains.

At the bottom of the shallow cup formed by the fusion of the bases of the perianth
segments, there occurs a parenchymatous protuberance, the cells of which resemble those
of the “arilloid” swelling which surrounds the base of the ovary. This protuberance is
continued into a short projection whose cells are filled with “ mystrin." This probably
represents the rudimentary ovary mentioned by Harvey and by Griffith, but none of my
sections exhibits any indication of distinct ovarial structure (fig. 16).

Тне FEMALE FLOWER.

Descriptions of the female flower of M. Thomii alone are given by Griffith and
Harvey, the latter of whom only figures it; Harvey also figures the female flower of
М. Polemanni but gives no account of it. Harvey's drawings show, approximately,
equal bracts, ciliate along the margins, and the anterior segment with a hairy
external median ridge. Centrally placed (in M. Polemanni) lies a swollen, cup-shaped
wrinkled “torus” from which springs the ovary, bearing at its apex a minute, trifid,
tubular perianth, and a long style, without апу prominent stigma. Harvey figures the
bracts of М. Polemanni both in “anthesis” and in the fruiting condition as
approximately of the same length, the anterior bract narrowing to its base and covered
with hairs, while the latero-posterior bracts are broader and only marginally ciliate.
The bracts of the female flower of М. Thomii, on the other hand, are, according to
Harvey's figures, of equal length in anthesis, but, in the fruiting condition, the latero-
posterior bracts are represented as about twice the length of the anterior bract.
Again, the perianth of the female flower of М. Thomiz is an almost spherical сар, with
a trilobed margin, contrasting, in this respect, with the narrow, tubular, trifid perianth
of М. Polemanni. Harvey also figures the ovary as basally pointed, the basal
projection rounded off and seated in the central depression of the “torus.” Griffith
describes the ovary of М. Thomii as “globosum, parce puberulum, basi attenuata” ;
Harvey also mentions the “ minutely hispidulous ” character of the wall of the ovary.

As my own observations differ in some respects from those of Harvey, it may be more
convenient to describe the whole flower de novo. The description, it may be prefaced,
is that of a flower whose ovule had been fertilised, although the style was still present
at the apex of the ovary.

The flowers are apparently proterogynous, for in those inflorescences which had
mature male flowers with anthers still closed or in the act of dehiscing, the female
flowers had lost their styles and perianths, and the ovary һай become transformed into a
small, almost black fruit with a sclerotic pericarp. In the single younger inflorescence
amongst my material the male flowers were still immature, while the bracts of the
female flowers already enclosed young fruits but with styles still attached.

In a letter from Mrs. Solly, accompanying the specimens, she makes an interesting
observation on the pollination of Mystropetalon. My correspondent writes as follows :—

150 PROF. Е. J. HARVEY-GIBSON ON THE MORPHOLOGY

“The female flowers open first. I found bees visiting the flowers in the orthodox
manner; I caught one and found it covered with pollen. This point I had not been
able to fix before.”

As already stated, the youngest female flower I could find contained an ovary whose
ovule had obviously been fertilised. The flower is surrounded, as in the case of the
male, by three bracts, the anterior being pointed linear and about one-fourth longer than
the two latero-posterior, which are pointed ovate. The anterior bract bears very
numerous unicellular hairs both over the abaxial surface and along the margins, while
the latero-posterior bracts are ciliate on their margins only. Arising centrally is a
circular, transparent disc (4 torus” of Harvey), from the cup-like depression in which
arises the ovoid or elliptical “ ovary " bearing apically a long filiform style. Just at
the junction of the ovary and style springs the extremely small cup-shaped perianth
(figs. 17 & 18). I have examined very many flowers and never found the tubular,
trifid form of perianth figured for M. Polemanni by Harvey. The form of the perianth,
on the other hand, invariably resembled that figured by the same author for M. Thomit.
The margin of the perianth was, however, not trifid, but was subdivided into six or more
lobes of approximately equal size. Finding this deviation from Harvey's figures of the
perianth of М. Polemanni, I at first thought that my material might include both species,
but that this was not the case was shown by an examination of male and female flowers
from the same inflorescence. It will be noted in the diagnoses of the two species given
- by Harvey that that author separates M. Thomii from М. Polemanni, on the ground
that the former has “anterior bractez broadly oblong” with a *'subglobose, obtusely
tridentate ” female perianth and “somewhat concave" laminz in the male perianth,
while M. Polemanni has its “anterior bractea (in the male flower) spathulate with a
narrow claw " and with “ very concave” laminse in the segments of the perianth, while
the female perianth is “tubular, trifid." The present material seems to combine these
characters, for the male flowers have, on the whole, the characters of M. Polemanni, while
the female flowers resemble more closely those cf М. Thomii. (The question of
identification is discussed below.)

The perianth cup has a vascular supply consisting of a few very delicate spiral and
reticulate tracheids arranged in two or three strands, and the rather regularly arranged
oblong cells of which the perianth is mainly composed are filled with **mystrin."
Arising from the base of the perianth cup towards its posterior side are two small
parenchymatous nodules representing the pair of abortive stamens mentioned by
Griffith (ig. 18). "They show no histological differentiation and are not present in all
flowers. Ко trace of an anterior stamen could be made out after examination of several
hundred serial sections of female flowers. The ovary is elliptical and sparsely covered
with small unicellular hairs, which apparently drop off soon after fertilisation, for
slighter older fruits are quite glabrous. No hairs occur on the perianth. From the
base of the perianth cup arises the relatively long filiform style. It is considerably
constricted at its point of origin from the ovary (fig. 18), but expands to its normal
diameter before reaching the rim of the perianth; thereafter it maintains a uniform

AND ANATOMY OF THE GENUS MYSTROPETALON. 151

thickness and finally ends in a papillose stigma which in most cases shows subdivision
into three more or less well-marked lobes. Harvey’s figures of the female flowers of
М. Polemanni do not show any prominent stigma at all. Usually three delicate
vascular bundles pass up the style ending in a brush-like tuft of reticulate tracheids in
each of the stigmatic lobes (fig. 19). The style is composed of oblong parenchyma
full of * mystrin," but I have been unable to see any indication of the three grooves
mentioned by Griffith—though it is possible that these may be visible іп the fresh
condition. On tracing the course of the bundles down the style it is found that they
anastomose from time to time, sometimes splitting into four or even five bundles and
then reuniting, finally fusing into two larger bundles in the constricted region just
inside the perianth. At the point of origin of the perianth these strands give off two or

sometimes three chains of very delicate spiral tracheids, which enter the perianth, while
the main strands break up into several fine branches which pass down through the
inner layer of the ovarial wall, finally uniting again into one larger bundle in the
conical base of the ovary, where, after passing through the torus, it unites with
the vascular system of the axis of inflorescence.

THE FRUIT.

The fruit remains enclosed within the persistent and accrescent bracts, its base
surrounded by the almost colourless swollen dise (4 torus" of Harvey). Тһе perianth
and style are deciduous, leaving a ragged scar at the apex of the fruit. The ripe fruit
is orange-shaped and about 2 mm. in diameter, and of a deep brown-black colour. It
tapers rapidly to its point of attachment in the swollen receptacle. In form it much
more closely resembles the fruit of M. Thomii as figured by Harvey than it does the
oblong elliptical fruit of М. Polemanni. The pericarp (fig. 20) consists externally of a
double layer of flattened cells whose cavities are completely filled with “ mystrin," thus
accounting for the intense brown-black colour of the fruit as seen with the naked eye.
These layers are followed internally by two layers of very thick-walled sclerotic cells,
deeply pitted and partly filled with “ mystrin.” Then follows а layer of polygonal or
oblong cells, rather loosely arranged, merging into a layer of more elongated cells,
among which run the delicate vascular strands already referred to. All of these cells
are filled with “ mystrin." Тһе pericarp encloses a mass of large polygonal cells whose
outermost layer is in the form of smaller and more flattened plates. These cells possess
large nuclei and finely reticulate and vacuolated protoplasmic contents which take on
a deep stain with the usual anilin dyes. This tissue, obviously endosperm, has no
** mystrin " in its cells, and for that reason is sharply marked off from the pericarp.

Engler (Pflanzenfamilien, ПІ. 1. 252) describes the fruit as “ nussartig, mit fleischigem
Epicarp, dünn krustigem Mesocarp und dünnwandigem Endocarp.” In no case have
I seen any evidence of “ fleshiness” in the external layers of the fruit-wall. Indeed,
sectionising the fruit isa matter of some difficulty owing to the brittleness of the
external wall. As Engler makes no mention of the fleshy disc subtending the fruit,

SECOND SERIES.—BOTANY, VOL. VIII. 2A

152 PROF, В. J. HARVEY-GIBSON ON THE MORPHOLOGY

опе is led to conclude that by “fleshy epicarp” he means to indicate this disc; but
seeing that it arises below the ovary and is quite independent of it (fig. 21), and further,
never reaches more than a quarter of the distance up the exterior of the fruit, it is
impossible to accept such an interpretation of its nature. If a differentiation into
distinct layers in the fruit-wall is to be recognised at all, then the three layers indicated
above must stand for epi- meso- and endocarp respectively, and the succulent “ torus”
must be regarded as an accessory development from the peduncle above the point of
origin of the bracts, analogous to that seen on а much larger scale in such fruits as
Anacardium.

None of the female flowers in my material was young enough to enable me to
confirm or refute Engler’s statement that there are three ovules in the ovary, each
reduced to an embryo-sac. There is certainly no confirmatory evidence to be obtained
from an examination either of the young or mature fruits.

The fruit is connected with the axis of inflorescence by an attenuated stalk composed
centrally of small deeply stainable parenchyma, through which runs the vascular cord
supplying the pericarp and style. These vascular elements are surrounded by several
patches of “ mystrin” cells. From this axis arises a massive swollen ring of large, thin-
walled parenchyma with little or no contents in the cells, forming the torus. It is
covered externally by a single strongly cutinised layer of flattened ceils whose contents
stain deeply. The fleshy disc adheres to the fruit when ripe, separation occurring at its
junction with the axis of inflorescence.

The embryo was first made out by Hooker in M. Thomii, who figures it as a minute
pear-shaped body imbedded in endosperm at the apex of the fruit, just below the point
of origin of the style. My sections confirm this position, but show an obconical embryo -
composed of uniform undifferentiated parenchyma lying in a minute cavity in the
relatively large-celled endosperm (fig. 21). None of my sections show any indications of
a placenta or funicle, nor does the seed possess anything in the way of a testa.

IDENTIFICATION.

Ав I have pointed out in the preceding pages, the plants on which the above
observations were made do not appear to conform strictly with Harvey’s diagnosis of
either of the two species named by him. In some respects they resemble JM. Tomi, in
other respects they suggest М. Polemanni, while in yet others they agree with neither
of these species. Sir David Prain informs me that specimens had been sent to Kew by
Mr. Bolus and labelled M. Toma, but that all of them on examination agreed with
Harvey's drawings of M. Polemanni.

. The following tabular statement shows the points of agreement and difference
between Harvey's descriptions and figures as compared with the characters shown by
the material on which the present paper is based.

AND ANATOMY OF THE GENUS MYSTROPETALON, 153

М. Тноми. M. Рогемаммг. Present MATERIAL.
Male flower. Male flower. Male flower.
1. Anterior bract oblong. l. Anterior bract spathulate. l. Anterior bract spathulate.
2. Posterior bracts 2 length of 2. Posterior bracts $ length of 2. Posterior bracts 1 length of

anterior bract. anterior bract. anterior bract.

3. Anterior perianth lobe 3. Anterior perianth lobe $ 3. Anterior perianth lobe rather
rather shorter than posterior. length of posterior lobes. shorter than posterior lobes.

4. Perianth lamine slightly 4. Perianth lamine very con- 4. Perianth laminæ slightly
concave. cave. concave.

5. Pollen cubical with fluted 5. Pollen cubical with fluted 5. Pollen cubical or pentagonal
angles. angles. without fluted angles.

Female flower. Female flower. Female flower.
6. Anterior bract oblong. 6. Anterior bract rather spath- 6. Anterior bract lanceolate.
ulate (in fig.).
7. Posterior bracts much 7. Posterior bracts equal to 7. Posterior bracts shorter than

longer than anterior in the anterior. anterior,
fruiting state.

8. Perianth ^ subglobose ог 8. Perianth tubular. 8. Perianth — subglobose ог
ellipsoid. campanulate.
9. Perianth margin 3-lobed. 9. Perianth margin markedly 9. Perianth margin multifid.
trifid.
10. Ovary ovoid. 10. Ovary oblong. 10. Ovary ovoid.
11. Stigma discoid. 11. Stigma not swollen. 11. Stigma trilobed.

In point 1 only my material agrees with M. Polemanni; in points 3, 4, 8 and 10 it
agrees with M. Thomii; in points 2, 5, 6, 7, 9 and 11 it agrees with neither.

I was able, through the kindness of Professor H. Dixon, F.R.S., of Trinity College,
Dublin, to explore Harvey’s Herbarium housed there, but unfortunately we were unable
to find the type-specimens, and since my return Professor Dixon writes me to the effect
that after another thorough investigation he has failed to discover them. Under: the
circumstances, and seeing that my material as above described, cannot be said to be
either of Harvey’s species, I am obliged to consider my plants as belonging to a new
species, to which I would desire to give the name of М. Sollyi in recognition of its
collector, Mrs. Solly, of Sir Lowry’s Pass, Cape Colony, with the specific characters
given in the above tuble.

BIBLIOGRAPHY.

Harvey, W. Н.--“ On two species of a new South African genus of the Natural Order Rhizanthe: of
Blume.” Annals of Natural History, vol. ii. no. 12 (1839) pp. 385-388, tt. 19, 20.

Grirrira, W.—“ On the Root Parasites referred by authors to Rhizanthee ; and on various plants
related to them.” ‘Trans. Linn. Soc. vol. хіх. (1845) pp. 303-347, tt. 34-39.

Ноокев, J. D.—*On the Structure and Atfinities of the Balanophoracee.” Trans. Linn. Soc.
vol. xxii. (1856) pp. 1-68, tt. 1-16.

Harvey, W. H., & О. W. Sonper.—Flora Capensis, vol. 11. 1861-2, р. 573.

Вемтнлы, G., & J. D. Hookex.—Genera Plantarum, vol. iu. pt. 1, р. 234.

15% MORPHOLOGY ОҒ THE GENUS MYSTROPETALON.

Barnrow, H.—Natural History of Plants. Eng. Ed., vol. vi. p. 505.

ENcrzn, E.— Pflanzenfamilien, vol. TIT. т. (1889) p. 252.

Млвготн, R.—* Das Kapland, insonderheit das Reich der Kapflora.” Wiss. Ergeb. d. deut. Tiefsee-
expedition auf dem Dampfer * Valdivia,’ 1898-99, vol. ii. pt. 3, p. 501.

Gorrrert, Н. R.—* Ueber den Bau der Balanophoreen.” Verh. Leopold.-Carol. Akad. d. Nat.
xviii. Supp. 1, 1841, pp. 231-272, tt. 1-8.

Зотеверкв, H.—Systematic Anatomy of the Dicotyledons, vol. ii. (1908) рр. 798-799.

Pierce, G. J.—“ On the Structure of the Haustoria of some Phanerogamic parasites." Ann. Bot.
vol. vii. (1893) pp. 291—327, tt. 13-15.

EXPLANATION OF THE PLATES.
PLATE 15.

Fig. 1. Rhizome and flowering axis of Mystropetalon Sollyi, n. sp., showing attachment to root of
Protea sp. (nat. size).

2. Young rhizomes, in various stages of development, on a Protea root (nat. size).

8. Transverse section of a young Protea root (P) with the rhizome and haustorium of
Mystropetalon. т, xylem and s, sclerotic fibres in host root. (х 16.)

4. Transverse section of a Protea root with entering haustorium. Centrally the deformed xylem
of the host root ; 0, “mystrin” cells ; bf, bast fibres of host root ; s, sclerotic tissue of the
cortex; ck, cork layer; p, phloem parenchyma and sieve-tubes of the haustorium ;
т, tracheids of the haustorium. (x 60.)

5. Tracheids mingled with “ mystrin ” cells from the haustorium. (х 850.)

т Tracheids (reticulate), “ mystrin ” cells, and phloem elements from the rhizome. (x 550.)

. Transverse section of the base of the flowering axis (x 16); a, vascular strand ; б, “ mystrin ”
есш "(X 15)
8. Superficial layers of the stem cortex showing the “ mystrin deposits. (х 550.)

PLATE 16.

Fig. 9. Transverse section of one of the smaller vascular bundles of the flowering axis. 6, “шузы”
cells; 2, xylem ; ph, phloem. (x 550.)
10. Transverse section of one of the larger vascular bundles of the flowering axis (lettering as in
fig. 9). (x 250.)
11. Transverse section of one of the scale leaves. (x 60.
12. Longitudinal section of one of the vascular strands of the floral axis, showing xylem elements,
phloem, and “© mystrin ” cells. ( x 550.)
13. Male flower. (x 35.)
14. Section of the wall of the anther with pollen grains. (x 350.)
15. Pollen grains, six- and seven-sided. (x 550.)
16. Abortive ovary of the male flower. (x мыл.
17. Female flower. (x 60.) |
18. Longitudinal section of the perianth and чн of the female flower, жү abortive stamens
| (semidiagrammatic). |
э Trilobed stigma. (х 350.)
20. Transverse section of fruit-wall and portion of the endosperm. ` (x 550.)
21. Longitudinal section e ma the fruit, showing the “ torus,” fruit-wall, endosperm, and €
( x 60.)

HARVEY-GIBSON.

TRANS.LINN.SOC. SER.2. BOT. VOL. VIII. PL. 15.

MYSTROPETALON, Harv.

HARVEY- GIBSON. TRANS. LINN. SOC. SER. 2. BOT. VOL. VIII. PL. 16.

> ai ЖС;

E.

O

i We 26
11 — ы — = :

"MI AVIS

*
СУС

4444 444546
каттата

ХА

(319584646744 465464

АА CUN

із

%%

D

M
|
y
4,

⁄ > #" 7
ӨРГЕ тлі

NM

әй те

шығы,
бары,

t гуы.

—

= T isi

еерее

London lepas epi Co. imp.
MYSTROPETALON, Наву. :

LINNEAN SOCIETY OF LONDON.

= MEMORANDA CONCERNING TRANSACTIONS.

22 The First Se £ the Transactions, containing both Botanical and Zoological contributions, has been completed |
in з Vols, and a few entire sets are still for sale. Only certain single volumes, or parts to complete sets, may -
at the original prices. The price of the Index to Vols, 1-25 is Өз. to the public, and 6s. to Fellows; to.
D te too pli m Se Fallow |
‚ Second Series of the Transactions is divided into Zoological and Botan ical sections. The prices of the `
parts: of these on have been pubis lished a are as undermentioned. (For the Zoological parts see Zoological -

Ѕксохр Serres.—Borany болшу.

: E When Price to the Price i.
fn Published. Public. Fellows.

Y т , Part |

о
с 6

=

баз

2ud Ser. BOTANY.) (VOL. VIII. PART 5.

THE

TRANSACTION

OF

ON THE CUTICLES OF SOME RECENT AND FOSSIL
OYCADEAN FRONDS. = =

- «^ 2M HAMSHAW THOMAS, мл FCB, =
| | =. | _ CURATOR or 1 THE Ei FAD MUSEUM, CAMBRIDGE, > 222

с КЕНЕ BANCROFT, BSe, ELS,

ie - 1851 RESBARCE SCHOLAR, NEWNHAM COLLEGE, CAMBRIDGE.

ERRATA.
Transactions of the Linnean Society, Ser. 2, Botany, Vol. VIII. Part 5.

т line 17 should read * PTILoPHYLLUM РЕСТЕМ (Phill.) Morris.”

182 ,, 5, for Stangeria read Ptilophyllum.

186 , 1 should read “ Orozamires GRAPHICUS (Leck.) Schimp.”

im > WW 35 “ ANOMOZAMITES NirssoNr (Phill.) Sew.”

190 ,, 24 А * NILSSONIA COMPTA (Phill.) Bronn."

191 ,, 4 from bottom should read “ NIrssoNIA MEDIANA (Leck.) Fox-Strangw.”

194 ,, Z should read “ Pritozamites Lecxensyi (Leck.) Nath.”

be 158221

VI. On the Cuticles of some Recent and Fossil Cycadean Fronds. Ву Н. НАМЗНАМ
Tuomas, M.A., F.G.S., Curator of the Botanical Museum, Cambridge, and NELLIE
Bancrort, B.Sc., F.L.S., 1851 Research Scholar, Newnham College, Cambridge.

(Plates 17-20. and 32 Text-figures. )

Read 6th November, 1913.

CONTENTS.
Page

L ИИИЦИЦЕГЕТСІІ еее ач S 52225550 ЕИ as 155
II. Cuticles of some Recent Cyeads .................................. 157
(a.) Material and Methods .................................... 157
(5.) Previous references ...................................... 157
(c.) Description of the Cuticles ................................ 158
L DONOR С... ена ене see etra hd rr kae 158
ii. Detailed description .................................. 158
iii. General summary of the structure of the recent Cuticles .... 174
III. Cuticles of Fossil Сусайеап Fronds ......... di Lx ELE IC 177
(a.) Previous observations on eutieular strueture .................. 177
(6.) Detailed description ............ ........................ 179
(c.) The Cuticles of some other types of Fossil Cycadean fronds ...... 194
(d.) Summary of the structure of the Fossil ОШИО ee eee ee 195

IV. Comparison of the Cuticular Structure of Recent and Fossil Cycadean
Fronds, with notes on that of other бушповрегтв.................. 197

I. INTRODUCTION.

IN an investigation of the cuticles of recent and fossil Cycads our attention is
principally concentrated on the form and structure of the stomata, a study of which has
attracted many observers. Тһе physiological functions of the euard-cells are so important,
and their shape and action are so uniform in plants of widely separated classes, that they
may be regarded as highly adapted to their special work, and as structures which were
evolved at a comparatively early period in phylogenetic development. But while the
morphology of the stomata is uniform, many differences of structure, especially of
thickening, do exist, rendering the work of classifying stomata by no means easy. One
of the few general attempts to make a survey of the whole range of these structures,
to classify them in groups, and to find broad general tendencies in their variation, is
that of Porsch (1905), who has introduced a conception of a phylogenetic ground-plan
modified by adaptations to external conditions. The idea that stomatal structures
SECOND SERIES.—BOTANY, VOL. VIII. 2B

156 МЕ. H. HAMSHAW THOMAS AND MISS N. BANCROFT ON

may be valuable from the phylogenetic standpoint has not yet been worked out in
any detail, nor has any attempt been made to compare modern plants with their
fossil ancestors with regard to this character. One of us (Thomas, 1911) has shown,
however, that some of the Carboniferous Calamites possessed the transverse bands of
thickening which are so characteristic of the guard-cells of the modern Equisetums,
and it seemed important to make further investigations of this kind. An especially
favourable opportunity is presented in the Cycadophyta; the modern representatives
of this group possess considerable variety of frond form, while the fossil fronds are also
varied, and are abundant in a fairly good state of preservation in Mesozoic rocks.
From a purely palzeobotanical standpoint the work offers some attractions, because we
have so often to deal with isolated fossil fronds which possess few points of distinction
apart from external form; the possibility of classifying them and obtaining information
as to their affinities by means of a study of their epidermal structure seems important.

So long ago as 1856 Bornemann made a brief comparative study of the epidermal
characters of recent Cycads for the purpose of identifying fragments of fossil Cyeadean
fronds from the Lettenkohle beds of Thuringia, but his studies of the fossil forms were
not very critical. He thought that the Mesozoic fronds were nearly related to if not
generically identical with, some of the modern forms, and therefore expected that the
study of the latter would at once facilitate the identification of the fossils. We are
now tending to regard the fossil Cycads as more and more widely separated from
their recent representatives, and we may at the outset expect the similarities in their
structure to be few.

Recent work on the fossil and living members of the Cycadophyta has raised them to
a very important position among the Gymnosperms. The investigation of the Carboni-
ferous Pteridosperms, and the recent discoveries with regard to the Bennettitales have
caused many botanists to seek for the origin of the Angiosperms amongst the ancestral
plexus of Cycadean forms, much attention having been focussed on the Mesozoic
plants of this affinity. It is, therefore, important to utilise every form of evidence
as to the relationships of these plants.

The method of enquiry adopted in the present paper is, first, the detailed investigation
of the epidermal structures found in the modern Cycads, and the attempt to find a type
of stomatal structure characteristic of the group, the departures in detail from this type
being noted. Subsequently, the epidermal structures of all the available Mesozoic
fronds are described, and an attempt is made to estimate their relations in this respect
to one another and to the modern fronds. In studying the fossil fronds, although we
have many beautifully preserved examples of their cuticles, yielding much information
in surface view, we have as yet been unable to get satisfactory preparations showing the
guard-cells of the stomata in section; we are thus расы! in an unfavourable position
for comparing them with the recent structures.

CUTICLES ОҒ CYCADEAN FRONDS. 157

П. CUTICLES OF SOME RECENT CYCADS.

(а.) MATERIAL AND Метнорв.--Мовб of the material used in the investigation of
recent cuticles was specially obtained from the Glasnevin Botanical Gardens, Dublin,
and from Kew, and was fixed in medium chromacetic acid. In some cases ordinary
laboratory material was used.

The types examined included Cycas (5 species), Dioon (2), Ceratozamia (3), Macro-
zamia (6), Zamia (8), Encephalartos (8), Bowenia (1 form), Stangeria (2 forms),
Microcycas being the only genus unrepresented.

The upper and lower cuticles of the leaves were examined in surface view and in
section. In many cases the epidermises were stripped off easily, but in others it was
necessary to use Schulze’s macerating fluid. Sometimes ршизе were treated with the
fluid and then washed in ammonia, as in the case of the fossil cuticles (vide infra); after
this treatment only the cutin layer remains, with little or no trace of underlying cells.

Permanent preparations of cuticles removed in either of the above ways were
made. Bismarck brown, Congo red, and diamant fuchsin were used for staining.
Lloyd (1908), in his researches on the Physiology of Stomata, seems to consider
Congo red more satisfactory as an epidermal stain than Bismarck brown; in the
present case diamant fuchsin was most used, as it is very rapid. The stain fades a
little, however, on exposure of the preparations. Unstained preparations of cuticles
were also largely used. In the case of the sections, one of the most satisfactory stains
was found to be Scharlach R (also used for macerated cuticles) prepared as directed by
Hill (1912). Ву it the cutin layer and cuticularised parts were clearly defined, staining
a bright brownish red, all the other tissues remaining unaffected. Phloroglucin with
hydrochloric acid was used to show the lignified lamellze which occur in every case on the
dorsal and ventral walls of the guard-cells. Serial sections stained with cotton red and
aniline blue, and with diamant fuchsin were employed in the correlation of surface views
and sections of stomata, Other tests were used, but an account of them need not be
repeated here, as descriptions are to be found in the literature quoted.

(5.) Previous REFERENCES.— The most important previous accounts of the epidermises
of Cycads include that given by Bornemann (1856), whose description is in connection
with some fossil types and includes remarks on the chemistry of the epidermal walls
and on the preparation of “artificial cuticles”; Kraus (1866) and Nestler (1895) also
give good descriptions of some Cycadean cuticles. These researches show that each
genus is fairly well characterised by a peculiar epidermis ; sometimes the genera may
be divided into sub-groups according to the structure of the epidermis (see Bornemann,
1856, р. 42, and Kraus, 1866, р. 13). Figures and descriptions of the stomata of some
Cyeads are given by Schwendener (1881), Strasburger (1866-67), Mahlert (1885), and
Porsch (1905). More detailed references will be given in the description of each genus.
The last-named author mentions the fact that though there is a general type of stoma
characteristic of the Gymnosperms as a whole, there is one special type for almost each

genus when the finer structure is taken into consideration.

MR. Н. HAMSHAW THOMAS AND MISS Х. BANCROFT ON

T
сл
Qo

(c. DESCRIPTION OF THE CUTICLES.
(1) Introduction.

In a consideration of the cuticles * of recent Cycads with a view to comparing
them with those of fossil fronds, the features which seem to be of greatest value
are those which may be determined from an examination of surface views. The actual
form, size, and arrangement of the guard-cells, and the thickening of their walls as seen
in section, are very constant throughout the group, and the variations mostly depend
upon the appearance in surface view. The chief features to be taken into consideration
are as follows :—

(1) The form and outline of the epidermal cells of upper and lower surfaces of the
leaves, both between and over the nerve-courses ; the pitting and thickness, straightness
or irregularity of the cell-walls.

(2) The distribution and position of the stomata, whether scattered or in definite
restricted areas ; whether irregularly placed or having their axes in a fixed direction with
regard to the axis of the pinna.

(3) The appearance of the stomata in surface view, depending upon the presence or
absence of an intercalated series of cells between the guard-cells and the general
epidermal cells, and the consequent degree of sinking of the guard-cells below the
epidermis ; the shape of the stoma-opening.

(4) The number and arrangement of the subsidiary cells T (the epidermal cells
surrounding the stoma, or in the case of much sunk guard-cells, those around the stoma-
opening at the surface of the leat).

(5) The presence or absence of crystals, either in the epidermal cells or in inter-
cellular spaces.

(6) Тһе presence or absence of hairs or hair-scars.

(7) The thickness of the cutin layer, and the presence or absence of foldings or
striations upon it.

(ii.) Detailed Description.
STANGERIA, Т. Moore. РІ. 17. figs. 1-6; text-fig. 1.
ВЕРЕВЕМСЕЗ :— Bornemann (1856), рр. 41, 45 ; Taf. 12. figs. 19, 20. Kraus (1866), p. 34; Taf. 23.
figs. 28, 29. Nestler (1895), p. 363.

The typical Stengeria paradoxa, Т. Moore, and its variety schizodon, were examined
and found to be exactly similar in epidermal and cuticular characters. The genus is
distinct from all other Cycads, in that the epidermal cells of both upper and lower
surfaces of the leaf-lamina have undulate lateral walls, like those of the epidermal cells
of ferns. These cells are irregular in form and arrangement, and are usually three- or
four-angular. Over the midrib of the pinnz the undulations are absent, and the cells

ж The term “cuticles” is used to include the layer of epidermal cells and the overlying layer of сойр. |
+ ‘ Wall-Zellen ” or * Nebenzellen” of the German authors.

CUTICLES OF CYCADEAN FRONDS. 159

are more regularly rectangular (text-fig. 1). Over the lateral veins also the cells are
somewhat rectangular, narrow, and elongated, but the walls are undulate. At the
pinna-margins the cells gradually become smaller and straight-walled. According to
Kraus the cell-walls are unequally thickened, more secondary material being laid down
on the outer side of a bend than the inner. Very little difference was noted in the
present case, perhaps owing to the comparative thinness of the walls; and after careful
preparation and staining, it seemed as if the slight inequality might be due to optical
effect rather than to any real difference in the amount of thickening.

J li
ВЕ ІП

Stangeria paradoxa : surface view of the cuticle from the midrib of the upper side of the leaf,
showing a stoma with characteristic thickening due to the dorsal (4) tral (v) lignified
lamelle of the guard-cells; g, the raised polar ends of the guard-cells. Note the absence
of folds of the cutin in the neighbourhood of the stoma. x 450.

Stomata occur scattered and irregularly-placed with regard to the pinna-axis, on the
lower side of the lamina; they are generally absent in the very narrow areas over the
lateral veins. They occur on both upper and lower mid-rib areas, their long axis being
parallel with that of the pinna (text-fig. 1). The upper, as well as the lower stomata
appear to be functional, as there is a well-developed air-space underlying each.

The stomata lie practically at the level of the epidermis (PI. 17. figs. 1 & 5); the
guard-cells are not sunk as in the case of Zamia (cf. Pl. 17. fig. 5 with РІ. 18. fig. 13).
Each stoma is surrounded by from four to six epidermal cells having "у thinner,
straighter walls than the rest of the epidermis. These “ subsidiary cells ’ are particularly
marked in the case of the lower stomata, and may usually be distinguished into polar
and lateral cells; Pl. 17. fig. 1 shows an example with two polar and two lateral

subsidiary cells. Division of the lateral cells often takes place, as text-fig. 1 shows.

160 MR. Н. HAMSHAW THOMAS AND MISS N. BANCROFT ON

The cavities of the subsidiary cells in many cases contain crystals of calcium oxalate *.
The stomata are elliptical in shape, and in surface view the lignified lamellee + of the
ventral and dorsal walls of the guard-cells give a very characteristic thickened
appearance (Pl. 17. fig. 1; text-fig. 1) to the area overlying the guard-cells. This
thickened area surrounding the pore of the stoma shows two thin patches at each pole,
owing to the raising up of the polar ends of the guard-cells, bringing them nearer
to the surface at these points, as described by Porsch for Bowenia. The projections of
thickening substance between each pair of thin areas strengthening the ends of the
stoma-opening are due to the extension of the ventral lignified lamellze of the guard-
cells (see text-fig. 26, p. 176). Laterally the cutin overlying the lamelle shows fine
striations (Pl. 17. fig. 1; text-fig. 1).

Hair-scars occur on the lower surface of the pinne, particularly on and near the
midrib and at the margins. They are only to be seen here and there on the upper mid-
rib. The scars are somewhat like a truncated cone in shape and are strongly cuticularised
(РІ. 17. figs. 2 & 6).

The cutin layer overlying the epidermal cells is fairly thick, particularly on the upper
surface of the ріппе. It has closely-lying striations which are seen from sections to be
due to folds (Pl. 17. fig. 3). "These folds are particularly marked on the upper surface,
and are irregular on the lamina, but are regular and run in a longitudinal direction on
the midrib (text-fig. 1). Тһеу radiate from the poles of the stomata and from the bases
of the hair-sears (Pl. 17. figs. 1 & 2; see also references quoted). Folding of the cutin
is not to be found in any other genus of Cycads, and constitutes, with the undulate
epidermal walls, a distinguishing feature of Stangeria.

Cycas, Linn. РІ. 17. figs. 10-15; text-figs. 2 & 8.

Rererences :— Bornemann (1856), рр. 40, 42 ; Taf. 11. figs. 3-6. Kraus (1866), p. 31; Taf. 19.
figs. 1-6; Taf. 20. figs. 10, 11. Strasburger (1866), Taf. 41. fig. 143. Mahlert (1885), Taf. 1.
fig. 20. Nestler (1895), р. 343.

SPECIES EXAMINED :—C. revoluta, Thunb.; С. circinalis, Linn. ; C. undulata, Desf.; C. Micholitzii,
Dyer ; C. media, R. Br.

The epidermal cells of Cycas are similar on both sides of the pinnz, and are more or
less irregular in shape and arrangement. They are usually somewhat elongated, and are
three- to six-angular. Over {һе midrib the cells are more regular and rectangular in
shape and run parallel with the pinna-axis. In these points there is general agreement
with Stangeria. In Cyeas, however, the lateral epidermal walls are not undulate,
being at the most somewhat curved. А distinguishing feature of all species of Cycas
examined $ is the pitting of the outer walls of the epidermal cells (РІ. 17. figs. 10-13,
fig. 15; text-fig. 2). In surface view and sections, these walls are seen to be perforated

* Cf. Kraus (1866), РІ. 23. fig. 29; the example figured in Pl. 17. fig. 1, does not show these crystals, owing to
previous maceration.

T Porsch (1905); see also Karzel (1907).

$ See also the lists given by Bornemann and Kraus.

CUTICLES ОЕ CYCADEAN FRONDS. 161

by more or less regularly-disposed pores, which occur principally towards the periphery
of the cells and which are covered by the cutin layer. This pitting occurs in both
the upper and lower epidermis, and is not known to be present in any other genus of the
Cycadaceze *. The pores vary considerably in size in the different species of Cycas:
the largest were observed in С. circinalis, the smallest in C. Micholitzir. Pitting of
the lateral walls also occurs, but in a varying degree (Pl. 17. figs. 10 & 13). It is well
marked in C. revoluta ; in C. undulata, however, pits occur only here and there.

Oycas | Micholitzii.—Lower cuticle in surface view: Cycas revoluta,—A stoma in surface view, focussed
s, а stoma-opening, surrounded by six subsidiary below the level shown in РІ. 17. fig. 12, so that
cells. Тһе thickened appearance round the the dorsal or upper surfaces (4) of the guard-cells
stoma-opening is due to the showing-through are seen surrounded by fifteen small cells, the
of the cutin layer lining the air-space and intercalated series shown in section in РІ, 17.
overlying the guard-cells. х 450. fig. 13. х 450.

Stomata occur only on the underside of the pinne, and are scattered and irregular in
position. As both Bornemann and Kraus have previously mentioned, the species of
Cycas fall into two groups, according to their epidermal and stomatal characters, which,
moreover, correspond with the external characters of the leaves. The first group contains
species such as C. revoluta, with narrow pinn:, the margins of which are inrolled +; the
second group, of which C. circinalis is an example, contains species with broader, flatter
pinnee, less xerophytic in texture than those of the first group. All the types examined,
with the exception of C. revoluta, belong to this group.

In the first type the epidermal cells are irregular in shape and disposition, particularly
on the under side of the ріппге, where, however, they are more isodiametric ; their
greatest diameter often lies transversely to the pinna-axis. The most distinctive feature
of the cuticle consists in the complication of the stomatal apparatus. The guard-cells
are well sunk below the surface, and between them and the subsidiary cells, which in this
case overarch the guard-cells and form an “ outer air-space," is an intercalated series
consisting of from tweive to fifteen small cells (РІ. 17. fig. 13; text-fig. 3T). These

* Microcycas has not yet been examined. |
+ Bornemann and Kraus have also examined C. inermis in this group.
> Cf. Kraus (1866), Taf. 19. figs. 4 & 5.

162 MR. H. HAMSHAW THOMAS AND MISS N. BANCROFT ON

intercalated cells appear to be derived from the same mother-cells as the subsidiary cells,
which are also from twelve to fifteen in number. The subsidiary cells are narrowly
and obliquely elongated (P1. 17. fig. 13), so that they arch over the depression containing
the guard-cells in such a way as to give the appearance, in surface view, of a somewhat
hemispherical elevation, perforated by a small rounded opening (Pl. 17. fig. 12) *. Тһе
intercalated cells and subsidiary cells are pitted in the same way as the general epidermal
cells (Pl. 17. fig. 13).

In the second group, the epidermal cells are irr egular in form and position, as in the
first type; the stomata, however, are more simple in organisation (Pl. 17. fig. 14). "They
are a little sunk below the general level of the epidermis, but there is no intercalated
series and the subsidiary cells do not overarch the depression ; hence (һе very different
appearance of the stomata as seen in surface view (text-fig. 2). Тһе subsidiary cells are
fewer in number, though very variable, usually between five and nine (text-fig. 2); in
shape they are from isodiametric to elliptical.

In C. circinalis and C. revoluta, fairly young stages were examined; these showed
very distinetly the presence of hair-scars on the under surface of the leaves.

The cutin layer is in all species fairly thick; it is not characterised by any special
foldings or markings, as in Stangeria.

ENCEPHALARTOS, Lehm. РІ. 18. figs. 10 & 11; text-figs. 4-8.
ligrERENCES :—Bornemann (1856), pp. 40, 44; Taf. 13. figs. 1-6. Kraus (1866), pp. 33, 84;
Taf. 20. figs. 12-15; Taf. 21. fig. 16. Mahlert (1885), Taf. 1. fig. 23. Nestler (1895), рр. 346-
348 ; Taf. 11. figs. 5-8 ; Taf. 12. fig. 9.
SPECIES EXAMINED :—E, villosus, Lemaire ; E. Woodii (? hort.) ; E. horridus, Lehm. ; E. caffer, Miq. ;
Е. lanuginosus, Lehm. ; E. Lehmanni, Eckl.; E. Ghellinckii, Lem. ; Е. жәнеде. Wilden.
(See also the lists of Bornemann and Kraus.)

Reviewing the species examined as a whole, there seems to be a considerable amount
of variation in the form of the epidermal cells, even in the cells of the same pinna (text-
fig. 5). Generally speaking, the cells are somewhat elongated and the longitudinal axes
run in the direction of the longitudinal axis of the pinna (Pl. 18. fig. 11; text-fig. 4).
The transverse walls are usually obliquely placed with regard to the piapa as the
cells are three- or four-angular and rhombic in shape.

In E. Lehmanni (text-fig. 5), Е. caffer, and E. lanuginosus, where the cells are small
and more nearly isodiametric, their greater diameter is usually transversely or obliquely
placed with regard to the axis of the ріппге; in Z. horridus the cells are irregular in form
and direction, particularly in the region of the lobe-bases.

The epidermal cells of the lower surface are of two kinds—those of the nerve-courses
and those of the stomatal areas, The cells of the nerve-courses are similar in form to
those of the upper surface. They are usually somewhat narrow and elongated and are
arranged in longitudinal lines. ‘The cells of the stomatal areas are typically shorter,
broader, and more isodiametric.

* Cf. Bornemann (1856), Taf. 11. fig. 4.

CUTICLES ОҒ CYCADEAN FRONDS. 163

The walls of all the epidermal cells are without undulations, but are often slightly
curved. In most of the cases examined, the lateral walls (never the outer walls) of both
upper and lower epidermal cells are pitted in the same way as the lateral walls of Cycas.
This is very marked in Z. Lehmanni (РІ. 18. fig. 10; see also fig. 11). In E. Wood
and on the lower surface of Ж. villosus pittings are infrequent. In the examined species
of. Encephalartos the inner and lateral walls of the epidermal cells are often slightly

` | |
М! A

Á

=

Encephalartos Woodii.—Upper cuticle in surface view, Е, Lehmanni.— Upper cuticle in surface view.
showing form and arrangement of epidermal cells. x 450
x 450.

thicker than the outer walls (РІ. 18. fig. 10) Towards the pinna-margins the cells
become smaller and thicker-walled. In some cases, as іп Ж. horridus and Е. villosus
(РІ. 18. fig. 11), the elongated epidermal cells are septate.

Stomata are confined to the lower surface of the ріппге, and more or less definitely to
the areas between the nerve-courses. Their longitudinal axis is typically coincident with
that of the pinna. They are in most cases considerably sunk below the general level of
the epidermis without overarching of the subsidiary cells; so that the outer air-space is
fairly long and narrowand tubular in shape (text-fig. 6). There is a somewhat irregular
system of intercalated cells separating the guard-cells from the general epidermal cells.
In median sections of the stomata, typically two intercalated cell-series appear (text-fig. 6);
the sections may not, however, pass through a cell of both series. Nestler (1895, p. 348)
mentions that one or two intercalated series are present ; while according to Kraus (1866,
p. 84) the guard-cells lie immediately under the epidermis except at the pinna-bases.
"The stoma-opening is somewhat elongated and elliptical to quadrangular in shape (text-
SECOND SERIES.— BOTANY, VOL. VIII. Әс

164 MR. Н. HAMSHAW THOMAS AND MISS N. BANCROFT ON

figs. 7 & 8). It is surrounded by from four to six subsidiary cells, and is strongly
cuticularised .

Simple hairs and hair-scars occur in Ж. Ghellinckii. In Е. caffer crystals of calcium

Encephalartos Lehmanni.—Section of the lower cuticle, showing a stoma in median section.
Note the presence of intercalated cells between the guard-cells and subsidiary cells,

and the heavy cuticularisation. х 450.
Fig. 8
: “А

villosus.—Lower cuticle in surface view; the
stoma-opening is surrounded by five subsidiary
cells, one of the polar cells being divided. х 450.
This cuticle was not removed by maceration, so
that the thickening of the guard-cells and the
lining of the outer air-space show through in
the surface view, giving a much thickened
appearance to the stoma-opening.

oxalate, such as figured by Nestler (1895, Taf. 11. figs. 6-8), were not observed; very
little material of this species was, however, available.

The cutin layer is of a considerable thickness in the case of both upper and lower
epidermises (РІ. 18. fig. 10; text-fig. 6).

Е, Lehmanni.—Lower cuticle (after maceration) in
surface view, showing a stoma-opening (s) sur-
rounded by six subsidiary cells (р, polar sub-
sidiary cells ; 1, lateral subsidiary cells). х 450.

я

CUTICLES OF CYCADEAN FRONDS. 165

Dtoon, Lindl. (Pl. 18. figs. 5-9; text-figs. 9 & 10.)
RzrERENCES :—Bornemann (1856), рр. 40, 45; Taf. 12. figs. 11-14. Kraus (1866), рр. 39, 40;
Taf. 22. figs. 21-23. Mahlert (1885), Taf. 1. fig. 22. Porsch (1905), p. 11; Taf. 1. fig. 7.

SPECIES EXAMINED :—D. edule, Lindl. ; D. spinulosum, Dyer.

The two species were found to be of a similar type, characteristic for the genus. In
D. edule the upper epidermal cells and those of the nerve-courses of the lower side of
the pinne are similar, and are of two kinds—thick-walled cells elongated fairly regularly
in the direction of the longitudinal pinna-axis, and more or less isodiametric thin-walled
cells. These two kinds of cells may be distinguished in transverse section (РІ. 18. fig. 7).
The thin-walled ceils appear to be the more numerous. The cells of the stomatal
regions between the nerve-courses are usually smaller than elsewhere. "They are almost
isodiametric and are uniformly thick-walled. In Р. spinuloswm the epidermal
characters are similar to those in D. edule, though all the cell-walls seem to be slightly
less thickened. In both cases the lateral walls of the cells are often slightly pitted
(Pl. 18. fig. 5) (Kraus, 1866, fig. 21), and all the walls are more or less heavily layered

Fig. 9. Fig. 10.

D. spinulosum,—A stoma in median section, showing
the dorsal (d) and ventral (v) lignified lamellw of
the guard-cells, and three intercalated series
of cells (2-42) between the guard-cells and sub-
sidiary cells. (Cf. РІ. 18. fig. 6.) х 450.

Dioon edule.—Lower euticle in surface view, showing
six subsidiary cells surrounding a stoma-opening, s.
Cf. the epidermal cells of the nerve-course (n)
with those of the upper surface shown in РІ. 18.
fip. 5. x 450.

(Pl.18.figs.5 & 9). In the case of the thick-walled cells of D. edule, the lumen is often
almost filled up.

Stomata occur very definitely limited to the areas between the nerve-courses on the
under surface of the leaves. Тһеу are much sunk in flask-shaped cavities, which are
slightly deeper in D. edule than in D. spinulosum. In both species there are typically

three intercalated series of cells between the guard and subsidiary cells TE fig. 6;
02

166 MR. H. HAMSHAW THOMAS AND MISS М. BANCROFT ON

text-fig. 10) (Porsch, 1905, Taf. 1. fig. 7). Тһе cavities are lined with a layer of cutin,
and the epidermal cells surrounding the stoma-openings (subsidiary cells) arch over
slightly. They are also a little raised beyond the surface of the pinna, and are rounded,
as seen in surface view. They are from six to eight in number and are much thickened,
especially in D. edule (text-fig. 9).

Crystals of calcium oxalate occur in D. edule on the lower side of the pinne in the
nerve-courses. They were not observed on the upper side. Kraus figures a crystal
from the lower epidermis, but does not specially note their occurrence there. He
mentions that “ here and there a thin-walled cell contains a crystal.” So far as could
be determined, however, from the appearance in surface view and in section, the
crystals occur in cavities or spaces between the cells rather than in the cells themselves
(Pl. 18. figs.8 & 9); for in surface view the rounded ends of the neighbouring cells
seem to project into a space, and in section Scharlach R shows the presence of a layer
of eutin entirely surrounding or lining the crystal-containing зрасе.* D. spinulosum
has no crystals.

The cutin layer in both species is very thick, especially in D. edule (РІ, 18.
figs. 7 & 8).

CERATOZAMIA, Brongn. РІ. 18. figs. 1-4; text-figs. 11 & 12.
Rererences :—Bornemann (1856), pp. 40, 45; Taf. 12. figs. 7-10. Kraus (1886), pp. 38, 39;

Taf. 21. figs. 17-20. Mahlert (1885), Taf. 1. fig. 21.
SPECIES EXAMINED :—C. mezicana, Brongn. ; С. fusco-viridis, D. Moore; C. latifolia, Miq.

These three species show the same general characters, all possessing two kinds of
epidermal cells, thick- and thin-walled. The thick-walled cells apparently predominate
іп Ceratozamia (cf. Dioon), though they are somewhat variable in number and in dis-
tribution in the three species examined. There is considerable variation, moreover, in
the relative thickness of the thick-walled cells; C. mexicana shows a very marked
contrast between the two kinds of cells (Pl. 18. figs. 1 & 3); C. latifolia shows com-
paratively little difference between them (text-fig. 11); while C. fusco-viridis is inter-
mediate in this respect. Тһе cells—particularly the thicker-walled ones—are much
elongated in the direction of the longitudinal axis of the pinna. The cells of the
nerve-courses, in the lower epidermis, are similar to those of the upper side; while
between the nerve-courses, as is usual, the cells are shorter and broader. Thin-walled
cells occur in these areas also, especially in C. mexicana (ef. Dioon).

In С. latifolia, the elongated thick-walled cells of both upper and lower epidermises
are often septate (text-fig. 11); Kraus notes the same fact in С. Kiisteriana.

С. mexicana shows pitting on the lateral walls of the thickened cells (РІ. 18. figs. 1 & 2).
This is more marked on the lower surface than on the upper, especially in the case of
the subsidiary cells of the stomatal apertures. In 6. fusco-viridis also, the lower
epidermal cells are much pitted; in C. latifolia, however, pitting was not noted, except

* A similar occurrence of erystal-containing cavities is quoted by Nestler in the case of Encephalartos caffer.
In D. edule the cavities are apparently larger than in that species,

CUTICLES ОЕ CYCADEAN FRONDS. 167

in the cells of the nerve-courses. The cell-walls, particularly in С. mexicana, are very
strongly layered (Pl. 18. figs. 1-4). Ceratozamia has wide stomatal regions and narrow
nerve-courses, the demarcation of the areas being most distinct in C. latifolia. The
stomata lie parallel to the pinna-axis, or somewhat obliquely to it, and are sunk below
the level of the epidermis, having one intercalated series of cells (Pl. 18. fig. 4). The
air-space is lined with a layer of cutin, extending as is usual to the ventral walls of the
euard-cells, and its somewhat oval opening is surrounded by from four to six subsidiary
cells; five were observed in О. latifolia (text-fig. 12), six in С. mexicana (РІ. 18. fig. 2),

.

Fig. 11. Fig. 12.

AS

Ceratozamia latifolia.—Upper cuticle in surface view,
showing thin-walled and thick-walled cells. At
ж, an elongated cell is divided by a septum.
x 450.

C. latifolia.— Lower cuticle (after maeeration) in surface
view, showing а stoma-opening (s) surrounded by
five subsidiary cells (p and 7). х 450.

and usually four in C. fusco-viridis. Kraus notes fourin C. Küsteriana. Тһе subsidiary

cells do not overarch the air-space to any extent.
The cutin layer is throughout fairly thick, partieula

leaves (Pl. 18. figs. 3 & 4).

rly on the upper surface of the

gs. 14 & 15; text-figs. 13-18.

Rererences :—Bornemann (1856), pp. 40, 44; Taf. 11. figs. 16, 17. Schwendener (1881), fig. 13.
Nestler (1895), pp. 352, 354; Taf. 12. figs. 10, 11; Taf. 13. fig. 16.

Species EXAMINED :-М. spiralis (= Fraseri, Мід. ?); М. corallipes, Hook. f.; M. Denisonii, F. Muell.:
M. Hopei, T. Hill ; M. cylindrica, C. Moore ; М. Macleayi, Miq. (= spiralis, Miq. ?).

MACROZAMIA, Miq. РІ. 18. fi

168 MR. H. HAMSHAW THOMAS AND MISS N. BANCROFT ON

The six species examined show agreement in the occurrence of thick-walled and thin-
walled epidermal cells. They fall into two groups, however, according to the arrange-
ment of the cells. М. Macleay? (=spiralis Р), М. Denisonii (text-fig. 13), and М. Hopei
(text-fig. 14.) differ from the other three species in the irregularly-shaped three- or
four-angular cells placed transversely or obliquely to the long axis of the pinna. In
this first group, moreover, the thick-walled cells are very marked, their cavities being
often almost filled up; there is also a heavy primary wall, particularly іп M. Denisonii
(text-fig. 13). Іп M. corallipes, М. cylindrica (Pl. 18. fig. 15), and М. spiralis
(= Fraseri Р) the cells are three- or more often four-angular ; they are generally elongated
in the direction of the longitudinal axis of the pinna, though the thinner-walled cells
are often squarish. In this second group, the walls are often much pitted, particularly
in the case of the thin-walled cells (Pl. 18. fig. 15). In the other type pitting is less

Macrozamia Denisonii.—Upper cuticle in surface view, M. Hopei.—Upper cuticle in surface view, showing

showing thin-walled and thick-walled cells, and thin-walled and thick-walled cells, and pitting of
the heavy primary cell-walls. Note the irregular the thin-walled cells. х 450.
arrangement of the cells. x 450.

evident. Layering of the cellulose is frequently marked, as in M. cylindrica (РІ. 18.
fig. 15). The cells of the upper surface and of the nerve-courses of the lower surface
are similar in both groups, but the cells between the nerve-courses are usually irregular
in shape and size, and are much more uniformly thickened (text-figs. 15 & 16), although
occasionally thin-walled cells occur. Sometimes, as in М. corallipes, the elongated
thiekened cells are septate.

Stomata occur only on the lower surface between the nerve-courses ; their long axis
is parallel with or slightly oblique to the long axis of the pinna. А regular longitudinal
arrangement, almost in lines, is seen in M. spiralis (= Fraseri?). In all cases the
stomata are sunk below the epidermis, an apparently rather irregular system of cells

CUTICLES ОЕ CYCADEAN FRONDS. 169

intervening between the guard-cells and the subsidiary cells. There seem to be typically
two series of these intercalated cells (Pl. 18. fig. 14), but the number of series as seen in
section through a stoma shows a little variation, apparently owing to irregularity in
position of the cells; this is seen in text-fig.17. In М. cylindrica the outer air-space
is long, and narrower towards the opening, owing to the slight overarching of the
subsidiary cells, which are slightly raised above the level of the epidermis (РІ. 18.
fig. 14). The opening as seen in surface view is oval or elliptical in shape. In
M. Denisonii the outer air-space is shallower and wider (text-fig. 17). In all cases the
air-space is fairly heavily lined with cutin. The subsidiary cells vary from five or six
to eight in number, but are typically six, as in M. Denisonii (text-fig. 15) and М. Hopei
(text-fig. 16).

Fig. 16.

i i 1 1 toma-openin,
Macrozamia Denisonii.—Lower cuticle in surface view (after maceration), showing 8 pening

surrounded by six subsidiary cells. х 52
М. MERE cuticle in а view, showing а stoma-opening with “ subsidiary cells (р and 1).
The dorsal surfaces of the guard-cells are indicated at 4. Х 450. (This cuticle was not removed by
maceration, so that all the thickening of the air-space and guard-cells поча through.)
М. Denisonii.—Lower cuticle, showing a section through a stoma. Note the irregular intercalated cells,

between the guard and subsidiary cells. х 450.

170 MR. H. HAMSHAW THOMAS AND MISS N. BANCROFT ON

In М. cylindrica, in the nerve-courses of the lower surface, and also here and there
on the upper surface of the pinnze, occur crystals of calcium oxalate (text-fig. 18); they
appear to be contained in small isodiametric cells, and not in cavities, as in Dioon edule,
and in М. spiralis (— Fraseri ?), where the crystal-containing spaces are very small and
not at all frequent.

Macrozamia cylindrica.—Lower cuticle (from a nerve-course) in surface view, showing a crystal-containing
сей. (Cf.the form and arrangement of the cells with those shown іп РІ, 18. fig. 15, from the
upper epidermis of М. cylindrica.) x 450.

The cutin layer is moderately thick (РІ. 18. fig. 14; text-fig. 17), but, as usual, it is
thinner on the lower than on the upper surface of the leaves.

ZAMIA, Linn. РІ. 18. figs. 12 & 13; text-figs. 19-28.
REFERENCES :—Bornemann (1856), рр. 40, 43, 44; Taf. 11. figs. 7-15. Kraus (1866), pp. 41, 42;
Taf. 22. figs. 24-27.
SPECIES EXAMINED :—Z. Skinneri, Warsz.; Z. muricata, Willd.; Z, Loddigesii, Miq.; Z. integrifolia,
Ait.; Z. Fischeri, Miq.; Z. Ottonis, Miq.; Z. Lindeni, Regel,=Z. cylindrica, Liebm. (= Leiboldii,
Miq. ?).

These eight species agree in essential characters of the epidermis. The cells in all
are elongated in the direction of the longitudinal axis of the pinn:e, and are three- or
four-angular in shape. Іп Z, Ottonis the cells are somewhat irregular, with groups of
smaller cells here and there. In some species, for example Z. Skinneri (text-figs. 20
& 23) and Z. integrifolia (Pl. 18. fig. 12), there is a marked distinction into two kinds
of cells, thin-walled and thick-walled. In Z. Skinneri and Z. Loddegesii, the thin-
walled cells are only occasional, while the other cells often show a considerable amount
of thickening. In 2. Fischeri (text-fig. 19) and 2. Ottonis the walls throughout are
fairly thin. 2. muricata has walls of intermediate thickness. Тһе cells of the nerve-
courses of the lower epidermis are similar to those of the upper surface. Іп 2. cylin-
drica and Z. Lindeni the cells of the nerve-courses are fairly thick-walled, elongated

CUTICLES OF CYCADEAN FRONDS. 171

and narrow, running quite straightly, and having somewhat pointed ends. The cells
between the nerve-courses are shorter and broader and more irregular, but they are
still slightly elongated, except in the immediate neighbourhood of the stomata. In
the stomatal areas the cell-walls are sometimes thinner than they are on the upper
surface; this is best seen іп Z. Skinneri and Z. Lindeni. In 2. Fischeri the upper
epidermal cells and those of the nerve-courses and stomatal areas are of about the same
thickness.

Pittings are sometimes plainly seen (РІ. 18. fig. 12), particularly in the case of the
thin-walled cells of nerve-courses. Occasionally the cellulose of the epidermal cells
is layered; and in some cases very much elongated cells are septate as іп Z. Ottonis
and Z. Skinner.

Fig. 19. Fig. 20.

U

и

M ДА (Y | A

Zamia Fischeri.—Upper cuticle in surface view. Z. Skinneri,—Upper cuticle in surface view. Gf. the
x 450. amount of variation in the thickness of the cell-
walls with that shown in text-fig. 19. x 450.

Stomata occur on the lower sides of the leaves in definite areas between the nerve-
courses; they are sunk only slightly below the level of the epidermis, there being no
intercalated series of cells between the guard-cells and subsidiary cells (Pl. 18. fig. 13).
The lignified lamelle of the guard-cells may be seen, as in Stangeria (and Bowenia),
on foeussing slightly below the level of the subsidiary cells. Тһе subsidiary cells are
characteristic in number and arrangement; there are from four to six, consisting of
an elongated cell at each pole of the stoma, and two (sometimes only one) lateral
cells at each side (text-fig. 22). Sometimes a second series of lateral cells is present
(text-fig. 21). |

SECOND SERIES.— BOTANY, VOL. VIII. 2»

172 MR. Н, HAMSHAW THOMAS AND MISS N. BANCROFT ON

Zamia Lindeni.—Lower cuticle,in surface view. The

thickened stoma-opening shows a double series of Z. muricata.—Lower cuticle in surface view.
lateral subsidiary cells at each side (7, 1). х 450. The stoma-opening in this case has five
(This cuticle was not removed by maceration, and subsidiary cells (p and 7) surrounding it.
thus shows the thickening substance of the guard- х 450. (Not macerated ; see text-fig. 21.)

cells and air-space, showing through to the surface.)

Z. Skinneri.—Upper cuticle in section, showing the variation in the
thickness of the cell-walls. (Cf. text-fig. 20.) x 450.

In Z. integrifolia hair-scars are occasionally seen, particularly on the under surface in
the nerve-courses. The scars seem to occur over a group of the thinner-walled cells.
They occur also in Z Ottonis and Z. Skinneri.

The eutin layer is similar in thickness to that of Macrozamia (cf. РІ. 18. figs. 13 & 14).

Bowenta, Hook. РІ. 17. figs. 7-9; text-figs. 24-26.

Rererences :—Nestler (1895), pp. 356, 357 ; Taf. 13. figs. 19-21. Porsch (1905), p. 13; Taf. 1.
figs. 1-4.
SPECIES EXAMINED :— Bowenia spectabilis, Hook., var. serrulata, André (= Bowenia serrulata (André),
Chamberlain (1912, p. 419)).
Bowenia serrulata agrees with such Zamias as 2. Fischeri in the arrangement of the
three- or four-angular epidermal cells and also in the thickness of their walls (cf. text-
figs. 19 & 25). The cells are elongated, with straight walls, and are slightly narrower

CUTICLES OF CYCADEAN FRONDS. 173

over the nerve-courses, even on the upper surface of the leaf. As is usual, on the lower
surface, the cells between the nerve-courses are shorter and broader. No pitting of
the walls was observed.

Stomata occur in both the upper and the lower epidermis (Pl. 17. figs. 7 & 8). They
are few and scattered on the upper side, occurring principally in the middle of the lower
half of the ріппге, between the veins; they decrease in number towards the upper half
and margins. They are fairly abundant on the lower side. In all cases their long
axis is approximately parallel with that of the ріппе. The guard-cells are situated
practically at the level of the epidermis (РІ. 17. fig. 9), there being no intercalated
series of cells between them land the subsidiary cells—they are less sunk than those
of any other genus except Stangeria (cf. Pl. 17. fig. 9, Bowenia, with fig. 5, Stangeria ;

Fig. 25.

М

TL Ln

(9999»99999999(999(9
АКА

Bowenia spectabilis var. serrulata.—Upper cuticle
in section. х 450,

|
ІМ

В. spectabilis var. serrulata.—Upper cuticle
in surface view. x 450

НН

fig. 14, Сусав circinalis; and РІ. 18. fig. 13, Zamia integrifolia, for degrees of sinking
of guard-cells without an intercalated series of cells). The stomata are surrounded by
from four to six subsidiary cells, the walls of which are fairly thin (Pl. 17. figs. 7 & 8),
particularly in the case of the stomata of the upper leaf-surface. Оп account of the
position of the guard-cells, ¿ е. almost at the general level of the epidermal cells,
the lignified lamellz of dorsal and ventral walls are plainly seen, giving a characteristic
appearance in surface view (РІ. 17. figs. 7 & 8; cf. Stangeria, fig. 1). As previously
described by Porsch in the case of B. spectabilis * the raising-up of the polar ends of the
guard-cells is indicated in surface view by the presence, at each pole of the stoma, of two

* Porsch (1905), p. 13; Taf. 1. figs. 1 & 3. See also Nestler (1895), Taf. 13. figs. 19 & 20, 5.
202

174 МВ. H. HAMSHAW THOMAS AND MISS N. BANCROFT ON

thin pear-shaped areas in the thickened area overlying the guard-cells. The projections
of thickening material formed by the extension of the ventral lignified lamelle are seen
at the poles of the stoma (see text-fig. 26, showing the correlation of serial sections
through a stoma with the surface view). The cutin layer overlying the lamelle is
faintly striated as in the case of Stangeria (Pl. 17. figs. 7 & 8).

Neither hair-sears nor erystal cavities were observed. The cutin layer, compared
with that of the other genera, is not very thick (Pl. 17. fig. 9; text-fig. 24).

(ii. General Summary of the Structure of the Recent Cuticles.

The epidermal cells of Cycad leaves have usually slightly curved or straight walls.
Stangeria has undulate walls somewhat like those of ferns*. The walls generally
attain a considerable thickness, and often show layering as in Ceratozamia, Dioon, and
some Macrozamias. Bowenia has comparatively thin walls, while Ceratozamia and
Dioon on the whole have the thickest, though these are characterised by the possession
of two kinds of epidermal cells, some elongated and thick-walled, and others shorter and
thin-walled. А variation in the thickness of the walls occurs also in Macrozamia, some
Zamias, and to a certain extent in Encephalartos. The epidermal cells are usually
elongated, and in Cycas and Stangeria are alike on both upper and under sides of the
leaf. In other types the cells of the under side following the course of the veins are
similar, or nearly similar, to those of the upper side, while the general epidermal cells
between the * nerve-courses ” are more isodiametric. Usually, in the upper epidermis,
the nerve-courses are not marked by any difference in the form of the epidermal cells ;
Stangeria shows a slight elongation and narrowing of the cells above the veins.

Stomata occur generally only on the under side of the leaves between the nerve-
courses, their long axes coinciding or forming a slight angle with the long axis of the
рп. The occurrence of stomata in these definite areas seems то be correlated with
the above mentioned difference in the form of the epidermal cells there, as compared
with those of the nerve-courses (see Zamia, Macrozamia, Ceratozamia, Dioon, and
Encephalartos). Іп Сусав and Stangeria the stomata are scattered and irregularly
placed; there are no fixed stomatal areas.

Bornemann (1856, p. 42) has recorded the occasional occurrence of stomata on the
upper side of the pinne іп Zamia muricata var. picía. Kraus (1866, fig. 15) figures
a stoma of the pinna-base in Encephalartos Altensteinii; it is apparently from the
upper side, though he does not mention the occurrence of stomata other than on
the under side of the pinnse. Nestler (1895, p. 362) was unable to find such stomata
in Encephalartos; in Bowenia, however, he records them from the basal parts of the
upper side of the pinnules, gradually decreasing in number and finally disappearing
towards the middle. The present investigations have revealed the presence of stomata
on the upper side of the ріппе in Stangeria. They are, however, confined to the
midrib (text-fig. 1).

* Bornemann (1856), p. 42; Kraus (1866), p. 13.

CUTICLES OF CYCADEAN FRONDS. 175

Nestler (1895, рр. 860, 364) gives a table showing the numbers of stomata per square
millimetre in a representative selection of species. The Ceratozamias (with from 24 to
31 per sq. mm.) and the Macrozamias (29 to 81) are seen to һауе the fewest, and the
Dioons (D. edule, 81, and D. angustifolia, 72) the most numerous stomata. The
investigated species of Cycas (with from 49 to 56 per sq. mm.), Encephalartos (41 to 52),
Stangeria (54), and Zamia (43 to 60) have very similar numerical distribution of
stomata, and are intermediate between the Ceratozamias and Macrozamias on the one
hand and the Dioons on the other. Bowenia spectabilis has 35 stomata per sq. mm. on
the under side and 14 on the upper side of the pinne.

The stomata of the Cycads as a group are larger than those of any other group
of Gymnosperms *. Their size seems to be fairly constant, being about "075 x 084 mm.,
while their structure conforms to a general plan—the “ Gymnosperm type" of Porsch.
The guard-cells in median and polar section show a constant form and position relative
to the general epidermal level.

The ventral and dorsal walls of the guard-cells have in every case lignified iamelle +,
though these are slightly variable in extent and in clearness of definition f. They give
to those stomata, the guard-cells of which are only slightly sunk, a very characteristic
thickened appearance in surface view (Pl. 17. fig. 1, Stangeria; figs. 7 & 8, Bowenia ;
text-fig. 26, Bowenia). In Bowenia, for example, the surface view of a stoma shows
the pore bounded on each side by a somewhat crescentic structure formed hy the
lignified lamellze on the upper or dorsal surface of the guard-cell. This is covered by
а layer of сойр. The lamellæ on the lower or ventral surface of the guard-cells are
only seen at the poles of the stoma, where the guard-cells are raised and the lamellze are
extended, appearing as thickened projections. Text-fig. 26 shows the correlation of
serial sections through a stoma with the surface view. "The gradual raising of the
guard-cells towards the poles and extension of the ventral lamellze with accompanying
diminution of the dorsal lamelle are plainly seen. Іп the case of guard-cells which are
much sunk, the surrounding, and usually overarching, epidermal cells obscure the
characteristic appearance caused by the lignified lamellae. In intermediate conditions
such as exhibited by Zamia (Pl. 18. fig. 13), the lamellz of the guard-cells may be seen
on focussing slightly below the epidermal cells. They are quite comparable with those
of Bowenia and Stangeria. е

The epidermal cells surrounding the guard-cells and stoma-opening—“ subsidiary
cells "—Aare usually fairly constant in number and arrangement in a genus. In no case
was material available young enough to trace the entire development of the guard-cells
and subsidiary cells from an epidermal mother-cell. Four seems to be the original
number of subsidiary cells, the typical arrangement of these being one at each pole of
the stoma and one at each side. It seems possible from a comparison with Gnetum and
Welwitschia $, together with a consideration of fairly young examples in Cycas revoluta

* Malhert (1885), p. 57. |
T These are not figured in every сазе; they are shown іп text-figs. 10 & 26 and in РІ. 18. figs. 4 & 6.
$ Karzel (1908). $ Takeda (1913), р. 351; & (1913), pp. 565, 366, text-fig. 8.

176 MR. Н. HAMSHAW THOMAS AND MISS N. BANCROFT ON

and Ceratozamia mexicana, that the polar subsidiary cells represent ordinary epidermal
cells, while the lateral ones are the product of division of the epidermal mother-cell
giving rise to the guard-cells also. Frequently the lateral cells are transversely
divided into two, so that six subsidiary cells—two polars and four laterals — аге
seen. Occasionally only one of the lateral cells is divided; division of the polar

Fig. 26.

Bowenia spectabilis var. serrulata.—A, surface views of stomata (cf. РІ. 17. figs. 7 & 8);
В, 1-4, serial sections through a stoma, corresponding to the positions 1-4 in the surface
view. These sections show the gradual raising of the guard-cells towards the polar ends
of the stoma, and the change in size and position of the dorsal (d) and ventral (v) lignified
lamelle of the guard-cells, these changes giving the characteristie surface view of the
stoma. В, 5, shows a polar section of a stoma in which the dorsal lamelle do not appear;
this section corresponds to position 5, in the surface view. (Diagrammatic.)

cells is comparatively infrequent. Ап increase in the number of subsidiary cells
is often seen where guard-cells are much depressed below the general level of the
epidermis, there being a mechanical necessity for this increase—the deeper the outer
air-space, the greater the number of cells required to ensure its rigidity, by means
of their lateral walls.

CUTICLES ОҒ CYCADEAN FRONDS. 177

In Bowenia and Stangeria, where the guard-cells are not sunk, there are from four
to six subsidiary cells, differentiated into polar and lateral (text-fig. 1; Pl. 17. figs. 1,
7, & 8). Occasionally there is some displacement of the cells, so that the differentiation
is not always quite definite (text-fig. 1; РІ. 17. fig. 8). Encephalartos (text-figs. 7
& 8), Ceratozamia (text-fig. 12; Pl. 18. fig. 2), Macrozamia (text-figs. 15 & 16), and
Zamia (text-figs. 21 & 22) typically show from four to six subsidiary cells, differentiated
into polars and laterals. Ап interesting distinction in Zamia is the frequent presence
of a second series of lateral subsidiary cells, parallel with the first, due perhaps to the
longitudinal division of the original lateral cells ж. this is well seen іп Zamia Lindeni
(text-fig. 21). The subsidiary cells in these four genera are ata different level from that
of the guard-cells, the latter being sunk to a greater or less degree, forming an outer
air-space. Тһе opening of the air-space at the surface of the leaf is elliptical to quad-
rangular in shape. In some cases, particularly in some Macrozamias and some species
of Encephalartos, there is а certain amount of overarching of the subsidiary cells. In
eases where the guard-cells are much sunk, the opening of the air-space on the leaf-
surface is more rounded in shape, and the subsidiary cells, which tend to increase in
number for the better strengthening of the large air-space, are not distinguished into
polars and laterals. This is seen in Dioon and Cycas revoluta, where all the cells are
of the same shape. Occasionally also the subsidiary cells are raised beyond the general
level of the epidermis, forming, small elevations perforated by the stoma -openings
(Pl. 17. figs. 12 & 13, Сусав revoluta).

In some cases crystals of calcium oxalate occur in the epidermal layers ; in Stangeria
they are to be found in the cells round the stomata ; in Macrozamia cylindrica crystals
occur on the upper leaf-surface and on the lower surface іп the cells of the nerve-
courses. In other cases, e.g. Dioon edule and Macrozamia spiralis (= Fraseri ?), they
occur in spaces between the epidermal cells.

Hair-scars occur in several cases, such as Encephalartos Ghellinckit, Cycas circinalis
and C. revoluta, Zamia integrifolia and Stangeria.

Тһе сайп layer overlying the epidermal cells is fairly thiek—usually thicker on the
upper than on the under side of the leaf. Dioon, Ceratozamia, and some species of
Encephalartos show the heaviest cutin layer, while that of Bowenia is comparatively thin.
The cutin of Stangeria has very characteristic folds.

ІШ. CUTICLES OF FOSSIL CYCADEAN FRONDS.

(а) Previous Observations on the Cuticular Structure of Fossil Cycadean Fronds.

Bornemann, whose work has been mentioned above, was the first to make observations
upon the structure of the cuticles of fossil fronds, and, in fact, made a comparative siudy,
based on the recent forms, of a type which has never since been attempted (1856). The

ж Of, Takeda (1913'), р. 351; (19137), pp. 365, 366 ; text-fig. 8.

178 ' МВ. Н. HAMSHAW THOMAS AND MISS N. BANCROFT ON

fossil remains with which he had to deal were, however, very fragmentary; little or
nothing could be made out as to the macroscopic characters of the fronds whose cuticular
structure is figured, and from his figures one is led to suppose that the preservation of
the fragments or the methods of treatment were not of the best. In the cases figured
the walls of the epidermal cells are all represented as straight, and the stomata figured
do not allow any close comparison with the examples now before us. It would appear
that the forms figured on his Taf. 7. figs. 1-6 are allied to Thinnfeldia, but in no other
cases are the stomata sufficiently well represented.

Schenk, in his Fossil Flora of Frankonia (1867), made a great advance in the study
of cuticles ; he described and gave accurate figures of the cuticles of a number of forms,
including Teniopteris stenoneura, Thinnfeldia spp., Nilssonia polymorpha, Otozamites
(Otopteris) Bucklandi, Pterophyllum spp., several conifers and ferns. He does not give
many good figures of preparations showing stomatal structure, but he was the first to
observe the sinuous walls of the cells in various species of Pterophyllum and in
Otozamites. He mentions that he has made preparations from the English species
O. acuminatus, О. Beani, and О. graphicus, also of the English species of Teniopteris,
which was found to be comparable with the Otozamitean structures*. He gave some
short descriptions of stomata, noticing that they occur usually on one side of the lamina
only; but there are many points of importance on which we can obtain no information
from his work. Тһе same author subsequently figured the cuticles of Déoonites
Dunkerianus (Goep.) 1.

Zeiller (1882) described and figured the epidermal structure of a Thinnfeldia
(Cycadopteris Brauniana) and Frenelopsis Hoheneggeri; he has also figured a small
piece of a cuticle of a Paleozoic Pterophyllum (Zeiller, 1906).

In his ‘ Jurassic Flora of the Yorkshire Coast’ Prof. Seward (1900, p. 234, fig. 43)
has given a good figure of part of a lower cuticle of Ctenis sp. The form of the
epidermal cells, often bearing papille, was well shown; the stomata were not clearly
seen, but there were depressions representing the cavities in which the stomata were
sunken.

Most of the recent observations on the epidermal structures of Mesozoic Cycadean
fronds have been made by Prof. Nathorst. He has described and figured the cuticles of
species of the following genera:— Dictyozamites, Pseudocycas, Ptilozamites ( Ctenozamites),
and Nilssoniat. He has given many excellent photographs of these, but has not yet paid
special attention to the structure of the stomata, which is one of the most important
features from the point of view of the present inquiry. As will be seen later, many of
the cuticles now described and figured can be very closely compared with the specimens
from the Scandinavian plant-beds. This is especially the case in the instance of Dictyo-
zamites Johnstrupi from Bornholm and Dictyozamites Hawelli from Marske. The close
similarity between these forms, and also the similarity between specimens of other types
occurring in both the Lower and Middle Estuarine beds of different parts of North

* Schenk (1867), p. 165.
Т Schenk (1871), pl. xv. figs. 2-4.
£ Nathorst, Palüobotanische Mitteilungen i.-iv., 1907-8; and * Ueber die Gattung Nilssonia, 1909.

CUTICLES OF CYCADEAN FRONDS. 179

Yorkshire, indicate that the cuticular structures are not essentially affected by purely
local conditions, but show characters which are related to the more general conditions
and also to questions of phylogeny.

More recently Prof. Nathorst (1912, p. 36) has given a short summary of the
results of his cuticular work, and has arranged the genera which have been studied
in five groups or orders characterised by the following forms:—(a) Pterophyllum,
(b) Otozamites, (c) Pseudocycas, (d) Nilssonia, (e) Ptilozamites. Не does not give a
detailed description of the characters of the cuticles of these orders, and they appear to
be somewhat unequal in value. Не thinks that classes (а), (b), and possibly (с), belong
to the Bennettitales, while (d) and (e) are more related to the modern Gymnosperms.

For the purposes of comparison with another group of Jurassic Gymnospermous
plants, the excellent figures of stomata of some genera of Ginkgoales given by
Prof. Seward * are very useful. The Ginkgoales possess a very definite and constant
type of stomatal structure which is somewhat similar to that found in some Nilssonias.

(^ Detailed Descriptions of the Cuticular Structure of some Fossil Cycadean Fronds.

I. Bennettitales.

PTILOPHYLLUM (WILLIAMSONIA) PECTEN (Phill.), n. comb.

The large and somewhat variable group of fronds usually placed in the aggregate
species Williamsonia pecten, can now be divided into several sections according to the
structure of their epidermal cells. The smaller type of frond with regular straight linear
or linear-lanceolate pinnze, closely approximated and attached by almost the whole length
of their bases, can be distinguished from the larger type with broader pinn:e, which are
slightly falcate, more truncated at the apex and contracted at the base, and with a some-
what more spreading venation. So far as can be seen from external appearance, these
two types seem to be connected by a series of gradations, as may be noticed in the figures
given in Prof. Seward's * Fossil Flora of the Yorkshire Coast,’ where pl. 3. fig. 1 resembles
the small type, and pl. 3. fig. 6 probably represents a frond of the larger type (the
differences mentioned above are, however, much more noticeable in other examples from
Marske and Whitby collected by one of us). When cuticular preparations are made, the
two types of frond can be readily distinguished by the presence or abseuce of large
numbers of thickened circular markings arranged in rows on the cuticles of the second
form. "These, as mentioned later, were probably of the nature of the thickened bases or
scars of hairs of which no trace is seen externally without considerable magnification.
The fronds possessing these curious markings, as well as other slight differences, must
probably be placed in a new species. The separation of these two types probably involves
the distinction of a third and larger type of frond, and the recognition, that the three
forms from Marske, figured recently by one of us (Thomas, 1913), are probably three

ж Seward (19117), pl. 4. figs. 56, 57; pl. 5. figs. 58-62 & 64.
SECOND SERIES.—BOTANY, VOL. VIII. 2E

180 MR. В. HAMSHAW THOMAS AND MISS N. BANCROFT ОХ

distinct species. We cannot here enter into this question in detail, but it may be
remarked that the existence of several species of fronds is rendered probable by the
discovery of a number of different species of male and female flowers which probably
belonged to some of the plants bearing these Ptilophyliwm fronds. Three or four forms
of male flowers and five or six species of female flowers of the Williamsonia type
are now known. Besides these we know of two or three other types of flower which
probably belonged to plants bearing fronds of some of the other forms placed-in the
Bennettitalean group in the present paper. In view of the uncertainty of connecting
*flowers" with fronds, it appears fo be better, at present, to use the generic term
Williamsonia to apply only to reproductive structures, while referring to the fronds
under the names of their form genera. Thus we shall refer to Williamsonia pecten
(Phill.} fronds as Ptilophyllum pecten and to Williamsonia gigas fronds as Zamia gigas,
Lindl. & Hutt. |

We may now proceed to the description of the cuticles of fronds of Ptilophyliwm
pecten (Phil) *. Тһе cuticles of the upper and lower epidermis are very distinct (вее
РІ. 20. fig. 1). Тһе upper epidermis is composed of uniform cells showing no traces of
veins and having no stomata. The walls of the cells show an extremely sinuous outline,
a characteristic in the Bennettitalean group; each wall is bent into a regular series of
folds whieh sometimes almost touch each other. Тһе appearance of this cuticle is seen
in the photograph reproduced on РІ. 20. fig. 2.

The lower epidermis is composed of cells with a somewhat similar outline, usually
rather thinner than those of the upper epidermis and with larger lumina. There is а
marked differentiation between the areas above the veins which possess no stomata and
the intermediate portions where the stomata occur in two or three rows (Pl. 20. fig. 1).
The cells above the veins exhibit a somewhat. more elongated form than the rest. А few |
circular hair-scars are sometimes seen. Near the margin of the leaf the cells become
more elongated and their walls thicker.

The Stomata.—In each square millimetre of the lower surface between 50 and 60
stomata occur; this number seems fairly constant in the specimens examined. Тһе
individual stomata are roughly circular, with a diameter of about "099 mm. Тһе guard-
cells are very regularly orientated, so that the slit lies transversely, at right angles to the
veins of the pinna. The regular arrangement of the stomata is a characteristic feature
of the Ptilophyllum-Otozamites group of fronds; in other genera they are irregularly
scattered. Тһе guard-cells do not seem to be appreciably sunk. Оп either side they
have a small subsidiary cell (s, text-fig. 27), оп to which the sinuous walls of the
epidermal cells abut. The actual structure of the guard-cells is somewhat difficult to
make out from the surface view of the remains of their cuticularised portions (Pl. 20.
fig. 3). From the analogy of the modern stomata (cf. text-fig. 26) it seems probable
that the structures seen in our preparations do not represent the outlines of the cells
round the stomatal pore but only parts of their walls which were differently thickened at

* For methods of obtaining cuticular preparations see Thomas (1912), also Nathorst (1912) and Bather (1908).

In this section the term cell-wall is used for the ridges on the cuticular film which now represent the walls of the
original cells,

CUTICLES ОЕ CYCADEAN FRONDS. 181

two levels. The appearance observed when preparations of stomata are seen in surface
view (text-fig. 27; РІ. 19. fig. 1) is as follows:—On either side of the central slit-like
pore are two elliptical or hemispherical structures; they are somewhat flattened where
they abut the pore, and have rounded ends, appearing, in fact, rather like an ordinary
euard-cell appears in surface view. Between these and the subsidiary cells lie two other
thickened patches, more or less hemispherical in shape, and apparently somewhat over-
lying the central structures. They possess a very characteristic outline, and have, like
the central ones, a deep brown colour; they often show peg-like extensions towards the
centre (see text-fig. 27). Very similar patches are seen in almost all the stomata of this
group. They often have a slightly shorter diameter (measured in the polar plane) than
the subsidiary cells, but are longer than the central structures; in some cases they
appear to be attached to the latter, but in other preparations appear to be free from
them at the ends, only lying over them in the centre (cf. РІ. 19. fig. 1).

Fig. 97.

Stoma of Ptilophyllum pecten, showing subsidiary cells (s)
and thickenings of guard-cells.

'The interpretation of these structures would be somewhat diffieult but for our know-
ledge of the thickenings possessed by the guard-cells of modern forms. In the surface
view of these we never see the outlines of the guard-cells, but only the form of their
thickenings or woody lamelle. Іп Bowenia the thickening on the upper face of each
guard-cell forms a hemispherical structure at each side of the pore, while the lamellz
on the lower faces contribute the projecting pieces seen at the poles (cf. text-fig. 26).
In the present case we seem to be dealing with a very similar structure. It is clear that
the central patches bordering on the slit do not represent the actual guard-cells. ‘They
are often extended into sharp points at the poles, are much smaller than the subsidiary
cells, and do not abut directly on to them; moreover, on this view the presence of the
intermediate patches cannot be easily explained. It seems likely that we see in the
fossil fronds two pairs of thickening patches or lamelle belonging to the upper and lower
sides of the original highly inclined guard-cells. These lamellze were somewhat similar to
the woody lamellz of the modern forms, but were more strongly cutinised. Тһе central
lamelle would be originally on the lower side of the cells and would be much shorter
than in the modern structures. Тһе outer and more thickened patches would represent
the upper lamellz of the guard-cells, which, however, overlapped the lower Іаше to a

much smaller extent than in Stangeria and Bowenia, but are almost identical in shape
2Е 2

182 MR. Н. HAMSHAW THOMAS AND MISS №. BANCROFT ON

and in their relations to the subsidiary cells, with the thickenings of Stangeria seen
after maceration (see text-fig. 32). Тһе figures below (text-fig. 28) show an attempt to
reconstruct the sectional views of the stomata at various points in accordance with the

97)

Diagrams representing surface view of stoma of Stangeria (1.), with hypothetical
sections (II.) at the different points indicated in I. The shaded and dotted
portions represent the probable positions of the thickened portions.

Fig. 28.

In some modern stomata we find that the upper lamellze do not extend unchanged to
the pore, but the latter is surrounded by a thinner oval area. А precisely similar
arrangement is seen in some of the Ptilophyllum and Zamia cuticles (see РІ. 19. fig. 2).

Unfortunately, it has proved impossible to obtain sections through the cuticles
showing the structure of these stomatal remains, but a comparison of the material
before us with the stomata figured by Lignier provides some interesting features and
gives considerable support to the interpretation just advanced. Іп his paper on
Bennettites Morierei, Lignier * observed that in some sections through the involucral
bracts, the stomatal guard-cells, cut through transversely, were seen. Іп one of his
figures, Pl. 1. fig. 9, which is reproduced below (text-fig. 29), two stomata are seen in

Section through the epidermis of a bract of Bennettites Morierei showing two stomata cut through.
The characteristic form of the guard-cells and subsidiary cells is well shown. (After Lignier.)

section cut through transversely somewhat near their poles. Тһе guard - cells are
seen to have the pear-shaped outline, so characteristic of the modern forms (of. text-
figs. 26, etc.), with a marked upward extension. Тһе upper walls of the guard-cells
are drawn as being somewhat thickened, the wall abutting on the slit being thin;

* Ligrier (1894), p. 18, pl. 1. figs. 8, 9.

CUTICLES ОЕ CYCADEAN FRONDS. 183

on either side of the guard-cells are two well-marked subsidiary cells. The figure
further shows that the guard-cells of adjacent stomata were approximately parallel to
each other, and that the inclined guard-cells became more vertical at the poles and
their upper or outer thickening became less. In figure 8 of the same plate another
stoma is shown cut through more or less longitudinally, but the section has not here
passed through the centre. No stomata were seen in surface view, but in a later paper *
surface views are given of stomata from the interseminal scales of the same specimen,
which had been obtained by collodion casts. Though the details are not very clearly
shown, it is evident that these stomata were of the same type as those which are
described here, the guard-cells and subsidiary cells of some of them being quite charac-
teristic. The structure of these stomata in Lignier’s species of Bennettites,a genus
closely allied to our Yorkshire Williamsonia, not only provides an interesting and very
close comparison with the modern Bowenia type, but also goes some way towards
demonstrating that the interpretation which we have given above, of the meaning of
the surface views of stomata in this group, is correct.

We have dealt at some length with the structure of these Ptilophyllum stomata
because the form of these organs is remarkably constant throughout the group of fronds
which we have assigned to the Bennettitales. Moreover, their appearance is so complex
that at first sight it is very difficult to understand them; no previous observers seem
to have made any attempt to grapple with the problem, but have merely spoken of the
thickening patches as the guard-cells.

| PTILOPHYLLUM sp.T

Typical examples of Ptilophyllum pecten, from the Lower Estuarine beds of
Whitby and Marske, show perfectly smooth cuticles with very few hair-scars. In the
larger fronds, however, from the same beds, a large number of very characteristic
structures are seen, which are also present on the fronds of the small pecten-form which
occur plentifully in the Cloughton Wyke beds of Middle Estuarine age.

These structures are shown in РІ. 20, fig. 4. They may be described as small annulate
projections from the surface of the cuticle ‘03 to ‘04 mm. in diameter, occurring in
numerous regular rows running parallel to the length of the pinna. The rows are very
close together, being often separated by a space less than their individual width. The
annulate structures of the same row are frequently contiguous and are sometimes
unevenly thickened, being broader and darker laterally and thinner at the points next
to the adjacent rings, so that an appearance is produced of pairs of crescentic patches.
They are so numerous as to obscure entirely the outlines of the epidermal cells and
make the stomata very indistinct.

We may regard these structures as the bases of hairs, with which the lower surface of
the frond was covered. Scattered hair-scars occur on the ordinary types of frond, which

* Lignier (1912), p. 426, fig. 1.
+ This species will later be described as Ptilophyllum hirsutum ; a figure of typical outline of a frond is given in
the ‘ Fossil Flora of the Marske Quarry,’ text-fig. 3 в. Thomas (1913).

184 МВ. H. HAMSHAW THOMAS AND MISS N. BANCROFT ON

do not differ markedly from those on some modern Cycad fronds, but they are not so
massive as in the present example, nor are they arranged in such a regular sequence.
When, however, the epidermal cells are arranged in regular rows, the outgrowths from
them may be expected to possess the regular arrangement which we see here. The
fronds just mentioned appear to possess stomata which are somewhat similar to those
above described (see text-fig. 30).

An interesting comparison can be drawn between the hair-scars seen so abundantly on
these fronds, and the hairs on the bracts and lower part of the strobilar-axis in the
flower of Williamsonia scotica, recently described from the Kimmeridge of Sutherland
by Prof. Seward *. The hairs in the latter formed a very striking feature; they were
very numerous, and arose from the thickly cuticularised outer wall of an epidermal cell
which formed a dark ring round the lower end of the hair. Higher up the walls of the
hair were much thinner. If leaves bearing hairs of this type had been preserved
as impressions in the same way as the material now before us, we could scarcely expect
the substance of the hair to be preserved, though the cuticularised base would remain as

Fig. 30.

т

e

Stomata of Ptilophillum sp., showing form of the guard-cell thickenings.

a clearly visible ring on the outside of the epidermis. It is possible, too, that hairs of
this type might have been deciduous. ‘The occurrence of these characteristic hair-scars
forms another link connecting the Péilophyllum fronds with the type of inflorescence
now known under the generic name of Williamsonia. |

ZAMITES (WILLIAMSONIA) GIGAS, Lindl. & Hutt.

There have yet been found few examples of this species from which good cuticular
preparations can be made. One specimen, however, from Marske, which has large
ріппге and apparently belongs to this species, yields preparations which show the cells
of the upper epidermis fairly clearly. Тһе lower epidermis is somewhat crumpled, and
the preparations made from it are always rather dirty ; its cells are very indistinct, and
the pits in which the stomata appear to have been sunk are usually filled with matrix.
We can, however, make out some of its epidermal characters. The upper epidermis is
thinly cuticularised, and is composed of cells of the usual type with very sinuous out-
lines. The lower cuticle is also very thin, and the outlines of the cells cannot be seen.
The stomata are arranged in fairly regular rows with their axes at right angles to the
rows; the guard-cells appear to be considerably sunken and are surrounded by a brown,
rather highly cuticularised area, probably the walls of the pit. The guard-cells possess

* Seward (1912), p. 108, pl. 12. figs. 40-42 & 45.

CUTICLES OF CYCADEAN FRONDS. : 185

the same characteristic thickenings as previously described; the crescentic patches аге
terminated in sharp points (cf. text-fig. 31) and partially cover the ventral patches,
while the central space on either side of the slit is thinner. The small projection
from the centre of the crescentic patch which is characteristically seen in Ptilophyllwm
pecten, Dictyozamites, ete., is here apparently absent.

A large frond from Marske with a form somewhat like that of Zamites gigas yielded,
_ after some difficulty, a cuticular preparation of the upper epidermis, in which the cells
are rounded in shape and quite distinct from the other examples in the present group.

ZAMITES (P) sp. Scale-leaves from stem.

In the Natural History Museum (Paleobotanical Department) of Stockholm, there
is an interesting specimen, collected at Whitby by Prof. Nathorst in 1909. It appears
to be the remains of a branching stem about 5 cm. wide; one part of it is fairly weil
preserved, and seems as if covered with numerous overlapping scales or leaves; it is

Fig. 31.

Stoma of Zamites gigas, Lindl. & Hutt., showing form of the thickening on the upper side of the guard-cells.
The circle surrounding the stoma represents the outline of the shallow pit in which the stoma was sunk.

somewhat similar to those of the James Yates Collection, figured by Wieland (1911,
figs. 2, 3).

The scale-leaves appear as if lanceolate in shape, of rather thick texture, and striated ;
they are very indistinctly outlined. They yield, however, cuticular preparatious which
show a certain amount of structure, though considerably obscured by a mass of hypo-
dermal or other tissue, which was evidently thick-walled and has resisted both decay
and maceration. The actual epidermal cells are not clearly seen, but some of the
stomata are quite visible; they appear to be similar in both form and arrangement
to those previously described (see РІ. 19. fig. 2).

186 MR. H. HAMSHAW THOMAS AND MISS N. BANCROFT ON

OTOZAMITES GRAPHICUS (Leckenby, ex Bean MS.) п. comb.

Fronds of this form from Marske yield good preparations, but the cuticles are very
thin and fragile, and hence only small pieces are obtainable. The similarity of its
cuticular structure to that of the foregoing types is very noticeable. The upper epidermis
has small squarish cells with rather thick walls showing closely approximated folds.
The lower cuticle is very thin, with the usual type of cells; the stomata are rather
sparsely scattered, only 40-50 per sq. mm. occurring in the specimens examined. In
this they contrast with Otozamites Feistmanteli from the same bed, in which the stomata
are twice as numerous. The stomata are arranged more or less in lines between the
veins which are marked by the greater elongation of the epidermal cells; they have
a fairly regular orientation, the pore being at right angles to the veins. They are not
deeply sunken.

The stomatal structure (see Pl. 19. fig. 4) is like that of Zamites. The thickening
patches are hemispherical to spindle-shaped, and the central area is thin. The sides
of the pore seem to be slightly thickened, and the two lateral subsidiary cells are also
somewhat thicker than the other epidermal cells.

OTOZAMITES F'EISTMANTELI, Zigno.

The dark shale bed at Marske has yielded good specimens of the smaller Ot¢ozamites,
which gives good cuticular preparations. These show some interesting points; the
stomata are clearly seen and are very numerous, being almost contiguous in lines
between the nerves. About 100 occur in each square millimetre. In structure and
form they are similar to the stomata of the larger species, but the lateral thickening
patches thin out gradually towards the slit, showing no sharp margin, but more or
less merging in towards the central area. Тһе latter is very slightly cuticularised
except at the edges of the pore, which stands out clearly as a dark central line (cf. Pl. 19.
fig. 3). Тһе subsidiary cells are again slightly thickened, and their relations to the
other epidermal cells are often well seen.

DICTYOZAMITES HAwELLI, Seward.

The cuticular structure of Dictyozamites Johustrupi, Nath. has already been described
in some detail by Prof. Nathorst (1907, p. 12, Taf. 3. figs. 2-8). The structure of the
present examples from Marske does not differ materially from that of the Bornholm
examples, but the structure of the stomata can now be further elucidated. As in the
former cases, the upper epidermis is devoid of stomata, and is composed of cells with
very sinuous walls, the reticular course of the veins being marked by the elongation,
narrowing, and thickening of the cells above them. The folds in the cell-walls of this
species are not so deep as those previously described. The lower epidermis is less
strongly cuticularised than the upper; the course of the veins is again marked by elon-
gated cells with less folded walls. The stomata occur only between the veins; they are
more or less irregularly scattered, but still show a distinct tendency to have the pore

CUTICLES OF CYCADEAN FRONDS. 187

placed at right angles to the adjacent veins. The stomata are numerous, about 120 per
sq. mm. Small papillze occur on many of the epidermal cells, but are not so conspicuous
as in D. Johnstrupi. The structure of the stomata appears somewhat complicated
owing to the presence of what seem like small thickened outgrowths or papille from
the surface above the guard-cells, and which are shown in Pl. 19. fig. 8, where the
euard-cells have been by some means pulled apart. Іп other cases they appear to
overlap the central space, as shown in Pl. 19. fig. 6, and РІ. 20. fig. 7, or give the
impression of the guard-cells interlocking with each other, as mentioned by Prof. Nathorst.
These projections are probably of a nature very similar to those observed in the case
of Ptilophyllum pecten (cf. Pl. 19. fig. 1). Sometimes they are not visible (cf. РІ. 19.
fig. 7), being probably on the under side of the preparation; in this case the thickening
patches are very comparable to those in Otozamites. On either side of each stoma are
subsidiary cells, somewhat thickened and often much elongated (cf. РІ. 19. figs. 6-8).
In preparations mounted with the lower side of the stomata uppermost, the latter seem
to have the same form as in several other examples of this group, the ventral patehes on
the guard-cells being somewhat thinner.

ANOMOZAMITES (WIELANDIELLA) NILSSONI (Phill.), n. comb.

Leaves of an Anomozamites form have been found in Yorkshire near Gristhorpe,
Marske, and Roseberry Topping. They are very similar in shape to those of
Wielandiella angustifolia, Nath., and may well have belonged to plants bearing floral
structures similar to those found by Nathorst in connection with his species; such
structures are not, however, yet known with certainty from England, though stems have
been found which are almost identical with those of the Rhætic plant. The cuticular
preparations obtainable from the Yorkshire Anomozamites leaves are exceedingly
beautiful, and show clearly the epidermal and stomatal characters. Тһе leaves from
the Lower Estuarine Series at Marske differ slightly from those of Gristhorpe in the
possession of numerous spherical papille on their lower euticles ; the stomata are also
slightly smaller.

In the specimens studied the upper epidermis is very uniform, and is composed of
more or less rectangular cells ; the courses of the veins are not clearly marked off, but
occasionally they are indicated by lines of slightly more elongated cells. The cell-walls
are again sinuous or zig-zag, but the folds have a different form from those in the other
genera, each of the ridges and grooves being angular rather than nail-shaped or rounded ;
each point appears to possess a small pad of cutin. Тһе upper epidermis is devoid of
stomata. In some preparations the cuticle from the rachis of the frond is also seen ;
this has regular rectangular cells with straight walls, and occasional structures like
rudimentary stomata. "The lower epidermis has a structure similar to that of the upper,
but the cells are less regular, and the stomata are irregularly scattered (cf. P1. 20. fig. 8).
The courses of the veins are seen, though not very clearly, and no stomata lie upon
them. There are about 70 stomata per sq. mm., and their form is very clearly seen in
most examples. It is shown іп РІ. 19. fig. 5, and РІ. 20. fig. 8. Тһе central slit is

SECOND SERIES.— BOTANY, VOL. VIII. 2r

188 MR. H. HAMSHAW THOMAS AND MISS N. BANCROFT ON

bordered by two fine linear membranes, beyond which are the thickening patches some-
what hemispherical in shape, with a straight inner surface and rounded or pointed ends.
The inner surface of each appears to be carried оп as a wall, a fact which is of some
importance in the interpretation of the structure. There are two crescentic subsidiary
cells.

In their structure these stomata resemble the examples seen іп Otozamites, especially
О. graphicus, and also in Teniopteris vittata (cf. РІ. 19. fig. 12).

TNIOPTERIS VITTATA, Brongn.

The entire linear-lanceolate fronds with a strong petiole, central midrib, and dichoto-
mising nervation which are so common in the Yorkshire plant-beds were designated
Teniopteris vittata by Brongniart and, as the name implies, regarded as fronds or
ріппге of a fern. Phillips in 1829 (p. 147) regarded it as allied to the recent Scolo-
pendrium, but subsequent authors * have for the most part considered it to be allied to
Marattiopsis and have placed it among the Marattiales. Another view was that it had
affinities with the modern genus Oleandra, since it possessed a petiole with a clean-cut
base, which suggested that it had become detached from its original stem or rhizome by
the formation of an absciss-layer 7; thus the generic name of Oleandridium was used by
Schimper (Гос. cit.), Feistmantel, and others.

On the other hand, the view was put forward by some authors, especially Bornemann
(1856, p. 59), that it was really the pinna of a pinnate Cycadean frond such as Stangeria,
and the name Stangerites was instituted to replace Teniopteris. Several subse-
quent authors have kept in mind this possibility, but no definite evidence has been
forthcoming in support of either view. T. vittata has never been found bearing synangia
or sporangia; it has also never been found attached as a pinna to a larger frond.
Recently, however, it has been found in association with Cycadean stems of the
Wielandiella type, and one specimen appears to show organic connection between a
frond and one of these stems. It is almost certain that we have to deal with entire
fronds and not with pinnz, and thus a very close comparison is offered with the
fronds of Anomozamites (Wielandiella); the latter possess the forking nervation of
Teniopteris and may be regarded as segmented fronds of the Teniopteris type.
T. vittata itself is sometimes found with the lamina divided into irregular segments.
When we come to examine the cuticular structure we obtain very strong confirmation
of the Cycadean view.

In its cuticular structure Teniopteris vittata approaches very closely to Anomozamites
Nilssoni. The upper epidermis is fairly strongly thickened and was composed of
squarish to hexagonal cells, with somewhat sinuous walls (РІ. 20. fig. 5). Тһе vein-
courses are only slightly differentiated, the cells being more regularly arranged in rows.
The outline of the cell-walls is interesting owing to the presence of a solid piece of
thickening at the point of every fold (this is just visible in the photograph); in some
cases it gives the appearance of an almost straight wall ornamented with projecting

* Schimper (1869), p. 607. Т Seward (1900), р. 158.

CUTICLES ОЕ CYCADEAN FRONDS. 189

tooth-like masses of thickening, but it can always be referred to the same general type
of wall-folding. We see the same tendency in Anomozamites. The lower epidermis
is composed of cells similar to those of the upper epidermis, but of a somewhat larger
size; their walls are thinner and may be sinuous or have the folds filled in, as just
mentioned. A few rounded hair-scars are present. The stomata are absent from the
nerve-courses and are irregularly scattered between (cf. РІ. 20. fig. 6); about 60-70
occur in each square millimetre. They have the same general form as previously
described, but are apparently rather more complex. Some of the examples from Marske
(РІ. 19. fig. 12) do not differ greatly from the Anomozamites stomata except in the
proportions of thickening, the areas abutting immediately on the slit being the more
strongly thickened. The stomata in the Gristhorpe examples appear to have been
slightly sunken ; in most cases the pore is clearly seen and is surrounded by a strongly
thickened margin running to a point at either pole (Pl. 19. fig. 11). The outer (upper)
thickening patches are often not so clearly defined as in the other genera, but have the
same relations to the auxiliary cells as those of Anomozamites. These outer patches are
often thinner than the inner ones and occasionally appear to project into the centre,
almost above the slit (Pl. 19. fig. 10); they sometimes develop a small projecting
hump, similar to that seen іп Ptilophyllum pecten and Dictyozamites. Though it is very
diffieult now to understand what the structures, which we see in the surface view, must
have meant in the solid, yet it is quite evident by comparison with the cuticles of other
groups, that Teniopteris vittata possessed essentially the same epidermal and stomatal
characters as the other fronds in the Bennettitalean group, being especially like
Anomozamites.

When this is considered in addition to the facts quoted above, it seems that we have
strong reasons for removing this species and its near allies from among the ferns and for
including it in the Bennettitales.

T ÆNIOPTERIS MAJOR, Lindl. & Hutt.

Some Teniopteris fronds occur at Gristhorpe which are somewhat larger than those of
T. vittata, but are readily distinguishable by their distinct and widely separated dicho-
tomising veins, which are one millimetre apart or more and are very characteristic ; these
fronds are probably distinct from T. vittata, and may be referable to the form described
by Lindley and Hutton as 7. major.

In its cuticular characters it does not differ much from the species just described; the
cells of the upper epidermis are rather large and have well-developed angularly folded
walls, slight additional thickenings being present on the tips. The lower epidermis has
somewhat scattered stomata which resemble very closely those of 7. vittata (РІ. 19.
fig. 9).

In this case the epidermal characters do not, so far as we have seen, present any
sharply marked features on which we can rely for the separation of the related forms

major and vittata.

190 MR. Н. HAMSHAW THOMAS AND MISS N. BANCROFT ОХ

ІІ. Nilssoniales.*

NirssoN1A, Brongn.

While the cuticles of the genera already described show a very considerable agree-
ment, both in the structure of their epidermal cells, and also in the form and arrange-
ment of the stomata, the cuticles of the remaining genera Nilssonia, Ctenis, and
Ptilozamites are quite distinct from them. In these forms the epidermal cells are some-
what irregular in shape, being rounded, hexagonal, or rectangular ; the walls are straight
without any indication of folding. The stomata appear to be somewhat sunken, and
when any trace of the guard-cells can be seen, they apparently lack special thickening
and have a more or less rounded contour. The appearance of these cuticles therefore
marks them at once as distinct from those of the Bennettitalean group.

The Nilssonias must have been important constituents of the Jurassic vegetation of
the Yorkshire area. Fronds of Nilssonia compta occur abundantly in the Middle
Estuarine bed in the neighbourhood of Gristhorpe Bay. Іп the same locality
N. mediana occurs ; this is almost identical in form with a plant which is very abundant
in the Lower Estuarine beds of Marske and Roseberry Topping, but which possesses
very much more delicate cuticles. The same form occurs at Whitby, while another type,
closely allied to М. compta, but probably a species new to England, is found at Roseberry
Topping. The cuticles of these three species or forms have been examined, and their
characters are briefly described below. The cuticles of the Swedish species brevis,
polymorpha, and pterophylloides have been described and figured by Prof. Nathorst.
They apparently agree somewhat closely with the English species described here, though
little or nothing could be made out about the form of their stomata.

NILSSONTA COMPTA, Phill.

The examples of this species from the Gristhorpe bed indicate that they originally
had a somewhat delicate texture, and owing to this and to the thinness of the cuticles it
is not easy to obtain good cuticular preparations from them. Even after portions of
the carbonised lamina have been successfully removed from the matrix they frequently
break up into numerous small fragments during the clearing process. It is believed
that this affords a character of real value in describing and identifying fronds. Іп many
descriptions of leaf-impressions, the lamina is described as leathery or coriaceous, from
the appearance of the specimen when still attached to the rock or even from the
appearance of its cast; in most of these cases little value сап be assigned to such a
statement. When, however, we are able to remove the carbonised or mummified plant
from the matrix and investigate the structure of its cuticle, we have definite grounds
for a statement as to the comparative texture of the lamina.

In the case of the plants preserved in the fine-grained mudstone of the Gristhorpe bed,
we can say with some confidence that such fronds as Zeniopteris vitiata, Т. major,
Ctenis falcata, and especially Ptilozamites Leckenbyi, had a firm and coriaceous lamina.
Others, such as Anomozamites Nilssoni and Nilssonia compta, were of intermediate

* See page 196.

CUTICLES ОЕ CYCADEAN FRONDS. 191

character, while Nilssonia mediana and №. orientalis had a distinctly delicate or
thin lamina, approaching the type possessed by the ferns and seldom separating from
the matrix in an entire condition.

The upper epidermis of N. compta is very distinct from the lower. It possesses small
rather irregularly-shaped cells arranged, to some extent,in rows. Тһе individual cells
vary greatly in shape; they may be isodiametrie with rounded or angular outlines, or
they may be more elongated, and of rectangular to fusiform shapes. The latter kind
probably occur above the nerves. АП the cell-walls are straight and not sinuous.
There are usually no stomata present, but in one or two cases, structures which may be
stomata, are seen near the margin of the lamina. А few rounded papillæ occur, these
being of the type common to most of the Nilssonias. Their form is better seen on the
rachis and in other species.

The cells of the lower epidermis are much more clearly seen; the nerve-courses are
distinctly marked by the occurrence of rows of elongated cells, often bearing papillæ, but
devoid of stomata. Тһе spaces between the nerves are occupied by rounded cells, with a
good deal of thickening in the corners. The cells surrounding the stomatal apertures
are much thickened. The stomata on these leaves are not numerous; estimates from
the very fragmentary cuticles indicate between 40 and 50 in each sq. mm. 16 is almost
impossible to make out the exact structure of the stomatal openings, but the following
features are recognisable. Each aperture is surrounded by crushed thickened cells, which
evidently formed, in their original condition, a chimney-like structure, being raised
beyond the surface of the leaf, and enclosing a pit, at the base of which the guard-cells
were placed. Amongst recent Cycadean cuticles, a similar condition is seen in Dioon
and in Cycas revoluta. The fossil cuticles do not show the projection of the cells sur-
rounding the stomatal apertures, owing to crushing in preservation. The cells are
seven or eight in number, and often show a star-like arrangement. Their relations to
the ordinary epidermal cells are not clear. When some of the preparations are mounted
so that their inner side (viz., that abutting on the mesophyll) is uppermost, the remains
of the guard-cells can often be distinguished (РІ. 19. figs. 14 & 18). The guard-cells
themselves may be somewhat indefinite, but the pore is usually quite distinct. The
thickening of the guard-cells seems to be of a simple type, and not complex as in the
Bennettitalean fronds; this may perhaps be correlated with their sunken position.

The cuticle from the rachis below the frond presents some distinctive features. It is
composed of elongated spindle-shaped cells, giving the appearance which sclerenchy-
matous cells show in longitudinal section. . At frequent intervals occur oval cells, which
are considerably more cuticularised than the rest. In the centre of each of these is a
rounded area, probably representing a short papilla. Structures of this nature are also
found sparingly on both surfaces of the lamina, as mentioned above.

NILSSONIA MEDIANA (Leckenby, ex Bean MS.), n. comb.
Examples of this species from Gristhorpe do not yield good preparations owing to the
delicate nature of the cuticle ; specimens from Marske, which are very similar in external
form, have a well-thickened cuticle, and preparations from these show the outlines of

192 MR. H. HAMSHAW THOMAS AND MISS №. BANCROFT ON

the cells very clearly. The cuticles of №. mediana differ from those of N. compta only
in detail. The cells of the upper epidermis are more elongated and spindle-shaped, and
those of the lower epidermis are very sharply differentiated according to their position
above or between the nerve-courses (Pl. 20. fig. 10). Some рарШ are present, and each
of the rounded cells from the areas between the nerves is thickened externally and forms
in consequence a slight convex projection on the surface of the cuticle. The stomata
can be more clearly seen than in N. compta. Тһе cells surrounding the stomatal
aperture are only slightly raised and thickened, and therefore they are less crushed.
The raised portions are in this species seen to be merely small outgrowths from the
ordinary epidermal cells, slightly thickened (cf. Pl. 19. fig. 15).

NILSSONIA sp.

А new form of Nilssonia has recently been discovered at Roseberry Topping, Yorkshire.
It is closely allied to N. compta of the Middle Estuarine beds, but has more rounded and
less regular leaf-segments, approaching the N. polymorpha type. It also has а very
marked wrinkling of the lamina. It yields good cuticular preparations (РІ. 20. fig. 9),
though these are somewhat brittle. 'The stomatal apertures in this species are sur-
rounded by several strongly thickened conical cells which originally must have formed
conspicuous projections, but which are now usually much crushed (see РІ. 19. fig. 13).
The papillate hairs are also well marked in this new form.

NILSSONIA ORIENTALIS, Heer.

Of the three species of Nilssonia described in Prof. Seward's ‘Jurassic Flora of
Yorkshire, NV. compta and №. mediana were well defined and represented by many
specimens in the then existing collections, while the third was only known from a few
examples. The name given to this was N. fenuinervis, a title which had originally been
proposed by Nathorst (1880, p. 35) for linear fronds, little divided and having numerous
fine veins. It now seems impossible to separate this English form from the Siberian
type described by Heer (1878, РІ. 4. figs. 4-9) ав N. orientalis *. In his original MS.
notes on the specimens amongst which №. Zemwinervis was found, Prof. Nathorst t
remarked on the great similarity in external form between the new species and some
examples of Tteniopteris vittata; and, in fact, they can only be distinguished readily in
well-preserved examples. In 1909, when working at the revision of the Swedish species
of Nilssonia and studying their epidermal characiers, Prof, Nathorst (1909, p. 27)
made cuticular preparations of some fronds from Cloughton Wyke, Yorkshire, which
were in his collection. These preparations showed a type of structure unlike that of
the other Nilssonias, and identical with that described above for Temopteris vittata.
Nathorst concluded that the Yorkshire frond was not a true Nilssonia, but was rather
allied to the ferns ; he accordingly instituted the new genus Wilssoniopteris, giving as

* With reference to the nomenclature of this form see Thomas (1913), p. 240.
t Our best thanks are due to Prof. Nathorst for the kind way in which he placed all his material at my disposal,

for showing me many of his manuscript notes and adding materially to the elucidation of this difficult
species, —H. H. T.

CUTICLES OF CYCADEAN FRONDS. 193

its characters the distinct cuticular structure and the dichotomising of some of the
secondary veins. In 1911, while working in Prof. Nathorst’s laboratory, one of us was
able to re-examine his specimens and preparations, and found that one of Prof. Nathorst's
fronds, that figured on РІ. 6. figs. 23 & 25, was really a specimen of Zænmiopteris
vittata, showing a forking venation. Аз this form yields good preparations it had been
used in the description of the examples figured on his РІ. 7. fig. 21.

From the other specimen only very fragmentary preparations could be obtained, but
they were sufficient to show that their structure was quite distinct, agreeing with that
of true Nilssonias, in which the cuticles are extremely delicate. Іп this form the
fragments of cuticle (one of them is shown in Pl. 20. fig. 11) show that the upper
epidermis is composed of elongated rectangular or tapering cells, with straight walls
similar to those of N. mediana. The lower epidermis has fine rectangular or squarish
cells with straight walls and small dark rounded structures like the papillate hairs of
N. compta and М. mediana. Хо stomata can be clearly seen.

Other examples of the same kind have been obtained from the Gristhorpe bed; they
agree in the characters given above, but do not add any information. "They show a well-
marked minute wrinkling of the cuticular film which rather obscures detail.

From these facts it appears that the name Nilssoniopteris must be abandoned, for the
cuticles of the Yorkshire specimens referable to Nilssonia orientalis (Lenwinervis) are
of the same type as those of other species of Nilssonia.

CTENIS FALCATA, Lindl. & Hutt.

Specimens of this species found in the Gristhorpe bed flake off very readily from the
matrix and yield fairly good cuticular preparations.

The upper epidermis is very strongly cuticularised, and the thickening evidently
originally extended down the radial walls to a considerable extent, for the cell outlines
are now represented by thick badly defined lines of cuticle. The cells above the nerves
are especially strongly cuticularised, often causing the nerve-courses to stand out as
dark, slightly raised lines; only two or three rows of cells, however, are affected, and
not a broad series. The anastomosing of the veins is clearly shown on both the upper
and the lower cuticles. The cells of the upper cuticle are irregular in outline, but are
more or less isodiametrie, varying between squares and hexagons. Some of them seem
slightly papillate, but are not markedly so. No stomata are present on the upper
surface. Тһе lower cuticle (РІ. 20. fig. 13) is somewhat similar to the upper in thickness,
but the cell outlines are larger and more distinct. The vein-courses are marked by a
more regular arrangement of cells, and the stomata occur somewhat scattered in the
intervening areas. Their numerical distribution is from 60-70 per square mm.

As in Nilssonia, the guard-cells are sunk and are not clearly seen. The cells
surrounding the stomata-openings are not much differentiated in form or thickening
from the ordinary epidermal cells; where they adjoin the guard-cells, however, there is
a solid ring of thickening round the stomatal cavity (РІ. 19. fig. 19). The latter is
apparently not very deep, and the guard-cells are often seen at its base. They are

194 MR. H. HAMSHAW THOMAS AND MISS N. BANCROFT ON

slightly thickened, and in many cases are not preserved, but the stomatal slit is
frequently visible.

Сеепів falcata presents a type of stomatal structure comparable with that seen
іп Nilssonia mediana. In the former species, however, there is little or no tendency
towards the papillate outgrowths from the subsidiary cells surrounding the stomatal
apertures, which is so pronounced a feature in some Nilssonias, for example М. compta.

PTILOZAMITES (CTENOZAMITES) LECKENBYI ( Leckenby, ex Bean MS.), n. comb.

The cuticles of this form have already been figured by Prof. Nathorst (1908, Taf. 1.
figs. 3-5) as providing an example of a frond with stomata on one side of the pinnæ
only. He did not enter into a detailed description of the structures seen.

In comparing the structure of the cuticle with that of other genera, the following
points may be noted. Тһе leaves have a very stout texture with thick cuticles so that
the leaflets. separate off whole from the matrix and can be easily handled. Іп most of
the specimens examined the serration of the margin is not pronounced, but a few teeth
are seen near the apex. Тһе upper cuticle is very thick and shows very clearly the
outlines of the epidermal cells which must originally have had their radial walls very
considerably cutieularised. Тһе cells are irregularly shaped, but are straight-walled
and somewhat rectangular. The cells above the nerves are more elongated than
the rest. Тһе lower cuticle is also thick and shows cells similar in form to those of the
upper epidermis. Тһе cells are slightly elongated above the veins, but the courses of
the latter are not clearly defined and are not always devoid of stomata. Тһе stomata
were numerous, their distribution being about 90 per square mm. (Pl. 20. fig. 12). In
form they show considerable resemblance to those in Ctenis, being slightly sunk, and
having the walls and opening of the stomatal cavity much thickened (Pl. 19. fig. 17) ;
the cuticle round the opening, in fact, forms a little ring, slightly raised beyond the
surface of the pinna. Тһе seven or eight subsidiary cells are somewhat thickened and
appear as a group darker than the rest. On mounting a lower cuticle so that the
mesophyll side is uppermost, the outlines of the guard-cells are faintly seen; these
appear to have some resemblance to those of the Williamsonia-Tweniopteris group. Тһе
slit lies at right angles to the longer diameter of the stomatal cavity, and on either side
of it there are distinct though faint indications of guard-cells (7) with pointed ends
(Pl. 19. fig. 16).

(c) The Cuticles of some other Types of Fossil Cycadean Fronds.

The cuticular structure of some other genera have been described by other authors,
but in several cases the figures given are not sufficiently detailed or accurate to allow of
their close comparison with the examples here described. Prof. Nathorst has figured
and described the cuticles of two species of Pseudocycas from the Lower Cretaceous
(1907, pp. 7, 9, Taf. 2. figs. 1-11). These cuticles show more or less rectangular cells
arranged in regular rows and having undulating walls. Тһе stomata are not clearly
recognisable, but from Taf. 2. fig. 11, they appear to be somewhat similar in general

CUTICLES OF CYCADEAN FRONDS. 195

character to those of Anomozamites and the Bennettitalean group. Nathorst compares
the structure of these forms with those of Ofozamites, Pterophyllum, &c., and it does
indeed seem probable that these Pseudocycas fronds should be included in the Bennetti-
talean group, though in the disposition of the stomata in a furrow they are distinct
from any of the known Jurassic forms.

Berry, in his account of the Lower Cretaceous flora of Maryland (1911, p. 334), gives
a diagrammatic figure of the cuticle of Dioonites (= Zamites) Buchianus, Ett. The
cells as іп Pseudocycas are arranged in regular rows. ‘hey are rectangular and
had somewhat undulating walls. ‘The stomata are all orientated in the same direction
(cf. Ptilophylium, &с.), presumably with their long axes parallel with the long axis of
the pinna. If the analogy with Ptilophyllwm held, however, the stomata should be
orientated at right angles to the long axis of the pinna. The structure of the guard-
cells is shown diagrammatically and cannot therefore be strictly compared with our
types ; there is considerable probability, however, that they are similar in form to those
of Otozamites and Teniopteris. Schenck has figured Dioonites Dunkerianus with
sinuous walls and stomata irregularly scattered. We can probably, therefore, include
Zomites (Dioonites) іп ће Bennettitalean group.

The cuticular structure of various species of Thinnfeldia has been described by Schenk,
Seward, Gothan, Zeiller, and other authors. It is very doubtful, however, whether
this genus should be included among the Cycadophyta. The cuticular structure seems
to be distinct from that described above, and is apparently more closely related to that
of the Ginkgoales, especially so far as the stomatal form is concerned. There is an
approach to the structure shown by Nilssonia mediana, but on the whole Thinnfeldia
seems to be quite distinct.

(4) Summary of the Structure of the Fossil Cuticles.

The comparison of the structures just described brings out very clearly one point,
viz. that the Cycad-like fronds from the Jurassic rocks of Yorkshire fall into two
sharply divided groups according to their epidermal structure. The fronds in the first
group have epidermal cells with sinuous walls and comparatively thin cuticles; the cells
are in general rectangular in shape, but in most cases their shape is not very apparent.
The stomata occur only on the lower side of the leaf and are usually absent below the
vein-courses; they exhibit a strong tendency towards arrangement with their axes at
right angles to the veins. Though there is some variation in the details of stomatal
structure, the general plan is as follows. The guard-cells were level with the surface or
only slightly sunk; the outlines of the guard-cells themselves are never seen but only
those of the thickenings which they possessed. Тһе thickenings of the guard-cells
usually appear as lamelle of definite shape and give the stomata а characteristic
appearance; on either side of the pore isa pair of smaller patches, probably representing
the lamella from the lower side of the guard-cells, and outside there is another pair of
lamellee, usually somewhat hemispherical in shape, representing the thickenings from
the upper side of the guard-cells. At the poles there are usually two clear spaces,

SECOND SERIES.—BOTANY, VOL. VIII. 26

196 MR. Н. HAMSHAW THOMAS AND MISS N. BANCROFT ОХ

looking sometimes like cells, but probably representing the thin places above the
ends of the guard-cells. More or less surrounding the stomata at the sides are two
large subsidiary cells, often somewhat thickened. The above description applies both
to (а) pinnate compound fronds, e.g. Ptilophyllum, Otozamites, Dictyozamites, and
Zamites, some of which are known to have belonged to plants bearing flowers of the
Williamsonia type, and to (b) simple fronds, е. g. Teniopteris, which is usually entire,
and Anomozamites, which is usually pinnately lobed. Fronds of the latter have been
found to belong to plants with a peculiar type of Bennettitalean flower (Wielandiella).
There are then, strong grounds for regarding the type of cuticular structure described
above as characteristic of the fronds of the Bennettitales.

The cuticles of the second series, which includes the genera Nilssonia, Ctenis, and
Ptilozamites, are sharply marked off from those described above. In this section the
cuticles may be thin (as in the Nilssonias) or very thick; they appear to have been
formed from cells of elongated, rectangular, rounded, or fusiform shapes, with straight
walls. The stomata are on the under side of the leaves and exhibit no regular
arrangement, the guard-cells are somewhat deeply sunken, and the six or eight
thickened subsidiary cells which surround them often form a somewhat overarching
canopy. The guard-cells are hemispherical or spindle-shaped, they are only thinly
cuticularised, and definite lamellar thickenings are seldom seen.

The cuticles in this group do not agree with one another so closely as do those in the
Bennettitalean group, Ctenis and Ptilozamites are similar, but the iélssonias ате not so
closely related. Тһе fronds were variously shaped, those of Cfenis and of some species
of Nilssonia were large and pinnate, Ptilozamites was bipinnate, while the fronds of
Nilssonia orientalis were simple and often entire or only slightly lobed. It may be
noticed that the thickness of the cuticle in these forms bears an apparently definite
relation to the size of the frond. Of the bipinnate Péilozamites no complete fronds have
been obtained, but they were probably of considerable size, their cuticles were very
thick; the large fronds of Ctenis, and also of the Lower Estuarine form of №8804
mediana had cuticles of considerable thickness. The smaller fronds of Nilssonia compta
from Gristhorpe had thinner cuticles, while the much smaller entire fronds of Nilssonia
orientalis had very delicate cuticles. We see no such gradation among the Bennetti-
talean group. We know very little about the affinities of the plants which bore the
fronds placed in this second section. Prof. Nathorst has found fronds of Nilssonia
pterophylloides associated with numerous seed-like bodies in such a way as to render
their original connection probable, but the stems and reproductive organs of the other
species and genera are as yet quite unknown. Though it is quite possible that some
of these plants were closely related to the recent Cycadales, there is as yet no definite
evidence in support of this view. It may be convenient if we institute the name of
NILSSONIALES for the group of fronds which has just been described.

CUTICLES ОҒ CYCADEAN FRONDS. 197

IV. COMPARISON OF RECENT AND FOSSIL CUTICLES.

The difficulty of comparing the epidermal structure of the recent and fossil fronds is
considerable owing to the impossibility of obtaining good sections through the latter.
It is, therefore, necessary in most cases to employ only the information obtained from
surface view, and since in most cases the different types of stomata are chiefly to be
distinguished by their appearance in transverse section, our work becomes much more
difficult. 1t is possible, however, to notice a number of points of resemblance, and іп
fact many interesting comparisons can be made.

In the first place we may notice that the fossil fronds, like the modern ones, exhibit
some xerophytic tendencies in the thickness of their cuticles, in the frequently sunken
stomata, and in the very small number of stomata per square millimetre, the numerical
distribution of stomata in both being remarkably similar. These facts may indicate
that the xerophytie characters of the modern Cycadean fronds are ancestral.

Comparisons can be drawn between the shape and structure of the guard-cells. The
sections through stomata of Bennettites Morierei figured by Lignier, indicate that the
Bennettitalean stomata had a structure of the general Gymnosperm type, and that the
euard-cells showed the characteristic Cycadean form. The appearance in surface view
of the guard-cells of several genera, when they are not obscured by overlying cells, is
somewhat characteristic. It is well seen in Stangeria, Bowenia, and sometimes in
Zamia, Macrozamia, and Encephalartos. The guard-cells are thickened in a peculiar
way with a small oval thin patch round the pore, on either side of which is а spindle-
shaped or crescentic thickened mass. There is a thin patch above each end of the
guard-cells, but a solid or Y-shaped thickening at the points of contact of the guard-
cells. The resulting shape, which may be seen from the figures, seems to be fairly
constant for all modern Cycads, and is very characteristic, being modified in the case of
the deeply sunken forms. ‘The same type of stoma is seen in most of the fossil fronds
examined. |

The Bennettitalean stomata possessed guard-cells with very peculiar thickenings, but
which are referable to the same general type as exhibited by the modern fronds. We
had a centralthin region round the pore and two fusiform or crescentic patches оп
either side (compare the figures of Zamites, РІ. 19. fig. 2, and Otozamites, РІ, 19. fig. 3,
with figures of Zamia, text-fig. 21, and Stangeria, РІ. 17. fig. 1).

We have to remember, however, that in looking at the recent stomata the most
noticeable thickening is due to the lignification of the guard-cells, while in the fossil
forms presumably nothing but cutin remains. Іп order to get preparations of the
recent struetures which are really comparable with the fossils, we may get rid of the
woody tissue by continued boiling in Schultz's solution; in the resulting preparations
little but cuticle will be left. Preparations made in this way show considerable
similarity with the fossil forms (çf. text-fig. 32 with figures on Plate 19), the woody
thickening seems to have been accompanied by a certain amount of cuticularisation
which is especially seen at the poles, while on either side of the pore the thickenings
on the upper surface of the guard-cells are seen as the spindle-shaped рое which

| G2

198 Mk. Н. HAMSHAW THOMAS AND MISS N. BANCROFT ON

are so characteristic in the fossil forms; the thin areas between the pores and the
cusp-shaped thickenings are the same as are seen in the fossil forms, especially in
Otozamites, Anomozamites, and Teniopteris.

The polar thickenings of the guard-cells which are seen in most of the recent forms,
and which appear especially marked in Stangeria (where they consist of both woody
tissue and cuticle), are almost entirely absent in the fossil fronds, and present a point
of difference.

It is probable that the relations of the guard-cells to the surrounding epidermal cells
and subsidiary cells, offer one of the most important points of comparison which can be
obtained in the surface view. In the Bennettitalean fronds we have a very constant
and characteristic arrangement. On either side of the guard-cells there are two large
subsidiary cells of exactly the same diameter as the guard-cells and probably developed
from the same mother-cell. {The small cells seen at the poles of the slit in some cases,
e. д. in Anomozamites, РІ. 19. fig. 5, are probably not polar subsidiary cells but thin
places above the poles of the actual guard-cells.] In the Nilssoniales group, however,
and in recent fronds, the subsidiary cells are very different. They are usually six or

Fig. 32.

Stomata of Stangeria paradowa as seen in surface view after prolonged maceration. Notice that the lateral upper
thickenings of the guard-cells show as spindle-shaped patches, while there appear to be empty cell-like spaces
at the poles. Cf. with figures of Ptilophyllum and Otozamites stomata.

more in number, and are arranged more or less radially round the guard-cells. In
Stangeria, Encephalartos, and other forms, we find two large lateral and two smaller
polar subsidiary cells, but they often become subdivided. Thus in this respect the
Bennettitales are quite distinct from the recent Cycads. On the other hand the fronds
inclnded in the Nilssoniales may be closely compared with the modern forms in this
character, the numerous overarching subsidiary cells of Nilssonia, with their strongly
cuticularised points, sometimes coalescing to form a solid ring, being comparable with
those of Dioon and Cycas. These overarching cells frequently are of the same type
as those forming the remainder of the epidermis in both the fossil and the recent forms.

It is unfortunate that we cannot get more evidence as to the form and structure of
the guard-cells in the Nilssoniales group, though in the case of the example of Ptilo-
zamites figured (РІ. 19. fig. 16) the зеи seem to have been of somewhat the same
type as those of the Bennettitales.

CUTICLES OF CYCADEAN FRONDS. 199

In comparing the general epidermal cells of the recent and fossil types we may
notice that there is nothing among the modern forms which is really comparable to the
very sinuous cells of Ptilophyllwm and Anomozamites. The cell-walls of the Stangeria
epidermis are somewhat sinuous, and this is of interest when taken with the forking
nervation of its pinne, another point of comparison with the Bennettitales. If the
sinuous cell-walls are to be regarded as a feature of any phylogenetic importance, it
may indicate the relationships of the Bennettitales with the Pteridosperms and ferns.
On the other hand, the straight walls of the Nilssonialean epidermal cells, together with
the shape of the cells, are additional points of agreement with the majority of the
modern Cycads.

We may summarize generally the relationship between the fossil and modern
Cycadean fronds as indicated by this enquiry, as follows. The cuticular structures
indicate that the Bennetittalean fronds form an easily recognizable group containing
the majority of the fronds previously described as Cycadean, which are quite distinct
from the modern fronds and from those of the Nilssoniales group. Тһе latter, which
includes the genera Nilssonia, Ctenis, and Ptilozamites, are somewhat closely allied to
the modern fronds, and may probably be regarded as the true Mesozoic Cycads and the
ancestors of the modern forms.

On the other hand, the Bennettitalean stomata show some important resemblances to
the Cycadean ones, and may well point to a parallel line of evolution from a common
stock.

It is scarcely possible at the present time to make any detailed comparison betweer
the epidermal structures which have been described in this paper and those of the other
groups of Gymnosperms, since in no case have the latter been studied in sufficient
detail. A few general comparisons may, however, be attempted. It seems probable
that of the other Gymnosperms, all of which possess euard-cells with the same charac-
teristic pear-shaped section, the Gnetales are the most closely allied in the form of their
cuticular structures. Іп fact, it would seem that the Bennettitalean cuticles are more
closely allied to those of some of the Gnetums than to any other group, the points of
comparison being the frequent arrangement of the stomata in rows, the relations of the
guard-cells to the subsidiary cells, and the method of thickening of the guard-cells.
In Gnetum Gnemon the walls of the epidermal cells are often sinuous. We have
been able to examine the cuticles of Welwitschia and Gnetum Gnemon, and the
stomata in these forms have also been described by "Takeda * and others, but other
species need to be examined before the comparison can be pressed.

The stomatal structures of Ginkgo T are somewhat similar to those of Nilssonia, but
differ in the smaller and more constant shape and number of the subsidiary cells and in
the very slight thickening of the guard-cells. The stomata of the Conifers 2, while
possessing guard-cells which are thickened in а somewhat similar way, seem to differ
from those of the Cycadales and Bennettitales in the uniformity of the thickening and

* Takeda (1913), p. 350, text-figs. 1-5, pl. xxix. figs. 2-6.
+ Seward (19117), pl. v. figs. 59-62.
t Mahlert (1885), figs. on Taf. 1.

200 MR. Н. HAMSHAW THOMAS AND MISS №. BANCROFT ОХ

in the spreading out of the woody lamelle at the poles into an almost continuous
investment without the characteristic thin patches. The relations of the guard-cells
to the surrounding epidermal cells also seem to be another important feature of
difference.

It would seem, then, from the foregoing study, that the characters presented by the
stomata and epidermal cells of the Gymnosperms can be regarded as important as
indicating to some extent their relationships. In the Cycadean alliance we find some
characters which have undergone comparatively little modification in the ages which
have elapsed between Jurassic times and the present day, and we are consequently
entitled to regard stomatal structure as being the expression of ancestral characters
rather than of purely loeal and temporary conditions of environment. Of course,
further study of other groups is needed, and we hope that it will be possible in the
future to make further researches upon this line of enquiry.

In conclusion, may we express our thanks to Professor A. C. Seward for his help and
advice during the course of this work, also to Professor Nathorst and other friends who
have given us information and assistance.

BIBLIOGRAPHY.

Влтнек, Е. А. (1908).--“ Nathorst's Methods of studying the cutinised portions of Fossil Plants.”
Geol. Mag., Dec. v. vol. v. p. 454.
Berry, E. W. (1911).— Maryland Geological Survey. Lower Cretaceous. Baltimore.
Bornemann, J. С. (1856).—“ Ueber organische Reste der Lettenkohlengruppe Thüringens.” Leipzig.
CHAMBERLAIN, C. J. (1912).— Two Species of Bowenia.” Bot. Gaz. vol. liv. p. 419.
Нии, T. С. (1912).—' Scharlach Е: а Microchemical Test for Oils.” New Phytologist, vol. xi. р. 72.
Karzer, К. (1908).—'' Die Verholzung der Spaltóffnungen bei Cycadeen." ^ Wiesner-Festschrift,
p. 510.
Kraus, С. (1866).—“ Ueber den Bau der Cycadeenfiedern." Prings. Jahrb. vol. iv. p. 305.
Ілехіен, О. (1894).—* Structure et Affinités du Bennettites Morierei, Sap. et Mar." Mém. Soc.
Linnéenne de Normandie, vol. xviii. Caen.
(1912).—' Stomates des Écailles interséminales chez le Bennettites Morierei 824 et Mar.).
Bull. бос. Bot. de France, vol. lix. р. 425.
Lrov», Е. E. (1908).—“ The Physiology of Stomata.” Carnegie Institute.
Mauterr, А. (1885).—'* Beiträge zur Kenntnis der Anatomie der Laubblätter der Coniferen, mit
besonderer Berücksichtigung des Spaltéffnungs-Apparates.” Bot. Centralbl. vol. xxiv. p. 54.
Narnonsr, А. С. (1880).—* Berättelse etc." Ofvers. К. Vetensk.-Akad. Férhandl. No. 5, p. 35.
(1907).-“ Paliobotanische Mitteilungen 1 © 2.” К. Svenska Vet.-Akad. Handi. Bd. xlii. No. 5.
---- (1908).—“ Palüobotanische Mitteilungen 4." К. Svenska Vet.-Akad. Handl. Bd. xliii. No. 6.
(1909).—'* Ueber die Gattung Nilssonia, Brongn.” К. Svenska Vet.-Akad. Handl. Bd. xliii.
No. 12.
(1912).—* Einige palaobotanische Untersuchungsmethoden." Paliobotanische Zeitschrift,
Bd. i. p. 26. :
Хевтьв, A. (1895). —“ Ein Beitrag zur Anatomie der Cycadeenfiedern.” Prings. Jahrb. vol. xxvii.
. 941

p. 941.
Prius, J. (1829).—Illustrations of the Geology of Yorkshire. York.

CUTICLES OF CYCADEAN FRONDS. 201

Ровзсн, О. (1905).—Der Spaltóffnungsapparat im Lichte der Phylogenie. Jena.
Ѕснемк, А. (1867).—Die fossile Flora der Grenzschichten des Keupers und Lias Frankens. Wies-
baden.
(1871).— Die fossile Flora der Nordwestdeutschen Wealdenformation. Cassel.
Ѕ$снімрЕк, W. Р. (1869).—Traité de Paléontologie végétale. Vol. i.
Scuwenpener, S. (1881).-“ Ueber Bau und Mechanik der Spaltóffnungen." Monatsber. d. Akad. d.
Wissenschaft. Berlin.
Srwarp, А. C. (1900).—The Jurassic Flora. I. The Yorkshire Coast. British Museum Catalogue.
London.
(1911!).—* Jurassic Plants from Chinese Dzungaria.” Mém. Comité Géologique, n. s., livr. lxxv.
St. Petersburg.
(19112).—* The Jurassic Flora of Sutherland." Trans. Roy. Soc. Edin, vol. xlvii. pt. ту. no. 23.
(1912).—* A petrified JVilliamsonia from Scotland.” Phil. Trans. ser. B. vol. ceii. p. 101.
Srraspureer, E. (1867).-“ Ein Beitrag zur Entwicklungsgeschichte der Spaltéffnungen.” ^ Prings.
Jahrb. v. p. 297.
Taxepa, Н. (1913!).—* Some Points in the Anatomy of the Leaf of Welwitschia mirabilis.” Annals of
Botany, vol. xxviii. p. 347.
---- (1913?).—' Development of the Stoma іп Gnetum Gnemon.” Annals of Botany, vol. xxvii. p. 365.
Tuomas, Н. Нлмвнам (1911).—“ On the Leaves of Calamites." Phil. Trans. ser. В, vol. се. p. 51.
(1912).— On some methods in Palzobotany." New Phytologist, vol. xi. p. 109.
—— (1913).—* The Fossil Flora of the Cleveland District of Yorkshire. І. The Flora of the Marske
Quarry.” Quart. Journ. Geol. Soc. vol. Іхіх. p. 223.
Wixraxp, G. В. (1011).--“ On the Williamsonian Tribe.” Amer. Journ. Sci. vol. xxxii. p. 433.
Тепле, В. (1882).-“ Observations sur quelques Cuticules fossiles." Ann. des Sci. Nat. sér. 6, Bot.
tome хш.
(1906).— Bassin houiller et permien de Blanzy et du Creusot. Fasc. 2. Flore fossile. Etudes
des gites minéraux de la France. Paris.

EXPLANATION OF PLATES 17-20.

[The figures in Plates 17-19 were drawn with the aid of a camera lucida, the magnification being
450 diameters, unless otherwise indicated. Тһе following abbreviations are used :—p, polar
subsidiary cells; /, lateral subsidiary cells; d, dorsal lignified lamelle of the guard-cells ;
v, ventral lignified lamelle ; g, guard-cells; s, stoma-opening ; $, intercalated subsidiary cells ;
а, air-space. |

PLATE 17.

Fig. 1. Stangeria paradoxa. Lower cuticle in surface view, showing a stoma with its characteristically
thickened appearance due to the lignified lamelle of the guard-cells. Note the radiating
strie at the poles of the stoma, due to folds in the cutin layer. -

Fig. 2. The same. А hair-scar from the under side of the leaf in surface view, showing the radiating
folds of the cutin.

Fig. 3. The same. Upper cuticle in surface view, showing the characteristic undulated walls of the

epidermal cells and the striations due to folds in the cutin.

4, The same. Upper cuticle in section, Note the irregular appearance of the cutin layer due to
folding.

Ez
м

Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12

7 59
оз %
> о

Fig. 8.

Fig. 9.
Fig. 10.

МВ. H. HAMSHAW THOMAS AND MISS N. BANCROFT ОМ

. The same.. Lower cuticle, showing a section through a stoma. Note the position of the

guard-cells, almost at the same level as the subsidiary cells.

. The same. Lower cuticle, showing a section through a hair-scar.
. Bowenia spectabilis, var. serrulata. Lower cuticle in surface уіеу, showing a stoma ; the

thickened appearance is due to the lignified lamellz of the guard-cells.

. The same. A stoma of the upper side of the leaf in surface view. Note the presence in this

case of six subsidiary cells, two of which are polars and four laterals.

. The same. Lower cuticle, showing a section through a stoma. Note the position of the

guard-cells with regard to the subsidiary cells.

. Cycas revoluta. Upper cuticle in section, showing the pittings of the lateral and outer walls

of the epidermal cells.

. The same. Upper cuticle in surface view, showing the pittings of the lateral and outer walls

of the epidermal cells.

. The same. Lower cuticle in surface view, showing the fifteen subsidiary cells surrounding a

stoma-opening. The shading indicates the raising of the subsidiary cells beyond the general
epidermal level (ef. fig. 13).

The same. Lower cuticle, showing a section through a stoma. Note the presence of an inter-
calated series of cells between the guard and subsidiary cells, and the large air-space formed
by the sinking of the guard-cells and the overarching of the subsidiary cells.

. Cycas circinalis. Lower cuticle, showing a section through a stoma. Note the sinking of the

guard-cells and the absence of an intercalated series of cells (cf. figs. 5, 9, & 13).

. The same. Upper cuticle in section, showing pitting of the outer walls of the cells.

PLATE 18.

. Ceratozamia mexicana. Upper cuticle in surface view, showing thin-walled and thick-walled

cells. Note the layering and pitting of the thickened walls.

. The same. Lower cuticle in surface view, showing a stoma-opening surrounded by six sub-

sidiary cells, two of which are polars and four laterals. The cell-walls are pitted and layered
as in fig. 1.

. The same. Upper cuticle, showing the thick- and thin-walled cells in section.
. The same. Lower cuticle: a section through a stoma, showing one intercalated series of cells

between the guard-cells and subsidiary cells.

. Dioon edule. Upper cuticle in surface view, showing the variation in thickness and the layering

of the cell-walls.

. The same. A section of astoma. Note the presence of three intercalated series of cells (i'-*’)

between the guard-cells and the subsidiary cells; also the dorsal and ventral lignified lamellæ
of the guard-cells.

. The same. Upper cuticle in section, showing variation in the thickness of the walls. Note

the heavy cutin layer.
The same. A section of the lower cuticle, showing a crystal-containing space between two
epidermal cells. The space is lined and covered by cutin.
The same. The lower cuticle, showing a crystal-containing space in surface view.
Encephalartos Lehmanni. Upper cuticle in section, showing pitting of the lateral walls.

Fig. 11. Е. villosus. Upper cuticle in surface view, showing slight pitting of the lateral wails of the

cells and septation of an elongated cell at г.

Fig. 12. Zamia integrifolia. Upper cuticle in surface view, showing variation in the thickness of the

_ walls and occasional pitting of the thin-walled cells.

CUTICLES ОЕ CYCADEAN. FRONDS. 203

Fig. 18. Zamia muricata. Lower cuticle, showing a stoma in section. Note the slightly sunk
guard-cells.

Fig. 14. Macrozamia cylindrica. Section through a stoma, showing two intercalated series of cells and
the long narrow air-space.

Fig. 15. The same. Upper cuticle in surface view, showing variation in the thickness of the cell-walls,
also pitting and layering of the walls.

PLATE 19.
[Camera lucida drawings of stomata of fossil fronds. |

Fig. 1. Ptilophyllum pecten. Showing the different thickenings of the guard-cells and the thin places
above the poles. Specimen from Marske. х 600.

Fig. 2. Zamites sp. Showing outlines of stomatal thickenings. From scale-leaves on stem. Specimen
in the Naturhistoriska Riksmuseum, Stockholm. x 500.

Fig. 3. Otozamites Feistmanteli. Stoma with differences of thickening less marked. The subsidiary cells
are not thickened. Specimen from Marske. х 500

Fig. 4. Otozamites graphicus. Тһе sides of the pore appear to have been slightly cuticularised.
Specimen from Marske. х 500.

Fig. 5. Anomozamites Nilssoni. Showing thin subsidiary cells and thin places at the poles. Specimen
from Gristhorpe. x 500.

Figs. 6-8. Dictyozamites Нашей. Subsidiary cells thickened: 6 and 8 show projections from upper
thickenings of guard-cells. Fig. 7. Stoma seen from mesophyll side. Specimens from
Marske. x 500

Fig. 9. Teniopteris major. Compare with T. vittata. Specimen from Gristhorpe. х 500.

Figs. 10-12. Teniopteris vittata. Showing the similarity with stomata of other Bennettitalean fronds.
Fig. 10 is seen from the mesophyll side. 10 and 11 from Gristhorpe, 12 from Marske.
x 500.

Figs. 13, 14, 18. Nilssonia sp. From aspecimen found at Roseberry Topping. х 500. Note in 18 the
series of overarching subsidiary cells. Figs. 14 and 18 show the guard-cells as seen from
the mesophyll side of the epidermis. -

Fig. 15. Nilssonia mediana. Subsidiary cells with thickened tips overarching the guard-cells.
Specimen from Marske. x 500

Figs. 16, 17. Ptilozamites Leckenbyi. Stomata as seen from upper (17) and lower side (16). In the
former the guard-cells can just be seen through the opening. Specimens from Gristhorpe.
x 500.

Fig. 19. Ctenis falcata. Showing thick wall of жыш round the stomatal cavity. Specimen from
Gristhorpe. х 500.

PLATE 20.
[Photomicrographs of cuticular preparations of fossil fronds. |

Fig. 1. Ptilophyllum pecten. Showing upper epidermis (to the right) and lower epidermis with the
stomata arranged in regular rows. From Marske. x 25.

Fig. 2. Ptilophyllum pecten. Upper epidermis more highly magnified, showing sinuous outlines of the
cells. x 170.

Fig. 3. у pecten. Part of lower epidermis more highly magnified, showing stomata.

Fig. 4. Pity sp. Showing rows of сода, hair-scars with epidermal cells below. From Marske.
x 200.

SECOND SERIES.— BOTANY, VOL. VIII. 2H

Fig. 12.

Fig. 13.

ON CUTICLES OF CYCADEAN FRONDS.

. Teniopteris vittata. Part of upper epidermis, showing characteristic folded cell-walls. From

Gristhorpe. х 200. `

. Teniopteris vittata. Part of lower epidermis, showing stomata which were slightly sunken, and

hence the walls of the epidermal cells are out of focus. x 200.

. Dictyozamites Нашей. Lower epidermis, showing stomata. The characteristic form of the

upper thickening lamellz is seen іп a few cases. From Marske. х 180.

. Anomozamites Nilssoni. Lower epidermis with stomata. From Gristhorpe. х 200.
‚ Nilssonia sp. From Roseberry Topping. Upper epidermis, showing characteristic papille.

x 180.

. Nilssonia mediana. Lower epidermis, showing elongated cells with papille occurring below

the veins, while the cells of the intermediate areas are rounded. From Marske. х 200.

‚ Nilssonia orientalis. Portion of upper epidermis from specimen No. 323 in the Natur-

historiska Riksmuseum, Stockholm. Found at Cloughton Муке. х 200. ‘his specimen
has been figured as Nilssoniopteris tenuinervis.

Ptilozamites Leckenbyi. Part of lower epidermis, showing straight-walled cells and stomata.
From Gristhorpe. х 200.

Ctenis falcata. Part of lower epidermis with stomata. From Gristhorpe. x 200.

Tuomas % BANCROFT.

TRANS. LINN. SOC. SER. 2. BOT. VOL. ҮШІ. P1.17.

CYCADEAN CUTICLES.

TRANS. LINN. SOC. SER. 2. BOT. VOL. VIII. Pl. 18.

Тномав & BANCROFT.

CYCADEAN CUTICLES.

Тномав & Ваксковт. TRANS. LINN. SOC. SER. 2. BOT. VOL. VIII. РІ. 19.

М

H. Н.Т. del. Grout, sc.
CYCADEAN CUTICLES.

Tuomas & Bancrorr. TRANS. LINN. SOC. SER. 2. BOT. VOL. VIII. PL 90.

Grout, se.
H.H.T. phot. rout, вс.

CYCADEAN CUTICLES.

LINNEAN SOCIETY OF LONDON.

MEMORANDA CONCERNING TRANSACTIONS. D

‘The First Series of the Transactions, containing both Botanical and Zoological contributions, has been completed —
- in 30 Vols, and a few entire sets are still for sale. Only certain single volumes, or parts 8 to complete д máy `
X be obtained at the original prices. The price of the Index to Vols. 1-25 is 8s. to the public, and бз. to Fellows; 0 =
сы 26-30, 4s. to the püblie, and 3s. to
e The Second Series of the Transactions is divided into Zoologie
zt “Botanical | pares ts of these which hare been ae are as undermentioned.
2.

al and Botanical sections. The prices of
(For the Zoological parts see Zoological

Secon SERIES.— BoTANY (continued).
eu When Price to the
Volume, — - — published. Public.

Ё s.

NI. Part E 1901. .... 1 10
Pat 1. 1901 .... 0 6
“Part IL 1002... .. 0 18
= RE ТҮ. 19083..... 0 8
Pat У. 1903. .... 0.8
Part УТ. 1903, .... 0 12
- Part VIL 1904. ..— 01%

2nd Ser, BOTANY.) (VOL. VIII. PART 6, 27

THE

TRANSACTIONS

OF

ГНЕ LINNEAN SOCIETY OF LONDON.

е URES а

_AN ANATOMICAL STUDY OF THE PALEOZOIC -
00% E-GEN US LEPIDOSTROBUS.-

AGNES ARBER, DSc, ELS,

: FELLOW OF NEWNHAM COLLEGE, ‘CAMBRIDGE.

ҮП. An Anatomical Study of the Paleozoie Cone-genus Lepidostrobus.
By AGNES ARBER, D.Sc., F.L.S., Fellow of Newnham College, Cambridge.

(Plates 21-27, and 4 Figures in the Text.)

Read 19th June, 1913.

CONTENTS.

Page
а en 205
2, А Survey of our Present Knowledge of the Structure of Lepidostrobus...... 206

3. On certain hitherto Unrecorded Features in the Structure of Lepidostrobus
oldhamius, Will., and Г. foliaceus, Maslen .......................... 208
4. A Description of some New Forms of Lepidostrobus .................... 212
(i) 2.-Втнеуалыу Sp. бу... 5 eee ehe erae es 212
(1. D. gracilis, ертті: ЕЕ, ae Бин eee 217
Gu.) L. oldhamius, WilL,f. minor ...................... ......... 920
(iv.) L. oldhamius, Will, f. pilosus .. ............................. 223

5. On a Case of Variation in the Anatomical Characters of Different Regions

ін а Singlo Comes Ое 228

6. On the Relation between Structure and Function in the Cone-genus
Lepidostrobus ...........2...................................... 230
Я. List of Memoirs edad ilas е ане eee oe еее Hie ee tse ore oe b> 232
8. Explanation of the Plates ........................................ 233

1. INTRODUCTION.

THE collection of fossil plants in the Sedgwick Museum, Cambridge, includes sections
of certain specimens of Lepidostrobus, which differ considerably from those members of
this cone-genus whose structure has been previously described. The present memoir is
based upon a study of these sections, and also of a large number of preparations from
other collections. In this connection I would express my great indebtedness to those
botanists who have kindly lent me sections of these cones and allowed me то
describe them,— Dr. W. J. Jongmans, Professor Е. W. Oliver, F.R.S., Dr. D. H. Scott,
F.R.S., Miss Winifred Smith, Mr. D. M. 8. Watson, and Professor Weiss. То Dr. Scott,
I must, in addition, offer my thanks for the help and information he has kindly given
me. I have also to thank Dr. A. Smith Woodward, F.R.S., the Keeper of the Department
of Geology, British Museum (Natural History), for allowing me opportunities of
studying the specimens under his charge. Certain sections, which have been of
use in this investigation, have been purehased by aid of а grant from the British
Association, for which I am much indebted.
SECOND SERIES.—BOTANY, VOL. VIII. 21

206 DR. AGNES ARBER ON THE

Finally, I wish to express my gratitude to Professor Т. МеКеппу Hughes, F.R.S.,
who has given me every facility for carrying out this work in the Sedgwick Museum.

9. А SURVEY OF OUR PRESENT KNOWLEDGE OF THE STRUCTURE
OF JLEPIDOSTROBUS.

The general characters of the cone-genus Lepidostrobus are too familiar to need
recapitulation here. In all essentials this genus is a remarkably uniform one. We are
fortunate in having a fairly complete knowledge of several species which have been
found in the petrified state. and 1% may be useful, before dealing with the observations
contained in the later sections of the present paper, to refer very briefly to these cones
and their structure. |

One of the first Lepidostrobi of which the internal structure became known, was
L. Brownii, (Unger) Schpr., a large cone from the Lower Carboniferous rocks, which
was originally described by Robert Brown *. We have recently been put into possession
of detailed knowledge of this species, and of one or two other cones closely related to it,
by means of Zeiller’s + exhaustive and beautifully illustrated monograph. The strobilus
has several peculiar features, the most striking of which are the following: the
lamina of the sporophyll terminates in a swollen cushion, and trabecule of sterile
tissue arise from the floor of the sporangium and penetrate among the spores. No
megaspores have been discovered.

Another Lower Carboniferous cone whose structure is preserved is Lepidostrobus
Veltheimianus, Scott, a heterosporous form from the Calciferous Sandstone of Burnt-
island, first described by Williamson £. This is a small and slender cone, similar in
general structure to the familiar L. oldhamius of the Coal Measures. А delicate radial
plate of sterile tissue arises from the floor of the microsporangium, and penetrates
some distance into the sporangial cavity.

From the Calciferous Sandstone we have also another heterosporous cone, Lepido-
strobus Wünschianus. Тһе structure of this strobilus was described and figured by
Binney $, not, however, in microscopie detail.

The best known Lepidostrobus, which occurs as а petrifaction in the Upper
Carboniferous rocks, is the Coal Measure form L. oldhamius, originally described by
‚ Williamson |. Binney Т, at an earlier date, had described and figured, under the name
of Lepidodendron Harcourtii, sections from the identical specimen upon which
Williamson afterwards based the species oldhamius. Binney also described other
sections of the same type of cone, attributing them to Г. vasculare. More recently
Maslen ** has redescribed Lepidostrobus oldhamius in detail, with special reference to
Williamson’s slides. The cone of this species is larger than that of L. Veltheimianus,
but smaller than that of Г. Brownii. No megasporangia have been up till now

* Brown, К. (48) and (51). For another early account of some members of the genus see Hooker, J. D. (’48)-
T ?еШег, В. (11). Williamson, W. C. (73), p. 294 and pl. 44. $ Binney, Е. W. (71), pl. 11.
|| Williamson, W. С. (94), pp. 27 & 28, pl. 9. fig. 58, pl. 6. figs. 61 & 62. :

Ч Binney, E. W. (71), р. 46 ete., and pls. 7 & 8. жж Maslen, A. J. (99).

PALZOZOIC CONE-GENUS LEPIDOSTROBUS. 207

described. Тһе sporophyll has a thin, leafy lamina, quite different from the cushion-
like body which terminates the sporophyll lobe of L. Brown. The distal end of the
sporangium is peculiar in arching over and concealing the ligule. Тһе latter organ,
which was first discovered by Maslen* in this species, now appears to be present
universally in connexion with the sporophylls of Lepidostrobus. Maslen 7 distinguishes
within the species Г. oldhamius three forms, which he calls (а), (В), and (y). Тһе
chief differences between these types lie in the anatomical structure of the axis.
Maslen's form (8) represents the species as originally described by Williamson.

Coulter and Land { have described a petrified cone from the Coal Measures of Iowa
which appears to belong to the type of L. oldhamius.

The structure of a large cone from the Upper Carboniferous (Westphalian) of the
Donetz basin has recently been discussed in great detail by Zalessky $, who has named
it Lepidostrobus Bertrandi. This fructification does not differ in any essential ana-
tomical feature from L. oldhamius, but, at the same time, there is no doubt that it
represents a distinct species.

Homosporous or microsporous cones of the type of L. oldhamius are not, however, the
only forms of Lepidostrobus met with in the Upper Carboniferous rocks. Williamson |
gives a figure of a section of a slender cone from the Coal Measures which differs
markedly from both L. oldhamius and L. Bertrandi. In Williamson's text this
section is dismissed with a brief mention, and we had no detailed information about
it until it was redescribed by Maslen 9 under the name of L. foliaceus. The form
of the sporophyll and the histology of the lamina are peculiar, and make it easy
to recognise the species аба glance. Williamson’s section contained only microspores,
but our knowledge of the cone has now been rendered much more complete by
Mrs. Scott's ** discovery that certain megaspores with peculiar appendages, with which
she had been acquainted for some time as isolated objects, were borne in sporangia which
belonged to L. foliaceus. This species thus ranks as a heterosporous Lepidostrobus.

Little is known with certainty on the subject of the attribution of these various cones
to the species of Lepidodendron which bore them. Williamson Tft pointed out that it
was extremely probable that the heterosporous cone, which he described from Burnt-
island, was the fructification of Lepidodendron Veltheimianum, and Scott ff has contirmed
this attribution and employed the name Lepidostrobus Veltheimianus for the cones. In
Lepidostrobus Wiinschianus, another heterosporous cone from the same horizon found
in Arran, we appear to have a second instance in which the structure both of the stem
and the strobilus are known §§. These two species seem, however, to be the only
instances in which we are definitely acquainted with the internal structure both
of the fructification and of the tree which bore it.

In certain cases it is possible to identify cones whose internal structure is preserved
+ Maslen, А. J. (99). + Coulter, J. M., & Land, W. J. G. (711).
| Williamson, W. С. (94), р. 27, aud pl. 9. fig. 57.

** Scott, К. (06).
$1 Scott, D. П. (00), p. 171.

* Maslen, A. J. (98).
$ Zalessky, M. (08).
Ч Maslen, А. J. (99), р. 373, and pl. 38. figs. 36-39.

tt Williamson, W. С. (73), p. 94.

$8 Binney, E. W. (71), p. 56, pl. 11, and Williamson (94), p. 29. а

| I

208 DR. AGNES ARBER ON THE

with others which we know only as impressions. It is probable, for instance, that
Lepidostrobus oldhamius may be identical with the cone-impressions known as
L. variabilis, Lindl. & Hutt., whicb. аге so common in the Coal Measures. Generally,
however, it is extremely difficult to correlate “impressions” and “ petrifactions” of
cones. Several reasons for this difficulty readily suggest themselves. The features,
about which we obtain most information from petrified material, are the anatomy of the
axis and of the sporophylls, and the internal structure of the sporangium. The general
shape and dimensions of such cones are often imperfectly known, and the external form
of the sporophylls and sporangia is frequently difficult to reconstruct. In “impressions ”
of cones, on the other hand, it is just these external features about which information
is easiest to obtain, while such material gives little clue to those characters which we
know in the greatest detail from petrified fructifications.

Another cause of difficulty in correlation appears to be that the cones of these
species, which lend themselves most readily to entire or partial survival as impressions,
are often ill-adapted to preservation as petrifactions. "We occasionally find, in the Coal
Measures, impressions of extremely large, detached sporophylls of Lepidostrobus. There
is, for example, a specimen in the Sedgwick Museum * in which the lamina reaches the
length of 8:9 cms. It seems obvious that a cone protected by the overlapping lobes of
such sporophylls would prove highly resistant to the infiltration of any petrifying
material This may perhaps explain why it is that none of the cones with well-preserved
internal structure which have been described from the Coal Measures, possess sporo-
phylls approaching these dimensions.

In a later section of this paper some account is given of certain forms of Lepidostrobus
which do not appear to have been hitherto described. In none of these cases is it at
present possible to correlate the petrified cone with the stem which bore it, or with
impressions showing surface features.

3. ON CERTAIN HITHERTO UNRECORDED FEATURES IN THR STRUCTURE OF
-LEPIDOSTROBUS OLDHAMIUS, WILL., AND L. FOLIACEUS, MASLEN.

(i) The Existence of a Sterile Plate in the Sporangium of L. oLpHAMIUS and
L. FOLIACEUS.

The type sections of Williamson’s species, Г. oldhamius, consist, as Maslen +
has pointed out, of two slides now in the Williamson Collection, British Museum
(C.N. 568 & 574); in addition there are certain slides in the Binney Collection,
Sedgwick Museum, Cambridge, which were cut from the same cone. Sections of
another cone in the Williamson Collection, known as “ Wild’s Cone” (C.N. 1776 А & В),
are also attributed by Maslen to the type form. The preparations of the latter cone are

* Lepidophyllum majus, Brongn, From the Upper Coal Measures, Coal Pit Heath Colliery, near Bristol.
Sedgwick Museum Coll., 2076. |
+ Maslen, А. J. (99), р. 359.

PALZOZOIC CONE-GENUS LEPIDOSTROBUS. 209

particularly valuable, on aecount of the good preservation of the structures external to
the axis.

When examining Wild's Cone, I was struck by a feature which has not been recorded
as occurring in this species—the presence, namely, of a delicate radial plate of sterile
tissue arising from the floor of the sporangium and appearing in tangential sections
merely as a streak * (Text-fig. 1). It resembles the outgrowth figured by Williamson f in

Text-fig. 1.

Radial section of а sporangium and sporophyll of Lepidostrobus oldhamius, Will., * Wild's Cone,” Williamson
Collection, Brit. Mus. (Nat. Hist.) C.N. 1776 В. ( x 32 circa.)
(l.p. = sporophyll pedicel; 1.1. = sporophyll trace ; spm.w. = sporangium wall; зёр. = plate of sterile tissue
arising from the floor of the sporangium.)

the case of L. Veltheimianus, except that it does not, as a rule, bifurcate, although it
may occasionally branch at the apex. It happens that, in the case of Williamson and
Binney's original cone, there are scarcely any good tangential sections passing through
à sporangium attached to its sporophyll; so we cannot get a view of the sterile plate
comparable with that obtained in the preparation of Wild's Cone. Іп one of the
sections of the original cone in the Binney Collection f, however, two of the sporangia
are саб with great exactness in the radial plane, and in both these cases there are
distinct indications of the presence of a plate of sterile tissue arising from the floor of
the sporangium and lying approximately in the plane of section. с

That this plate of sterile tissue within the sporangium is actually characteristic of the
type form of Г. oldhamius is confirmed by certain sections in Dr. Scott’s Collection.
А sporophyl with a sporangium from one of these § is shown in РІ. 27. fig. 55.

* Other sections of Wild’s Cone are preserved in the Manchester University Collection, and here the sterile
plates can also be observed ; е. g., В. 384 5.

t Williamson, W. С. (73), р. 295 & pl. 44. figs. 23, 24, 25.

2 Binney Coll., No. 87. § Scott Coll., 2456.

210 DR. AGNES ARBER ON THE

In this case the sterile upgrowth (sé. р.) is usually well preserved and can be seen to
consist of delicate cellular tissue. The sterile plate is still better displayed in certain
sections of sporangia near the apex of a cone of L. oldhamius which occur in a
preparation in the University College Collection * (РІ, 24. fig. 17). The most remarkable
instances of sterile plates with which I am acquainted, however, are to be found ina
series of obliquely transverse sections of a cone from Shore, Littleborough, preserved
in the Leiden Herbarium (Slides 211-214). A single section, which obviously, from
internal evidence, belongs to the same series and must have been cut from this identical
cone, has been presented by Miss Winifred Smith to the Sedgwick Museum, Cambridge $.
From a study of the latter section alone, I originally drew the conclusion that this
cone represented a new species, to which, in a preliminary note i, I gave the name
Lepidostrobus laminatus. Ат examination of the Leiden slides, whieh I have recently
had the opportunity of making, through the kindness of Dr. Jongmans, has сопуіпсей
me that the cone in question is not specifically different from L. oldhamius ; the name
L.laminatus must hence be discarded. Тһе sections of the cone are useful because
they pass obliquely through the apex and display the internal structure of immature
sporangia with great clearness. The sterile plates in these young spore-sacs are massive
and well preserved, and give rise with great regularity to two smaller lateral processes,
one on either side (st.p., Pl. 21. figs. 1 & 3 and Pl. 24. figs. 15 & 16). Тһе branches are not
directly opposite to one another, and the main process is continued above its branches
for a considerable distance. That these outgrowths are of the nature of plates, running
in the direction of the long axis of the sporangium, and are not merely peg-like
structures, is evidenced by the fact that they present a general similarity of pur.
in the seven sporangia in which they are visible in the section represented in РІ. 21. fig. 1
These sporangia, owing to the fortunate obliquity of the section, are cut at varying
distances from the main axis, and we are hence enabled to visualise the structure of
a complete individual sporangium. Another point, which is established by the comparison
of the different sporangia, is that the sterile plate died out towards the distal end of the
spore-sac. The lowest sporangium figured in РІ, 24, fig. 15, which is also shown photo-
graphically in Pl. 21. fig. 2, is obviously cut through near the distal end, since the ligule
(1д.) is visible. In this case there is no sterile upgrowth from the “ sub-archesporial
pad” ў. "һе structure of one of the sterile processes is shown in detail in Pl. 24. fig. 16.
In the older sporangia the sterile plates are relatively less рова the lateral branch-
plates seem to shrivel and disappear quite early.

The type section of Lepidostrobus foliaceus, Maslen (Williamson Coll., C.N. 1614), is
approximately radial, but more or less tangential in the middle region. In at least two
sporangia there are indications of a sterile plate similar to that just described in

* U.C.L. Coll., С. 19,

+ Slide 281, Sedgwick Muscum Coll. This section probably belongs between Slides 213 and 214 of the Leiden
Herbarium Coll.

+ Arber, A. (13). $ Bower, Е. О. (94), р. 513.

PALJEOZOIC CONE-GENUS LEPIDOSTROBUS. 211

Т, oldhamius. The same feature is also observable in certain sections of the species
in the Manchester University Collection *.

The structure which we have been considering is not confined to L. oldhamius and
L. foliaceus, but is found, in addition, in the new forms of Lepidostrobus which will
be described later in the present paper. It has also been previously recorded in
L. Veltheimianus t and in an unnamed species $ from the Coal Measures; while sterile
upgrowths, taking the form of pegs rather than plates, are known in L. Browni $.
I am һепсе inclined to think that the occurrence of some form of sterile upgrowth
from the sporangial floor may eventually prove to be characteristic of all forms of
Lepidostrobus || which are microsporous or homosporous ; such a structure does not seem
at present to have been recognised in any megasporangium. The fact that these sterile
plates have been hitherto overlooked in so well-known a type as L. oldhamius is no doubt
due to the cireumstance that the delicate tissue of which they are composed is liable to
shrivel beyond recognition as the sporangia ripen.

(iL) The Occurrence of Periderm in the Sporophylls of LEPIDOSTROBUS
OLDHAMIUS, Will.

In Williamson's original transverse section of Lepidostrobus oldhamius Ч there are
distinct indications of a radial arrangement of the cells of the sporophyll lamina towards
the dorsal surface. This grouping is sufficiently regular to suggest that periderm is in
process of formation. Тһе same appearance of cork at the dorsal surface of the sporo-
phyll lobe occurs in various sections of cones belonging to L. oldhamius in Dr. Scott's
Collection, e. 0. 8.1900, a part of a sporophyll from which is shown in transverse section
in РІ. 27. fig. 56. |

It seems probable that such cork formation was widely prevalent in the sporophylls of
Lepidostrobus, since it also occurs in three of the forms described in the present paper—
L. Binneyanus, sp. nov., L. gracilis, sp. nov., and L. oldhamius f. pilosus (cf. Pl. 94.
fig. 20, Pl. 25. figs. 29 & 36, and Text-fig. 3, В & С). As will be shown on p. 227, the
periderm in Z. oldhamius f. pilosus appears to have been hypodermal in origin, and it is
not improbable that this was also true in the case of other species. РІ. 25. fig. 36
represents a transverse section through the dorsal surface of a sporophyll of L. gracilis.
Неге the cork has quite the appearance of being hypod 1. In L. Binneyanus also,
where the periderm seems at first glance to be epidermal (РІ. 25. fig. 29), it is possible
that the true epidermis may have been discarded. i

* Manchester University Coll., 9.452; Q. 456; Q. 457.

t Williamson, W. C. (73), p. 295 & pl. 44. figs. 23, 24, 25.

t Bower, F. О. (94), р. 529 & pl. 48. figs. 101 & 102.

$ Brown, R. (51), p. 471; Bower, F. О. (94), p. 597 & pl. 48. figs. 96-100 ; Zeiller, R. (711), p. 36.
| A similar conclusion seems to have been reached by Bower, F. О. (08), p. 325.

є Williamson Coll. Brit. Mus. (Nat. Hist.), C.N. 568.

212 DR. AGNES ARBER ON THE

4. А DESCRIPTION OF SOME NEw FORMS OF LEPIDOSTROBUS.

(1.) LEPIDOSTROBUS BINNEYANUS, sp. nov.
(РІ. 21. fig. 4; РІ. 22. figs. 5-8; РІ. 24. figs. 18-25; РІ. 25. figs. 26-30.)

(1) Material.

The Binney Collection, Sedgwick Museum, Cambridge, included, amongst its Lower
Coal Measure blocks, one which contained, in addition to other fossils, two cones of a
Lepidostrobus showing certain peculiar features. A series of sections has now been
cut from each of the cones*. The following account is based upon these two fructi-
fications, which may be distinguished as Q.a and 9. 8. They appear to belong to
species hitherto undescribed, for which I propose the name Lepidostrobus Binneyanus.

(2) General Form and Dimensions of the Cone and Sporophylis.

We can only arrive at very rough measurements for the cone, owing to the fact that
both specimens are fragmentary and much crushed. ‘The fructification was slender,
with a diameter of about 1:5 cm., of which the axis occupied 3 mm. The approximate
measurements of the upturned limb of the sporophyll are 6 mm. in length and 5 mm. in
breadth at the base. It should be remembered, however, that, owing to imperfections
of preservation and to distortion of the more delicate parts of the cone, measurements
of the length of the upturned limb of the sporophyll of any Zepidostrobus, taken
from sections of petrified material, are relatively of little value. They are likely to fall
considerably below the actual dimensions, and should merely be accepted as minimum
measurements. In order to obtain the full measurement, the sporophyll-limb would
not only need to be perfectly preserved and uninjured, but the section would require
to traverse the whole length of the limb in the median plane, an event which would
rarely happen.

In Lepidostrobus Binneyanus there was a downward heel at the junction of limb and
pedicel, which appears to have sometimes reached a length of about 3 mm. (l.h., РІ. 24.
fig. 20). Тһе basai part of the sporophyll lamina was somewhat thick, measuring not
less than 1 mm. from dorsal to ventral surface. The sporophyll pedicel, which was
5 to 6 mm. long, was slightly winged (p.w., Pl. 22. fig.7 and Pl. 25. figs. 26 & 26 a), but,
owing to the way in which the cones have been crushed, it is exceptionally difficult to
determine its exact width. The midrib appears to have been 0:2 to 0:3 mm. wide, while
the wing extended on either side for at ieast 1 mm. |

(3) The Detailed Structure of the Cone Axis.

The slender xylem cylinder of the cone axis is “5 mm. (or less) in diameter, and is
almost solid (v.c., Pl. 21. fig. 4). There are, however, a few thin-walled pith elements
(p.) in a central position. These are obliterated by crushing in the transverse section

* Q. a. (1-11) and Q. 6. (1-5).

PALZOZOIC CONE-GENUS LEPIDOSTROBUS. 213

photographed in РІ. 21. fig. 4, but their position can be made out. They are visible in
longitudinal section in Pl. 22. fig. 5. The number of tracheides on any radius between
the pith and the protoxylem (рх.) is about four. The contrast in size between the
scalariform metaxylem (mz.) and the spiral protoxylem elements is unusually great
(РІ. 21. fig. 4 and РІ. 24. figs. 18 & 19). Тһе horizontal bars of thickening on the walls
of the metaxylem tracheides are connected by delicate vertical striations (РІ. 24. fig. 18).
There seems to be a complete or almost complete absence of the short barred tracheides
characteristic of the peripheral zone of the xylem in L. oldhamius, Will. (a)*. The
phloem is not preserved. Тһе firm inner cortex (i.c. Pl. 21. fig. 4) is four to five
elements in thickness. As is usually the case in Lepidostrobus, the middle cortex has
perished, but the outer cortex survives.

(4) The Sporophyll Traces in the Axis.

We only possess a single radial longitudinal section of L. Binneyanus T, and this is
somewhat imperfect. Тһе sporophyll traces, as shown in this section, were peculiar in
following an obliquely downward path in the outer part of their course from central
cylinder to sporophyll (/.7., Pl. 24. fig. 21).

РІ. 24. fig. 22 shows a sporophyll trace, as it appears in transverse section while
traversing the space immediately outside the central cylinder. The xylem, which is
alone preserved, consists of a compact group of seventeen tracheides ; there are two or
three small elements oceupying a central position which possibly may be the protoxylem.
As the sporophyll traces pass through the inner cortex, they each become invested with a
sclerised sheath, which accompanies them on their outward journey (i.c.s., Pl. 21. fig. 4
and Pl. 24. fig. 23). Asa rule, the xylem alone is preserved, but occasionally a small
detached group of elements is also visible on the outer side of the xylem; this may
perhaps represent the phloem. The xylem group, which is roughly circular while the
trace is still close to the vascular cylinder (РІ. 24. fig. 22), becomes narrowed radially
and widened tangentially as it passes further out (РІ. 24. fig. 23). The sporophyll traces
in the middle cortical sheath are too much distorted for measurement, but before they
become free from the inner cortex their tangential width (including the sheath) is about
0°15 mm.

(5) The Detailed Structure of the Sporophyll.

The dimensions and general form of the sporophyll have already been briefly mentioned.
A noticeable feature of the winged sporophyll pedicel, when viewed in tangential section,
is the occurrence of a cushion of sclerised elements on either side of the median attachment
of the sporangium. Many of the cells of the cushion are so thick-walled that their
cavities are almost or entirely obliterated (sc.c., Pl. 22. figs. 7 & 8 and РІ. 25. figs. 26, 26 а,
& 27). On the lower side of the pedicel there isa well-marked keel, of which the histology
is somewhat unusual. The epidermal layer (ер.) of small cells is succeeded internally bya
hypoderma (Ay.) of large cells with firm walls. These are markedly distinct from the

ж Maslen, A. J. (99), p. 362. T Q.a. 6.

913
SECOND SERIES.—BOTANY, VOL. VITI. 4K

214 DR. AGNES ARBER ON THE

thin-walled cells forming the inner tissue of the keel, which are often much crushed.
The epidermis and also the hypoderma seem to consist, in places, of more than one layer
of cells; but in spite of this irregularity, the general appearance, as seen in tangential
sections, is highly characteristic. It contrasts markedly, for instance, with the large-
celled epidermis and absence of hypoderma in L. oldhamius, Will, f. minor (of. Pl. 99,
figs. 7 & 8 with Pl. 26. fig. 40).

The upturned lamina of the sporophyll consists, especially in its basal region, of thin-
walled parenchymatous tissue (7., Pl. 22. fig. 6), but sclerenchyma is not altogether absent,
and increases in amount higher in thelobe. Тһе appearance of the lamina suggests that
it was slightly succulent. Тһе cells towards the dorsal surface are relatively small, and
their radial arrangement, as well as the angular form which they assume in cases where
the tissue has become crushed or shrunken, indicate that periderm formation has come
into play (pd., РІ. 22. fig. 6, Pl. 24. fig. 20, and РІ. 25. fig. 29). There is occasionally a
slight appearance of a similar formation at the ventral surface of the sporophyll lamina.

An interesting feature is the occurrence of delicate cellular outgrowths (/.o., Pl. 22.
fig. 6 and РІ. 25. fig. 80), which appear to arise from the inner surface of the downward
heel (or dorsal flap) which occurs at the base of the sporophyll lamina. Unfortunately
the preservation is not sufficiently perfect to make it possible to give an exact account
of the nature and arrangement of these structures. However, from a comparison of
the two sections in which they can be distinguished, I am inclined to think that they
consisted of cell-plates, two or three elements in thickness.

The anatomy of the sporophyll trace during its passage through the pedicel is
illustrated іп Pl. 22. fig. 8 and РІ. 95. fig. 28. Тһе small compact xylem group,
consisting, in three cases in which the elements were counted, of less than a dozen
tracheides, is surrounded by large transfusion elements, which are especially abundant
on the upper side of the bundle. These transfusion tracheides are very well developed
at the junction of lobe and pedicel. In the upturned lobe, the bundle presents the
same general appearance as in the pedicel, but the tracheides are fewer in number.

(6) The Ligule.

A ligular pit is visible in more than one case, and the body marked “? lg.” in РІ. 24.
fig. 20, may perhaps represent the ligule itself,

(7) The Sporangium.

The sporangium has the radially elongated attachment to the pedicel, which is one of
the characteristics of the genus Lepidostrobus. Тһе distal and proximal ends of the
sporangial base are represented in Pl. 24. figs. 20 & 21. The attachment is of
considerable breadth, and à massive ridge of tissue is continued up into the base of the
sporangial cavity (84.ғ., Pl. 22. fig. 7, РІ. 24. fig. 21, and РІ. 25. figs. 26 & 27). This
may perhaps be interpreted as of the nature of sterilised sporangial tissue, since it 18
bounded on either side by a continuation of the sporangium wall*, but, at the same
time, it is perfectly continuous with the parenchyma of the sporophyll.

* Maslen draws attention to the same feature in the case of L. oldhamius (В) and L. foliaceus. Maslen, А. 2.
(99), рр. 370 & 374, pl. 38. figs. 32 6 38

PALZOZOIC CONE-GENUS LEPIDOSTROBUS. 215

The cells, which compose the sterile ridge, are often somewhat elongated in a vertical
direction. The most interesting feature of the tissue is that, in some cases, the walls of
these elongated cells are marked by delicate bars of thickening, indicating that they
were transfusion elements (РІ. 25. fig. 27). The distribution of these elements seems to
be very irregular. In certain tangential sections none can be detected (e. g., the one
represented in РІ. 25. fig. 26a), in others there are a few, while in others again (e. g.,
Pl. 25. fig. 27), the tissue of the ridge appears to be almost entirely composed of such
elements, which are continuous below with the transfusion tracheides of the sporophyll
pedicel.

The sporangia, as seen in tangential section, are much crushed and distorted,
dehiscence having already taken place. Тһе structures inside the sporangium are
hence difficult to interpret, but there are indications of the existence of an outgrowth of
cellular tissue arising from the sterile ridge and penetrating into the cavity of the
sporangium (s£.p., РІ. 25. Во. 27).

The sporangium wall is of the usual columnar type (РІ. 25. figs. 26 a, 37, & 38), lined
internally by a narrow zone of delicate thin-walled tissue which is poorly preserved.

(8) The Spores.

The spores are somewhat more mature іп Q. В. than in ©. я. In the former cone they
are often still arranged in tetrads (РІ. 24. fig. 24). Тһе ripe spores show the triradiate
marking with great distinctness, and the wall is also decorated with a narrow wing in
the equatorial plane (Pl. 24. fig. 25). The greatest diameter of the spores is about
‘03 mm , showing that the cone was either homosporous or microsporangiate.

(9) Comparison with other Species of Lepidostrobus.

The cone of Lepidostrobus Binneyanus is considerably smaller in diameter than that of
the type form of Г. oldhamius* от L. Bertrandit. It also differs from L. oldhamius in
having no well-marked groove on the outer surface of the sporophyll lamina at its junction
with the heel {. Тһе tangential width of the sporophyll trace, including its sheath, just
as it emerges from the inner cortex, is 0:15 mm. ‘This is similar to the corresponding
dimension in the case of L. oldhamius, Will. (а) of Maslen $, but slightly less than that
of the type form of L. oldhamius; in the case of one of the original sections of the
latter figured by Binney ||, I have found the sporophyll traces in this region to be about
0:25 mm. in diameter. From Г. Bertrandi, Г. Binneyanus also differs anatomically in
possessing a strongly thickened inner cortex, and in having delicate vertical striations
between the horizontal thickening bars of the scalariform tracheides (РІ. 24. fig. 18).

* The diameter of Г. oldhamius, measured from sections 87 and 88, Binney Collection, Sedgwick Museum,
is about 3 ems.

+ The cone of L. Bertrandi is as much as 5 cms. in diameter. See Zalessky, M. (708), p. 33.

t Compare РІ. 24. fig. 20 of the present paper with Seward, A. C. (10), fig. 191 в, p. 188. Maslen's restoration
of the type form of L. oldhamius fails to show this characteristic groove, Maslen, A. J. (99), pl. 37. fig. 22.

$ Maslen, А. J. (99), p. 364.

| Binney Coll., Sedgwick Museum, Cambridge, Section 88, figured Binney, E. W. (71), pl. 7. figs. 2, 4, 5.

2к2

216 DR. AGNES ARBER ON THE

It further differs both from L. oldhamius and L. Bertrandi in its extremely small pith,
and in the fact that the sporangium is attached to the upper surface of the pedicel by
a well-marked longitudinal ridge, which may possibly represent a sterilised region of the
sporangial tissue (sé.7., Pl. 22. fig. 7, Pl. 25. figs. 26, 26 а, & 27). In this feature, as well
as in the occurrence of sclerised lateral pads (86.с.) on the upper surface of the pedicel,
it approaches Z. foliaceus, Maslen *, from which, however, it differs in various important
points. The cone is larger than that of L. foliaceus, and the laminse of the sporophylls
are not constricted at the base, and do not consist of the large-celled parenchyma with
secretory cells which is so characteristic of L. foliaceus.

The small, fragile cone of Г. Binneyanus belongs, obviously, to an entirely different
type from the massive fructification of the Lower Carboniferous L. Brownii, with its
cushion-like sporophyll lobes. It resembles it, however, in one feature, namely that the
sporophyll trace is accompanied, apparently throughout the entire length of the pedicel,
by transfusion tracheides +, which are especially аа on the ventral side (РІ. 22.
fig. 8, Pl. 24. fig. 20, Pl. 25. fig. 28).

Tn addition to the characters, to which we have just referred, in which L. Binneyanus
resembles one ог. other of the known species of Lepidostrobus, this cone also shows
certain features which cannot be paralleled іп any member of the genus hitherto
described. The sporophyll pedicel is more broadly winged than that of any other
Lepidostrobus known to us in the petrified state (РІ. 25. fig. 26), though these wings sink
into insignificance when they are compared with those of various so-called ** bracts” of
Lepidostrobus, which occur as impressions {. As regards internal structure, a highly
characteristic feature is the histology of the keel of the pedicel, which is notable for the
large-celled hypoderma which is situated beneath the small-celled epidermis (Pl. 22.
figs. 7 & 8). The occurrence of abundant transfusion elements in the ridge of sterile
tissue which forms the floor of the sporangium (РІ. 25. fig. 27), is also a character which
has not been recorded for any other Upper Carboniferous species.

Another point, in which L. Binneyanus appears to be unique, is the course of the
sporophyll traces, which pursue an obliquely downward path in the outer part for their
journey from vascular cylinder to pedicel (Pl. 24. fig. 21). While this arrangement
differs from that which prevails in the other species of Lepidostrobus, in which, as is
well known, the traces pass obliquely upwards from stele to sporophyll §, it recalls
the anatomy of Mesostrobus |.

(10) Diagnosis.
LEPIDOSTROBUS BINNEYANUS, sp. nov.
Horizon. Lower Coal Measures.
Cone slender, about 1:5 cms. in diameter, with axis 3 mm. in diameter.

‚ * Maslen, A. J. (90), p. 373 & pl. 38. figs. 36-39.
t Zeiller, В. (11), р. 32.
£ L. fimbriatus. See Kidston, В. (°83), pl. 31. figs. 2, 3, 4.
$ Williamson, W. С. (94), pl. 6. fig. 62; Zalessky, М. (08), pl. 3. figs. 1 & 1a; Bower, Е. 0. (98), p. 339.
| Watson, D. М. S. (09), text-fig. 1, p. 381.

PALZOZOIC CONE-GENUS LEPIDOSTROBUS. 217

Vascular cylinder of cone-axis containing a very small pith.

Metaxylem tracheides scalariform, with delicate vertical striations between the
bars of thickening.

Sporophyli traces follow an obliquely downward direction in the distal portion of their
course from stele to sporophyll pedicel. The bundle is accompanied by transfusion
tracheides throughout the pedicel.

Sporophyll pedicel about 6 mm. in length, winged, with a sclerised ridge on either
side of the median line. The pedicel has a keel which is characterised by an epidermis
of small cells, followed internally by a large-celled hypoderma.

Sporangium attached to pedicel by means of a well-marked ridge of tissue, containing
transfusion elements. There appears to be a radial plate of sterile tissue arising from
the floor of the sporangium.

Spores about 0:03 mm. in diameter, including the equatorial border.

(ii.) LEPIDOSTROBUS GRACILIS, sp. nov.
(РІ. 22. fig. 9; РІ. 23. figs. 10-12; РІ. 25. figs. 31-39; РІ. 26. figs. 40-42.)

(1) Material.

The material on which the following account is based consists primarily of five
sections * (four longitudinal and one transverse) cut from coal-balls obtained by
Mr. Lomax from Hough Hill, Stalybridge. Two of these sections + were cut from the
same cone, and the three others, though cut from different cones, appear to belong to
the same type.

A single section in the Williamson Collection, British Museum (C. N. 1776 D.), which
is described in Williamson's manuscript catalogue as “ап apparently distinct form,"
seems also to belong to this species. Іп addition, the Leiden Herbarium possesses
-a series of sections f, including considerable portions of two cones, which are probably
specifically identical with L. gracilis. These latter cones were obtained from the Seam
Fine Frau-Nebenbank, Rheinpreussen Colliery, near Duisburg, Germany.

(2) General Form and Dimensions of the Cone and Sporophylls.

Some idea of the general construction of the strobilus may be obtained from the
photograph reproduced in РІ. 28. fig. 10, which is tangential as regards the middle
region, but nearly radial at the sides. Іп the two principal longitudinal sections, the
cone is incomplete at the apex $ and possibly at the base also. The dimensions of
the longest fragment are 4'6 cms. in length by 0'9 em. in diameter, but, as there
is no truly radial section, the actual width of the cone must be estimated as

* Slides 276, 977, 278, 279, 280, Sedgwick Museum Coll., Cambridge.

T Slides 276 & 277.

£ Slides C 1, C 1 g 1-5, g 10-16, Leiden Herbarium Coll.

$ Slide C 1, g 1, Leiden Herbarium Coll., shows the apex of a cone which appears to belong to this species. It is
acutely pointed and formed of overlapping sporophyll laminze. The sporophyll pedicels towards the tip of the cone
. leave the axis at an acute angle, whereas, in the main part of the cone, they leave it almost at right angles.

218 DR. AGNES ARBER ON THE

somewhat greater than would appear from these measurements. It is clear that, even
allowing for the full width not being seen in the sections available, the fructification
was unusually slender in proportion to its length. The pedicel, which was not less than
4 mm. long, was about 1 mm. in width near the axis, and slightly broader further out.
The lamina appears to have attained a length of 12 mm. or more, with, in addition,
a downward keel 2 mm. in length. None of the preparations are truly radial and the
apices of the sporophylls are, as a rule, either not preserved or missed in the sections,
with the result that a superficial examination would lead one to suppose that the lamina
was only about 5 mm. in length. The width of the sporophyll at the level of insertion
of the ligule was about 6 mm. It will be recognised from this description that the
sporophyll, if found as a detached object in “impression material,” would closely
resemble one of the spade-like examples of Lepidophylium with their narrow pedicels,
such as L. brevifoliwm, Lesq., which has recently been recorded from the British
Coal Measures *. |

(3) The Detailed Structure of the Cone Axis.

Our knowledge of the anatomy of the cone axis is derived from a single transverse
section. Тһе axis was traversed by a vascular cylinder enclosing a pith (РІ. 23. figs. 11
& 12). The cylinder is small, its diameter being about 0:3 mm. Between the proto-
xylems and pith there are, on any radius, 3 to 4 large xylem elements. А marked
characteristic of the axis is that the outer zone of the inner cortex (¿e.) is formed of
darker and more resistant tissue than the corresponding zone in the type specimens
of Lepidostrobus oldhamius. The middle cortex is absent, and the firm outer cortex
is of the usual type.

(4) The Sporophyll Traces in the Axis.

The preservation is not sufficiently good to allow the number of sporophyll traces,
cut across in a transverse section of the axis, to be determined with certainty, but it is
clear that they were relatively few, probably not more than 20.

Sporophyll traces, just on the point of leaving the axis, are shown in tangential
section in Pl. 25. fig. 31. Each is accompanied by a large parichnos space. The trace
consists of a mass of xylem, in which no protoxylem can be recognised with certainty,
and a ventral crescent of ** phloem," separated from the xylem by a zone of parenchyma
(РІ. 25. fig. 34). Тһе * phloem” evidently consisted of delicate thin-walled cells, for it
has suffered almost complete destruction. 'There are no transfusion elements in this
region.

(5) The Detailed Structure of the Sporophyll.

Near the axis of the cone the pedicel is triangular in section, without an elongated

dorsal keel (Pl. 25. figs. 31 & 32). In this region it is composed, externally, of several

layers of dark, thick-walled cells, followed internally by a thin-walled zone, surrounding
the parichnos space (par.). The vascular bundle is embedded in this thin-walled tissue

* Arber, E. A. N. (12), p. 252.

PALZOZOIC CONE-GENUS LEPIDOSTROBUS. 219

on the ventral side of the pedicel. Enclosed in a sheath of this tissue, it protrudes into
the parichnos space, which becomes reniform when seen in tangential section.

Further from the axis, the pedicel widens to about 1:5 mm., the triangular form
becomes slightly modified, and the parichnos space disappears (РІ. 25. fig. 88 and РІ. 26.
fig. 40). А very short dorsal keel is developed, and the dark scierised tissue becomes
concentrated into two ventral cushions (sc.c.) lying to right and left of the attachment
of the sporophyll. Transfusion tracheides do not occur in the proximal region of the
pedicel, but they are abundant towards the distal end (Pl. 25. fig. 35) and in the
upturned limb (РІ. 26. fig. 42). |

Тһе sporophyll thickens considerably, where the lamina joins the pedicel, and the
upturned portion appears to have a thick median rib. In the sporophyll lamina the
tissue on the outer or dorsal side is dark and resistant, while that on the ventral side,
which includes the bundle, consists of thin-walled cells. The arrangement of the cells
near both surfaces of the lamina suggests that periderm formation may have occurred
here to a slight extent (РІ. 25. fig. 36).

| (6) The Ligule.
А ligule is seen in one case in transverse section. It is enclosed in a ligular pit, and
its attachment is shown (РІ. 26. fig. 41).

(7) The Sporangium.

The sporangium is attached to the pedicel by a slender median ridge of tissue which
rises into the cavity of the sporangium (80.7. Pl. 25. fig. 33 and РІ. 26. fig. 40). This
ridge is bounded laterally by the wall of the sporangium, and may perhaps, as in
Г. Binneyanus, be regarded as consisting of sterilised sporangial tissue (see p. 214).
Since dehiscence has taken place, it is difficult to examine the structures inside the
sporangium, but there is no doubt that an upgrowth of sterile tissue of plate-like form
arose from the floor of the sporangium (st.p., Pl. 25. fig. 33).

The outer cells of the sporangial wall are of the usual columnar type, and sometimes
_ one or two layers of thin-walled cells are preserved internal to them. Тһе columnar
layer is thicker where it approaches the supporting pedicel (РІ. 25. fig. 37 6) aud thinner
on the ventral side (Pl. 25. fig. 37 а). Dehiscence took place by means of a longitudinal
split along the median line of the upper surface (Pl. 25. fig. 32).

(8) The Spores.
Very few of the spores are retained, since the preparations are from cones that have
already dehisced. Some of the spores are shown in РІ. 25. fig. 39. ‘They are slightly
bordered in the equatorial plane. Their maximum diameter is about 0:028 mm.

(9) Comparison with other Species of Lepidostrobus.
The cone of Lepidostrobus gracilis is more slender in form than that of any other
member of the genus hitherto described as a petrifaction. The diameter is less than
1 em., even in the case of a cone which was certainly mature, as most of the sporangia

220 DR. AGNES ARBER ON THE

һауе dehisced. The number of sporophyll traces cut across in a single transverse section
of the axis does not appear to exceed 20, whereas in the forms of Lepidostrobus oldhamius
described by Maslen * the number ranges from 30 to 60. This difference is no doubt
correlated with the small dimensions of the axis іп Г. gracilis. Тһе ligule, or at least
its base, is sunk in a definite ligular pit (Pl. 26. fig. 41), while, according to Maslen 7,
“ the complete absence of a ligular chamber” characterises L. oldhamius. The structure
of the sporophyll pedicel also differs from that of L. oldhamius, and more closely recalls
that of Г. foliaceus t and L. Binneyanus $ in possessing two lateral, sclerised cushions,
between which the sporangium is attached by means of a ridge of sterile tissue. In the
details of its histology, however, the pedicel differs markedly from that of L. Binneyanus
(see р. 213), while the lamina differs no less widely from that of L. foliaceus with its
large-celled parenchyma and secretory cells. |

(10) Diagnosis.
LEPIDOSTROBUS GRACILIS, sp. nov.
Horizon. Lower Coal Measures.
Cone very slender in proportion to its length ; about 1 em. in diameter.

Vascular cylinder of cone axis about 0:8 mm. in diameter, enclosing a pith.

Outer zone of inner cortex, a resistant, dark-coloured tissue.

Sporophyll pedicel about 4 mm. in length. Short dorsal keel and a ventral sclerised
cushion on either side of the median line.

Sporophyll lamina not less than 12 mm. іп length. Tissue dark and resistant towards
the dorsal surface.

Ligule sunk in a ligular pit, at least as regards its base.

Sporangium attached to the pedicel by a slender median ridge of tissue, which is
carried up into the cavity of the sporangium as a sterile plate. |

Spores, diameter about 0:028 mm., slightly bordered in the equatorial plane.

(iii.) LEPIDOSTROBUS OLDHAMIUS, Will., f. MINOR.
(РІ, 23. figs. 18 & 14; РІ.26. figs. 43-49.)

(1) Material.

One of the Lower Coal Measure blocks in the Binney Collection, Sedgwick Museum.
contained a cone which, although it bears a general resemblance to Lepidostrobus old-
hamius, Will., yet diverges from it sufficiently to entitle it to rank at least as a “ form”
of this species. Тһе cone has been cut up into five longitudinal sections ||, of which
РІ. 23. fig. 13 represents a typical example.

* Maslen, A. J. (99). T Maslen, А. J. (98), p. 259.

£ Maslen, А. J. (799), р. 374, and pl. 38. figs. 37 4 38.

$ See Pl. 22. fig. 7 and Pl. 25. figs. 26 & 27 of the present paper.
| Binney Coll, Sedgwick Museum, АМ. (25-29).

PALZOZOIC CONE-GENUS LEPIDOSTROBUS. 221

(2) General Form and Dimensions of the Cone and Sporophylis. .

The specimen consists of the apical portion.of a cone of which the base has been
destroyed, so that it is impossible to determine the dimensions of the fructification as
а whole. Тһе length of the existing fragment is nearly 25 ems. The sporophylls to
one side are slightly damaged; if they were complete the greatest diameter would be
1:5 cms. The axis is 3 to 4 mm. in width. The apices, both of the cone as a whole
and of the axis, are noticeably obtuse and rounded. The laminæ of the sporophylls
curve over and protect the apex (Pl. 23. figs. 18 & 14).

Тһе sporophylls spring from the axis at an angle very slightly less than a right angle
(РІ. 23. fig. 18 and РІ. 26. fig. 43). It should be recalled, however, that the angular
relation of sporophyll to axis may vary in different regions of the same cone. We
are here dealing only with the apical part of the strobilus, so it is unsafe to lay much
stress on this character. Тһе lamins of the sporophylls are relatively short. Тһе
pedicel is nearly 6 mm. long and apparently from 1-2 mm. wide, whereas the upturned
part of the lamina seems to be only 3-4 mm. long, and the downward heel about 1 mm.
The width of the lamina appears to have been 3-4 mm., but the absence of transverse
sections makes it impossible to check the accuracy of this statement. The uncertainty
connected with measurements of the sporophyll in petrified material has been referred
to on p. 212, but it is at least clear that in this case as in Lepidostrobus gracilis, it was
of the “spade” type, without a broadly winged pedicel. In 4. oldhamius f. minor the
pedicel appears to be longer and the lamina smaller than in L. gracilis.

(3) The Cone Axis.

The detailed structure of the cone axis cannot be ascertained, owing to the absence
of transverse sections and to the fact that in the longitudinal sections the vascular
cylinder is not in any case preserved. The outer cortex consists of the usual elongated
cells with firm walls.

-

(A) The Sporophyll Traces in the Axis.

On examining the section which is photographed in Pl. 23. fig. 13, the outgoing
sporophyll traces are found to be cut almost longitudinally, although the outer cortex is
traversed tangentially. "This indicates that the course of the sporopbyll traces, at least
in the outer region of the axis, was markedly steep.

(5) The Detailed Structure of the Sporophylis.

The sporophyll pedicel is triangular in section, and the dorsal keel is scarcely
developed (РІ. 23. fig. 14 and РІ. 26. figs. 46 & 47). There are slightly raised sclerised
cushions on the ventral surface to right and left of the median line. "The outer layers
of the pedicel consist of thick-walled sclerised tissue, and the inner thin-walled zone is
not preserved. "The result is that the parichnos appears to be larger than it was during

SECOND SERIES.— BOTANY. VOL. VIII. 21,

292 DR. AGNES ARBER ON THE

life; the space occurs throughout the pedicel and occupies part of the heel of the
lamina (РІ. 26. fig. 43).

When the pedicel leaves the axis, the xylem portion of its trace consists of 16 or more
elements. There seems to be a tendency for the smaller elements (? protoxylem) to be
placed internally towards the ventral side of the bundle. The general structure of the
pedicel is clearly shown in the case of some sporophylls bearing young or abortive
sporangia, which are cut tangentially near the apex of the cone (РІ. 26. fig. 47). Inthe
region of the junction of pedicel and lamina, the bundle 18 accompanied by a considerable
mass of large transfusion tracheides. These lie chiefly on the ventral side, but a few
occur on the dorsal side (Pl. 26. fig. 44) and also on the flanks of the bundle.

In the upturned part of the lamina the outer tissues are dark in colour and appear
to consist of thick-walled cells, but the bundle traverses a clear-looking, parenchymatous,
central zone. The bundle is accompanied by transfusion tracheides for some distance
up the lamina on the ventral side. The epidermis on the inner face of the lamina
forms a particularly well-marked layer.

(6) The Ligule.

A ligule and a narrow ligular pit appear to be preserved in the case of one sporophyll
(Pl. 26. fig. 45).

(7) The Sporangium.

The sporangium is attached directly to the upper surface of the pedicel, without the
intervention of a ridge of sterilised sporangial tissue. There is, however, a delicate
radial plate of sterile tissue rising into the cavity of the sporangium (s¢.p., РІ. 26.
fig. 46). The distal wall of the sporangium slopes sharply backward from the region of
its attachment to the pedicel, thus leaving a considerable space between the upturned
limb of the sporophyll and the end of the sporangium (РІ. 26. fig. 43). Тһе sporan-
gium wall consists of an outer layer of columnar dark cells followed internally by one
or two layers of thin-walled cells. In РІ. 26. fig. 48 these two regions of the wall are
seen in succession in an obliquely tangential section.

(8) The Spores.

The sporangia contain spores whose greatest diameter is 0"028 to 0-03 mm., from which
we may infer that the cone was microsporangiate or homosporous. Some of the spores
are shown in РІ. 26. fig. 49. They were bordered in the equatorial plane and the tri-
radiate marking is particularly distinct.

(9) Comparison with other Forms of Lepidostrobus oldhamius, Will.
It seems clear that the cone under discussion cannot be assigned to the type form of
L. oldhamius. It differs therefrom in its relatively small size, in the backward slope of
the distal walls of the sporangia (Pl. 26. fig. 43), which show no tendency to curve over

PALZOZOIC CONE-GENUS LEPIDOSTROBUS. 223

and hide the region of insertion of the ligule, and in the absence of a groove on the
dorsal surface of the sporophyll between lamina and heel. Тһе absence of any pad of
sterile tissue on the floor of the sporangium, and the presence of a ridge of sclerised
tissue to right and left of the median line of the sporophyll pedicel, in place of fairly
uniform sclerisation, are other characters which distinguish .L. oldhamius f. minor from
the type form.

It is possible, however, that this cone may eventually prove to be identical with
Maslen's L. oldhamius (a).* The two cones agree in the diameter of the axis and
apparently also in the number of xylem elements which pass into the sporophyll trace.
L. oldhamius f. minor is somewhat more slender than /,. oldhamius (a), but this may
be due to the fact that in the former case we possess only the apical part of the cone,
whereas in L. oldhamius (а) the apex is unknown. 16 is impossible, however, to
compare L. oldhamius (а) and L. oldhamius f. minor in detail because, in the specimens
described under the former title, the structures outside the axis have scarcely survived,
while, in the latter, the sporophylls and sporangia are well preserved, though nothing
can be seen of the internal structure of the axis.

(iv.) LEPIDOSTROBUS OLDHAMIUS, Will., f. PILOSUS.
(Plate 27. figs. 50-54; Text-figs. 2, 3 (A, B, & C), 4.)

(1) Material.

The British Museum (Nat. Hist.) possesses a single section of a cone from Oldham
(У. 8871, General Collection) which Williamson t figured, and to which he devoted a
few lines of description. This section shows remarkable resemblances to two cones from
Shore, Littleborough, preparations of which are included in the collections belonging
to University College, London, to Manchester University, and to Mr. D. M. 8. Watson.
The seven transverse sections in the University College and Watson Collections have
obviously been cut from the same cone, which, for purposes of description, we may distin-
guish as Cone A. Taken in the following order they form a series passing downwards
from the apex of the cone:—U. С. L., C. 16.4; Watson, А. 92; U.C. 1., C.16.c;
Watson, A. 93; U. C. L., C. 16. 5; Watson, A. 94; U.C. L., С.16.а. А setof transverse
sections of a second cone, which we may call Cone В, is preserved in the Manchester
University Collection. The following is the order of the sections from the apex of the
сопе downwards:—R. 387; В. 386.6; В. 886.а; R.385.5; В. 385.а. In both А
and B the extreme top of the fructification is preserved, but А is the more complete,
since it includes a lower region of the cone than is seen in В. Сопез А and B and the
Williamson section referred to above resemble Lepidostrobus oldhamius, but they have,
at the same time, certain peculiar features which seem to render it advisable to distin-
guish them in some way. For this purpose Г propose to use the provisional name
Lepidostrobus oldhamius, Will., f. pilosus.

* Maslen, A. J. (799), р. 361.
+ Williamson, W. С. (794), pl. 8. fig. 56, & p. %.
21,2

224 DR. AGNES ARBER ON THE

(2) General Form and Dimensions of the Cone and Sporophylls.

As no longitudinal sections are known, it is impossible to estimate the length of the
cone. The maximum diameter of the whole fructification was about 2 cms. and of
the axis about 4 mm. However, as no specimens are at present available showing |
the base of the cone, it is probable that these numbers do not give the full dimensions.

"Towards the apex the cone was of tapering form, and the laminz of the sporophylls
extended for some distance above the tip of the axis. Numerous lamin:e are shown in
Pl. 27. fig. 52 which passes through the apex of a сопе. The outline of the apex, as
seen in transverse section, was rendered irregular by the very prominent bases of the
sporophyll pedicels (Pl. 27. fig. 50). Тһе vertical height of the sporangia must have
been extremely small, since a very slight obliquity in a transverse section—so slight as
to be scarcely recognisable in the case of the axial region—is sufficient to make the
sporangia appear almost as in tangential section, with the pedicel proximally placed on
one side of the cone and distally placed on the other side (cf. the sporangia to the upper
right-hand side and the lower left-hand side of the photograph, РІ. 27. fig. 50).

The width of the sporophyll lamina at the level of insertion of the ligule is 7 mm.
The length of the lamina must have been considerable, judging by the number of over-
lapping sporophylls seen outside the sporangia in any transverse section. The unusual
development of the sporophyll laminse is the only feature of this cone to which
Williamson's description draws attention. Не remarks on the “ well defined sections of
the overlapping free ends of the sporangiophores, which have somewhat the form
of similar sections of leaves of Lepidophloios” *, and points out that they are more
foliaceous than those of Lepidostrobus Brownit.

(3) Detailed Structure of the Cone Axis.

The xylem cylinder of the cone axis does not exceed 0:7 mm. in diameter. It has a
large pith, 0:45 mm. across. Тһе xylem is peculiar in including a relatively small
number of metaxylem elements. Williamson does not figure or mention the vascular
cylinder, but on examining the section + which he describes, I have found that there
is no metaxylem present, only the protoxylem being developed. The lowest section of
Cone A 1 is the one in which we should expect to find the xylem at its fullest develop- |
ment, since it is cut further from the cone apex than any other section of the species
which we possess, but even here the metaxylem is interrupted at intervals, and does not
form a continuous hollow cylinder, whereas the small elements of the protoxylem
external to it are arranged in an unbroken sheath (Pl. 27. figs. 50 & 51). The pith,
which survives in more than one section, consists externally of delicate collenchymatous
elements and internally of larger cells with firmer walls.

The inner cortex, when it is preserved, forms a narrow zone of thin-walled cells (?.c.,
РІ. 27. fig. 51). Тһе outer cortex is a firm layer, which may reach a thickness of about
0:7 mm. between the leaf bases (o.c., Pl. 27. fig. 50).

* Williamson, W. C. ('94), p. 27. ;
t Brit. Mus. (Nat. Hist.) General Collection, V. 8871.
$ 0. C. Г. Coll, C. 16. a.

PALZOZOIC CONE-GENUS LEPIDOSTROBUS. 225

The great interest of the anatomy of the cone axis of this species lies in the variations
in structure which arise as the apex is approached. This point will be dealt with more
fully in Section 5 of the present paper (see pp. 228-230 and Text-fig. 4).

(4) The Sporophyll Traces in the Axis.

Numerous sporophyll traces are visible in the transverse sections. In one case* I
have counted about 50. Тһе xylem of these traces is fairly massive ; immediately after

Text-fig. 2.

Lepidostrobus oldhamius, Will., f. pilosus.

А. Sporophyll trace in middle cortex, from a transverse section of the cone axis. Watson
Coll., 4.94. (x318.)

B. Sporophyll trace in outer cortex, from a transverse section of the cone axis. Watson Coll.,
А.93. (х318).

м (2у.= xylem ; 4.6.8, = inner cortical sheath ; o.c. = outer cortex.)

C. A tetrad of spores and two single spores, one of which shows the triradiate marking

and the other the equatorial border. U. C. L. Coll., С. 16.5. (х 318.)

leaving the central cylinder it consists of 14 to 19 tracheides. As usual, the bundles
acquire a sheath of delicate tissue in passing through the inner cortex. The diameter
of sporophyll trace and sheath while in the middle cortical space is 0:15 to 0:18 mm.
Traces from the middle and outer cortical regions respectively are shown in Text-fig. 2,
А & B. Near the cone apex the inner cortex is no longer differentiated, and in conse-
quence the sporophyll traces possess no sheath.

* U. C. Г, Coll, C. 16. 0.

226 DR, AGNES ARBER ОХ THE

(5) The detailed Structure of the Sporophyll.

Owing to the absence of longitudinal sections it is difficult to elucidate the structure
of the sporophyll pedicel. Some of the transverse sections, which are slightly oblique,
show, however, that the pedicel bore two lateral sclerised pads. The sporophyll trace con-
sisted as usual of spiral tracheides. Transfusion elements can occasionally be recognised
in the neighbourhood of the bundle not far from the proximal end of the pedicel, but
they do not become abundant until the distal end is approached. At the base of the
upturned lobe of the sporophyll the transfusion tracheides are large and conspicuous,

Text-fig. 3.

A and В. Lepidostrobus oldhamius, Will., f. pilosus. Part of the dorsal surfaces of two sporo-
phylls from a transverse section traversing the sporophylls above the apex of the cone axis.
А is from a very young sporophyll in which no periderm has yet developed. В is from an
older sporophyll, in which there are indications of hypodermal periderm. Manchester
University Coll., В. 387. ( х 258.)

C. Lepidostrobus oldhamius, Will., f. pilosus. Part of the dorsal surface of a sporophyll lamina
from a transverse section through the sporophyll at the level of the ligule. U. C. Г. Coll.,
С.16.а. (x258.) Apparently a hypodermal periderm is just being initiated, and cracking
has occurred during petrifaction along the plane of weakness caused by the first formed walls.

D. Paulownia imperialis, S. & Z. (Scrophulariacee). Part of dorsal surface of a calyx segment
in transverse section to show hypodermal periderm for comparison with figs. B & C. ( x 258.)
(h = base of a hair.)

and they remain so throughout the entire length of the lamina. Неге they form 4
discontinuous sheath surrounding the small xylem group, but generally placed at
some little distance from it.

The lamina was flat and leafy, becoming narrower and thicker towards the tip, where
it was almost triangular in section. In the apical region the tissue was darker and more
sclerotic than at a lower level. An interesting feature was the remarkable hairiness of
the very young sporophylllaminse. The hairs, which are shown in РІ. 27. fig. 53, arose

PALZEOZOIC CONE-GENUS LEPIDOSTROBUS. 227

from the epidermis of both dorsal and ventral surfaces of the young lamine, and
formed a tangled mass filling the spaces between the sporophylls. They are of con-
siderable size, varying in diameter from 0:014 to 0:095 mm. Their length cannot be
estimated, as they are so much curved and twisted that any section cuts them into
numerous segments, but since no septa can be detected, it is probable that they were
unicellular. These hairs appear to have been most abundant in the basal region of the
lamins. They were ephemeral structures, disappearing in the case of the older sporo-
phylls; those growing from the lateral margins were retained longest. The hairs are
so noticeable a feature that I propose, on their account, to distinguish these cones as
Lepidostrobus oldhamius, Will., f. pilosus.

The sporophylls of L. oldhamius f. pilosus throw considerable light upon the origin
of the periderm, which, as I have shown on p. 211, commonly occurs in the
cone scales of this genus. In this connection one of the sections f of Cone B in the
Manchester University Collection is particularly valuable, since it traverses the apex
of the cone above the tip of the axis, and provides a series of transverse sections
of sporophylls. Those in the centre are quite young, but in passing towards the
periphery of the section, we meet with progressively older examples. The younger
sporophylls are bounded, both dorsally and ventrally, by a very distinct and regular
epidermis. In the older sporophylls, on the other hand, the dorsal surface is protected
by a layer of periderm. A comparison of such stages as those represented in Text-
fig. 8, A & В, renders it highly probable that the phellogen was hypodermal in origin.
This suggestion is confirmed by an examination of Cone A. Here, in a transverse
section { which passes through the base of a sporophyll lamina at the level of the ligule,
we meet with an appearance (Text-fig. 3,¢) which exactly suggests, at first sight, an
early stage in the development of hypodermal periderm in various modern plants. It is
possible, however, that the diagrammatic character of this section may be a little
deceptive, and that the apparent tangential strip of “ phellogen " may be due merely to
the separation of the cell-walls, owing to splitting during or before petrifaction. But
even if this is so, the crack probably followed the plane of weakness of the earliest walls
which were laid down when periderm formation was initiated. As an example, from
the Phanerogams, of a similar formation of cork for the protection of the reproductive
organs, we may take the hypodermal periderm, which is developed in the sepals of the
Scrophulariaceous tree, Paulownia imperialis, Sieb. & Zuce., even before the flower-buds
begin to unfold (Text-fig. 3, D).

(6) The Ligule.

In one of the transverse sections cut near the cone apex six ligules are visible
(19., РІ. 27. fig. 54). Тһе immaturity of this region of the cone no doubt accounts for
the preservation of so large a number of these organs.

* These hairs may be compared with those occurring at the base of the sporophyll pedicel іп ZL. Brownii.
Zeiller, R. (711), pl. 13. figs. 7-9.

T Manchester University Coll, В. 987. Figured as a low power drawing, Seward, A. C. (’10), fig. 189, p. 184.

$ 0. С. L. Coll., C. 16. a.

$ Manchester University Coll., В. 386.5.

228 DR. AGNES ARBER ON THE

(7) The Sporangium.

The extreme tangential width of the sporangium is 3 mm. and its radial length is 5 to
6mm. Asalready pointed out, it appears to have been shallow in its vertical dimension.
А pad of sterile tissue occupies the floor of the sporangium, and from this a plate-like
process arises into the cavity in the median plane. The sporangium wall consists of
· the usual columnar outer layer, with a lining of soft tissue.

(8) The Spores.

The spores are about 0:023 mm. in diameter. They have а narrow border in the
equatorial plane, and in some cases the characteristic triradiate marking is visible
(Text-fig. 2, C).

(9) Comparison with other forms of Lepidostrobus oldhamius, Will.

The dimensions of the whole cone, the axis, the vascular cylinder, and the sporophyll
traces of Г. oldhamius f. pilosus agree closely with those of the form which Maslen *
distinguishes as L. oldhamius, Will. (a). It differs from the latter, however, in several
respects. Тһе number of xylem elements in the sporophyll traces is somewhat greater,
and the traces cut across in any given transverse section are more numerous. ‘The
thiek continuous zone of protoxylem, the relatively small amount of metaxylem, and
the absence of peripheral barred tracheides are also features in which the form which
we have described diverges from Maslen's (a) form.

When we compare L. oldhamius f. pilosus with the type form of L. oldhamius
(Maslen's (8) form) we find that the cone of the former is more fragile and the sporangia
considerably smaller and shallower, while the free overlapping laminz of the sporophylls
form a more conspicuous feature. Тһе anatomy also offers a marked contrast, since the
metaxylem, instead of being massive as in Г. oldhamius, hardly forms a continuous
zone.

It should be pointed out, however, that all the characters which distinguish
Г. oldhamius f. pilosus from the other forms of Г. oldhamius, including even the
occurrence of hairs upon the lamine—a unique feature as far as petrified material
of Lepidostrobus cones are concerned— may possibly be due to the fact that the sections
of this form, which we have described, are all cut near the apical region of the cone.

5. ON A CASE ОҒ VARIATION IN THE ANATOMICAL CHARACTERS OF DIFFERENT REGIONS
IN А SINGLE CONE.

In his memoir on Zepidostrobus, Maslen + drew attention to the need for “ complete
cones which can be cut up throughout their entire length so as to show what variations
in structure, if any, take place in different regions in the same cone." No series of
preparations, so far as I am aware, has yet been made with this object definitely in view ;

* Maslen, A. J. (99), р. 361. _ + Maslen, А. J. (99), p. 371.

PALZEOZOIC CONE-GENUS LEPIDOSTROBUS.

bS
bo
кә,

Text-fig. 4.

Portions of xylem ring of Герідовіғобив oldhamius, Will., f. pilosus, from three successive sections
passing through the same cone axis at different levels. (x 387.)

A. Section near cone apex ; only the protoxylem is developed. U. С. Г. Coll., C. 16. b.

В. Section further from apex ; metaxylem is beginning to appear. U. С. L. Coll., C. 16.5.

C. Section still further from apex ; metaxylem relatively well developed. U. С. Г. Coll., C. 16. a.
(ma.-metaxylem ; px.=protoxylem ; 1.4. =врогорву! trace.)

ә
SECOND SERIES.—BOTANY, VOL. VIII. “sM

230 DR. AGNES ARBER ON THE

but the two cones, which I have described in the preceding section of this paper, under
the name of L. oldhamius f. pilosus, throw an interesting light on the question. There
are no fewer than seven transverse sections of one of these cones (see p. 223), the two
uppermost of which pass through the sporophylls above the apex of the axis. The third
section (U. С. L. Coll., С. 16. с) traverses the tip of the axis. Here we find a vascular
cylinder consisting, as seen in transverse section, of a continuous ring of protoxylem
elements enclosing a pith, the metaxylem elements being completely absent (Text-
fig. 4, А). In this section neither outer nor inner cortex is differentiated, and the
sporophyll traces are devoid of the usual sheath of inner cortical tissue. Іп the next
section (Watson Coll, A. 93) there is still no continuous inner cortex, but the outer
cortex is differentiated, and there are a few elements of metaxylem inside the proto-
xylem ring. Тһе sporophyll traces each have an inner cortical sheath, acquired at a
lower level. Still further from the apex we come to a section (U. C. L. Coll., C. 16. 0)
іп which the metaxylem is a little stronger (Text-fig. 4, B) and there are traces of the
inner cortex, while in the next (Watson Coll, A. 94) there is more metaxylem, and the
inner cortex is visible as a fragile but approximately continuous zone. The lowest
section of all (U. С. L. Coll., С.16.а) has the metaxylem slightly better developed
(Text-fig. 4, C and Pl. 27. figs. 50 & 51), while the outer cortex has become fairly massive.
It is uncertain how far distant the lowest sections were from the base of the cone,
and it is possible that а more complete series would exhibit still further changes. -
Unfortunately the single section * of this form figured by Williamson, and the series
of seetions t through a cone in the Manchester University Collection resemble the series
which we have just discussed in showing only the characters of the apical region of
the fructification.

It is clear that, in the case of Lepidostrobus oldhamius f. pilosus, the radial width of
the metaxylem and the degree of differentiation of the cortical layers varied very
markedly according to the distance from the apex. This example illustrates the danger
of using the anatomical features of isolated sections as diagnostic characters in distin-
guishing the species of Lepidostrobus. It is necessary to lay particular stress upon this
point, since the attribution of these cones to the Lepidodendroid stems which bore them
must depend very largely on anatomical evidence, for lack of petrified specimens showing
actual continuity between the vegetative and reproductive organs.

6. ON THE RELATION BETWEEN STRUCTURE AND FUNCTION IN THE
CONE-GENUS LEPIDOSTROBUS.

The cones of Lepidostrobus are markedly uniform in the main features of their
structure, but there is, at the same time, considerable variation both in size and form.
These fructifications range from a few centimetres to about half a metre in length, and
there is a degree of diversity in the form of the sporophylls recalling that to which
Brongniart $ long ago drew attention in the modern genus Lycopodium.

* Gen. Coll. Brit. Mus. (Nat. Hist.), V. 8871.
T Manchester Univ. Coll., R. 385 a & b, В. 386 а & b, Б. 387.
+ Brongniart, А. (37), p. 28 & pl. 12.

PALZOZOIC CONE-GENUS LEPIDOSTROBUS. 231

The principal modifications displayed in the cone structure of Lepidostrobus may be
regarded as falling under two heads—first, those adaptations which appear to be primarily
concerned with the protection and support of the sporangium, and, secondly, those
which facilitate the nutrition of the spores.

Modifications for the protection and support of the individual sporangium were
evidently rendered highly necessary by the great bulk and the delicate structure of this
organ, and its “ knife-edge ” attachment to the sporophyll. The whole mass of sporangia
must also have been greatly in need of devices to increase its rigidity. A well-developed
sporophyll lamina, provided with a dorsal “heel” or flap, obviously affords very efficient
protection, both to its own sporangium and to those immediately below it. This
protective effect is no doubt increased by the presence of periderm, which, as we have
shown in the present paper, is frequently developed towards the outer surface of the
sporophyll lamina. Тһе sporophyll pedicel is in some cases a slender stalk, triangular
in section, but in other cases it is provided with lateral membranous wings. These wings
support the bulging sides of the sporangium, and this support becomes even more
effective when a cushion of sclerised tissue occurs on either side of the median attach-
ment of the sporangium (sc.c., РІ. 25. fig. 26a). These cushions are often very well
marked in cases where there is a ridge of sterile tissue forming the floor of the
sporangium, and appearing, in tangential sections, as a stalk to the sporangium. In
such cases the sporangium is slightly raised above the sporophyll pedicel, and hence the
support given by the sclerised cushions on either side would probably become particularly
valuable. Another feature, which increases the rigidity of each individual sporangium,
is the radial plate of sterile tissue which so often arises from the sporangial floor
(cf. Pl. 24. fig. 15, &c.).

The resisting power of the whole mass of sporangia is materially increased by the
occurrence, in some species, of a plate-like downward keel to the sporophyll pedicel,
which forms a firm barrier between the adjacent sporangia lying below it, and seems, in
Williamson's words, “designed to steady the several segments of the strobilus” *
(cf. PL 22. fig. 7).

The large size of the sporangia must have rendered it a matter of some difficulty to
ensure an adequate supply of food and water to the developing spores. The occurrence
of a ridge or pad of sterile tissue, uniting the sporangium with the sporophyll, was
probably of great service in this connexion, since it increased the area of communication
between archesporium and sporophyll (cf. Pl. 25. figs. 26 & 26 а). This ridge, in one of
the species described in the present paper, contained numerous transfusion tracheides, by
means of which water could be conveyed from the sporophyll trace to the developing
spores (РІ. 25. fig. 27). А still further modification of the floor of the sporangium
consists in the development of sterile trabeculs, or a median sterile plate, penetrating for
а considerable distance into the cavity. Such a structure, as Bower + has pointed out in
the case of L. Brownii, no doubt facilitated the nutrition of the spores. In L. Brownii,
moreover, as Zeiller { has recently shown, transfusion tracheides occur on the upper side

ж Williamson, W. С. (73), р. 295. Т Bower, F. О. (94), р. 528.
$ Zeiller, В. (11), p. 32.
2м2

282 DR. AGNES ARBER ON THE

of the vascular bundle throughout the sporophyll pedicel. By their means, water could
easily pass from the sporophyll trace to the trabecule. In Lepidostrobus oldhamius
the sterile plates are best developed in the young sporangia whose spores have not com-
pleted their development, and in which the need for the conduction of food and water
must be at its maximum (see РІ. 24, fig. 15).

Inall the cases in Lepidostrobus in which transfusion tracheides occur in the sporophyll,
these elements are noticeably abundant at the junction of limb and pedicel. Here they
are chiefly massed on the ventral side of the sporophyll trace, in the neighbourhood of the
ligule (cf. РІ. 26. figs. 41, 43, 44, 45). Among living plants we find an analogy for this
development of transfusion tissue in the tracheidal cup which surrounds the base of the .
ligule in Selaginella levigata, Baker, var. Lyallit, Spring.* Harvey Gibson has suggested t
that the ligule in Selaginella is a mucilage organ, whose function is the temporary one
of keeping the growing point and young leaves moist. It seems quite possible that it
played the same part in the cones of Lepidostrobus, and that, іп both cases, the perform-
ance of this function is facilitated by the tracheides which occur in this region of the leaf.
The idea that the office of the ligule may merely be to keep the delicate organs of the
developing cone from desiccation, and that it is functionless when the fructification is
mature, gains some confirmation from Zeiller’s £ discovery that, in L. Brownii, this
organ is quite ephemeral, and only survives in the young apical region of the petrified
cone.

7. List оғ MEMOIRS CITED.

ARBER, А. ((13).—^ On the Structure of Lepidostrobus laminatus, sp. nov.” Rep. Brit. Assoc. Dundee,
(1912) p. 674, 1913.

ARBER, E. А. М. (12).--“ On the Fossil Flora of the Forest of Dean Coalfield (Gloucestershire), and
the Relationships of the Coalfields of the West of England and South Wales." Phil. Trans.
Roy. Soc. London, Ser. B, vol. 202. (1912) p. 233.

Binney, E. W. (’71).—Observations on the Structure of Fossil Plants found in the Carboniferous
Strata. Part П. Lepidostrobus and some Allied Cones, Palzontographical Society, London,
1871. ;

Bower, F. O. (93).-“Оп the Structure of the Axis of Lepidostrobus Brownii, Schpr." Ann. Bot.
vol. vii. (1893) p. 329.

Bower, Е. О. (94).-“ Studies in the Morphology of Spore-producing Members.—Equisetinee and
Lycopodinem." Phil. Trans. Roy. Soc. London, Ser. B, vol. 185. Pt. 1. (1894) p. 473.

Bower, F. О. ((08).—The Origin of a Land Flora. London, 1908.

Broneniart, А. (37).—Histoire des Végétaux Fossiles. Vol. ii., 1837-38.

Brown, В. (’48).—“ Some Account of an undescribed Fossil Fruit." Ann. & Mag. Nat. Hist. 2nd ser.
vol. i. (1848) p. 376.

Brown, К. (51).--“ Some Account of an undescribed Fossil Fruit." Trans. Linn. Soc. London, vol. хх.
(1851) p. 469.

Courter, J. M., & W. J. G. Lano (711).-“ An American Lepidostrobus.” Bot. Gaz. vol. li. (1911)
p. 449.

— h

* Gibson, В. J. H. (96), pl. 8. fig. 21. t Gibson, R. J. H. (96), p. 87.
£ Zeiller, R. (11), pp. 33 & 34.

PALZOZOIC CONE-GENUS LEPIDOSTROBUS. 233

Отвѕом, R. J. H. (96). “Contributions towards a Knowledge of the Anatomy of the genus Sela-
ginella, Spr[ing]. Pt. II. The Ligule." Ann. Bot. vol. x. (1896) p. 77.

Нооккк, J. D. (48).—* Remarks on the Structure and Affinities of some Lepidostrobi.” Mem. Geol.
Survey Gt. Brit. vol. ii. pt. ii. p. 440, 1848.

Kinston, В. (783).--“ Report on Fossil Plants, collected by the Geological Survey of Scotland in
Eskdale and Liddesdale." Trans. Roy. Soc. Edinb. vol. xxx. (1883) p. 531.

Masten, A. J. ('98).—" The Шеше in Lepidostrobus.” Ann. Bot. vol. xii. (1898) p. 256.

Masten, А. J. (799).--“ The Structure of Lepidostrobus.” Trans. Linn. Soc. London, 2nd Ser. Bot.
vol. v. (1899) p. 357.

Scorr, D. H. (’00).—Studies in Fossil Botany, Еа. І. London, 1900.

Scorr, R. (Mrs. D. H. Scorr) (”06).-“ On the Megaspore of Lepidostrobus foliaceus.” New Phyt.
vol. v. (1906) p. 116.

Sewarp, А. C. (^10).—Fossil Plants. Vol. ii., Cambridge, 1910.

Warson, D. M. S. (09).—* On Mesostrobus, a New Genus of Lycopodiaceous Cones from the Lower
Coal Measures, with a Note on the Systematic Position of Spencerites.” Ann. Bot. vol. xxiii.
(1909) p. 379.

Милламзом, W. С. (’73).— On the Organization of the Fossil Plants of the Coal-Measures. Part ПІ.
Lycopodiacez (continued).” Phil. Trans. Roy. Soc. London, vol. 162, Pt. п. (1873) р. 283 (for
1872).

Милллмзом, W. С. (94) .—“ On the Organization of the Fossil Plants of the Coal-Measures. Part
XIX." Phil. Trans. Roy. Soc. London, Ser. B, vol. 184, 1894 (for 1893) p. 1. ;

ZaLEsskv, M. (708).-“ Végétaux fossiles du terrain carbonifère du Bassin du Donetz. II. Étude sur
la structure anatomique d'un Lepidostrobus." Mém. Comité Géol. St. Pétersb., N. S. Livr. 46,
1908.

?еплжк, В. (11).—* Étude sur le Lepidostrobus Brownii (Unger) Schimper.” Mém. de l'Acad. des
Sci. Paris, t. 52, 1911.

EXPLANATION OF THE PLATES.

Plates 21, 22, 23, 27 are from Photographs by W. Tams, Cambridge.
Plates 94, 25, 26 are from Drawings by the author.

PLATE 21.
Leprposrrosus OLDHAMIUS, Will. Figs. 1-3.

Fig. 1. General view of an oblique section through the apex of a cone to show plates of sterile tissue
arising from the. floors of the sporangia, (st.p.). Sedgwick Mus. Coll., Slide 281 (before
grinding down). (X74 circa.) The sporangium marked * is shown on a larger scale in
fig. 2, and the sterile plate ¢ in Pl 24. fig. 16.

(ap.=cone apex; sp.=spores ; /.==sporophyll ; Ig.= ligule.)

Fig. 2. Enlargement of the sporophyll and sporangium, marked жіп fig. 1, to show ligule. Sedgwick
Mus. Coll., Slide 281. (х 21 circa.)

(spm.w. = sporangium wall; sp. = spores ; s.a.p.=sub-archesporial рай; /.— sporophyll ;
1.4. = vascular bundle of sporophyll ; /g. —ligule.)

284 DR. AGNES ARBER ON THE

Fig. 3. General view of the same oblique section reproduced in fig. 1 after it had been further
ground down to render it more transparent. Sedgwick Mus. Coll., Slide 281. (x74 circa.)
Тһе sporophyll and ligule shown in fig. 2 were destroyed in the grinding.
(ар. = apex of сопе ; /.—sporophyll; sp.=spores ; spm.w.=sporangium wall; s.a.p.=sub-
archesporial pad ; s¢.p.=plate of sterile tissue arising from the floor of the sporangium.)

LrPrposrRoBus BInNEYANUS, sp. nov. Fig. 4.
Fig. 4. Transverse section of central part of cone axis. Binney Coll., Ө. a.1. (x45 circa.) The
sporophyll trace marked /.£.* is shown on a larger scale in Pl. 94. fig. 23, and that marked
1.4.1 in Pl. 24. fig. 22.
(v.c.— vascular cylinder ; р. = pith ; mz. = metaxylem ; pz. = protoxylem ; i.c. — inner
cortex; i.c.s.=inner cortical sheath surrounding the sporophyll trace ; 14. —sporophyll
trace.)

PLATE 22.
Шеріровтковов BINNEYANUS, sp. nov. Figs. 5-8.

Fig. 5. Radial longitudinal section through vascular cylinder of cone axis. Binney Coll, Q.a.6.
(x 105 circa.)
(mz.=metaxylem; px.=protoxylem ; p.— pith.)

Fig. 6. Part of a transverse section of a cone passing through two sporophylls near the base of their
upturned lobes and through sporangia. Binney Coll, Q.B.2. (x35 circa.) The cellular
outgrowth marked /.o. is shown on a larger scale in РІ. 25. fig. 30.

(l. = sporophyll; 41.4. = vascular bundle of sporophyl ; spm.w.= sporangium wall;
sp.—spores ; pd.— periderm ; 1.0. —cellular outgrowth from sporophyll.) à;

7. Two sporophylls seen in tangential section. Binney Coll., Q. a. 4. (х 47 circa.)

(p.w.=wing of pedicel; sc.c.=lateral sclerised cushion of pedicel ; s.r.— median ridge of
tissue by which the sporangium is attached; k.p.=keel of pedicel; ep. = epidermis ;
hy.=hypoderma; 1.2. = уаѕсшаг bundle of pedicel; spm.w.=sporangium wall.)

Fig. 8. Tangential section of the keel of another sporophyll from the same preparation as fig. 7, in

which the conducting tissue is better shown. Binney Coll., Q.a.4. (x47 circa.)
(k.p.=keel of pedicel; sc.c.— lateral sclerised cushion of pedicel ; v.=xylem; t.t. =trans-
fusion tracheides ; ep.=epidermis ; hy.=hypoderma.)

Fig.

93

LEPIDOSTROBUS GRACILIS, sp. nov. Fig. 9.
Fig. 9. A single sporophyll, cut almost radially, incomplete at the proximal end. Sedgwick Mus. Coll.,
Slide 276. (x14 circa.)
(l.=sporophyll lamina ; sc.¢.=dorsal sclerised tissue of sporophyll ; l.h.—heel of sporo-
phyll ; /.p.=pedicel of sporophyll.)

а

PLATE 23.
LEPIDOSTROBUS GRACILIS, sp. nov. Figs. 10-12.

Fig. 10. General view of a longitudinal section through a cone, which is tangential in the middle and
approximately radial at the sides. Sedgwick Mus. Coll., Slide 276. (x12 circa.)

(l.=sporophyll; LA.-—heel of sporophyll; /.t.=sporophyll trace; k.p.—keel of pedicel ;

рр. = wing of pedicel ; sc.c. = lateral sclerised cushion of pedicel ; spm.w. = sporangium

wall.)

Fig. 13.

Fig. 15.

Fig. 16.

5:

Fig. 18.

Fig. 19.

Fig. 20.

Fig. 21.

PALZOZOIC CONE-GENUS LEPIDOSTROBUS. 235

. Transverse section of cone axis. Sedgwick Mus. Coll., Slide 278. (х 47 circa.)

(p.=pith ; ma. =шеахует; pa.—protoxylem ; /.t.=sporophyll trace; i.c. —inner cortex;
0.с. = outer cortex ; /.p.b.— base of sporophyll pedicel.)

. Central part of the axis shown in fig. 11, enlarged. Sedgwick Mus. Coll., Slide 278. (х140

circa.)
(р. = pith; mz. = metaxylem ; pz. = protoxylem ; lt. = sporophyll trace; i.c. = inner
cortex.)

LEPIDOSTROBUS OLDHAMIUS, Will., f. minor. Figs. 13, 14.

General view of a section, which is radial as regards most of the sporophylls, but tangential
as regards the axis. Binney Coll., AM.27. (X441 circa.)
(/.=sporophyll; 4p.— pedicel of sporophyll ; sp.=spores ; /.t.=sporophyll trace.)

. Apical portion of another section of the same cone as that seen in fig. 13, showing a number

of sporophylls and sporangia in tangential section. Binney Coll, AM.26. (х 44 circa.)
(l. =sporophyil ; /.p.—sporophyll pedicel; spm.-sporangium.)

PLATE 24.
Leripostrosus OLDHAMIUS, Will. Figs. 15-17.

General view of part of the cone photographed in РІ. 21. fig. 1. Sedgwick Mus. Coll., Slide
дер (х 14.) The sporangium which is marked * is that indicated by an asterisk in РІ. 21.
fig.

xe ‚рөк 14. =уазсшаг bundle of sporophyll; par.=parichnos space ; /g.=ligule ;
spm.w. = sporangium wall; sp. = spores; s.a.p. = sub-archesporial pad ; st.p. = plate of
sterile tissue arising from the floor of the sporangium.)

Enlarged drawing of the sterile plate from the sporangium marked t in РІ. 21. fig. 1. Sedgwick
Mus. Coll., Slide 281. (X 77 circa.)

. Sporophyll раней and base of sporangium іп tangential section to show the sterile plate (st.p.)

in a sporangium older than that shown in fig. 16. U. С. L. Coll., С.19. (х 23.)
(s.a.p.=sub-archesporial рай; spm.w.=sporangium wall; Lic ioter pedicel.)

Lepipostrosus BINNEYANUS, sp. nov. Figs. 18—25.

Metaxylem tracheides from the radial section of the cone axis photographed іп РІ. 22. fig. 5, to
show the vertical striations connecting the scalariform thickenings. Binney Coll., Q. a. 6.
(x 318.)

Кыны of а spiral protoxylem tracheid seen іп longitudinal section, from the ваше рге-
paration аз fig. 18. Binney Coll., Q.a.6. (х 318.)

Radial longitudinal section, not quite median, passing iugi the distal end of a sporophyll
pedicel and showing the distal attachment of the sporangium. Binney Coll., Q. 8.4. (х 47.)

(l. = sporophyll lamina ; 1.1. = heel of sporophyll lamina; lp. = ойуп pedicel ;
t.t. = transfusion tracheides; spm.w. = sporangium wall; sp. = spores; lg. =? ligule;
pd. = periderm.)

Radial longitudinal section through the proximal end of a sporophyll pedicel, showing the
proximal attachment of the sporangium and the oblique downwardly directed course of the
sporophyll trace in the axis. Binney Coll, Q.a.6. (x 47.)

(az.—cone axis; 1.1. =уазсшаг bundle destined for sporophyll; Jp.=sporophyll pedicel ;
spm.w.=sporangium wall; si.r.=ridge of sterile tissue rising into base of sporangium ;
$p.=spores.)

286 DR. AGNES АЕБЕВ ON THE

Fig. 22. The xylem of a sporophyll trace in transverse section shortly after it has left the central
cylinder. The same bundle is marked 14.4 in the photograph reproduced in РІ. 21. fig. 4.
Binney Coll.,Q.a.1. (х 318.)
(In this figure and in fig. 23 the lower side of the trace faces the centre of the axis.)
Fig. 23. A sporophyll trace in transverse section, passing through the sclerised inner cortex. The same
bundle is marked /./. in the photograph reproduced in Pl. 21. fig. 4. Binney Coll., Q.a.1.
(x 818.)
(zy.—xylem; i.c.=inner cortex ; 1.c.s. — inner cortical sheath enclosing sporophyll trace.)
Fig. 24. Tetrad of spores. Binney Coll, Q.a.1. (х 318.)
Fig. 25. Mature spores showing the triradiate marking jand equatorial border. Binney Coll., ©. 8.4.
(x 318.)

PLATE 295.

LzPrposrRoBUs BINNEYANUS, sp. nov. (continued). Figs. 26-30.

Fig. 26. Tangential section of a sporophyli and sporangium somewhat crushed and distorted. Binney
Coll., Q.a.4. (x 28.)

(ро. = wing of pedicel ; k.p.=keel of pedicel; sc.c.—sclerised cushion ; /./.— sporophyll

trace ; spm.w.=sporangium wall; s/.r.—ridge of sterile tissue forming floor of sporangium.)
Fig. 26 а. Part of the sporophyll and sporangium shown in fig. 26, more highly magnified. Binney Coll.,
Q.a.4. (x 47.)

(p.w.— wing of pedicel ; k.p.=keel of pedicel; sc.c.—sclerised cushion ; 1.7. — sporophyll
trace ; ¢.t.=transfusion tracheides ; ep.—epidermis ; Ay. —hypoderma; spm.w.=sporangium
wall ; s/.r.—ridge of sterile tissue forming floor of sporangium.)

Fig. 27. A small part of a tangential section passing through a sporophyll pedicel, similar to that shown
in fig. 26 а. The sterile ridge (st.r.) in this case contains numerous transfusion tracheides
(2.2.), and there is apparently a median plate of sterile tissue (s¢.p.) arising from the sterile
ridge. Binney Coll., Q.a.3. (х 77.)

(spm.w.=continuation of sporangium wall down the side of the sterile ridge ; sc.c.=
sclerised cushion of sporophyll pedicel.)

Fig. 28. Xylem elements and some transfusion tracheides from a tangential section of a sporophyll similar
to that shown in figs. 26 and 26 а. Binney Coll., Q.a.4. (x 818.)

(zy.—xylem ; ¢.t.=transfusion tracheides.)

Fig. 29. Part of the dorsal surface of one of the transverse sections of sporophyll lamine photographed
in Pl. 22. fig. 6, to show radial arrangement of cells indicating periderm formation (pd.).
Binney Coll., Q. 8.2. (x 318.)

Fig. 30. Cellular outgrowth from the ventral surface of a sporophyll lamina. This outgrowth is seen on
a smaller scale, marked /.o.«, in the transverse section photoéraphed in Pl. 22. fig. 6.
Binney Coll., Q. 8.2. (х 318.)

LEPIDOSTROBUS GRACILIs, 8р. nov. Figs. 31-39.

Fig. 31. Obliquely tangential section through three sporophyll bases. Тһе two lower sporophylls (5 and с)
are still connected by tissue belonging to the outer cortex, but the upper one (a) has become
free from the cone axis. Sedgwick Mus. Coll., Slide 280. (x 14.)

(Lp.b.— base of sporophyll pedicel; zy.—xylem; ph.=so-called phloem ; рат. = parichnos
space.) (

PAL/EOZOIC CONE-GENUS LEPIDOSTROBUS. 237

Fig. 32. Tangential section through a sporophyll pedicel, a little further from the axis than a in fig. 31,
but not so far out as to show the attachment of the sporangium. Sedgwick Mus. Coll.,
Slide 280. (x 14.)
· (spm:w.-sporangium wall; zy.--xylem ; pA. =so-called phloem ; per. „оа space.)
Fig. 88. Tangential section through a sporophyll pedicel further from the axis than fig. 32. The
attachment of the base of the sporangium is shown. The sporangium has dehisced and the
walls are displaced. Sedgwick Mus. Coli., Slide 276. (х 14.)

(pw.=wing of pedicel; sc.c.=sclerised cushion; k.p.=keel of pedicel; /./.=vascular
bundle of sporophyll ; spm.w.=sporangium wall; st.r.=ridge of sterile tissue. forming floor
of sporangium ; st.p.==sterile upgrowth from floor of sporangium.)

Fig. 34. Vascular bundle from the sporophyll base marked /.p.5.(c)in fig. 31. Sedgwick Mus. Coll.,
Slide 280. (x 318.)

(zy. =хует ; ph.=so-called phloem.)

Fig. 35. Obliquely transverse section through the vascular bundle of a sporophyll near the distal end
of the pedicel, showing xylem (ху.) flanked by transfusion tracheides (¢.t.). Sedgwick Mus.

: Coll., Slide 279. (x 318.)

Fig. 36. Part of a transverse section of a sporophyll janisa to show traces of periderm formation near
the dorsal surface. Sedgwick Mus. Coll., Slide 278. (x 818.)

(pd.=periderm ; ep.?=layer which may represent remains of epidermis.)

Fig. 37 a & b. Transverse sections through the columnar layer of the wall of the sporangium shown in
fig. 32. Тһе upperside represents the inner surface. Sedgwick Mus.Coll., Slide 280, (x 318.)

Fig. 37 a represents а part of the wall close to the dehiscence line, while fig. 37 6 represents
part of the thicker region of the wall which is adjacent to the sporophyll pedicel.

Fig. 38. Tangential section through the columnar layer of the sporangium wall. Sedgwick Mus. Coll.,
Slide 279. (x 318.)
Fig. 39. Spores. Sedgwick Mus. Coll., Slide 279. (x 318.)

PLATE 26,
LEPIDOSTROBUS GRACILIS, sp. nov. (continued). Figs. 40-42.
Fig. 40. Tangential section passing through а sporophyll and the base of а sporangium. Sedgwick Mus.
Coll., Slide 279. (х 193.)

(kp.=keel of pedicel; p.w.=wing of pedicel ; sc.c.=lateral sclerised cushion of pedicel ;
2y.= xylem ; ph.=so-called phloem ; spm.w.=sporangium wall ; st.r.= ridge of sterile tissue
occupying base of sporangium.)

Fig. 41. Sporophyll from transverse section of а сопе to show the ligule (/g.) in the ligular pit (1g.p.).
Тһе sporophyll pedicel is torn. Sedgwick Mus. Coll., Slide 278. (x 14.)
(1. — lamina of sporophyll ; zy.—xylem ; ¢.t.= transfusion tracheides.)
Fig. 42. Transverse section of the sporophyll trace near the base of the lamina. The tissue surrounding
the bundle is broken, and probably some transfusion tissue із lost on tue right-hand side.
Sedgwick Mus. Coll., Slide 279. (x 818.)
(zy. xylem ; 2.2. —transfusion tracheides.)

Т.вріроѕтковоѕ oLDHAMIUS, Will. f, Minor, Figs. 43-49.
Fig. 43. Radial longitudinal sections of two sporophylls. Binney Coll., AM. 27. (х 14.)

(az.—axis; lp.=sporophyll pedicel; /.=sporophyll lamina; Lh.=heel of sporophyll
lamina ; .р.=Нешаг pit; par.=parichnos ; spm.w.—sporangium wall; зр. =зрогез ;
ay.=xylem ; ¢.t.=transfusion tracheides. The region marked with an asterisk is reproduced
on a larger scale in fig. 44.) |

SECOND SERIES.—BOTANY, VOL. ҮШ. | 2%

288 ON THE PAL;EOZOIC CONE-GENUS LEPIDOSTROBUS.

Fig.44. High-power drawing of xylém (zy.) and transfusion tracheides (¢,¢.) from the region marked
with an asterisk in the үр of the upturned lamina of the lower sporophyll shown in fig. 43.
Binney Coll., AM.27. (x 318.)
Fig. 45. Tangential section passing through a sporophyll and showing what appears to be the ligula ë y. J.
Binney Coll., АМ. 27. (x 14.)
(ay.=xylem ; t.t. = transfusion tracheides.) |
Fig. 46. Tangential section passing eee a sporophyll pedicel and sporangium. Binney Coll.,
АМ. 29. (x 14.)
(spm.w.=sporangium wall; sc.c.=lateral sclerised cushion; sé Pec =plate of sterile tissue
arising from the sporangium dies )
Fig. 47. Tangential section passing through a sporophyll pedicel bearing an immature or abortive
sporangium. Binney Coll., AM.28. (x 193.)
(p.w.=wing of pedicel ; с.с. жалма cushion of pedicel; zy.—xylem; spm.w. [ро
angium wall.)
Fig. 48. Obliquely tangential section of a sporangium wall, passing through outer columnar thee (o.l.)
and inner thin-walled layer (2./.). Binney Coll., AM.26. (x 818.)
Fig. 49. Spores.: Binney Coll., AM.26. (x 318.)

PLATE 27.
Т.ерІрозтвовоѕ OLDHAMIUS, Will. f. prnosus. Figs. 50—54.

Fig. 50. Transverse section through the cone at some distance from the apex. U. C.L. Coll.,C.16.a. (x 6.)
(v.c.=central cylinder of cone axis; 0.c.=outer cortex of cone axis; p.w.=wing of sporo-
phyll pedicel ; spm.w.=sporangium wall; sp.=spores ; Ig.=ligule; /.=upturned lamina of
` sporophyll.)
Fig. 51. Central part of the section shown in fig. 50 more highly magnified. U.C. L.Coll., C.16.a. (x 61.)
(v.c.=central cylinder of cone axis; m.=space from which the pith has disappeared ;
mz.—metaxylem ; px.=protoxylem ; /.t.=sporophyll trace ; i.c.=inner cortex.)
Fig. 52. Transverse section through the apex of the same cone as that from which the sections repre-
sented in figs. 50 and 51 were taken, passing through the sporophylls above the top of the
axis. Towards the outside the sporophylls are cut through near their apices, and are more
triangular in section and more sclerised than those further in, which are cut nearer the base
of the upturned lobe. Watson Coll., A.92. (x 6.)
Fig. 53. Central part of the section SANE in fig. 52, more highly magnified to show, the tangled
hairs (4.). Watson Coll., A.92. (x 77 circa.)
(?.=lamina of sporophyll.)
Fig. 54. Transverse section through another cone, so near the apex of the cone axis that theo outer and
inner cortex and the metaxylem are scarcely differentiated. Manchester University Coll.,
R.386.5. (x 6.)
(v.c = vascular cylinder; /.=upturned limb of sporophyll; spm.w.—sporangium wall;
lg. = опе of the six ligules which occur in this section.)

Lepipostrosvs OLDHAMIUS, Will. Figs. 55 & 56.
Fig. 55. Tangential section of sporophyll and sporangium from an obliquely transverse section of a cone
to show the plate of sterile tissue (sf.p.) rising into the cavity of the sporangium. Scott Coll.,
2456. (x 12 circa.)
222 (Lt.—sporophyll trace ; p. —sporophyll pedicel; sym.w. =sporangium wall ; sp.=spores.)
Fig. 56. Transverse section of the outer edge of the upturned limb of а sporophyll to show perice
_(р4.). Scott Coll., 1200. (x 82.)

TRANS.LINN.SOC.SER.2.BoT. VoL VI, PL 21.

Arber.

W. Tams photo

LEPIDOSTROBUS OLDHAMIUS Wil Fig! 13, L.BINNEYANUS Fig. 4

TRANS. Linn.Soc.SER2 Вот. УогУШІРІ22.

„©
7

->
£
H

tar r> *
W. Tams phot.

LEPIDOSTROBUS BINNEYANUS

Huth coll
P.

Fis? 5 '"RACILIS nove. Её.
sp.nova Fig? 5-8, L. GRACILIS sp.nove 18.9.

Arber.

TRANS.LINN.Soc.SER 2BOTVOLVILP123.

ж--

ur p

ж, бъ к<

` а" ^.
ә eg. :
Й Ф; š f
` 2.4; * L3
^ aX
| #2

LJ
2 №, гь
| А Жалау
>, "d U^ +
А CaN Z
ree А

}
27 Е

4 Lorene
қаты oF mt
Ка
ьэм

ТАГ Cr
W.Tams phot.

Huth coll
LEPIDOSTROBUS GRACILIS spnova Fig$ 10-12, L.OLDHAMIUS WilLfminor Fi$? 13,14.

Arber. TRANS.LINN.Soc.SER.2 Bor. Vor VIll,P1.24.

EA
<.
os

x

ә;

Om
BS

С
"a

С
ЫЫ? КУЗ
SN

a?
Gh

2 Ç “Z, ENSE
CA

MuR T
«ШН пн

Be

T
ay: MU
Жыл.
ЫТТЫ 777
Bm «nc dm ПТ.
АЛТ. -
К и, -
и» C jm

ya aie —
«арғын EU

LLLI Те —

ВИЗА auro meus. 24

En» эжиу, "UE 59
ayn diii |

р!
унлн «ННН
К, ањ
a

CANA ec

e
9

^
рах 27

20.
Ael.

A.Arberdel x ; дір Hath,London.

— LEPIDOSTROBUS OLDHAMIUS Fié?15-17, L.BINNEYANUS fig? 18-25.

Arber. TRANS.LiNN.Soc.SEn. 2 Bor Vor VIII PL 25.

ЖУУ

х

%,

J X X PI
2 N. -NER ` š
` 1 ыы а

БРТ —
CROP
tt 35.

ЖҮ”:

38.
3 7b 8 Huth, London

A.Arber del.
LEPIDOSTROBUS BINNEYANUS sp.nova. Figs 26-30, L. GRACILIS sp.nova Fig? 31-39.

Arber. TRANS.LINN, Soc.SeR.2. Bot Vol VIILP1.28.

T
AMN
MA

46.

"Huth,London.

-48,L.OLDHAMIUS Will.f minor Fig? 43 49.

oe A.Arber del. | ee

Arber. TRANS.LINN.Soc.SER.2.Bor. Vor VIII, P127.

W.Tems phat. Huth coll.

LEPIDOSTROBUS OLDHAMIUS PILOSUS М.ғ sp.nova Figs 50-54, L. OLDHAMIUS Will Fig? 55,56.

| | CONCERNING < TRANSACTIONS.

| Zoological ив, Вад been completed а
п single volumes, or parts to complete sets, a
ols. 1221 is s Be to the public, aad 6s. to Fe lows ;

and tan nical section - The prices of the
Е the Zoological parts see ынсы

is divided into — cal Bot
) i entioned. | 555

2nd Ser. BOTANY.) | (VOL. ҮШІ. PART 7.

THE

OF

a DE | CONTRIBUTION TO THE STUDY OF THE EVOLUTION OF. THE |
— — FLOWER, WITH SPECIAL REFEREN СЕ ТО ТНЕ HAMAMELIDACEAE
_ CAPRIFOLIACES, AND CORNACEA: |

[2980]

VIII. A Contribution to the Study of the Evolution of the Flower, with special
reference to the НататеНдасез, Caprifoliacese, and Cornaceex. Бу ARTHUR
SAMUEL HORNE, B.Sc., F.L.S, F.G.S. Lecturer оп Botany т the Royal
Horticultural Society's School of Horticulture.

(Plates 28-30 and 13 Text-figures.)

Read 4th December, 1913.

Сохтихтз.
Page
I SOMME тыда ата n К 239
7 БИШИ СОМАЛИ сары а Туш Vu s Л. 245
U^ ООО СТА ға AWE ier VIR 951
жал ВОИС OCA OC EU EA 4 GENE n D a 261
(ау Соир а rdi ow eras VONMAVEMA Иа уы Vs Cs іш 261
(b) Жатаған 800 Маша сқа ilo уу... 7... ws 552 MOD
(OJ ы іс сег” ана cx AR 269
(ау бағына адау аа Ы T. QUI VENE EN re 274
(e) Ноа асу S LII 22 соно 278
(0) Cove са ЫЗ ни 282
GO Melanophglia io voe ny oe Vs A тота ыы ла ORE 284
(л) емнен ауле Ca dU Le Ed UT 284
1 NAME CRUSE ее, ее ra ШЕЕ 287
Rat T BRE USLELL у..,......... тае n s 291
Argument :—The Interpretation of Resemblances and Differences.
Origins. Phylogeny of the Hamamelidacee. Phylogeny of the
. Caprifoliaeese. Phylogeny of the Сотпасев. Тһе Vascular
Structure of the Flower in Angiosperms. The Ovule in Angio-
sperms. Ovular connections. Ovular Vascular System. The
Ру одолу OL ANGIORDOTION Lees ua ic ho sass 291-302 ,

1. INTRODUCTION.
THE structure of the flower of Davidia, a genus referred to the Согпасее in ‘Die
Natürliche. Pflanzenfamilien, has been already described by the author * in a paper
dealing with the structure and affinities of Davidia involucrata, published in the
‘Transactions of the Linnean Society, in 1909. During the study of Davidia there
was occasion to examine the flowers of several genera of Cornacez, and in so doing
striking structural differences were discovered not only between representative genera
of the subfamilies but also between the genera that have been grouped ina particular
subfamily, such as, for instance, the Cornoideæ. At the same time, as reference had

* А. б. Horne in Linn. Soc. Trans. 2nd ser., Dot. vii. (1909) 303-326, t. 31-33.
SECOND SERIES.— BOTANY, VOL. VIII. 20

240 MR. А. 8. HORNE—A CONTRIBUTION ТО

been repeatedly made to the Hamamelidacee and Caprifoliacez in the attempt to arrive
аб an interpretation of the floral morphology and a solution of the phylogeny of
Davidia, it soon became evident that a close scrutiny of the evolutionary tendencies
exhibited by these families would lead to a broader conception of the significance of the
new discoveries, and shed considerable light upon the origin of the flower in several
genera of Согпасее. Influenced by these considerations, the author has undertaken an
elaborate and extended investigation of the Hamamelidaces, Caprifoliaceze, and the
Cornacez (as defined by Harms), notwithstanding the fact that these families are
arranged in different cohorts in Engler’s System—with the object of elucidating the true
phylogeny of the Cornacee, and in the belief that the investigation would have a direct
bearing upon the whole question of the phylogeny of the Angiosperms.

It is the object then of the present paper to record the facts that have been observed,
and to present the conclusions that have been deduced from a comparative study of the
evolutionary changes which are evident among the Hamamelidaces, Cajprifoliace:,
Согпасеге, Araliace:e, and other nearly associated families.

The composition of the family Hamamelidaceze varies but little in the chief systematic
works. Of the eighteen genera enumerated by Niedenzu *, іп * Die Pflanzenfamilien,’
fifteen are recorded by Bentham and Hooker t, sixteen by Baillon $, and nine of the ten
genera known to science at the time, by Endlicher $. Our knowledge of the general
morphology is largely due to Baillon's | researches, which are described in ‘Adansonia.
Daniel Oliver § contributed a valuable paper on Sycopsis, a new genus of Hamameli-
dace:e, in 1860, and a note ** relating to the structure of the anther in Hamamelis and
Loropetalum in 1862. Тһе embryology of several genera was studied by Shoemaker tt
in 1905.

The composition of the family Caprifoliacezee has also varied little in the different
systems. Of the eleven genera enumerated by Fritsch £t in ‘Die Pflanzenfamilien,’
all are recorded by Bentham and Hooker, and nine by Endlicher. Adora, appearing
in Bentham and Hooker's list, is not retained by Fritsch. Е. Hóck $$ proposed the
transference of Sambucus to a new family, the Sambucacez, but his suggestion has not
received general acceptance.

The literature is chiefly systematie, or of a strictly morphological character. Maxi-
mowicz ||| contributed an extensive monograph on the Caprifoliaceze to the ‘ Bulletin de
l'Académie Impériale des Sciences de Saint Pétersbourg.’ Oersted Ч] revised the grouping

* Е. Niedenzu in Engl. u. Prantl, Nat. Pfi., Teil iii. Abt. 2a (1891) 115-130.
T G. Bentham et J. D. Hooker, Gen. Pl. ii. 664—669.

$ H. Baillon, Histoire des Plantes, iii. (1872) 456-461. § 8. Endlicher, Gen. РІ. (1836-1840) 803.
|| H. Baillon in Adansonia, i., v., xi. Ч D. Oliver in Linn. Soc. Trans., Bot. xxiii. (1860) 83.
** D. Oliver, 2. с. (1862) 457. +t D. N. Shoemaker in Bot. Gaz. xxxix. (1905) 248.

if Fritsch in Engl. а. Prantl, Nat. Pfi., Teil iv. Abt. 4 (1891) 156.

$$ F. Hóck in Bot. Centr. li. (1892) 233.

ШІ C. J. Maximowiez in Bull. Acad. Imp. Sci. St. Pétersbourg: xxiv., Lonicera (1878); xxvi, Viburnum
(1880); xxvii, Zriosteum (1881); xxxi., Linnea, Diervilla (1887).

44 Oersted in Vidensk. Meddel. Naturh. Foren. Kóbenh. vi. & vii. (1862) 267-305.

THE STUDY OF THE EVOLUTION ОЕ THE FLOWER. 241

of the species of Viburnum, according to the character of the fruit, in 1861. Е. A. Newell
Arber* interpreted several cases of synanthy in Lonicera, in 1903. The embryology
of the family has been neglected since the records that appear in the memoirs of
Hofmeister + (1858, 1859), Vesque { (1878, 1879), and Guignard $ (1882).

The composition of the family Согпасеге, unlike that in the above-mentioned families,
does vary considerably in different systems. With the object, therefore, of making the
exposition more intelligible, a table (see p. 242) has been drawn up, to collate the
classification of the Cornaceae adopted by Harms in ‘Die Pilanzenfamilien’ with those
of other systematists.

It will be noted that the genera aggregated by Harms in the Cornaceze were
distributed among several families in the systems of Endlicher, De Candolle, and
Lindley ; thus the Cornacee comprise only five genera in De Candolle’s || * Prodromus’
(Cornus, Aucuba, Griselinia, Mastixia, and Curtisia); while other genera known to
science at the time were segregated in the families Garryacew, Alangiacese, Helwingi-
aces, and АтаПасезе.

Benjamin Clarke є described some observations upon the flowers of Marlea in a
paper contributed to Hooker’s ‘Journal of Botany’ so long ago as 1850, and came to
the conclusion that the Alangiez (Alangium and Marlea) are closely related to Cornus.
Clarke’s paper seems to have been overlooked.

Bentham and Hooker** include the Согпеғ, Alangiesz, Garryacesz, and Nyssacez of
earlier systematists in the Cornacez. They define the Cornace in terms of the ovule.
Thus the genera are characterized by possessing ovules of the ventral type (with dorsal
raphe), іп contradistinction to the АгаНасеж where the ovules are of the dorsal type
(with ventral raphe). But this criterion has proved to be unsound for two reasons:
first, because the structure of the ovule in certain genera (Helwingia, Nyssa) was at
the time imperfectly known, and, secondly, in that the so-called dorsal ovules are not
necessarily homologous and do not strictly conform to a single type.

Baillon ++ contributed important additions to our knowledge of the morphology of
the family. He was the first to point out the peculiar morphology of the supposed
calyx іп Garrya, and worked out the development {$ of the flower in both Garrya
elliptica and the hybrid, Garrya Thuretii. But his interpretations of floral structure and
affinities in the Согпасеге were profoundly influenced by the results accruing from his
researches §§ relating to the organogeny of the flower in the Amentifere. He found

* Е. А. Newell Arber in Journ. Linn. Soc., Bot. xxxv. (1903) 463-474.

+ W. Hofmeister in Pringsh. Jahrb. i. (1858) 120-121; and in Abhandl, Kónigl.-Süchs. Gesell. Wiss. (1859),
533-672, t. 1-97. |

£ J. Vesque in Ann. Sci. Nat., Bot. бте вет. vi. (1878) 237-285, viii. (1879) 261-390.

$ Guignard іп Ann. Sci. Nat., Bot. бте sér. xiii. (1882) 136-198.

|| A. de Candolle, Prodromus, iv. (1830) 271.

Ч Benjamin Clarke in Hooker's Kew Journ. vol. ii. (1850) 129.

** Bentham et Hooker, ‘Genera Plantarum, vol. i. part iii. (1867) 947.

++ H. Baillon in Adansonia, v. (1864) 179-203 ; x. (1871-1873) 112-117.

+t Н. Baillon in Bull. Soc. Linn. Paris (1877), 139.

55 Н. Baillon in Adansonia, xi. (1873-1876) 163-174 ; xii. (1876-1879) 1-20.

Tanır L— THE ARRANGEMENT OF THE CORNACEJE OF HARMS IN OTHER SYSTEMS. °

бб

OL NOLLOSBRLILNOO УҮ--ЯХЯОН `S `Y “AW

Wangerin, Baillon, Bentham & Hooker, n Endlicher, De Candolle,
1910 *. Harms, 1898 f. 1877-1879 +. 1867 6. Endlicher, 1847 ||. | Lindley, 1847 4. 1836-1840 **. 1830 ++.
Garry ced, IL Garrya. Согпасеве ser. II.) Garryacee. Garryacee. Garryaces ante- Garryace:.
: desmeis affines.
_ Nyssacew. П. ( Nyssa. | D rsatan soq H, Хуввасеге. Alangiaceæ. Nyssaceæ ; gener: -- --
| 3 Ба асе affinia.
Camptotheca. е -- -- — — -- —
с ПІ, Davidia. > Combretacee. — -- E -- -- M. уе а
Alangium. 4 Corner. Alangiacee. Alangieæ. Alangiaceæ.
Alangiaceg. | у.)
Магіев, T Cornaces sec. Т.
>
V. Mastixia. Umbelliferse. Cornee Aucu- genera dubia Corna-
Е em. ceis affinia cum
5; : ? ) Я
VI. Curtisia. (АтаПеге.) J |3 vere, foot | Polyosmate Voto- | poe
mita.
(Helwingia. 3 Araliacesg. Helwingiacew. _ Helwingiaces. Helwingiaces ; Helwingiacece.
genera Santalaceis
айта.
: Corneæ genera Rhamnaceis
1 ——
Corokia. Aucubes. affinia,
| ^ Cornacem.
г Cornacew ser. І. | (1огпасев вес. I. 1 |
Cornus. 1 Corner. Согпеғ. Cornese.
Согпасеве 4 Согпеғе ver, |
Benthamia. d
VIL - |
Toricellia. - Araliaceg. Araliaces. АтаПасез. Araliacee.
Melanophylla.: -- -- TÉ -- — A a — — — —
Kaliphora. n i een LM. a ds d Mam ok
Corneæ Aucubese жі кер oe
Aucuba. Сотпасе ser. I. » Cornacee вес. II. cum Quadriala г Согпасеге. Cornee.
Decostea.
: & Decostea Tj =
. Griselinia. Araliaceer. -- — genera dubis sedis. -- --
W. Wang erin in ‘ Das лес шеді 2 Heft, 1910. § G. Bentham et J. D. Hooker, Gen. PE i. M = JEn
4 Harms in Engl. u. Prantl, Nat il iii. Abt. 8,1 || S. Endlicher, Gen. Suppl. iv. pts. ii. & iii
22 E. Baillon Histoire des Plantes, a 1879 ite ca "end Umbelliferge), and vi. 1877 * J. Lindley, The Veg. King. 2nd ed. teats
Combretacese e). жж S. En wes Gen. Pl. 1836-1840. ++ А. De Candolle, Prodr. iv. 1830

THE STUDY OF THE EVOLUTION OF THE FLOWER. 243

that the ovary of Quercus, Castanea, and other genera becomes multilocular during
development, owing to the inward or centripetal growth of the parietal placentze, and,
hence, was led to the belief that the bilocular ovaries found in certain genera of
Cornacez originated from unilocular types. But the ovary, in Сотпасеге, as will be
subsequently shown, has advanced beyond the evolutionary phase described by Baillon
for the Amentiferze, and presents instead advanced stages in progressive abortion.

Harms* brings together the largest grouping of genera ever included in the
Согпасее. Не asserts that intermediate forms exist between ovules of the dorsal and
ventral types, and hence there is no adequate reason for excluding genera with the
Aralian type of ovule (Nyssa and Davidia) from the Cornacez. But Harms's opinions
carry little weight, owing to the fact that the criteria upon which he relies are
insufficient to express the diverse structural forms of ovule obtaining in this family.

Walther Wangerin + contributed a lengthy paper, entitled “ Die Umgrenzung und
Gliederung der Согпасеғе,” to Engler's ‘Jahrbücher, in 1906. This paper describes іп
detail the chief morphological and anatomical characters of the family, gathered from
the works of Sertorius, Baillon, and others, and from an examination of herbarium
material. This paper forms the basis of Wangerin’s monograph { contributed to ‘ Das
Pflanzenreich’ in 1910. Тһе Cornaceze of Harms are, in this monograph, resolved into
four old families, the баттуасеге, Alangiaceze, Әуввасеге, and Cornacew. Two of these,
the Хуввасеге (Nyssa and Davidia) and Alangiaceæ, according to Wangerin, are
- distinguished by two integuments. The present author $, however, recorded only a
single integument for Davidia in 1909, and is now able to show that Nyssa also is
unitegumentary. Hence Wangerin has relied upon a morphological error with
regard to the Хуввасеге, and, in consequence, has failed to present a convincing case
for the re-establishment of both families.

The present investigation reveals several hitherto-unrecorded details in the floral
organization, the special morphology of the ovary and ovule, and the course of the
vascular tissue in the families under consideration. A striking evolutionary series has
been traced in the Caprifoliacez, where, owing to progressive abortion, the parietal
portion of the ovary that is well represented in Leycesteria has become reduced until, in
Sambucus, only vestiges remain. Progressive abortion, in the Caprifoliacese, is accom-
panied by progressive sterilization—the ovules are not only reduced in number, but
transitional stages occur from ovules physiologically sterile to be found in Lonicera to
the merest rudiments discovered in Sambucus. Owing to reduction the ovary or loculus,
in several genera, becomes monospermous, with a terminal seed, but in each case the
one-seeded condition has been attained by a different method. Comparison with an
analogous sterilization series in the Hamamelidaceze shows that the terminal seed in the
Caprifoliaceze is constantly axially suspended, or in the exceptional case of Viburnum
axially borne by origin, whereas in the latter family it is situated parietally.

* Harms in Engl. п. Prantl, Nat. РЯ., Teil iii. Abt. 8.

t W. Wangerin in Engl. Bot. Jahrb. Band xxxviii. Heft 2 (1906).
t W. Wangerin in Engl. Das Pflanzenreich, Heft 41 (1910).

$ А. S. Horne in Trans. Linn. Soc. ser. 2, Bot. vii. (1909) 323.

244 MR. А. 8. HORNE—A CONTRIBUTION ТО

Passing then to the Cornaces, the author selected Corokia and Marlea, genera with
unilocular or biloeular ovaries, and traced the structural changes accompanying the
reduction from two loculi to one. With the aid of all the available facts it was then
possible to analyse and interpret the uniovulate Cornace:e, and the single terminal ovule
was found to be of multiple origin. Тһе sterilization series evident in the Caprifoliaceze
extends by way of Sambucus to the Cornaces, where vestiges of ovules are unknown, and
the conclusion reached that the ovaries or loculi of the Согпасее were primitively
pluriovulate is supported by the direct evidence deduced from their own structural
peculiarities.

А series has been traced in the direction of the loss of carpellary autonomy by way of
the Araliacese, Caprifoliaceee, and Согпасе that culminates in the reduced ovary of
Aucuba, which in appearance and structure simulates a single inferior carpel.

Structural synthesis in the flower is accompanied by a corresponding economie
advance in the development of the vascular system. Тһе dorsal and lateral bundles of
the carpels, and the traces of the perianth and stamens that exist as independent series
in Aralia are in the Caprifoliaceze represented by one series, and the latter family
exhibits a still further economical advance reaching in Viburnum the condition obtaining
in the Compositze, where the bundles in the single series are reduced to half the sum
of the petals and stamens. Two lines of advance towards the evolution of the single
series have been traced, one presented by Aralia, Davidia, and Marlea, and the other
exemplified by the Hamamelidacez.

The ovule in the Caprifoliacee and Сотпасеге has advanced far beyond the evolu-
tionary stage exhibited by the Hamamelidacee, but whereas there is little deviation
from a single type in either the Hamamelidacee or Caprifoliacez, the ovules of
Cornacez present diverse forms and structures. Тһе nucellus of Cornacez presents
some resemblance to that of the Rosaceze on the one side, and on the other through
Helwingia to the reduced nucellus of Lilium or Inula.

Considerable emphasis is laid upon the fact that similar stages and methods of floral
construction recur, not only in the same but different families, and upon the fact that
similar ends are attained in the same or different families by dissimilar methods. The
separation of the sexes occurs in several genera of Hamamelidaceze, Rosacee, and -
numerous other families. Epigynous structures have arisen in numerous groups and
even in the same group by more than one device. The centripetal development of
parietal placentæ, rendering the ovary multilocular, occurs in Castanea, Quercus,
Rhodoleia, and many other genera. Similar stages in the evolution of the vascular
system have recurred in Castanea, the Caprifoliacez, and Cornaces, whilst similar
arrangements of ovules are found and the details of the vascular supply are nearly
identical in different families. Neither is the ovule by any means exempt. The
separation of one type from another; the transition from bi- to unitegumentary forms,
and the evolution of the exalbuminous seed has happened again and again in the racial
experience of Angiosperms, Convergence and parallel development are established
doctrines, but many authors have failed to grasp their significance. Convergence in

THE STUDY OF THE EVOLUTION OF THE FLOWER. 245

one or a few characters may not prove of much importance, but when the action of
evolutionary forces has achieved a like resultant or effect, and similar combinations
of characters are found, the problem of origins becomes almost insoluble.

The author is led, from the facts presented here, to express the strong conviction that
the structural peculiarities to be perceived in both the Caprifoliaceze and Cornacee are
indicative of different origins; that certain genera in each family have pursued inde-
pendent but similar lines of development, and hence both families are convergent and
polyphyletie. Тһе case for the polyphyletie origin of Cornaceze has already been stated
before the Botanical Section of the British Association at Portsmouth in 1911 *.

Тізе author, in conclusion, indicates the bearing of modern experimental work upon
the field of enquiry pursued here, discusses the affinities of the three families specially
studied, outlines the general features in the evolution of the floral vascular system and
of the ovule, and, finally, sums up the position created by this investigation in relation
to the natural classification of Angiosperms.

This research was commenced at the Royal College of Science, London, and finished
at the Wisley Laboratory, and my thanks are due to the President and Council of the
Royal Horticultural Society and the Director of the Laboratory for the opportunities
they have given me for pursuing my study on this subject at Wisley.

I am chiefly indebted to Professor J. В. Farmer, F.R.S., and to Mr. А. W. Hill, F.L.S.,
Assistant Director of the Royal Botanic Gardens, Kew, for the material used.
Specimens of Griselinia lucida were kindly collected and preserved for the purpose
of this investigation by Miss Heywood, in New Zealand, at the request of Dr. O.
V. Darbishire; and flowers of Griselinia littoralis were contributed by Mr. J.
W. Hamilton of Hamwood, Dunboyne, Ireland. For the rare female catkins of Garrya
elliptica, the author is indebted to Messrs. Veitch, and inflorescences of several
rare Caprifoliaceze were obtained from the Gardens of the Royal Horticultural Society.

2. HAMAMELIDACES.

Since the writer aims at a comparison of origins, considerable importance 1s attached

to details of floral organization, and for this reason special attention is given to variability
in the position occupied by the ovary in relation to the remaining floral whorls, to details
in the internal structure of the ovary, to the number and structure of the ovules, and to
sterilization in the ovary; special interest also attaches to an instance of the trans-
formation of the petals into stamens that he has observed in Corylopsis and to the
elucidation of the: morphology of the bracts found in Parrotia, whilst the course of
the vascular tissue in the flower has also been traced.

The variability in the position occupied by the ovary in relation to the outer whorls
will be considered first. Distyliwm is hypogynous, Hamamelis and Corylopsis are
perigynous, and the perianth and stamens are seated upon a circular ridge of tissue
attached to the ovary below mid-height. Trichocladus also is perigynous, but the ridge

* A. S. Horne in Proc. Brit. Assoc. Adv. Sci. Portsmouth (1911), p. 585.

246 MR. А. 8. HORNE—A CONTRIBUTION TO

is inserted at a higher level. Hustigma, from the illustration accompanying Seeman's *
* Botany of the Voyage of H.M.S. ‘ Herald, ” is subepigynous. Thus the ovaries of the
above-mentioned genera fall into a complete series varying from a hypogynous to an
epigynous structure. Sycopsis, as originally noted by D. Oliver f, differs from all the
remaining Hamamelidacez. Oliver states that the ovary appears to be inferior, owing
to the fact that it is surrounded by a closely adpressed tube, which bears at its margin
the remaining floral whorls. Не remarks that when the tube is laid open by dissection
the ovary is clearly superior. Hence Sycopsis forms the only other member of a second
series, beginning with a Distylium-like form in which the perigynous ridge has become
specially developed and adpressed to the ovary-wall. Very frequently, organs originally
adpressed become congenitally united, so that the step from the structure in Sycopsis
to a truly epigynous flower is not a great one. Loropetalum is epigynous, according to
the figures of Hamamelis chinensis given by Robert Brown in Abel's { * Narrative of a
Journey to China’ and to the published illustration of Loropetalum chinense$. The
probable origin of the epigynous condition has not yet been investigated. In proceeding
to pass in review the structure of the ovary itself, ii may be remarked that the
Hamamelidacez; are divided by Niedenzu || into two subfamilies, the Bucklandioide:e
(ovaries with pluriovulate loculi) and the Hamamelidoidez (ovaries with uniovulate
loculi)—a method of grouping that corresponds to the one adopted by Endlicher 9 in
1836. Now the ovary of Rhodoleia, a genus of the Bucklandioides, is specially
noteworthy : that of R. Championi is described as bilocular іп Walpers's ** * Annales
Botanices Systematice ' and in Curtis's tt * Botanical Magazine, but Bentham {f, in his
‘Flora Hongkongensis,’ records that it consists of two carpels united at the base into a
one-celled ovary with parietal placentee. Baillon §§ states that the ovary of Rhodoleia
is either completely or incompletely two-celled, that it is especially incomplete below,
where often the placente do not even touch; that in the Sumatran species
(В. Teysmannia) they are more or less fused higher up, but in the Chinese (№. Championi)
they are only in contact and may be separated without rupture; thus they are really
parietal as in so many of the Saxifraginez. J. G. Boerlage ||, in ‘Flora van Neder-
landsch Indie,’ records that the ovary is bilocular and sometimes unilocular owing to
the disappearance of the dissepiment, and, finally, Niedenzu 44Г states that it is either

Lj

* Seeman, Botany of the Voyage of H.M.S. * Herald,’ Tab. 95.
т D. Oliver in Trans. Linn. Soc., Bot. xxiii. (1860) 83; Sycopsis sinensis in Hook. Ic. 1931 (1891).
$ Robert Brown іп Abel’s ‘ Narrative of a Journey to China,’ fig. 4.
§ In Icon. Select. Horti Thenen. pl. 58. fig. 8.
|| Niedenzu іп Engl. u. Prantl, Nat. Pfl., Teil iii. Abt. 2 (1891) 121.
4 8. Endlicher, Gen. РІ. (1836-1840) 803.
** In Walpers's ‘ Annales Botanices Systematice,’ Tom. iii. 842 (1850-1853).
. Tt In Curtis's Bot. Mag. vi. Tab. 4509 (1850).
$$ Bentham, * Fl. Hongkongensis,' p. 131 (1861).
$8 H. Baillon, Histoire des Plantes, iii. 393 (1872).
ШІ 7 . G. Boerlage, “ Flora тап Nederlandsch Indie,’ i. Polypetalw, 454 (1890).
44 Niedenzu, 1. с. p. 122; fig. 68 (p. 193).

ТНЕ STUDY OF THE EVOLUTION OF THE FLOWER. 247

unilocular, completely bilocular, or bilocular in the superior portion only. The above-
mentioned descriptions appear to the writer to afford clear evidence that the ovary of
Rhodoleia is varying in the degree of centripetal development of the placentz, and in so
doing exhibits a tendency to break into different types distinguished by the presence and
absence of a septum respectively. Тһе ovaries of the Hamamelidoide:e possess a septum
which renders them bilocular.

The ovules in the Bucklandioides are arranged in series attached, іп unilocular
forms of Rhodoleia, to parietal placente. Іп Liquidambar, according to Shoemaker *,
they are developed on marginal placente. The writer attaches some importance to the
method of placentation in the Hamamelidoidez. Sargent f states that the ovules іп
Hamamelis ате suspended from the septum. Niedenzu $ gives a representation of them
attached to the septum in Hamamelis and Corylopsis spicata, but the writer, after а
careful examination of several species (notably, Hamamelis virginica, arborea, and
Zuccariniana, Corylopsis spicata, Parrotia Jacquemontiana, and Fothergilla Gardeni and
major), finds that they are not thus attached, but borne upon parietal placentæ that are
situated immediately above the septum.

There are some interesting points connected with the number of ovules, and it is
possible to arrange certain genera in a series from those in which the ovules are
numerous in each loculus to those with uniovulate loculi. Shoemaker $ records two
series in each loculus of Liquidambar styraciflua. Le Maout and Decaisne || note that
the ovules of the Hamamelidez (including Bucklandia) are either solitary or numerous,
but in the latter case only the lower are fertile, whilst in Disanthus cercidifolius, from
Baillon's є description, the number of ovules in each series is reduced to three.
Reduction reaches an extreme in the Hamamelidoidese. Неге it is generally stated that
the ovaries possess uniovulate loculi, although Baillon** long ago described two
instances in which the loculi are, аб first, biovulate, but later uniovulate, owing to the
arrestment of one of the pair, thus in Hamamelis virginica, “ Dans langle interne de
chaque loge, il y a un placenta vers le haut duquel sinsérent un ou deux ovules
descendants. Dans ce dernier cas, l'un d'eux s'arréte de bonne heure dans son
développement"; and in Distylium “ Il y a dans chaque loge deux ovules dont un seul
prend tout son développement." Le Maout and Decaisne ff even figure a four-seeded
fruit of Hamamelis. Shoemaker tt found only one example of an ovary with biovulate
carpels іп .H. virginiana, although he examined several hundred flowers and suggests

* D. N. Shoemaker in Bot. Gaz. xxxix. (1905) 262.
t Sargent, Silva of N. Amer. v. (1893) 1.
+ Niedenzu in Engl. u. Prantl, Nat. PA., Teil iii. Abt. 2 (1891), figs. 65 (Hamamelis) and 72 (Corylopsis spicata).
$ D. М. Shoemaker, 7. с. 262.
i| Le Maout and Decaisne, Gen. Syst. Bot. ed. Hook. 410 (1876).
«| H. Baillon, Histoire des Plantes, iii. 396 (1872).
++ H Baillon, І. с. 390 (Hamamelis); 395 (Пізіу ит).
Tt Le Maout and Decaisne, l. с. 409.
t£ D. М. Shoemaker, 1. с. 251.
SECOND SERIES.—BOTANY, VOL. VIII. 2P

248 МВ. А. S. HORNE—A CONTRIBUTION ТО

that the condition described by Baillon may occur more frequently in places from which
he had not obtained material. The writer has hitherto failed to find even traces of a
second ovule in H. virginica, Zuccariniana, and arborea (Text-fig. 1, A 8, А 4), but he has
observed biovulate loculi in Fothergilla Gardeni amd major, Corylopsis spicata, and
Parrotia Jacquemontiana (Text-fig. 1, С 2). Parallel instances to these occur among the
Araliacese and Umbelliferze: thus two ovules suspended from terminal parietal placentze
develop in each loculus of Aralia and Acanthopanax and, according to Tanfani *, of
Apium; and one of each pair aborts as it does in the Hamamelidoidee cited above.

According to Daillon's f description, reduction in the ovary proceeds still further,
thus in Hamamelis the ovary is sometimes completely sterile :—‘‘ Dans un grand nombre
de fleurs, le gynécée demeure petit, stérile, оп ne renfermant que des rudiment
d'ovules" ; again, in Corylopsis, “Il y a tous les passages quant au gynécée entre des
ovaires complètement stériles et vides; et ceux que contiennent des ovules bien
constitués, dans les fleurs femelles ou hermaphrodites.” Неге, in two distinct genera,
separation of the sexes is recorded as actually taking place, a feature highly significant
when it is remembered that several genera of Hamamelidaces; possess polygamous
flowers and in certain others the flowers are unisexual.

The development of the ovule of Hamamelis virginiana and Liquidambar styraciflua
was described by Shoemaker f іп 1905, who found a nucellus of considerable bulk
invested by two integuments and a parietal layer of cells similar to that formed in the
nucellus of the Saxifragaceee and Коѕасее. Тһе writer has examined the ovules of
Fothergilla, Corylopsis, and Parrotia, and finds that they are constructed like those
of Hamamelis.

Comparisons have been made from time to time by various writers between the
genera of the Hamamelidacezee and Согпасее that are provided with conspicuous
involucral bracts. The writer has examined the bracts of Parrotia, and finds that they
differ morphologically from those of the Согпасее. Тһе bracts of Parrotia are
tripartite, but this character cannot be easily observed, owing to the special development
of one or other of the three segments. Since the foliage leaves and the lowermost
bracts are stipulate, it appears that the coloured portion of the bract may correspond
either to a stipule or the lamina of a leaf, and so differs from the bracts of the
exstipulate genera, Cornus and Davidia.

The disposition and transformations of the stamens furnish other characters of
interest from an evolutionary point of view. According to Baillon 6, eight stamens are
present in Dicoryphe, but the four alternipetalous ones are sterile, and again, in
Tetrathyrium subcordatum and Corylopsis, glands occur in the interval between the
stamens. Shoemaker || points out that abortive pollen is sometimes developed in the
nectaries present in the flowers of Liquidambar styraciflua, and he regards these

* Е. Tanfani in Nuov. Giorn. Ital. vol. xxiii. (1891) 451-469, t. 4-7.

T H. Baillon, Histoire des Plantes, iii. 390 (Hamamelis); 391 (Corylopsis).
$ D. М. Shoemaker, in Bot. Gaz. xxxix. (1905) 262.

$ Н. Baillon, 7. с. 392 (Dicoryphe) ; 391 (Tetrathyrium).

| D. N. Shoemaker, 1 c. 262.

THE STUDY OF THE EVOLUTION OF THE FLOWER. 249

Text-fig. 1.

Diagrams illustrating the course of the floral vascular tiseue in four genera of the Hamamelidacesw :—
Al, Hamamelis arborea; A 2-A 4, Н. Zuccariniana; B1-B3, Fothergilla major; C1, C2, Parrotia
Jacquemontiana; D 1-0 3, Corylopsis spicata. (Abbreviations, see р. 308.)

ОР 2

950 MR. А. 8. HORNE—A CONTRIBUTION TO

nectaries as reduced stamens. Oliver* states that the ‘‘didynamous squame” in
‚ -Eustigma replace petals. An instance of the transformation of petals into stamens
was observed by the writer in flowers of Corylopsis spicata gathered from Kew plants
in 1909. The normal flowers possess five petals and five stamens. A complete
transitional series occurred between the normal and the entirely apetalous flowers. The
transitional flowers possessed foliose stamens, each consisting of a broad filament
and rudimentary anther lobes. The genus Fothergilla possesses a number of stamens
arranged in a single verticil. These emerge from a perigynous collar, which surrounds
the ovary. The writer finds from seventeen to twenty-two іп F, Gardeni and major.
Baillon + investigated the development of the stamens and found that they represent
several distinct whorls. Since Fothergilla is apetalous, it is possible that the outer
whorls are really transformed sepals and petals.

The course of the vascular tissue has been followed in four genera :— Нататейз,
Fothergilla, Corylopsis, and Parrotia. The vascular cylinder (Text-fig. 1, А 1) separates,
at the base of the ovary, into an outer and an inner series of bundles (D1). Тһе bundles
of the outer series branch just below the level of the insertion of the perigynous
ridge of tissue, to form externally the stemonal and perianth traces, s£ and pe; and
internally the dorsal and lateral carpellary bundles, d and Z (B1, B2). Тһе former
then leave the ovary. The inner bundles vary in number and disposition in different
genera. In Fothergilla, two bundles are present in the septum as shown іп C1
(Parrotia), and each divides before leaving 15 to form two, which traverse the corres-
ponding placente, B3. "Thus the bundle m (C 1) forms m1 and m 2, which traverse pl 1
and pl2 respectively (B3). А single branch from m1 supplies an ovule in the left-
hand loculus and one from m 2 an ovule in the right-hand loculus. But in Hamamelis
Zuccariniana, where the loculi are uniovulate, only one placenta corresponding to pl1
and p/2 is found, and the septal bundles, m, pass into it unbranched (A3, А4). In
this case, one branch passes to the ovule from the bundle, m.

The above-described system is of a similar type to that which is found in the genus
Saxifraga. It resembles in some details that of Carpinus betulus, described by Benson $
and Welsford.

Summary.

1. A complete series from hypogynous to epigynous forms of flower exists in the
Hamamelidacee, ranging from Distylium, by way of the perigynous genera Hamamelis,
Corylopsis,and Trichocladus to Eustigma. Another series ranging from a Distylium-like
form is only represented by Sycopsis, a genus which simulates epigyny owing to the
fact that the ovary is invested by a closely adpressed perigynous tube. Loropetalum is
epigynous, but the origin of this condition needs investigation.

2. The genus Rhodoleia, according to published Бес іі етеді exhibits variability in
the internal structure of the ovary.

* D. Oliver in Trans. Linn. Soc., Bot. xxiii. (1860) 83. t H. Baillon in Adansonia, x. (1871) 127.
i Benson and Welsford in Ann. of Bot, xxiii. (1909) 630. :

THE STUDY OF THE EVOLUTION OF THE FLOWER. 251

3. The ovules in Hamamelis, Fothergilla, Corylopsis, and Parrotia are suspended from
parietal placente that are situated immediately above the septum in each.

4. The Hamamelidace: exhibit a complete series іп the reduction of the number of
ovules, ranging from Liquidambar, with numerous ovules in each loculus, by way of
Disanthus, in which Baillon records three ovules in each series, and Fothergilla,
Corylopsis, and Parrotia, with two ovules (one abortive) in each loculus, as in several
Araliacexe and Umbellifere, to Hamamelis, in which each loculus is uniovulate.

5. According to Baillon, the ovaries of Hamamelis and Corylopsis are often arrested
in development and remain small and sterile. This feature is of special interest, owing
to the fact that genera with polygamous and with unisexual flowers are frequent in
the Hamamelidacee.

6. The ovules of the Hamamelidacez possess a bulky nucellus which becomes invested
by two integuments and resembles in structure that of the Saxifragacex and Возасезе.

7. The bracts of Parrotia are tripartite and correspond morphologically to a stipulate
leaf. One or other of the segments becomes specially developed and conspicuous.
Hence the conspicuous portion may be homologous with either a stipule or the lamina
of a leaf. |

8. A complete series in the metamorphosis of petals into stamens has been observed
in Corylopsis spicata.

9. The course of the vascular tissue in the flower has been followed in four genera—
Hamamelis, Fothergilla, Corylopsis, and Parrotia. The same type prevails in them all
and is similar to that which is found in Saz¢fraga. x

3. CAPRIFOLIACES.

The general details of floral organization in the Caprifoliacez are so well known that
they need not be recapitulated. Hence the morphology and vascular structure of the
ovary and ovule will receive special consideration.

Morphology of the Ovary.

The ovary of Leycesteria, the type genus of the family, is five-chambered below,
owing to the fact that the margins of the carpels are in this region congenitally united,
forming a solid axial portion with radiating septa (Text-fig. 2, A1), but the margins
are free above, so that the accretion is incomplete (A 2). Hence the biseriate ovules are
axially or parietally borne according to their situation. The fruit contains many seeds.
Lonicera possesses a trilocular ovary, which is structurally similar to that of Leycesteria
(B1-B3), but although the ovary is pluriovulate, few seeds are formed. Frequently in
L. Caprifolium only one seed reaches maturity, the septa being ruptured or pushed
aside by the developing seed, which eventually occupies the whole cavity of the ovary.
Here the ovary does not undergo structural modification during growth in connection
with the habit of producing few seeds. Abelia and Linnea also possess trilocular ovaries
of the Leycesteria type, but in each case one loculus is uniovulate, the ovule being

252 MR. А. 8. НОВХЕ--А CONTRIBUTION ТО

Text-fig. 2.

pl

Diagrams illustrati E! Еа сєз ЁЗ
каз stra че course of the floral vascular tissue and the arrangements of the ovules in Caprifoliaces :—
Ñ № Leycesteria formosa; B1-B3, Lonicera Periclymenum ; C1-03, Linnea boreaiis, D. Abelia; Е1-Е4,
Symphoricarpus racemosus (continued in text-fig. 3).

THE STUDY OF THE EVOLUTION OF THE FLOWER. 253

suspended axially (D, C1-O3; 4061). The structural agreement in this respect between
Abelia and Linnea is so close that it is impossible to distinguish between them. The
genus Dipelta, established by Maximowiez *, possesses an ovary with two pluriovulate
and two uniovulate compartments. Seeds are not formed in the pluriovulate loculi.

Symphoricarpus forms the next member of a remarkable series in which that portion
of the ovary corresponding to the superior half in Leycesteria fails to develop. Тһе
adult flowers of S. racemosus are four-chambered below (E 1—E 3), but two-chambered in
the very small upper portion, owing to the fact that the upper portion of the lateral
earpel is incompletely developed (Е 4; /oc2, [oc 4). Each lateral chamber, (ос 1, loc 8,
contains a single axially suspended ovule. "The loculi at right angles to these are pluri-
ovulate, and the ovules which they contain are arranged in biseriate rows, suspended
from the small adpressed parietal placentz (E 3, pl) present in the upper portion of the
ovary, with the exception of the lowermost ovule, which is axially borne. Only a
single specimen of an entirely pentamerous flower was found ; in this case the inferior
portion of the ovary was five-chambered and the additional loculus uniovulate.

Abortion in the parietal portion of the ovary is carried still further in Viburnum,
where, if the very young flowers of V. Lantana be examined, three united pairs of
parietal placente may be observed (Text-fig. 3, F2, ЕЗ; р/1, 2, 3). These are almost
in contact with one another, so that they separate three compartments (loc 1, 2, 3).
During subsequent growth only a single loculus (F 1-F 3; loc 2) develops as indicated
in Text-fig. 4, leaving two upper arrested compartments (loc 1, 3). Тһе latter are pluri-
ovulate as previously recorded by Vidal T, and the ovules in each are arranged in two or
more series (Text-fig. 3, F 4; serl,ser2). At the same time Vidal, who made a special
study of the development of the ovary in the Caprifoliaces, entirely overlooked the
fact that the fertile ovule (F 1, ov) does not extend as it normally should do into either
of the adjacent compartments (F1-F3; 1001,8), but into the loculus which has
developed from the originally opposite compartment (F1-F3; /oc2). Again, the
placentz (pl 2, p13) do not bear ovules as they normally should do on the sides adjacent
to loc2. Finally, the fertile ovule does not actually form a member of the series of
sterile ovules that arise from p/1 and extend in /ocl and 3. These features furnish
an important clue to the interpretation of the peculiar morphology of Viburnum, which
will be discussed after the vascular system has been described.

Sambucus forms the extreme member of the series under discussion. The number of
loculi in S. nigra varies from two to five, but there are usually three. А single terminal
axially situated ovule occurs in each loculus (G 2). The area above the ovules, which
appears to form a part of the lower tissue region of the style, is in reality an exceed-
ingly rudimentary parietal portion of the ovary, wherein the ovules are reduced to tiny
swellings, arranged in short biseriate rows above the normal ovules. Figs. С 4-G 8
show the ovular rudiments arranged in the series а, а’; 6,6’; and с, с’ above the ovules
А, B, and C, respectively (G2). Hence it appears that the terminal ovules of Sambucus

* Maximowiez in Bull. Acad. Imp. Sei. Saint Pétersbourg, xxiv. (1878) 50.
+ Vidal in Bull. Soc. de Statistique, Grenoble, 4 ser. t. iv. (1899) 157.

bo

MR. А. 8. HORNE—A CONTRIBUTION ТО

Text-fig. 3.

D v» Cup"

pli

loral strue ture in F 1-F 4, Viburnum Lantana; G 1-G 3, Sambucus nigra; G 4-G 8, the same, biseriate

ovular rudiments; H 1-H 3, Diervilla Lonicera. (Abbreviations, see p. 308.)

THE STUDY OF THE EVOLUTION OF THE FLOWER. 255

occupy in the ovary au equivalent morphological position to those of the pluriovulate
loculi in Symphoricarpus.

In Diervilla the portion of the ovary corresponding to the parietal region in Leycesteria
is specially developed, the axial portion (Н 1) being relatively small. Hence the greater
number of the ovules are parietally borne. ‘The opposite placente are in close contact,
so that the ovary is rendered bilocular (Н 3). In Diervilla Lonicera the ovules in each
loculus are multiseriate (Н 2, Н 3, ov 1, 2, 8, 4), but in other species this arrangement is
less definite. Diervilla, in the adpressed placentze and multiseriate ovules, differs from
all other СаргіѓоПасеге.

Тһе Ovule.

Hooker and Thomson * describe the ovule of Viburnum Tinus as consisting in its early
stages “ of a minute curved nucleus encircled with an annulus which is the only integu-
ment.” This description applies equally well to the Caprifoliaceze considered in this
paper. The nucellus develops more or less obliquely, and is of comparatively small bulk.
At maturity the ovules tend to occupy a tangential position in the loculus—that is to say,
the nucellus and the vascular bundle of the raphe rest in a plane at right angles to a
radial plane. They are not displaced from this position owing to abortion, but in
Symphoricarpus the lowermost ovule of each biseriate row may be radially situated with
the raphe abaxial, not adaxial as in the АтаПасеге.

Vesque in 1878 and Guignard in 1882 record that in Lonicera the embryo-sac is
produced from a linear row of three cells. The writer finds that the single megaspore-
mother-cell in the sterile ovules of Symphoricarpus, which are morphologically similar to
the fertile ovules, produces not a linear row but a group of megaspores (РІ. 29. figs, 51-
54, m.sp). Complete embryo-sacs were not observed in the sterile ovules. This suggests
that the sterile ovules in Viburnum have become structurally modified in consequence of
the loss of fertility, and that the raphe, the integument, and the nucellus have become
degenerate. Тһе megaspores of the normal ovule of V. Lantana are arranged in a linear
row, as described by Hofmeister in his memoir of 1858 ; but the sporogenous tissue of the
imperfect ovules originates from one or more subepidermal cells, of which each may
produce a group of four cells (Pl. 29. figs. 55-57). Each megaspore, аз а result of
germination, produces an even number of nuclei, usually four or eight; but it is often
difficult to trace them to their respective “ megaspores,” since the nuclei, from eight to
sixteen and even more in number, appear to be situated in a single mass of protoplasm
(РІ. 29. figs. 58-60). Тһе energids resulting from the division of the ^ megaspores ”
do not become specially modified as in normal embryo-sacs. At the time when the
perfect ovule has reached maturity, the energids are grouped to form hollow proto-
plasmic sacs which subsequently disintegrate.

Progressive reduction reaches an extreme in Sambucus. The tissue swellings, which
correspond in situation to biseriate rows of ovules as shown in Text-fig. 3, G 4-G 8, are
caused by the radial elongation beyond the normal of a small plate of sub-epidermal cells

* Hooker and Thomson, in “ Priecursores ad Floram Indicam " (1855), Journ. Linn, Soc., Bot, ii. (1858) 163,
SECOND SERIES.—BOTANY, VOL. VIII. 209

256 MR. А. 8. HORNE—A CONTRIBUTION ТО

(РІ, 29. fig. 62). At the time of megaspore formation in the ovule each cell of the
plate is filled with a finely reticulate cytoplasm and contains а resting nucleus. Тһе
nucleus subsequently divides to form two or four daughter nuclei, which are evident
during the early stages of endosperm formation in the young seed, but soon disintegrate
(РІ. 29. figs. 63, 64).

Course of the Vascular Tissue.

The vascular structure of the flower-stalk of Leycesteria formosa separates into an
inner or axial (Text-fig. 2, А1) and an outer or peripheral series of bundles. The
bundles of the outer series extend upward in the wall of the ovary—those alternating
with the loculi toward the stamens (82), those opposite the loculi toward the petals (pe)—
and emerge from the terminal neck-like portion into the stamens and petals without
entering into connection with the axial vascular strands. The presence of a single
peripheral series of bundles is a feature which is constant for all the Caprifoliacese— -
in Aralia two distinct peripheral series are present, an outer relating to the perianth and
stamens, and an inner composed of the dorsal and lateral carpellary bundles. But
the number of bundles in the series is not constant: there are ten in Leycesteria,
being equal to the sum of the petals and stamens; but in Linnea (СТ) and Dier-
villa (Text-fig. 3, H) they are less than the sum, and only one-half the sum in Viburnum
(Text-fig. 4, В, а, 6, с, а, е).

Тһе inner axial bundles are numerically equal to and alternate with the loculi in
Leycesteria. They extend axially upward into the parietal placente (Text-fig. 2, A 2, az)
and finally into the style, where they alternate with the lobes of the stylar canal, as
shown іп C3 (Linnea). А similar grouping of the axial bundles obtains in the remaining
genera. In Lonicera, where the ovary is trilocular, there are three axial bundles (B).
In Sambucus, where the number of loculi varies from two to five, there is a corresponding
variability in the bundles. In Aralia the absence of a loculus involves the absence of
the corresponding axial and of the inner peripheral ovarian bundies, but in Sambucus
only the axial bundles vary correspondingly.

Two ranks of ovules occur in each loculus of Leycesteria. It follows, therefore, that
each axial bundle (such as, for instance, ax 1 in А 1) is adjacent to the rows of ovules
(ov 1, ov 2’), the nearer series of each of the adjacent loculi, and receives corresponding
series of vascular connections—one series from each row. Each ovule receives only a
single bundle. А similar arrangement obtains in relation to the ovules in the pluri-
ovulate loculi of Symphoricarpus, but the single ovnule in each uniovulate loculus receives
two connections, being in this respect the equivalent of а pair of ovules in the adjoining
о Тһе vascular structure of the single ovule in Sambucus is similar to
this.

Special attention has been given to the vascular structure of the flower of Viburnum,
on account of its peculiar morphology. Тһе floral vascular skeleton has been recon-
structed from serial sections in a wax model (РІ. 30). This model will now be described, -
but the description should be followed with the aid of the diagrams of transverse sections

THE STUDY OF THE EVOLUTION OF THE FLOWER. 257

Text-fig. 4.

Diagrams illustrating the course of the vascular tissue in Viburnum Lantana.

(Abbreviations, see p. 308.)

bo
©
12

958 MR. А. 8. HORNE—A CONTRIBUTION ТО

in Text-fig.4. Тһе position of the bracteole is marked br. Only five bundles, indicated
by the letters a-e, extend throughout the ovary-wall The bundles а, 6 are on the side
opposite the bracteole, and the remainder on the side adjacent to it (A-D). Thebundlea
is connected with the stamen s£ 1; b with the petal pe 1 and stamen st 2; c with pe 2, st 8,
and реЗ; d with s£4; and e with pe 4, st 5, and pe 5 (E, Е, G). It is evident, therefore,
that all the petals and stamens, with the exception of 8/1, pe 1, and s£ 2, are connected
with the bundles extending in one side of the ovary-wall—the side bearing the placentee ;
so that a special enlargement of one part of the original wall—that opposite the placentee—
seems to have taken place. Besides the principal bundles, two smaller ones, z and y (В),
are present, situated between 0 and с and e and а respectively. These die out upwards
after branching and anastomosing with the adjacent principal bundles. Branches are
shown at 2 and y (C). Figure G represents a transverse section of the model at the
level а-8 shown in Pl. 30. It will be observed that small branches extend inward
- from the principal bundles just below the point of insertion of the stamens and petals—
for instance, the branch d from pel and 42 from s¢2. These branches represent the
originally free dorsal and lateral carpellary bundles of the primitive flower. Similar
branches occur in Sambucus, but they are entirely absent in Leycesteria, as mentioned
above. Within the single series made up of principal and auxiliary bundles (т, y) there
is another bundle, az (A-D), which extends towards the placentze and divides just below
the ovule to form several small strands (D). Some of these collect to form the single
ovular bundle. Тһе remainder collect to form three strands, ov 1, ov 2, and ov3, which
extend upwards into the three placentz, p? 1, pl 2, and рі 3 respectively (E). Thence they
extend into the style, where they alternate with the above-mentioned dorsal branches.
Hence it is clear that the bundle az, owing to the extension of its branches into the
placentz, does not follow the course normally pursued by a peripheral bundle that
relates specially to a parietally-borne ovule, but it is, on the contrary, analogous to the
entire axial system of the Caprifoliacese with multilocular ovaries (Text-fig. 2). It is,
therefore, not unreasonable to suppose that the bundle has been actually displaced from
an axial to a peripheral position during the progressive abortion of the now rudimentary
compartments. Assuming this to be the case, and taking also into consideration the
morphological details already described relating to the ovule itself, together with the
facts that rudimentary ovules are sometimes situated immediately above the fertile ovule
and that rudimentary ovules do not arise from the placente (рі 2, pl 3) on the side of
the fertile loculus, there is not the least doubt that the ovule now parietally situated was
originally axially borne, but has suffered displacement concurrently with the displacement
of the septa and axially extending vascular tissue. This solution of the morphology
of the ovary brings Viburnum into line with the remaining Caprifoliacez.

Summary.
1. The Caprifoliaceze show a complete series in the direction of abortion in that part
of the ovary corresponding to the parietal portion in Leycesteria. The parietal region
is rudimentary in the pluriovulate carpels of Symphoricarpus and in Viburnum. -It is

THE STUDY OF THE EVOLUTION OF THE FLOWER. 259

extremely rudimentary in Sambucus, and obsolete in the uniovulate loculi of Symphori-
carpus. The parietal region is also small in Avalia and the Hamamelidacee.

2. Progressive abortion in the ovary is associated with progressive sterilization.
Physiological sterility occurs in Lonicera where although the loculi are pluriovulate, only a
few seeds are formed in the ovary. The sterility of particular ovules is not pre-determined. .
In Symphoricarpus the ovules of the pluriovulate loculi, which are similar in form to the
fertile ovules, are definitely sterile. Тһе megaspores are not arranged in a linear row,
as in the fertile ovule, and the embryo-sacs are incomplete. In Viburnum the sterile
ovules are very imperfect. The integument and raphe are considerably reduced, and the
sporogenous tissue develops abnormally. The ovules in Sambucus are represented only
by a few cells. The Hamamelidacez also show stages in reduction in the number of
ovules, but the stages are not similar in the two families.

3. The vascular structure of the flower exhibits a definite advance upon that of Aralia
in the direction of economy. Instead of the double peripheral series of bundles which
exists in Aralia, only a single series is present. Ву a further economical advance
the peripheral bundles, which in Leycesteria are numerically equal to the sum of the
petals and stamens, are reduced to one-half the number in Viburnum, arriving at a
similar phase in the evolution of the vascular system of the flower to that attained by
the Compositze.

4. The Caprifoliaceze mark an advance upon the typical АтаШПасеге in the direction of
progressive loss of carpellary autonomy. The styles are united, except in Sambucus
and Viburnum. Again, the synthetic development of the vascular system has led
to the disappearance in the body of the ovary of a distinct series of bundles corre-
sponding to the inner peripheral series of dorsal and lateral carpellary bundles that exists
in Aralia, Acanthopanax, and Hedera. This series is represented in the superior portion
of the ovary of Sambucus and Viburnum by strands which connect the ovarian bundles
with the stylar prolongations of the axial (marginal) bundles. But the connecting
strands are absent in Leycesteria and Lonicera, so that the final traces of the dorsal
bundles have disappeared.

5. The ovules of the Caprifoliaceze are similar in structure, and take up a tangential
position in the loculus. This position is unaffected by abortion. In Symphoricarpus,
however, the lowermost sterile ovule in each pluriovulate loculus is radially situated,
the raphe being abaxial.

6. The morphology of the terminal seed in the Caprifoliacez is rendered somewhat
obscure, owing to abortion and sterilization in theovary. The seed is axially suspended in
Sambucus, and is in this respect analogous to the lowermost ovules in the pluriovulate loculi
of Symphoricarpus. Тһе seed, in Viburnum, has through abortion become displaced from
the axial position which it originally occupied to a parietal position. Since rudimentary
ovules occur immediately above the placental area of the seed, the latter also appears to
correspond to the lowermost member of a series, as in Sambucus. Тһе seed produced
from the single ovule of the uniovulate loculus in Symphoricarpus may correspond to
the lowermost ovule of the pluriovulate loculi; but, if so, the fact is not evident, since
the parietal portion of the uniovulate carpels, perhaps owing to a sudden mutation, has

260 MR. А. 8. НОВХЕ--А CONTRIBUTION ТО

become entirely obsolete. It is clear that the terminal ovules of the Саргіѓоһасеғ have
not all originated in the same manner, but they are either axially situated or of axial
origin, whereas those of the Hamamelidacee and of Aralia, Acanthopanaz, and Hedera
are situated parietally.

7. The vascular structure of the single axially-suspended ovule in the genera with
uniovulate loculi is similar to that of a pair of ovules in the pluriovulate loculi—two
bundles enter the ovule and join to form one in the raphe. But in Aralia, Acanthopanaz,
and Hedera, and in the Hamamelidaceze, where the terminal fertile ovule is parietally
suspended, only a single bundle enters the ovule.

8. It is evident that the ovaries of several genera of Caprifoliaceze have reached their
present stage of construction in different ways, and in all probability from more than
one prototype. Leycesteria and JDiervilla present strong resemblances—both genera
possess a tubular corolla with an irregular limb, the loculi of the ovary are in each case
pluriovulate and the fruit polyspermous—nevertheless, the ovaries are structurally
different. In Diervilla the placentz are adpressed throughout the greater portion of
the ovary; in Leycesteria they are congenitally united in the lower portion. Now,
aecording to published descriptions, ovaries of both the above-mentioned types occur in
the flower of the species of the Hamamelid, Rhodoleia. Hence it is not improbable that
the structures obtaining in Diervilla and Leycesteria originated from prototypic ovaries
fluctuating like those of Rhodoleia in the degree of centripetal development of the
сагре!з. But it would be unwise to consider the prototypes necessarily identical, since
the loculi in Leycesteria contain biseriate rows of ovules, whereas those of Diervilla may
contain multiseriate rows. The latter condition recalls that of Polyosma, a genus
included among the Saxifragaceze іп Engler’s System.

The remaining Lonicerez, with the exception of Lonicera itself and Triosteum,
possess dimorphic carpels. Abelia and Linnea differ in form and habit, but the ovaries
are structurally similar; the same is the case with regard to Symphoricarpus and
Dipelta. The ovaries of Symphoricarpus and Triosteum are structurally dissimilar, and
must be regarded as having originated independently from an ovary constructed like
that of Leycesteria.

The seriate rudimentary ovules of Viburnum and the biseriate rudiments present
in Sambucus indicate that the ovaries in both genera were originally pluriovulate, whilst
the structural characteristics of the ovary itself suggest that each has passed through
the Leycesteria stage of construction. Nevertheless, the Sambuces have long been
recognized as forming a distinct subsection of the Caprifoliacez, distinguished from the
Lonicereze, which possess congenitally united styles and a tubular corolla, by their rotate
corolla and free styles. Indeed, Hóck established a separate family for Sambucus—the
Sambucacee. Moreover, from a study of the ovary in Sambucus and Viburnum, it seems
clear that the process of abortion has been in each case a different one and dissimilar to
the processes that have affected the construction of the ovary of the Гопісегее. Hence
the actual origin of the Sambuceze remains obscure, and it is by no means certain
whether the descent of either Sambucus or Viburnum could be legitimately traced to
prototypic forms of the Lonicerez.

THE STUDY ОҒ THE EVOLUTION OF THE FLOWER. 261

4. СОҺХАСЕ Ж.

Nine genera of Согпасеж have been studied :—Cornus, Marlea, Aucuba, Garrya,
Helwingia, Corokia, Melanophylla, Griselinia, and Nyssa.
Each genus is described separately, and the points are taken in the following
order :—
Morphology of the flower and ovule.
Vascular structure.

Interpretations and comparisons.

(а) Совхоз.

Morphology of the Flower and Ovule.

The genus Cornus was made the type of a new family, the Согпеге, by De Candolle *,
and is characterized by possessing polypetalous, epigynous flowers with an inferior,
bilocular ovary (rarely trilocular, according to Bentham and Hooker), surmounted
by alternating tetramerous whorls of sepals, petals, and stamens. The genus Benthamia,
established by Lindley T in 1833, which produces flowers of the Cornus type grouped to
form a globular capitulum, is now regarded as a subgenus of Cornus, since the kind of
inflorescence no longer forms a distinctive character.

The writer has devoted special attention to the following points :—Variability in the
ovary, the structure of the style, the insertion of the ovules, and the structure of the
nucellus. |

Тһе flowers of several species of Cornus, including C. mas, C. sanguinea, C. Атотит,
C. alba, have been examined, but the ovaries have proved to be bilocular without
exception. Benjamin Clarke $, however, records that he found several loculi in cultivated
specimens of C. sanguinea. Thus he states :—‘‘I have found, when cultivated in a nursery,
the number of petals and stamens frequently increased to five, six, seven, or more—the
greatest increase observed being eight petals and ten stamens, the ovary of the flower
being four-celled ; the carpels, however, sometimes increased to seven without any
. further increase of the petals and stamens.”

The ovary of Cornus, which is terminated by а single style, might not be
strictly bicarpellary, for several instances have been recorded in other families
where the loculi and the styles present in the ovary belong to different carpels.
Or, again, the style might be composed of the styles of four carpels. UN ow the
structure of the style in C. mas does not afford a clue as to its composition. The
Stylar canal is ribbon-shaped and opens out eventually into the loculi. Clarke, however,
observed important structural peculiarities in the stvles of C. alba and C. sanguinea.
He found that the canal in the upper part of the style is four-lobed—the lobes

* De Candolle, Prodromus, iv. 271. + J. Lindley in Bot. Reg. t. 1579.
+ B. Clarke іп Hooker's Kew Journ. Bot. ii. (1850) 130.

262 MR. А. 8. HORNE—CONTRIBUTION ТО

alternating with the four-rayed stigma. This fact led to the expectation that either two
or four units enter into the composition of the style. He then discovered that the lobes
of the canal in C. alba and C. sanguinea are not orientated alike, heing situated, in the
latter species, opposite stamens instead of petals. In both species, the lobed canal
passes into a ribbon-shaped canal, which is homologous with that in C. mas. Clarke’s
observations led him to the conclusion that the style is composed of the styles of two
carpels, that two of the canal lobes are marginal and two sutural, and that the
anomalous orientation in C. sanguinea is due to the imbricate disposition of the
congenitally united styles. The writer’s observations are in complete accord with those
made by Clarke. Clarke also found in the case of a flower with a trilocular ovary that
the number of canal lobes was increased by two—one marginal and one sutural—for :
each additional loculus.

The development of the ovary takes place іп the manner described by Payer *—the
projection is first of all flask-shaped; later, a basal ridge develops which divides the
single cavity into two. The ovules arise from this ridge or septum, one on either side.
The ovular mamelon grows downwards. During the downward growth the nucellus
differentiates at the tip and is first directed towards the septum; later, it bends away
from the septum, and turns obliquely upwards so that it assumes a tangential position.
At the same time it is invested by a single integument.

The development of the ovular mamelon and differentiation of the nucellus were
observed іп C. alba and sanguinea. Тһе structure of the young nucellus is shown in
РІ. 29. fig. 67. Pl. 29. fig. 65 represents a vertical section through the ovule of
Benthamia, in which the nucellus and an early stage in the growth of the single
integument are evident. The nucellus resembles а sporangium, since the arche-
sporial cells are much alike, but the sporangial appearance is partly due to viewing
the nucellus in an oblique plane, for in fig. 66 (Benthamia) and fig. 67 (C. sanguinea)
the cells are clearly arranged in rows and, in the latter case, one row (ат) is more
noticeable than the rest. А complete series of preparations leading up to the fully-
formed embryo-sac has not been obtained. The embryo-sac, however, appears to be
formed in the normal manner from one cell of a row of four. А formation of parietal
tissue has not been observed.

Vascular Structure.

The vascular tissue of the pedicel, іп mature flowers of C. sanguinea, separates into
an outer series (ош? ser) of eight and two inner bundles (in), all of which traverse the
wall of the ovary (Text-fig. 5, A). Of the outer series, alternate bundles (s/) extend
towards the stamens and the remainder (pe) towards the petals B-F. Small inner
branches, ovy (B, C), arise from the outer bundles before the latter emerge from the
ovary—of which some unite to form the *dorsal" bundles, d (E, F), present in the
i le. The bundles s£ 1 and s? 3 (A, B) opposite the inner bundles (in) do not

ranch.

* J. B. Payer, Traité d'organogénie comparée de la fleur (1854-1857), 418.

THE STUDY ОҒ THE EVOLUTION OF THE FLOWER. 263

The inner bundles (in 1, in 2) are situated one at each end of the septum. Each
divides into three (C, o, m, о), the opposite lateral branches (o) unite and extend towards

Text-fig. 5.

Diagrams illustrating the course of the vascular tissue in Cornus sanguinea. (For abbreviations see p. 308.)

ovules on either side. The median branches (m) are joined by small branches; o b and
o! b from о and 0! respectively (D) pass into the style, one on each side of the canal (E),
but soon die out, leaving only “ dorsal" bundles present there (F).

SECOND SERIES.— BOTANY, VOL. VIII. 2R

264 MR. А. S. HORNE—A CONTRIBUTION TO

Comparison with the Caprifoliacee and Hamamelidacee.

Kunth *, influenced by the resemblance between Cornus and Viburnum, placed the
Cornez, typified by Cornus, in the Caprifoliaceze. But the Corne were soon afterwards
separated from the latter to form a new family, the Corne: (De Candolle +, Endlicher T)
or Cornaces (Lindley $). Lindley considered that their habit and general characters,
and especially their polypetaly, justified this removal. Bentham and Hooker note that
the polypetalous condition is the only real distinctive character. It is, in the writer's
opinion, a valuable one, for the Caprifoliaceze are definitely gamopetalous, notwithstanding
the fact that the corolla is highly variable in different genera.

In variance with Kunth, Lindley and Oliver | are of opinion that the Cornesz more
nearly approach the Hamamelidaceze. Oliver considers even that the latter family may
not improbably be regarded as a section of the former. Oliver is guided by general
morphological resemblances, such as arrangement of the leaves, the form of inflorescence
(capitate іп Benthamia and Nyssa capitata), disposition of the stamens, the * inferior,
bicarpellary and bilocular ovary," the ovules solitary and pendulous, ete. But the facts
regarding the embryology of the Hamamelidacez revealed іп Shoemaker’s $| paper on
Hamamelis, and those deduced from the study of the ovule in the Cornacex, are against |
the approximation of these families: in the Hamamelidacese, the nucellus is bulky, it is
invested by two integuments and an extensive parietal tissue is formed, as in the
Saxifragacez; whereas, in Cornus, Ше nucellus is small, without or with very little
parietal tissue and invested by a single integument.

The writer finds a close correspondence in the floral organization between Cornus
and the Sambucev, and Cornus and Sambucus may be profitably compared. Тһе whorls
in S. canadensis are pentamerous, those in Cornus tetramerous with the exception of
the ovary. In both genera, the loculi are uniovulate, the ovules terminal, axially borne
and tangentially orientated; the nucellus is very similar and ‘invested by a single
integument; with regard to the vascular structure, a single series of bundles is present
in the ovary, dorsal branches pass into the style, and the vascular connections of the
ovule are similar. Now the variability in the number of units contributing to the ovary
in Sambucus is accompanied by a corresponding variability in the number of loculi, axial
bundles, and styles, and the writer has expressed the opinion that the dorsal portions of
the carpels have become definitely incorporated as an integral and invariable portion
of the ovary-wall. Hence the ovary is not strictly bi-, tri-, or quadri-carpellary in bi-, tri-,
or quadrilocular ovaries. Passing to Cornus, the presence of an additional unit, as
shown by Clarke, is asserted by the appearance of an extra loculus, a peripheral ovular
bundle and two additional lobes to the stylar canal, the structure of the ovary-wall
remaining unchanged, as in Sambucus; so that it does not seem unreasonable to inter-
pret the organization in terms of that of Sambucus, nor to consider the ovary bicarpellary

x З Kunth, N оу. бел. et Sp. їп. 430, ТА. de Candolle, Ргоаг. iv. 271.
+ 8. Endlicher, Gen. РІ. (1836) 798. J. Lindley, Veg. King. 2nd ed. (1847) 782.

o $
| D. Oliver in Trans. Linn. Soc., Bot. xxiii. (1860) 83. D.N. Shoemaker in Bot. Gaz. xxxix, (1905) 262.

THE STUDY OF THE EVOLUTION OF THE FLOWER. 265

by reduction from a quadricarpellary form, just as the ovary of Sambucus nigra may be
considered bi-, tri-, or quadricarpellary by reduction from the symmetrical condition in
S. canadensis. In the Hamamelidacez, the ovary is bicarpellary in the hypogynous
state, and the bicarpellary condition is constant along the line of advance pursued by the
flower in this family in the direction of epigyny.

If sympetaly does indeed constitute a real barrier between Cornus and the
Sambuce:e, one is struck by the parallelism evident in the construction of the ovary,
lessened only a little in importance by the occurrence of ovular rudiments and free
styles in Sambucus. ”

(b) MARLEA AND ALANGIUM.

Morphology of the Flower and Ovule.

The flowers of Marlea are hermaphrodite, and consist of an inferior ovary surmounted
by isomerous whorls of from four to eight sepals, petals, and stamens—usually six,
according to Clarke*, in М. begonifolia, seven in flowers of М. platanifolia growing
at Kew, in 1910.

The flowers of Alangium resemble those of Marlea, except that the stamens may be
equal to, or twice or four times as many as the petals. Тһе ovary is unilocular.

The ovary of Marlea, according to Bentham and Hooker +T, may contain from one to
three loculi, or it may be unilocular in the upper and trilocular in the lower portion.
Wangerin { states that it consists of two carpels and may be either uni- or bilocular ;
he records a second smaller sterile loculus in M. barbatum. The writer has found only
one-celled ovaries in M. platanifolia. Clarke $ recorded an occasional increase in the
number of carpels in M. begonifolia when growing under confined circumstances, “ in
one instance there were two additional imperfect cells separating the original cell and
in another three fully formed fertile carpels.” The ovary of Marlea terminates іп а
single style. Тһе canal and the stigma are four-lobed in M. begonifolia, leading
to the expectation that the style is made up of two or four units. Clarke compared
the style with that of Cornus alba and interpreted it accordingly.

Bentham || and Hooker state that the ovules are suspended from a central column in
Marlea—that is, axially. Clarke’s Т original figures of М. begonifolia confirm this
statement.

The ovular mamelon, in the very young flower of M. platanifolia, almost entirely
occupies the exceedingly small loculus. The mamelon grows downwards and the
nucellus develops from it obliquely. The mature ovule is tangentially orientated in
multilocular ovaries of M. begonifolia, from Clarke’s ** figures. Harms and Wangerin

* В. Clarke in Hook. Kew Journ. Bot. ii, (1859) 131.
+ Bentham et Hooker, Gen. Pl. i. pt. iii. 949.
+ W. Wangerin in Das Pflanzenreich, 41 Heft, Alangiacee, 1.
|| Bentham et Hooker, Gen. РІ. ii. 949.
** В. Clarke, 1. с. pl. 5. A. figs. 8, 9.

$ В. Clarke, 1. с. 130.
Ч B. Clarke, I. с. pl. 5. А. figs. 2, 8, 9.

2%2

266 MR. А. 8. HORNE—A CONTRIBUTION ТО

state that two integuments are present and two are figured in ‘Die Pflanzenfamilien’
and ‘Das Pflanzenreich’ respectively. Тһе writer has only observed the primordium

Text-fig. 6.

Diagrams showing the course of the vascular tissue in Marlea platanifolia. (For abbreviations see p. 308.)

of one integument, but the question cannot be definitely settled until slightly older
flowers haye been examined.

THE STUDY OF THE EVOLUTION OF THE FLOWER. 267

Vascular Structure.

The sections examined were obtained from very young flowers. The following
description relates to the series through the oldest flower obtained prior to the upgrowth
of the integument.

The toral vascular cylinder separates to form an outer series of fourteen bundles,
which extend upward in the ovary-wall toward the petals and stamens (Text-fig. 6,
A-G, st and pe), and an inner series of fine strands (B-E, in ser).

Тһе bundles of the outer series branch before leaving the ovary to form series of
inwardly extending strands (G, ovy). The latter behave іп an irregular manner, but
could not be traced into the style, and clearly correspond to the middle ring of bundles—
equivalent to dorsal and lateral carpellary bundles— present in Davidia.

The bundles of the inner series are feebly differentiated, and extend as far as the
point of insertion of the ovule (E, i» ser): some originate directly from the toral
cylinder (B, ? ser), passing internally as far as the axis of the flower, then outward
and upward in the ovary-wall; the rest arise from the outer bundles at a higher
level (C, D).

The connection with the ovule could not be definitely settled. А small branch, о,
could be traced from ov 1 (E), so that it is possible that branches from ov 1 and ov 2 unite
to form the ovular bundle. Higher up in the series (F) the bundles ov 1 and ov 2 are
still evident and ov 1 is present in the style (H). Тһе remaining bundles of this series
aggregate into two (F) which extend into the style (H). The bundles of the inner
series are clearly concerned with supplying the ovule with nutriment, and appear to
correspond with the innermost or septal bundles of Davidia, Clarke's* view of the
bicarpellary condition of Marlea partly depended upon tracing the dorsal ribs of the
carpels into the style, but it is now evident that the so-called dorsal ribs are merely
colleetions of small ovular strands.

'The axial extension of certain branches of the inner series at the base of the single
loculus in М. platanifolia is curious, but М. begonifolia is often bilocular, and it is
not improbable that the branches which are infralocular in M. platanifolia pursue а
septal or axial course іп bilocular forms of M. begonifolia.

Nature and Origin of the Flower.

The position of Alangium and Мачеа, in the systems of different authorities,
oscillates between the Combretacez and Cornacez. De Candolle +, who was followed
by Lindley ¢, and for a time by Endlicher$, established these genera in a separate
family, the Alangieze, on the Combretacean side—an arrangement that, at the time,
seemed consistent with the occurrence of multiseriate stamens and a unilocular ovary
in both the Alangiez and Combretacee.

* В. Clarke in Hook. Kew Journ. ii. (1850) 129. + A. P. De Candolle, Prodr. iii. 203.
1 J. Lindley, Veg. King. 2nd ed. 719 (1847). § 8. Endlicher, Gen. РІ. 1184.

268 МВ. А. 8. НОКХЕ--А CONTRIBUTION TO

Baillon *, who was undoubtedly influenced by his researches + on the organogeny of
the flower in the Amentiferz, postulated a closer affinity. Не had discovered that the
ovary in Quercus, Betula, and Castanea, which is at first unilocular, becomes ultimately
multilocular, owing to the slow inward or centripetal growth of the parietal placente,
and compared the development of the ovary in Castanea with that in Terminalia, a
genus of Combretaceze. Moreover, Payer I, in describing the growth of the terminal
placent: in the latter family, had already demonstrated their parietal and centripetal
nature. ВаШоп, accordingly and not unwisely, conceived the origin of the bilocular
ovary in Marlea from the unilocular Alangium, supposing that the tendency towards
‘centripetal development exhibited by the Combretaceæ had advanced further in
Alangium and Marlea. Не then inferred that a close affinity existed between the
Combretaceze and the Alangiez, and literally swept the Alangiez into the
Combretaceze.

Endlicher §, in 1847, discarded the idea of Combretacean affinity and transferred the
Alangiez to the Cornacez. Clarke |, in 1850, from his comparative study of Marlea
and Cornus, advocated а very close affinity with the latter genus, stating that there
is: but little left to distinguish the Alangiess and Cornacez. The view of Cornacean
affinity held almost general acceptance until the advent of Wangerin's paper € .
with this, the balance again inclines to the Combretaces. Engler at first accepted
the opinions of Harms **, but later tf brings the Alangiacez into juxtaposition with
the Муззасее between the Rhizophoraceze and Combretacez in the Myrtiflorze.

The facts accruing from this investigation do not, in the writer's opinion, favour one
side more than the other. Baillon’s hypothesis is plainly untenable for the following
reasons:—Only one placenta is present in the ovary when unilocular, not two ; the ovary-
wall is much thicker on one side than on the other, leading to the expectation that a
. loculus has failed to develop; and, finally, the infra-locular vascular tissue, under the
circumstances already discussed by the writer, suggests an originally axial or septal
extension. Тһе evidence, on the contrary, shows conclusively that the ovary is reduced.

Although Marlea agrees with Cornus in the placentation ‘and orientation of the ovule,
the resemblance between them is less close than Clarke supposed, and even the styles
possess distinetive characters ; the vascular systems, too, are dissimilar.

А thorough comparison cannot be made until the structure of the ovule and nucellus
in Marlea is better understood.

* H. Baillon in Adansonia, xii. (Castaneacées), 1-20.
+ H. Baillon in Adansonia, xi. (Corylées), 163.
t J. B. Payer, Traité d'organogénie comparée de la fleur (1854-1857).
$ S. Endlicher, Gen, Suppl. iv. part iii. 17 (1847).
| B. Clarke in Hook. Kew Journ. ii. (1850) 139.
ЧТ У. Wangerin in Das Pflanzenreich, 41 Heft, 6 (1910).
** Harms in Engl. u. Prantl, Nat. PA., Teil iii. Abt. 8.
tt А. Engler, Syll. der Pñ. ed. VI. 179 (1909). ж

THE STUDY OF THE EVOLUTION OF THE FLOWER. 269

(с) AUCUBA.
Morphology of the Flower and Ovule.

The plants are unisexual, and the bracteate flowers are borne in compound cymose
racemes with branches arranged decussately. The female inflorescence bears a number
of minute rudimentary, as well as perfect flowers. The male flower consists of
alternating tetramerous whorls of sepals, petals, and stamens, the female of a
unilocular ovary with alternating epigynous tetramerous whorls of sepals and petals.
A few hermaphrodite flowers bearing two or three stamens were found on the female’
plants of A. himalaica growing at Kew in 1910. The ovary terminates in a style with
an oblique stigma, which points towards the placental side of the ovary. Payer and
Eichler state that the placenta is directed away from the bract ; Baillon, that it may be
directed away from the bract, towards it, or laterally ; the writer finds that it may
occupy any one of the four antepetalous or alternipetalous positions. Hence the
orientation of the ovary relative to the perianth is variable. The development of the
flower has been accurately described by Payer—its primordium resembles an incomplete
circular ridge (this becomes circular with an oblique opening); as growth proceeds, the
mamelon becomes ovoid enclosing a cavity—the loculus,—and the narrow orifice forms
eventually the slit-shaped stylar canal. The placenta, which is terminal and bilobed in
transverse sections of mature ovaries, arises as a well-defined bulge from the ovary-wall.
The single descending ovule develops at the end of this protuberance. The ovaries are
very rarely biovulate. One biovulate ovary was found in A. japonica—the additional
ovule developed from a placenta on the opposite wall of the loculus, but the ovary
remained unilocular; the stigma was unequally bilobed. Two out of twenty ovaries
were biovulate in A. himalaica ; the stigmas were irregularly lobed—perhaps an abnormal
feature, since a capitate stigma is figured in Hooker's * illustrations of this species.

The earliest stages in the development of the ovule, prior to the differentiation of the
nucellus, are shown, in vertical section, in Text-fig. 13, Е, G. Тһе pivot-cells at the
point where the mamelon bends away from the ovary-wall are indicated at p. Іп
G the portion of the mamelon below the pivot-cells shown in F has elongated con-
siderably, owing to cell-division, and three regions of special merismatic activity are to
be observed. "The lowermost of these is localised at «n int, which is the primordium of
the lower rim of the integument; the next is confined to the primordium of the -
nucellus, mw; the third and uppermost, wp int, extends from the upper edge of the
nucellus to the pivot-cells on the adplacental side of the ovule. Неге the cells are
providing for horizontal extension. The next stage (D) shows the ovule after con-
siderable growth has taken place; the chalazal region is now clearly defined and
the nucellus has turned through an angle of 45°. This has been accomplished by
the longitudinal elongation of the abplacental (ab) end of the ovule during chalazal
differentiation, whilst the adplacental (ad) side has undergone less longitudinal, but

* J. D. Hooker, Illustr. Himal. Plants, 12.

270 МБ. A. 8. HORNE—A CONTRIBUTION ТО

more horizontal, extension. The under-margin of the integument has now enclosed and
extended beyond the lower end of the nucellus, and its upper concave end forms the
floor of the micropyle. The upper portion of the integument, which Baillon styles an
obturator, is of considerable bulk and of an unusual type. It invests the upper surface
of the nucellus and forms the roof of the micropyle. It consists of cells arranged in
curved rows (intr), clearly shown in D, E, that have originated from the short rows
shown іп С. ‘This curvature is due to the growth-changes, which cause the rotation
of the nucellus. The nucelius is entirely free from the integument. Subsequent
growth-changes place the axis of the nucellus towards the vertical, and are more
concerned with the development of the lower portion of the ovule.

A complete series of stages in the development of the nucellus has been obtained.
When the ovule has reached the stage shown in G, the nucellus already consists of
several short axial rows of cells. From РІ. 29, fig. 69 it is evident that the primary
archesporial cells have divided to form primary parietal and sporogenous cells. In one
row further divisions have ensued, resulting in the formation of three parietal and two
sporogenous cells. The next stage was obtained in 4. himalaica, where a linear row
of four megaspores and a definite layer of parietal tissue had been formed. The
embryo-sac, in the case shown in fig. 70 (4. japonica), will evidently be formed from the
lowermost member of.the linear row, being relatively deep-seated. Sometimes more
than one megaspore prepares to germinate, as in fig. 71 (А. japonica), megaspores of
adjacent rows; and in fig. 72 (4. himalaica) a second megaspore is situated near the
embryo-sac, which, in this case, appears to be the third cell of the row. During the
formation of the megaspores, the nucellus increases in bulk by division of the cells
forming the axial rows. Тһе epidermal cells also undergo division (fig. 70). The
embryo-sae develops in the normal manner.

Vascular Structure.

A reconstruction of the vascular system of the flower of 4. japonica was made
in wax from serial transverse sections of a particular flower. The vascular system of
this flower will now be described, and can be best followed by diagrams of transverse
sections (Text-fig. 7). Three bundles of different sizes are present in the pedicel (A):
of these the largest (A, 1, 2, 3, 4) divides first to form two and ultimately five bundles
(B, 1, 2, 2, 8, 4); another (A, 5-6) divides once—one branch forms a principal bundle
(B, 5), the other (B, 6) extends towards the axis a little and divides into two small
branches (C), which extend outward again and join the principal bundles (C, 5, 7) on
either side; the third (A, 7) forms two principal bundles (D, E, 7,7). Rudimentary axial
branches are formed from the other bundles evident in the pedicel (B, C, rud. ax.)

The principal bundles extend as far as the point of insertion of the ovule without
branching (F), and then each divides to form two or three smaller branches (G); four
of these leave the ovary—one to each petal (G, pe),—the remainder form first an arc-
shaped group (H) and higher up a single bundle in the style.

THE STUDY ОҒ THE EVOLUTION OF THE FLOWER.

Text-fig. 7.

3

/
нр

ЧГ
ұз

Diagrams showing the course of the vascular tissue in the female flower of Aucuba ;
x ovary; I-L, biovulate ovary. (For abbreviations see p. 808.)

SECOND SERIES.—BOTANY, VOL. VIII.

А-Н, uniovulate

25

271

272 MR. A. S. HORNE—A CONTRIBUTION TO

The ovular bundle is directly connected with a principal bundle; and the latter
divides into three, of which the median extends into the ovule (Е). In another flower,
branches from the adjacent principal bundles unite to form the ovular bundle.

The diagrams I, J, K, L are drawn from sections of a biovulate ovary. The ovular
bundles originate from the principal bundles 1 and 5 (I, J).

The Nature and Origin of the Flower.

Aucuba was first of all placed among the “ genera Corneis айша” of Endlicher’s
System *, but later included among the Согпе +, under a subfamily, the Aucubez. It
has remained closely associated with Oornus in subsequent systems.

Diverse interpretations have been made of the nature and origin of the unilocular
ovary. ВаШоп % argued the probable derivation of the ovary of Cornus (bilocular) from
that of Aucuba by the centripetal ingrowth of parietal placentae, pointing out that
the ovary of Garrya formed a connecting-link between the two. This argument would
imply that the ovary is polymerous, consisting of carpels united by their margins.

According to Eichler$, on the other hand, the ovary is monomerous. Wangerin ||
considers it to be monomerous by reduction :—“ Der understündige Fruchtknoten wird
hier von einen einzigen Karpellblatt gebildet ” and ** Cornus bildet das Zentrum dieser
Gruppe, von dem sich Aucuba hinsichtlich der Reduktion in Gynzeciums am weitesten
entfernt."

Now the ovary of Aucuba could be easily synthesized from carpels which become
united by their margins, and, if this were so, Baillon's hypothesis would carry weight,
since fhe carpels retain the potentiality for centripetal ingrowth. But the vascular
evidence at present turns the scale in favour of a different explanation. The rudimentary
branches at the base of the loculus recall the similarly situated infra-locular branches in
Marlea platanifolia. It was suggested in this case that the infra-locular branches
extend as axial or septal bundles in septate forms of Marlea begonifolia. Hence, in the
case of Aucuba, it seems not unreasonable to suppose that the curious infra-locular rudi-
menís represent an axial or septal system that existed in the primitive flower. If this
were so, the ovary was probably quadrilocular and hence tetramerous. The primitive
ovary of Aucuba may be likened to that of Viburnum, the effect brought about by
abortion, however, being different in the two cases. It was deduced for Viburnum, from
the peculiar floral morphology and vascular structure, that the parietally placed ovule
was originally axially suspended—the abortion of septa being evidently accompanied by
the displacement of the axially borne ovule and axial vascular tissue to à peripheral.
position. But anomalies of the kind described for Viburnum do not occur in Aucuba,
so we may conjecture that the axially suspended ovules present іп the primitive ducuba
. disappeared with the septa, leaving only a single parietally borne ovule in the ovary. In

м 5. Endlicher, Gen. РІ, 798. + $. Endlicher, Gen. Pl. Suppl. iv. part 3. 18.
t H. Baillon in Bull. Soc. Linn. Paris (1877) 139. $ Eichler, Blütendiagr. ii. (1878) 415.
| W. W angerin іш Engl. Bot. Jahrb. xxxviii. Heft 9 (1906) 6, 50.

THE STUDY OF THE EVOLUTION OF THE FLOWER. 273

biovulate ovaries, the ovules from their situation belong to different carpels, but the
manifestations of the second unit are confined to the additional ovule and a small
additional stigmatic lobe. Here, again, as in Sambucus, the dorsal portions of the
carpels have become a permanent portion of the ovary-wall and are non-variable. The
entity of the carpel is almost completely lost, so that the disappearance of the septa
must be regarded as only modifying the polymerous condition of the ovary ; the ovary
is not reduced to a single carpel.

The polymeric hypothesis outlined above receives contributory support from the
evidence that has been deduced from studying the course taken by the vascular tissue in
the flower. 'The principal bundles are clearly composite—that is to say, they exist or
existed in relation to more than one organ, as they do, for instance, in Viburnum. Thus,
in the upper part of the ovary, the bundles connected with the petals branch. These
branches are comparable to the carpellary bundles, which traverse the style in Sambucus
and Cornus; whilst the remaining alternate bundles occupy the site of stemonal bundles,
although stamens are absent (when а stamen, normally present, is absent, as in Aralia,
Davidia, and Helwingia, the corresponding bundle is absent also). Hence each of the
eight bundles, in Aucuba, is partly of сатреПагу origin. If the carpellary element in
each is alternately of “ventral” and “ dorsal” origin, in correspondence with the bundles
of Garrya, this would indicate a tetramerous ovary. But, on the other hand, the car-
pellary elements might have originated from a siugle carpel of the type found in Davidia
and the АтаПасеге, having several bundles. Тһе evidence afforded by biovulate ovaries,
however, is opposed to the latter possibility, since the appearance of an additional unit
is not accompanied by a corresponding increase in the number of carpellary branches
in the style—an increase which, according to Benjamin Clarke, happens with the
manifestation of an additional unit in Cornus.

If these speculations are correct, the flower was originally hermaphrodite and the
ovary polymerous—multilocular in its lower, uniloeular in its upper portion, as in
Leycesteria. By analogy with the Caprifoliacee and Hamamelidaces, but perhaps
prior to the attainment of epigyny, the loculi were pluriovulate—the ovules in the lower
portion axially, those in the upper portion parietally borne. Then the septa disappeared
and ihe ovules were numerically reduced, leaving only a single parietally suspended
ovule analogous to the single fertile ovule present in the loculi of Aralia and
Hamamelis.

Now it is evident, from the facts that have been enumerated for both Cornus and
Aucuba, that considerable differences exist between them. First, the ovule of Cornus
is suspended axially—a situation that is constant for the Caprifoliaceæ with terminal
ovules; whereas, in Aucuba, it is borne upon a parietal placenta—a placentation
apparently constant for the terminal ovules of the Hamamelidaceæ. This difference is
one of fact, and does not depend upon the hypothesis of origin outlined above.
Secondly, the nucellus in Aucuba exhibits distinctive characters, and in its primary
bulk, definite “calotte” epidermal cap, and subsequent girth-extension, owing to
cell-multiplication, is closely comparable to that of the Rosaceze.

282

974 MR. А. 8. HORNE—A CONTRIBUTION ТО

(d) GARRYA.
Morphology of the Flower and Ovule.

The plants of Garrya are diccious. Both the male and female inflorescences are
catkin-like. The flowers are situated in the axils of decussate bracts, either singly as
in G. Fadyenii and the majority of the species, or in groups of three as in G. elliptica,
The bracts are free in G. laurifolia, free and foliaceous in G. ovata, but connate at their
bases and foliaceous toward the lower end of the inflorescence іп 6. Wrightii.

The male flower of С. elliptica consists of an outer whorl of perianth leaves, an inner
whorl of four stamens which alternate with the perianth divisions and centrally situated
rudiments of an ovary. Payer*, who studied the development of the flower in
С. elliptica, states that two whorls of perianth leaves are present :—‘‘ Le périanth est
double et composé: l'externe d'un petit bourrelet à quatre dents dont deux sont latérales
et un peu plus grande que les deux autres situées, l'une en avant et l'autre en arriere ;
Vinterne, de quatre divisions alternes libres entre elles jusqu'à la base." Similar
observations were made by Baillon +, who also described the development of the flower
in this species. Harms discovered a small tooth-like process alternating with the
perianth divisions—also in 6. eiliptica,—which he identified as a rudimentary calyx leaf.
But Wangerin $, from a study of several species, points out that the projections observed
by Harms are homologous with bracts. Тһе writer has not reinvestigated the develop-
ment, but is unable to detect the presence “ a second whorl in the mature flowers
of G. elliptica.

Bentham and Hooker |, in * Genera Plantarum,’ and De Candolle, in * Prodromus, note
that the stamens are opposite the calyx lobes in some species of Garrya. “ Stamina
calycis lobis in aliis speciebus alternata, in aliis certe opposita." The information is
evidently derived from Bentham's Ч description of the new Mexican species, С. laurifolia
and G. ovata, in * Plantzte Hartwegian:e ' :—' Stamina in utraque specie laciniis perigonii
opposita, sic etiam ut videtur G. elliptica.” But no reference to opposite stamens occurs
in Harms's diagnosis nor in Wangerin’s monograph on the Cornaces. For this reason it
became eminently desirable to settle the point definitely. Accordingly, Dr. Rendle very
kindly allowed the writer to examine specimens of the male flowers of G. laurifolia
(one of the species in question) from the Herbarium of the British Museum of Natural
History. Two portions were dissected. Each consisted of a terminal flower and the
lateral axillary flowers situated immediately below. In both specimens the uppermost
bracts were small and resembled the perianth segments. Owing to this circumstance,
one of the terminal flowers (РІ. 28. fig. 22) appeared to consist of an outer whorl of
six perianth segments and an inner whorl of four stamens; of the latier, two were
opposite the perianth-like bracts (br). The other terminal flower (fig. 24) consisted

* J. B. Payer, * Lecons sur les Familles naturelles des Plantes? (1860-1865) 124.

T Н. Baillon in Bull. бос. Linn. Paris, i. (1877) 139. + Harms in Ber. Deutsch. Bot. Gesell. ху. (1897) 19.
$ W. Wangerin in Engl. Bot. Jahrb. xxxviii. Beibl. n. 86 (1906) 52-53.

1 Bentham е et t Hooker, беп, Pl. і. 951. 4 G. Bentham, Plant. Hartwegianw, 14.

THE STUDY OF THE EVOLUTION OF THE FLOWER. 275

of an outer whorl of five leaves and an inner of three single and one branched stamens
(st), so that it simulated a pentamerous flower. Опе of the stamens (st) was situated
opposite the bract. But the lateral flowers (fig. 25) of both specimens resembled those of
6. elliptica—the stamens were alternate with the perianth segments. Bentham seems to
have fallen into error through examining the terminal flowers of the inflorescence.

The male inflorescence of G. Thuretii, a hybrid between G. elliptica (male) and
С. Fadyenii (female), called in error 6. Macfaydiana by Carriére, discovered by
M. Gustave Thuret at Antibes about 1862, combines the characters of both parents
This point was overlooked by Carriére * in his descriptions published in the “Revue
Hortieole, of 1869 and 1879, and subsequently by Baillon +, who states that the male
flowers resemble those of С. elliptica. Тһе material upon which the following observa-
tions are founded was gathered at Kew from plants obtained from France about
the year 1896. Тһе “flowers” occur singly in the axil of each bract, agreeing thereby
with the female parent, but each stalk bears in reality three, more or less congenitally-
united flower-heads, so that there is a marked correspondence with the male parent.
There were from four to six perianth leaves in the specimens examined, and in some
cases two or three were congenitally united to form a lobed structure. The stamens
varied in number from four to twelve: the outer were usually of normal size, the
inner smaller and often rudimentary.

The female flower consists of a bicarpellary ovary terminated by two, almost free,
styles and surmounted in б. elliptica by a pair of minute folioles, which are situated
almost at right angles to the styles. Lindley T regarded the folioles as a calyx and they
are described as such by Harms $ in * Die Pflanzenfamilien. Baillon |, on the other
hand, points out that the terminal flowers of the inflorescence often bear on the sides of
the ovary near mid-height two bracts which alternate with the two preceding normally
orientated ones; for this reason he considers that the supposed sepals are nothing more
than displaced bracts. Wangerin investigated the matter in greater detail, and his
original figures show a complete series in the displacement of bracts, thus:—The folioles
are absent in б. gracilis (Wang., fig. 5, D) ; they are situated at about one-third the height
of the ovary іп G. Wright (Wang., fig. 4, D); they are depicted higher up still in
С. laurifolia (Wang., fig. 2, D), and terminal in Œ. Fremont (Wang., fig. 2, C). Having
examined herbarium specimens of several species of Garrya, notably С. Рейс and
G. laurifolia, the writer is able to confirm Wangerin's observations.

The development of the female flower of the hybrid, С. Thuretii, was described by
Baillon ** as follows :—* Sa fleur femelle est-elle representée au début par un mamelon
plein qui occupe l'aiselle d'une bractée de l'inflorescence. Bientôt sur се mamelon se
montre, à droite et à gauche, une feuille carpellaire. Toutes deux s'elévent en devenant
connées et constituent la cavité de l'ovaire, que leurs sommets couronnent ensuite en

* Е.А. Carriére in Revue Hort. 1869, 17, and 1879, 154.

t H. Baillon in Bull. Soc. Linn. Paris, i. (1877) 139.

+ J. Lindley, Veg. King. 2nd ed. 295 (1847).

$ Harms in Engl. u. Prantl, Nat. Pf., Teil iii, Abt. 8, 256. II H. Baillon, 7. c.

% W. Wangerin in Das Pñanzenr. Heft 41 (1910), Garryacee, 1-17. ** H. Baillon, 1.с. 140,

976 МВ. А. S. HORNE—A CONTRIBUTION TO

divergeant de deux branches stylaires latérales." The parietal placentz are terminal and
situated opposite one to the other, They are orientated laterally, not antero-posteriorly
as described by Baillon, but they may occupy either position in the species with singly
situated flowers, according as the single flower represents a median or a lateral member
of the axillary groups of three present in G. elliptica. The styles should rest normally
at right angles to the placentz, but they are slightly twisted from this position in
G. elliptica. They rest in the plane of the placentz in G. Рейс апа G. longifolia.

A single descending ovule with “ dorsal” raphe develops from each placenta. It
becomes ultimately tangentially orientated. The nucellus is invested by a single
integument.

The writer has been unable to follow the development of the nucellus, owing to the
lack of suitable flowers. Тһе first material examined consisted of fruits of the hybrid,
G. Thuretii. Later, fertile fruits of G. elliptica were obtained. Finally, some flowers
of this species were examined and the ovules removed and microtomed. These were in
ап advanced stage of development. Тһе embryo-saes were considerably elongated and
sterile, having escaped fertilization. Тһе basal part of the nucellus consisted of about
five rows of cells; so that, in all probability, the young nucellus is comparable in bulk
to that of Cornus.

Text-fig. 8.

Diagrams showing the course of the vascular tissue in the female flower of Garrya elliptica.
А, young fruit; B-F, flower. (For abbreviations see p. 308.)

Vascular Structure.
The vascular tissue of the pedicel consists of two opposite simple bundles (d) and two
groups of small strands, which alternate with the former. The small bundles of each

THE STUDY OF THE EVOLUTION OF THE FLOWER. 977

group unite to form опе (%), во that four (41, 42, m1, m2) traverse the ovary-wall
(Text-fig. 8, В). Each of the bundles m forms an ovular branch o (С), and then extending
upward divides to form two (E, %, т"), which become the opposite marginal bundles
of the free styles (Е); 41 and d 2 become the median bundles of the corresponding
styles (F).

The ovary corresponds closely to that of Hamamelis and other Hamamelidaceze, con-
sisting of two carpels with their respective traces, the adjacent marginal bundles being
eongenitally united.

Nature and Origin of the Flower.

Notwithstanding his interpretation of the morphology of Garrya, Lindley * found some
difficulty in assigning it a systematic position. Не first selected the “ Urtical Alliance."
but later decided in favour of the Helwingiaces, then recently established by Decaisne, .
and accordingly associated the monotypic family Garryaceze with the Helwingiacew to
form the alliance or cohort Garryales. Both families agree, writes Lindley, in possessing
an epigynous calyx and a minute embryo embedded in copious endosperm.

Garrya, bearing the hall-mark of epigyny, is henceforth compared with epigynous
forms, even by Baillon, who sceptically regarded Lindley’s calyx. Baillon t included
Garrya in his Согпасе». He supposed that the ovary, from an evolutionary point of
view, is intermediate in structure between those of Aucuba and Cornus. This Aucuba-
Garrya-Cornus series of his is analogous to that which he believed could be traced
between the Combretacez and bilocular forms of Jfarlea.

It is now evident that the true nature of the flower has long been masked, owing to
the vagaries of bracts. Bracts are inserted on the ovary-wall in several species, and in
the common species, б. е ір са, reach an epigynous position. Іп С. Fadyenii, where
the folioles do not exist, the perianth was supposed to have disappeared ; thus Endlicher 7
states, “perigonii in floribus pistilligeris vestigium nullum." Bracts, again, caused
Bentham to err when he described opposite stamens іп С. laurifolia: the stamens are
alternate with the perianth segments. Finally, the close proximity of a pair of bracts
to the perianth segments in the male flower led Harms astray, who believed that they
represented a calyx.

The loss of the epigynous calyx or perianth, which had been considered homologous
with the perianth of Cornace:e, involves a total reconsideration of the whole problem.

Wangerin, influenced by the amentaceous inflorescence, dicecism, apetalous and
hypogynous flowers, compares Garrya with Salix, Betula, Juglans, and Myrica, and
removes it from the Cornaceæ to the restored Garryaceze. This family, in Engler's
revised Syllabus, is re-established in the Garryales, now situated between the Salicales
and Myricales. Both Wangerin and Engler seem to have regarded certain characters
as of little consequence—for instance, the basal *orthotropous " ovules of Juglans and

* J. Lindley, Veg. King. 2nd ed. 295.
t H. Baillon, Histoire des Plantes, vii. 82.
1 S. Endlicher, Gen. РІ. Suppl. ii. 90.

978 MR. А. 5. HORNE—A CONTRIBUTION ТО

Myrica and the exalbuminous seeds of Salix, Juglans, Betula, and Myrica. The ovules
of Garrya are anatropous, and the seeds contain copious endosperm.

Again, since the embryology of Garrya is not yet understood, it would be wise
to reserve a final judgment until a more thorough comparison with the Amentiferz
could be made, the embryology in the latter group being already known from the work
of Benson and others.

Now genera with amentaceous inflorescences and diccious, apetalous, hypogynous
flowers occur also among albuminous families with anatropous ovules, and their ovaries
are comparable in many cases to those of Garrya. Thus the resemblance to the Hama-
melidaceze is comparatively close. The structure of the ovule forms the only considerable
difference. Garrya, in many features, is not unlike Scepa, now incorporated with
Aporosa, a genus of the Phyllanthez ; but the Phyllanthez are defined as possessing
ovules with a ventral raphe, which is not the case in Garrya. Lastly, Garrya may be
favourably compared with Forstiera, a genus of Oleacez.

(е) HELWINGTA.
Morphology of the Flower and Ovule.

The flowers are unisexual and seated upon the midrib of a foliose bract. The male
inflorescence consists of a group of separate pedicellate flowers, found to be either
trimerous or tetramerous, both kinds being present in the same inflorescence. Each
flower possesses a polyphyllous perianth and a single whorl of stamens, which alternate
with the perianth leaves. The female inflorescence comprises a single, minutely pedicel-
late, epigynous flower, which bears at its base a small foliole. The flower consists of an
inferior polycarpellary ovary terminated by free styles and surmounted by a single whorl
of perianth leaves, resembling those of the male flowers, which alternate with the stigmas.
Stamens, or the rudiments thereof, are entirely absent. Both whorls are usually tetra-
or pentamerous, but in one exceptional case a pentamerous ovary was surmounted by a
tetramerous perianth. The ovary is multilocular, each loculus containing a single,
terminal, axially suspended, descending, anatropous ovule with abaxial raphe. The
ovule is radially orientated, and since the nucellus of the mature ovule rests between
the axial portion of the ovary and the abaxial raphe, the orientation is the inverse of
that characteristic of the Araliaceze, where the ovules are also radially orientated.

Owing to its radial orientation, vertical sections of the ovule are easily obtained.
Such a section of the youngest ovule found is shown in Pl. 29. fig. 73. The
integument shows no sign of dual origin. The nucellus approximates closely to
the condition in Lilium, and consists of three short axial rows made up each
of two or three cells. РІ. 29. fig. 73 shows a single enlarged subepidermal archesporial
cell, of which the subsequent development has not been traced. The nucellus soon
assumes the vertical position, owing to the activity of meristem at its base, and by the
time the embryo-sac is formed the ovule has elongated considerably (Text-fig. 13, B).

THE STUDY OF THE EVOLUTION OF THE FLOWER. ` 279

This elongation is due to the activity of the lower cells of the nucellus, which, owing to
the continuous formation of transverse partitions, give rise to vertical tiers of cells, so

Text-fig. 9.

N
X

SC E
2 Е) 5; ч

v)

амы A B

Diagrams of transverse sections showing the course of the vascular tissue in the female flower

of Helwingia ruscifolia. (For abbreviations see р. 308.)

that the nucellus expands longitudinally and provides for subsequent extension. Mean-

while, the lower rim of the integument increases by cell-division to form oblique rows

(inf), whilst the upper rim obtains only a small extension; hence the integument,
SECOND SERIES.—BOTANY, VOL. VIII. 2T

280 MR. А, S. HORNE—A CONTRIBUTION ТО

as а separate organ, forms only a small portion of the ovule, as in Davidia. Its
boundary, /, is indicated in Text-fig. 13, B. Тһе embryo-sae, as shown in РІ. 29. fig. 74,
is immediately subjacent to the epidermis of the nucellus. Only one epidermal cell is
shown in the figure, and no trace of a cap formed from sterile megaspores nor traces of
parietal tissue are evident. Тһе embryo-sac itself exhibits no noteworthy departure
from the normal.

Vaseular Structure.

The vascular tissue of the pedicel in the female flower separates into а peripheral
and an axial (ax) series of bundles (Text-fig. 9, A). "There are ten outer and five
inner bundles in pentamerous flowers, but the number in each series varies with the
carpels; in the case of the exceptional flower noted above ( Helwingia, text, p. 278), the
bundle corresponding to the missing sepal was absent. Тһе bundles of the outer series
opposite the loculi (B, d) take an upward course, without branching, into the style (F)
and stylar lobes—one into each, corresponding to dorsal carpellary bundles; those
alternating with the loculi (B, se) extend upward also without branching and enter the
sepals.

The bundles of the axial series (B, ax) are opposite the loculi. Each one extends
axially as far as the corresponding placenta, forms three branches of which the median
becomes the ovular bundle, whilst the lateral branches (C, E, /) pass outwards, and
ultimately join the corresponding dorsal bundle (F).

Two interpretations of the vascular structure may be presented, and until more
extended investigations have been made claim equal attention. The first is suggested
by a comparison with Aralia and Nyssa. Ав in Aralia, the traces are free, and when
the number of carpels and perianth segments deviates from the normal the traces vary
correspondingly. А transition from hermaphrodite to female flowers occurs in Nyssa,
when both the stamens and stemonal traces disappear; so Helwingia, by analogy with
Nyssa, may have lost both stamens and petals, together with their traces, when epi-
супопв. The second and perhaps a more tempting solution is to suppose, by analogy
with Garrya, that the flower was already apetalous and unisexual in its hypogynous
state, in which case petal and stemonal traces have never existed in the ovary-wall.

Morphology of the Perianth.

Lindley * compares the perianth with the supposed calyx of Garrya. Bentham and
Hooker 4, on the contrary, hold that it is homologous with the corolla of Meryta, a genus
of the Araliacee. Both genera are placed side by side in the Araliacee for the following
reasons :—(1) The perianth leaves alternate with the stamens in both genera, and the
wstivation is valvate; (2) in Meryta both calyx and corolla are present in the male
flower, although the calyx is absent in the female flower. Bentham and Hooker deduce
that the calyx, already obsolete in the female flower of Мегуа, has entirely disappeared

* J. Lindley, Veg. King. 2nd ed. 295 (1847).
+ G. Bentham et J. D. Hooker, Gen. Pl. i. (1867) 940.

THE STUDY OF THE EVOLUTION OF THE FLOWER. 281

in Helwingia. Now Bentham and Hooker state that the Соғпасеге and Araliaces are
to be distinguished by the type of ovule prevailing in the family ; so they were evidently
unaware that the ovule, in Helwingia, does not conform to the Aralian type, with
ventral raphe, and hence could not belong to the latter family. If, then, the genera
are not closely related, the arguments bearing on the homology of the perianth diminish
in value.

Another solution of the puzzle presents itself, however. It will be observed that the
perianth leaves alternate with stamens in the male flower, corresponding in position to
petals; but in the female flower they alternate with the stigmas, and might represent
sepals. Now if the perianths of the male and female flowers are homologous, it is
conceivable that some anomaly exists with regard to the stamens. If it were supposed
that the flower was primitively ob- or diplostemonous, but has lost the inner or outer
whorl of stamens, as the case may be, then the perianth segments could be interpreted
as sepals and the orientation of the whorls in both the male and female flowers could be
explained. Among Согпасег a double whorl of stamens has been recorded in Nyssa
sessiliflora.

Comparison between Helwingia and Cornus.

Morren and Decaisne * established Helwingia in a new family, the Helwingiaceæ.
This arrangement was subsequently criticised by Baillon f, who states that Decaisne was
misled by artificial characters. Baillon points out that the female flower is constructed
very like the flower of Cornus, and the ovules of both genera are of the same type. He
considered that Bentham and Hooker, in placing Helwingia among the Araliaceæ, were
led astray owing to the numerous agreements in character and organization between
Helwingia and Meryta.

Now the ovules of Cornus (with dorsal raphe) are orientated tangentially—an orienta-
tion that appears to be constant for the Caprifoliaceze ; whereas those of Helwingia (with
dorsal raphe) are radially orientated, and the position taken up in the loculus seems to be
the outcome or result of a genetic difference in the ovule, as explained by the writer in
а previous paper. Hence the ovules of the two genera аге not necessarily homologous,
as considered by Baillon.

Again, the vascular systems do not correspond. That of Helwingia, owing to the
existence of separate perianth and dorsal carpellary bundles, agrees more nearly with the
Aralian type. The vascular supply to the ovule is different in the two cases. In
Helwingia each axial bundle divides to form three, of which the median enters the
ovule ; in Cornus, lateral branches from a pair of opposite peripheral bundles unite to
form the ovular trace. These details show that the resemblances between the two
genera are not so real as ВаШоп supposed, and that he himself was influenced by

structural parallelisms.

* Morren et Decaisne in Bull. Acad. Brux. v. (1836) 169.
+ Н. Baillon in Bull. Soc. Linn. Paris, i, (1877) 137.

282 ; MR. А. S; HORNE—A CONTRIBUTION ТО

(f) CoRoKTA.
Morphology of the Flower and Qvule.

The flowers of Corokia Cotoneaster are hermaphrodite, with epigynous pentamerous
whorls of sepals, petals, and stamens. The ovary is usually unilocular, sometimes

Text-fig. 10,

ос ыр

Diagrams showing the course of the vascular tissue in the flower of Corokia Cotoneaster.
(For abbreviations see p. 308.)

< bilocular. The style consists of two congenitally united units in bilocular forms.

we.

THE STUDY OF THE EVOLUTION ОЕ THE FLOWER, 283

stylar canal is cross-shaped (in bilocular forms), and upon tracing it downwards the
longer arm passes into communication with the loculi and the shorter arm is distinctly
sutural. The canal is sometimes three-lobed in unilocular ovaries.

The terminal ovules in bilocular ovaries are axially suspended, one in each loculus ;
but traces of parietal placentee, absent in Cornus, exist immediately above the point of |
insertion of the ovules (Text-fig. 10, G, pl). The ovules are radially orientated, as in
Helwingia. The orientation is not disturbed in relation to the absence of the septum.

Vascular Structure.

Several hand-cut series and one by the microtome were prepared through unilocular
ovaries, whilst two series by hand and one by the microtome were taken through
bilocular ovaries. The last-mentioned microtomed series will now be described. Six
peripheral bundles (owt ser) and one centrally-situated or axial bundle were present
below the loculus (Text-fig. 10, A). The peripheral bundles, branching and anastomosing,
increase upward to form an outer series of twelve bundles which are the traces of six
petals and six stamens, and some small inner branches (B-E). The majority of the
inner branches arise from the bundles near the septum on either side(D). They extend
into the septum, and strands from opposite sides (ov 1, ov 1’, and ov 2, ov 2') augmented by
the axial bundle (az) unite to form the ovular trace. Two bundles extend upward from
this level, but soon die out (Е). In another flower four bundles were present at a
corresponding level, one in each placenta (б, ov). Two of the above-mentioned inner
branches (E, d) extend upward and oceupy a dorsal position in the style (F).

In uniloeular ovaries, the. *axial" bundle is situated near the side of the loeulus
(Н, Т). It extends upwards and is joined by inner branches corresponding to those
marked ov in D, so that an irregular arc or collar of vascular tissue is formed (J, К)
from which the ovular trace arises. |

The change to the unilocular condition is accompanied by the displacement of the
axial bundle, and by the diversion of the vascular tissue pertaining to the — ovule
to the present ovule.

Comparison with Cornus.

Corokia was described by A. Cunningham * in 1839. Endlicher t at first placed it near
the Rhamnaces, but later f transferred it to the Согпасеге, and it remains in this family
in the systems of Bentham and Hooker, Baillon, Harms, and Wangerin. Corokia,
however, differs from Cornus in the following details:—(1) the presence of parietal
placentze immediately above the situation of the ovule; (2) the existence of an axial
bundle; (3) the method of vascular supply to the ovule, and (4) orientation of the
oyule. Hence the structural agreement is not quite so close as hitherto supposed.

* A. Cunningham in Ann. Nat. Hist. iii. (1839) 249.

t S. Endlicher, Gen. РІ. 1108 (1840).
+ S. Endlicher, Gen. Pl. Suppl. iv. pt. 3, 18 (1850).

284 МВ. А. 8. HORNE—A CONTRIBUTION TO

(9) MELANOPHYLLA.

This genus was first described by Baker * in 1886. Two species— М. alnifolia and
JM. aucubifolia—were recorded. According to Baker’s description, the flowers are
hermaphrodite with alternate pentamerous whorls of sepals, petals, and stamens. Тһе
ovary is bi- rarely trilocular and the styles bi- or trifid. Тһе loculi are uniovulate,
Baker t described a new species, M. crenata, in 1896 with illustrations. Figures from
specimens of this species in the Herbarium Boissier are given by Wangerin £ in * Das
Pflanzenreich. Only опе loculus is shown in both Bakers and Wangerin’s
illustrations.

Baker, in 1886, states that two species of Melanophylla are comprised in Humblot’s
collection from the north-east portion of the island of Madagascar. - Now in * Plantes
de Madagascar’ $, a new species, M. Humbloti, is figured which is evidently one of those
mentioned by Baker. This species is not recorded in ‘Das Pflanzenreich’ and no
written description seems to have been published, but it is mentioned in Palacky’s ||
‘Plantarum Madagascariensium. According to the figures of М. Humbloti the
stigmas, which ате tongue-shaped, are not recurved as in Baker's species—the calyx is
depicted with a waved rim and not with definite segments as in JM. crenata, and there
are slight differences in the ovary and the stamens. Only one loculus and one ovule
are figured.

Wangerin associates Мейалорһу а with Griselinia to form a new tribe of Cornoidez,
the Griselineze.

(h) GRISELINIA.

Morphology of the Flower and Ovule.

The plants and flowers of Griselinia are unisexual. The male flowers of G. littoralis
and G. lucida consist of sepals, petals, and stamens in alternating pentamerous whorls,
which surround a flat circular dise with five or six grooves radiating outward from
the centre. Ovarian rudiments are absent. The female flowers are epigynous, and
surmounted by sepals and petals in G. littoralis and by sepals only in G. lucida. The
petals are imbricate. Stamens are absent in all the species. Bentham & Hooker Ч
and Wangerin ** state that there are no rudiments of stamens, and none have hitherto
been recorded. The writer finds not only rudiments of stamens in G. littoralis and
G. lucida, but rudiments of the absent petals in G. lucida. These are very minute, but
they can be observed with the aid of a good pocket-lens. Text-fig. 11, K, represents a
transverse section of G. littoralis; here the positions of the petals are denoted by

* 7.6. Baker in Journ. Linn. Soc., Bot. xxi. (1886) 352.
t J. G. Baker in Hook. Ic. Pl. xxv. (1896) pl. 2499.
t W. Wangerin in Das Pflanzenr. Heft 41, 100.
$ Н. Baillon іп“ Plantes de Madagascar,’ Atlas, iii. pl. 400 (1896).
i Palacky in Plant. Madagase. fasc. iv. (1908) 50.
Ч G. Bentham ct J. D. Hooker, Gen, Pl. i. 951.
** W. Wangerin іп 1, c. 95.

THE STUDY OF THE EVOLUTION OF THE FLOWER. 285

pe, and the rudiments, five in number, occur just within the sepals, occupying the
position of stamens (7 st). The rudiments of stamens and petals in б. lucida are

Text-fig. 11.

Diagrams showing the course of the vascular tissue in the flower of Griselinia: A-J, б. lucida; K, G. littoralis.
(For abbreviations see p. 308.)

Shown in J. The ovary is terminated by three styles. It is unilocular or bilocular
according to Bentham and Hooker*. Wangerin + describes only one loculus. The

* G. Bentham et J. D. Hooker, Gen. Pl. i. 951. T W. Wangerin in Das Pflanzenr. Heft 41, 95.

286 МВ. А. 8. HORNE—A CONTRIBUTION TO

writer finds по bilocular ovaries іп his material of G. lucida. The ovular mamelon
grows downwards and develops into an ovule of the abaxial type. The orientation in
bilocular ovaries has, of course, not been observed. Тһе nucellus, which is com-
paratively long, is of very small bulk—a median section shows three rows of cells within
the nucellar epidermis ; it is invested by a single integument (РІ. 29. figs. 75, 76).

Vascular Structure.

Serial sections of the flowers of both 6. lucida and С. littoralis were prepared. Іп
G. lucida six or seven bundles were present at the base of the ovary (Text-fig. 11, A).
These branch to form a single series, consisting of a variable number of bundles—
twelve or more—in the ovary-wall (B-F). Some of the smaller bundles extend inward,
and in the neighbourhood of the placenta anastomose, presenting in thick sections the
appearance of an incomplete vascular collar (F-I); from this “collar” as many as five
separate bundles, of which three (ov 1, ov 2, ov 3) are shown (F), extend into the
© placenta" and unite there to form the single ovular bundle. Since the petals and
stamens are absent in this species, there are no outgoing branches to these organs,
and all the principal bundles, together with a few which proceed from the “ collar,”
extend into the basal portion of the style (I, J), diminishing in number until only а
few remain. The vascular structure is similar in б. littoralis (petals present), except
that the bundles are more numerous—nine at the base of the loculus and sixteen
higher up in one specimen.

The vascular structure in the neighbourhood of the placenta recalls the somewhat
similar collar-like arrangement of vascular tissue in the reduced unilocular ovary of
Corokia. In Griselinia, however, branches from every part of the ovary-wall contribute
to the bundles entering the “ placenta."

Nature of the Ovary.

Baillon considered the ovary of Griselinia to be incompletely tricarpellary, the body
of the ovary consisting of the lower portion of one carpel. But the peculiar vascular

structure justifies a modification of this view. When we consider the cases already
_ discussed, it is found that where the terminal ovule is by origin axially situated the
vascular tissue presents an anomaly, but where it is parietally situated by origin the
conditions are not the same and the vascular tissue need not necessarily be diverted
from the normal course. Thus, in Viburnum, the ovary was primitively septate, but,
owing to the loss of the septa, the vascular tissue that had originally an axial extension
became diverted to a peripheral position. Again, in Corokia, where in bilocular forms
the ovules are suspended from the septum, the absence of a loculus disturbs the normal
disposition of the vascular tissue in relation to the ovule. But in Aucuba, also originally
septate, where the оуще is suspended from a parietal placenta and comparable to the
parietally situated ovules in the Combretacese, no vascular anomaly exists in relation
to the position occupied by the ovule. Now, in Griselinia, the placenta is abnormally
large, and the numerous branches contributing to the ovular trace arise from bundles

THE STUDY ОЕ THE EVOLUTION ОЕ THE FLOWER. 287

extending in all parts of the ovary-wall, and appear to have been diverted from the
original disposition. Hence not only сап we deduce that the ovary was primitively
septate, but also that the ovule was axially placed.

Now, abortion, in Viburnum, originally tricarpellary, has conduced to the special
development of the body of one carpel, a fact that is evident from the asymmetrical
disposition of the vascular tissue in the flower and from the arrested loculi. But, in
Griselinia, as also in Corokia and Aucuba, the original course taken by the floral
vascular tissue does not appear to have been affected by abortion. Hence the conclusion
is reached that the septa disappeared without a special enlargement of the lower portion
of any one carpel.

The writer’s interpretation of the floral structure may be summarized briefly as
follows:—The flower was primitively hermaphrodite, and became unisexual after the
attainment of epigyny; it exhibits an apetalous tendency, definitely expressed in
G. lucida, the ovary was originally tricarpellary and trilocular, and the loculi, by
analogy with the Caprifoliacese, were at first pluriovulate. Subsequent abortion resulted
in the disappearance of the septa, but did not induce the special enlargement of the
lower portion of any one carpel as in Viburnum; concurrently, the ovules were
numerically reduced to one, the equivalent of an axially suspended ovule in the
primitive ovary. ІҒ Baker's record of bi- and trilocular ovaries in Melanophylla be
correct, we may conceive that Griselinia has proceeded а stage further than J#elano-
phylla in the direction of abortion.

($) NYSSA.

Morphology of the Flower and Ovule.

The flowers of Nyssa are hermaphrodite or unisexual. The male flowers consist of
sepals, petals, and stamens, which vary from five to an indefinite number. According to
Bentham and Hooker *, ovarian rudiments are absent or a subulate process exists in the
centre of the disc. The female flowers consist of an inferior ovary surmounted by a
calyx—petals are absent or minute and there are no rudiments of stamens—and
terminated by one or two styles. The flowers of М. sylvatica, growing at Kew, which
the writer examined, proved to be either hermaphrodite or female. The hermaphrodite
flowers possessed one, two, or three stamens, so that a complete series existed between
the two kinds. ;

Bentham and Hooker + state that the ovary is unilocular, but Baillon $ records that
it may be sometimes bilocular. The writer has observed only unilocular ovaries in
N. sylvatica. The ovule is terminal in mature ovaries, but this is not strictly so in very
young flowers; in one particular case measurements were made with the aid of a series
of sections, each “015 mm. thick; the height of the loculus was 0:24 mm.; the total
length of the ovule 0:165 mm. Тһе ovule was attached to the ovary-wall at a point

* G. Bentham et J. D. Hooker, Gen. РІ. 1. (1867) 952.
+ С. Bentham et J. D. Hooker, l. с. i Н. Baillon, * Histoire des Plantes,’ vi. 267.

SECOND SERIES.— BOTANY, VOL. VIII. 20

288 МЕ. А. 8. HORNE—A CONTRIBUTION ТО

about 0:08 mm. from its own summit and the top of Ше loculus. It is of the adaxial
type and comparable to that of Davidia. Wangerin states that two integuments are
present, but the writer finds only one (РІ. 29. fig. 77). The nucellus is of small bulk

Text-fig. 12.

out ser

Diagrams showing the course of the vascular tissue in the flower of Nyssa sylvatica.
A-G, one flower; Н, I, another flower; J, fruit. (For abbreviations sce p. 308.)

ami comparable to that of Davidia, but its further development has not been followed,
Owing to the lack of suitable material.

* W. Wangerin in Das Pflanzenreich, Heft 41, p. 9.

THE STUDY OF THE EVOLUTION OF THE FLOWER. 289

Vascular Structure.

Serial sections through very young flowers and hand-sections of sterile “fruits” were
prepared from only one species—W. sylvatica.

In one flower six bundles were present in the lower portion of the ovary (Text-fig. 12,
A), these extended upward toward four petals, pe, and two stamens, в/ (В-Е); in another,
nine bundles occurred (H), but only seven of these supplied stamens (three) and petals (1).

An inner series of fine branches originates from the cylinder of bundles near the
point of attachment of the ovules (D, E) and extends into the style (F). The branch d
occupies a dorsal position in the style, but the flowers were too young to show the
vascular connections of the ovule, hence the nature of the remaining branches could not
be determined. From sections of fruits it appears that two or three strands unite to
form the ovular bundle (J). |

Owing to the variability in the flower іп the genus, а comparison needs to be made
with the remaining species before the significance of the vascular peculiarities can be
adequately discussed.

Comparisons.

Endlicher * placed Nyssa in а new family, the Nyssacez, near the Santalacez.
Baillon ł transferred it to the Combretacew for reasons similar to those which influenced
him in attaching Alangium to this family. But the occasional presence of two styles,
one of which is rudimentary in N. sylvatica, and the recorded case of a second incomplete
loculus indicate that the ovary is unilocular by abortion. Bentham & Hooker { and
Harms$ include Nyssa among the Cornace:e, but it differs from Cornus in possessing the
Aralian type of ovule, in its variable flowers, imbricate corolla, free styles, and reduced
ovary. Wangerin | separates it from the Согпасе on the ground that the ovule is
bitegumentary, but since the ovule possesses only one integument this criterion fails.

Nyssa approaches Aralia in the form of ovule (adaxial). But the ovules of Aralia
ате parietally borne, whereas the ovular bundles of Nyssa, by analogy with Griselinia,
might indicate that the ovule was originally axially suspended. A closer structural
agreement exists between Nyssa and Davidia, both genera possessing polygamous

flowers, free styles, and numerous stamens.

SuMMARY.
1. The following details of floral morphology are recorded :—
a. Тһе occasional occurrence of hermaphrodite flowers in Aucuba himalaica.
b. Biovulate ovaries in A. japonica and himalaica.
с. Anomalous flowers in the hybrid Garrya Thuretii—these consist of three
congenitally united flower-heads borne upon one “ pedicel,” thus combining

* 8. Endlicher, Gen. РІ. 328. + H. Baillon, Bull. бос. Linn. Paris, i. (1877) 271.
+ Bentham et Hooker, Gen. Pl. i. 952. § Harms in Engl. u. Prantl, Nat. Pfl. Teil iii. Abt. 8, 257.

| W. Wangerin in Das Planzenr. Heft 41, 9. dvi
U

290 MR. А. S. HORNE—A CONTRIBUTION TO

the characters of both parents, viz., 6. elliptica in which the flowers oceur in
axillary trimerous groups, and G. Fadyenii in which they are situated singly
in the axil of each bract.

d. Kudiments of a single whorl of stamens in the female flowers of Griselinia
littoralis and lucida and of the absent petals in G. lucida.

е. Rudimentary terminal parietal placentz in Corokia Cotoneaster.

2. The terminal ovules are suspended axially in Davidia, Cornus, Helwingia, and in
bilocular forms of Marlea and Corokia; probably axially by origin in Nyssa and
Griselinia, and parietally borne in Aucuba and Garrya.

3. The ovule exhibits considerable variety in organization :—

а. It is in Nyssa and Davidia of the dorsal type (raphe ventral or adaxial) with
radial orientation ; of the ventral type in Helwingia and Corokia (raphe dorsal
or abaxial) with radial orientation, but the inverse of the former; and of the
ventral type with tangential orientation in bilocular forms of Cornus and
Marlea.

0. The nucellus in the genera investigated (including Davidia and Nyssa deseribed
as bitegumentary by Harms and Wangerin) becomes invested with a single
integument.

с. The integument is completely free from the nucellus in Aucuba and Griselinia,
but in the specialized ovules of Davidia and -Helwingia is only found as а
separate organ towards the summit of the nucellus.

d. The nucellus consists of about seven axial rows of cells in Aucuba; four or five
in Cornus, Davidia, Griselinia, and Garrya ; and three in Helwingia, where it
approximates to the condition obtaining in Liliwm and Inula.

e. The nucellus of Aucuba, in the formation of a definite layer of parietal tissue, a
small epidermal cap, and in subsequent bulk-extension owing to cell-division,
approaches the type described by Péchoutre for the Rosacez.

J. The germination of more than one megaspore, already recorded by the writer
for Davidia, occurs also, but not to the same extent, in Aucuba.

4. The floral vascular systems of (һе Согпасеге do not conform to one type as in the
Caprifoliacee. It is in Cornus of the Caprifoliacean type. The systems of Griselinia,

| > Nyssa, апа Helwingia exhibit Araliacean tendencies; those of Davidia and Marleu are
. transitional in character between that found in Aralia and the synthetic systems of

Cornus and the Caprifoliaces. Тһе stage of economical advance reached by Viburnum
has not been attained.

а. Infra-locular extensions of vascular tissue occur in Aucuba and unilocular forms
of Marlea.

б. The ovular trace is anomalous in Griselinia, Nyssa, and unilocular forms of
Corokia. |

с. Branches from the ovular trace traverse the peripheral region of the integument
in Дала and Corokia.

THE STUDY OF THE EVOLUTION OF THE FLOWER. 291

5. The flowers of certain genera admit of the following interpretation :—

a. Cornus: Тһе ovary is bilocular by reduction from а quadricarpellary primitive
type. It is analogous to that of Sambucus. :

b. Alangium: The ovary is unilocular by reduction from the bilocular type found
in Marlea begonifolia, and reduced from a form primitively polycarpellary.

c. Aucuba: The flowers have become unisexual since the attainment of epigyny.
The ovary was primitively quadricarpellary and quadrilocular. The septa
subsequently disappeared, but the lower portion of the carpels remained
unmodified. Тһе ovules, originally numerous, became numerically reduced
until only a single parietally suspended one remained.

d. Garrya: Тһе flower is hypogynous as already noted by Wangerin.

e. Griselinia : The flowers have become unisexual since the attainment of epigyny,
and a tendency towards apetaly finds its expression іп G. lucida. The ovary
was primitively tricarpellary and trilocular, at all events іп its upper portion,
but the septa disappeared, leaving the lower portion of the carpels unmodified,
but involving the displacement of an ovule that was primarily axially situated.

6. Progressive abortion in the ovary reaches its limit in this family—the ovaries of
Davidia and Helwingia ате multilocular, the change from the bi- to the unilocular
condition can be traced in Corokia, and the ovaries of Aucuba, Nyssa, Alangium, and
Griselinia are unilocular.

7. Progressive sterilization has proceeded still further than in the Hamamelidaceze
and Caprifoliacez, nearly all the genera have reached the level of the extreme cases іп
these families, and the ovary or loculus is absolutely uniovulate.

8. The general series that has been traced in the direction of the loss of carpellary
autonomy culminates in Awcuba, where, owing to the absence of septa and to the
parietal situation of the ovule, the gynzecium appears to consist of a single carpel.

9. Both the Cornaceze of Harms and the Cornaceze of Wangerin are heterotypic,
but this fact is masked, owing to parallel development and convergent organization.

5. FINAL DISCUSSION.

THE INTERPRETATION OF RESEMBLANCES AND DiFFERENCES.—Hitherto that which
may be called the Principle of Resemblances has held amazing sway. Drawing са
breath, as it were, іп ће days of Ray, virile before the birth of the * Origin of Species,
it has evolved through the length of two centuries, and reached its zenith as presented by
Adolf Engler in * Die Natürlichen Pflanzenfamilien ” just over twenty years ago. Almost
on the eve of the publication of Engler's mighty work, a new era in botanical history
had dawned. In 1889, De Vries’s ‘ Intracellulare Pangenesis’ had been given to the
world, an essay which Bateson * declares “ remarkable as a clear foreshadowing of that
conception of wnit-characters which is destined to play so large a part in the development
of genetics,” to be swiftly followed by ‘ Die Mutationstheorie "+; whilst the year 1900
brought with it the triple rediscovery of Gregor Mendel.

* W. Bateson, Mendel’s Principles of Heredity,’ 5 (1913).
+ H. De Vries, ‘ Die Mutationstheorie ’ (1901-3).

292 MR. А. 8. HORNE—A CONTRIBUTION TO

As the incidence of the ‘Origin of Species’ shed transcendent light upon the
mystery of the Natural System in 1859, so the work of these great protagonists in
experimental evolution, and the expositions emanating from the schools of experiment
and thought founded by them, cannot fail in time to illuminate Systematic Botany
and inaugurate a new phase in the evolution of the idea of the Natural System.

De Vries by а rigid analysis of Gnothera separated numerous diverse forms breeding
true, some exhibiting comparatively great differences. Тһе Mutation theory postulates
the evolution of forms progressively divergent. Hence it is conceivable that forms—for
instance, genera grouped in different families—are in some cases comparatively closely
related. Again, the characters exhibited by De Vries's mutant GEnotheras are paralleled in
other mutating races—for example, giantism in GZnothera gigas also occurs in Lathyrus
odoratus (Bateson, 1909), Pisum sativum (Mendel), and in Primula sinensis (Keeble,
1912). Facts of this kind serve to redirect attention to the importance of the doctrines
of parallel development and convergence in relation to floral evolution, especially since
the surviving natural races trend in the direction of synthetic and economic advance.
Convergence in reproductive characters may occur even more frequently than has yet
been contemplated, so that like forms, such as genera in the same family, may be less
closely related than others are that differ more pronouncedly. Hence the Principle of
Resemblances must now be qualified by a greater regard for the significance of differences,
even if it be not shaken to its foundation by the doctrine of convergence and the
existence of insidious parallelisms.

Before proceeding to attempt to solve the phylogenies of the families considered here,
it will be convenient to review the evolutionary processes that have been discussed in
relation to the results obtained by the experimentalists. But the prospect is by no
means encouraging, since De Vries* writes: * The great lines of evolution of whole
families and even of genera and of many larger species obviously lie outside the limits
of experimental observation. They are the outcome of the history of the ancestors of
the present types, and a repetition of their history is far beyond human powers."

The carpels in Davidia vary from five to ten, with a mean number of six or seven;
they vary in number in the forms of Hedera Helix; and, in Sambucus, species are
characterized by five, and a range of from two to four with a mean of three respectively.

Mutation in the carpels has not yet been specially mentioned as occurring in the races
produced by experiment.

Syncarpy supervenes upon apocarpy, and the advance towards зупсагру is accompanied
or followed by that towards perigyny and epigyny, but here again there is as yet little
direct experimental evidence.

The most fascinating phenomena are those attendant upon sterilization. A vast
number of instances are known in which sterility is probably due to seasonal conditions,
or to diversion or poorness of the nutrient supply, including the very numerous cases
where only few of very many flowers develop in the inflorescence, - Davidia, where all
_ the ovules in the ovary may be fertile or sterile, may be included in this category.
_ Thence we pass to numerous other cases in which it is doubtful whether sterility has
* H. De Vries, ‘ Species and Varieties ’ (1905) p. 993,

THE STUDY OF THE EVOLUTION OF THE FLOWER. 293

become definitely fixed or not; as in Lonicera, where the majority of the ovules
habitually abort ; in Polyosma, where, according to Blume, only one of the numerous
ovules forms a seed; in Liquidambar, where the lowermost ovules are stated to be
always sterile; in Aralia and Corylopsis, where one of a pair, and in Quisqualis, where
one of three ovules is sterile. And, finally, cases occur, exemplified by Symphoricarpus,
where the sterility of particular ovules is no longer a matter of circumstance, having
become fixed and hereditary.

Sterilization proceeding to an extreme leads to unisexuality ; thus in the Hamameli-
daceze series exist between forms in which the ovary may or may not be sterile, to
others with definitely polygamous flowers and others where the unisexual condition is
attained; in Nyssa, series сап be traced between hermaphrodite and female flowers
owing to the loss of stamens; in Griselinia, Aucuba, and Garrya rudiments of the absent
sex are retained, but the last traces have disappeared in Helwingia and in the male
flowers of Davidia.

А great deal of light has been thrown on the series just enumerated by the results
obtained in the experimental fields of enquiry. De Vries*, by his analysis of
ever-sporting varieties, was able to show how profoundly the environment affects
development. Неге the primordium of the flower in Polycephalic Poppies was
influenced by changed conditions applied within the fifth or sixth week after the
beginning of germination, and fewer anomalous stamens were produced than under
normal conditions, whilst the occurrence of polycotyly and polycephaly in polycotylous
and polycephalous half-races did not become definitely fixed during cultivation and
selection. The phenomena concerned variability, not mutability. Нет also found that
the development of the anther and pollen in (Z. scintillans was dependent on external
conditions to a high degree. Пе Vries's + Gznothera mutants show clearly grades of
sterility. Some are rich in pollen, whilst scintillans is poor; others form many seeds,
some few, as in elliptica, whilst in breristylis many plants set no seed at all, but he
believes the degree of fertility in many mutants will ultimately be found to follow
Quetelet’s law of fluctuating variability. Finally, with regard to sex sterility,
unisexuality is exhibited by ever-sporting varieties, mutants, and Mendelian. hybrids.
R. P. Gregory § records that no stamens occur in one of the double forms thrown off by
Primula sinensis; and Е. В. Saunders | has shown that the gynccium in double
Petunias is more or less deformed and when fertilized yields по seed, whilst double $
stocks are always sterile and are always obtained from seed set from ever-sporting
singles. De Vries ** obtained а male sterile mutant, GZnothera lata. Lastly, Bateson Tt,
Punnett, and Saunders, working with Sweet Pea, and Salaman 11, with Potato, have
isolated male-sterile Mendelian forms.

+ Н. De Vries, ‘ Species and Varieties’ (1905) 369, and * Matation Theory,’ ii. (1910) 343, 450.

+ H. De Vries, ‘ Mutation Theory,’ i. (1910) 383.

$ Н. De Vries, 1. c., scintillans, 383 ; elliptica, 396; brevistylis, 316.

§ R. P. Gregory in Journ. Gen. i. (1911) 90. | E. R. Saunders in Journ. Gen. 1. (1911) 63.
q Е. В. Saunders in Journ. Gen. i. (1911) 304. ** H. De Vries, L с. 409.
Tt Bateson, Punnett, & Saunders in Rept. Govt. Com. Roy. Soc. iv. 16,
tt Redcliffe N. Salaman, in Journ. Linn. Soc., Bot. xxxix. (1910) 301.

294 i MR. A. 8. HORNE—A CONTRIBUTION ТО

The facts that the number of chromosomes may vary in mutants, as recorded by
Gates * (1907) and later by Anna M. Lutz f (1907) for Œnothera, and that giantism
is accompanied by increase іп cell-size as first shown by Gates { ((nothera), and later
by Keeble § (Primuia), show the depth as well as breadth of the new aspect of things.
It is hardly possible, therefore, to exclude the ovule from the possibilities of mutation.
The change and grades of change from orthotropy to anatropy, the origin of the single
and double integument, and even the change from many to few axial rows may yield in
turn to patient and persistent experimental endeavour.

We are now able to conceive the origin of mutants in floral characters. Mutants
might arise, for example, from apocarpous forms bearing some few and some many
carpels. These could in time produce new species. If both the floral and vegetative
organization alter considerably during the continuance of mutation, forms would arise
exhibiting still greater divergence. But the advent of syncarpy and epigyny again and
again in different families, and the frequent occurrence of sterilization in both Mono-
cotyledons and Dicotyledons, culminating time after time in the evolution of uniovulate
ovaries, whilst increasing enormously the tale of floral diversity, nevertheless contribute
toward the evolution of synthetic and convergent structures.

Овтвтхв.--Тһе interpretation of phylogenies must be influenced by views of origin,
and the origin of new forms, as postulated by the Mutation Theory, is to render the
recognition of affinities more problematic than ever. It is the custom to sort plants
into divisions and groups according to this or that bias, and to construct a phylogenetic
twig or branch of the great tree of descent ; but the Mutation Theory brings home to
us that that which is actually accomplished in Systematie Botany is the aggregation or
linking of species, genera, and even families, as the ease may be, that have attained a
similar level in evolution, a similar resultant effect of racial experiences. These aggre-
gations may or may not be related. The Composite, for example, form a homogeneous
group consisting of genera in the same phase of synthetic and economic development ;
but is it certain that the members of the family are bound by the bonds of relationship,
and are traceable to a single progenitor ?

A general device in Comparative Morphology is to seek for clues to an interpretation
of the origin of a structure from morphological series, and thence deduce a concept of
affinity. But spacial series may not necessarily coincide with series in time, as
exemplified by the sterilization series in the Caprifoliaceee and De Vries's побһегав,
where one form has not passed into another form, but each one has originated separately.
Some points, however, carefully chosen, may establish unlooked-for differences. Thus
the situation of the sole surviving seed is constantly parietal in the Hamamelidacez and
axial in the Caprifoliacee. Hence in the Согпасезе, where genera occur with either the
parietally or axially borne seed, there is a stron g bias towards the use of this character
as an indicator of different origins. Again, the tangentially and radially orientated

* В. R. Gates in Bot. Gaz. xliii, (1907) 81-114. Т Anna М. Lutz, Science, xxvi. (1907) 151.
+ В. В. Gates in Arch. f. Zell. iii. (1909) 525-552. § F. Keeble in Journ. Gen. ii. (1912) 163.

THE STUDY OF THE EVOLUTION OF THE FLOWER. 295

ovule, as exemplified by Cornus and Helwingia respectively, may prove a useful
character, the sympathetic development of the ovary in relation to the form of ovule
having evidently evolved independently in each genus during the period of abortion
through which the ovary has conceivably passed.

PHYLOGENY OF THE HAMAMELIDACEAE. — Robert. Brown* considered that the
Hamamelidaceze possessed affinities with Bruniacee on the one hand and the
Cornacez on the other. Affinity with the Bruniaceze and БахИтасасезе is more
nearly consistent with well-balanced modern opinion.

Lindley t placed Altingia and Liquidambar between the Betulacez and Salicacee.
Hallier £, in 1905, reverts to this affinity, conceiving the descent of the Betulaceze from
the Hamamelidacez. This plausible supposition is not supported by the embryology,
and the endosperm as a crucial character has been too lightly regarded. L. S. Gibbs ў
suggests that the Alsinoidez illustrate an intermediate stage in the development of the
exalbuminous seed by a progressive reduction in the functions of the nucellus; and
Péchoutre's | investigations point to a similar tendency in the Rosacee. But even
if the evolution of the exalbuminous seed has occurred in several families, it scarcely
follows, especially when the embryology is different, that the Betulacee are virtually
exalbuminous Hamamelidoidee. Тһе families are probably convergent, with separate
and distinct lines of origin.

The Hamamelidacez, whilst exhibiting diverse characters of floral organization, form
nevertheless a fairly homogeneous group. 1% is difficult to subdivide the family into
well-defined sections, and every criterion hitherto used has failed. The characters are in
harmony with Saxifragacean characteristics, and the Hamamelidacee pass presumably
into a pro-Saxifragacean plexus forming an older line originating perhaps from more
than one progenitor.

PHYLOGENY oF THE CAPRIFOLIACEJE.— The Caprifoliaceze are remarkable, owing to а
singular correspondence in several important and deep-seated characters—the epigynous
ovary, the biseriate ovules, the evolution of an axially suspended seed owing to abortion,

and the tangentially. orientated and similarly organized ovule. Неге, again, the
characters that have been adopted to separate the genera into groups or even families
have proved futile, except in the case of Adoxa.

Diervilla temptingly recalls Polyosma, but the resemblance reveals divergencies from
the Caprifoliacese rather than a possible prototype; the differences, however, are not
incompatible with affinity on the Saxifragacean side.

Lindley § suggested a possible connection between Sambucus and Hydrangea, partly

* Robert Brown, ‘ Characters and Descriptions of Three new Species of Plants’ (Lendon, 1818).
+ J. Lindley, Veg. King. 2nd ed. (1847) 253.
+ Hans Hallier in New Phyt. iv. (1905) 151.
$ І. S. Gibbs in Ann. Bot. xxi. (1907) 51.
|| Е. Péchoutre in Ann. Sci. Nat., Bot. 8™¢ sér, xvi. (1902) 156.
@ J. Lindley, Veg. King. 767.
SECOND SERIES.—BOTANY, VOL. VIII. 2x

296 МВ. А. 8. HORNE—A CONTRIBUTION TO

influenced by the sterile ray-florets in both genera. Неге, although the reasons given
may now be thought unsound, and the differences between the types compared are
greater than in the former case, a second reasonable line of descent on the Saxifragacean
side may be contemplated.

The organization of the ovary of Leycesteria is not far removed from that of Pyrus,
whilst biseriate ovules are found in Spiræa, but the structure of the ovule is not com-
patible with direct descent on the Rosacean side. Leycesteria perchance originated
from an issue of the pro-Rosacean plexus, in which the ovule itself progressed and the
biseriate rows, reduced in Pyrus, were retained.

PHYLOGENY OF THE CoRNACEX.—The Cornacez of both Harms and Wangerin are
polyphyletic, and the family consists chiefly of outlying forms, representatives of stocks
that have become obsolete.

There is no warrant from recent discoveries for the recession of Garrya to the vicinage
of the orthotropal cohorts. Garrya is more reasonably linked by slender and far distant
threads to the Bruniacee and Hamamelidace:e. :

Aucuba is probably a scion from pro-Rosacean stock. Неге the descending form of
ovule found in Pyrus aud Prunus has become specialized, and the ovary considerably
reduced. |

It is curious that, whilst this morphological analysis dissociates Garrya and Aucuba,
the most successful reciprocal effects in cross-grafting have been obtained with these
genera. Owing to the kindness of Sir Harry Veitch, the private records of a long series
of experiments, records of successes and failures in cross-grafting, have been placed at
the writer’s disposal. Here, whilst the successful grafts are nearly always between
genera of the same family, it must be admitted that failures between nearly related
genera are numerous. Іп one case alone a graft succeeded, and that only for a time, іп
which genera of different families were used—the reciprocal grafting of Ilex Perneyii and
Buxus balearica. But genera from different families were infrequently chosen for
experiment. Тһе following reciprocal grafts failed:—Cornus and Aucuba, Griselinia
and Aucuba, Sambucus and Viburnum, Loropetalum and Hamamelis, Distylium and
Sycopsis, Nyssa and Aucuba, Aralia and Hedera.

Helwingia is difficult to relate. It may be, perhaps, regarded as an epigynous repre-
sentative of a parallel line to the Sapindacee.

Davidia and Nyssa are isolated forms abutting on the Araliaceze, but scions from
different progenitors belonging to a pro-Araliacean stock. Неге obtains a specialized
form of the Rosacean ascending ovule ; so that this plexus might extend through issues
now extinct into the primitive Rosace:e.

Alangiace:e still await a keen scrutiny of the ovule, and leave the respective claims
of the Cornacee, Combretaces, and Polyosma unsatisfied. Alangium appears to be a
reduced analogue of Davidia, possessing the Cornean type of ovule.

Corokia and Griselinia are provisionally regarded as outlying forms to Cornus that
have had an independent origin.

THE STUDY ОЕ THE EVOLUTION OF THE FLOWER. 297

VASCULAR STRUCTURE OF THE FLOWER IN ANGIOSPERMS.—Since Van Tieghem’s *
memoir of 1868, when the vascular systems of several evolutionary types of both mono-
cotyledonous and dicotyledonous flowers were examined, this subject has progressed very
little beyond the purely descriptive stage. Hence the leading features of economical
advance may be outlined here. The simplest trace of a single carpel consists of a median
or dorsal and two marginal or ventral bundles. In the Hellebore, where the carpels are
united at their bases, bivalent traces occur in the basal portion. Pronounced cases of
syncarpy reveal two series of bundles in the ovary-wall—the outer homologous with
dorsal, the inner with marginal bundles, and in Geranium longipes the marginal
bundles coalesce in pairs. The vascular changes which supervene with perigyny and
epigyny vary in different families. In Aralia the stemonal and perianthic traces, now
included in the ovary-wall, form a third and outermost series of bundles. The two
outer series, distinct in Aralia, have become synthesized to form one in Castanea, several
genera of Согпасеге, and the epigynous Sympetale. The amalgamation is incomplete
in Davidia and Marlea. Another mode of origin of the single outer series is indicated
іп the Hamamelidaceze, where the single series of bundles, evident as far as the insertion
of the perigynous ridge, might have a further upward extension as the ovary becomes
inferior.

The last traces of the dorsal carpellary bundles may disappear, as in Lonicera and
Leycesteria, but here the bundles passing to stamens and petals are perhaps not strictly
homologous with stemonal and petal traces respectively.

By a further economical advance the bundles in the single series, which in Cornus,
Leycesteria, and many other genera are equal to the sum of the petals and stamens, are
in Viburnum, the Campanulacese, and Composite reduced to one-half the number.

Lines of asymmetrical development are exemplified by Viburnum, a precursor of the
irregular vascular systems which prevail in the Valerianacee.

No phylogenetic significance can be attached to a particular form of vascular system ;
but vascular details may afford a clue in support of an interpretation of origin, as in
Viburnum, Aucuba, etc., but they should always be used with the utmost reserve.

Tug OVULE IN ANGIOSPERMS.—Mirbelf wrote in 1829: “On peut dire qu'à cette
premiére époque tous les ovules sont orthotropes."

Тһе orthotropous ovule is primitive to the anatropous types.

The ovule, in Juglans regia, according to Benson and Welsford f, is not strictly
orthotropous, but exhibits “опе mode of the phylogenetic origin of the orthotropous
basal ovule from an anatropous parietal type.” Benson and Welsford describe two
types of flower: in one there are two opposite basal parietal placente, of which one bears
a single erect ovule and the other is barren ; in the other, the ovule is apparently borne
by the two placentz and becomes basal and orthotropous. Now, in the first case, the

ж Ph. Van Tieghem in Ann. Sei. Nat., Bot. 5"* ser. ix. (1868) 127-226, t, 9-12.
+ Mirbel in Mém. de l'Acad. des Sciences, ix. (1830) 611.
+ Benson and Welsford in Ann. Bot. xxiii. (1909) 629.
2 £ 2

298 МВ. А. 8. HORNE—A CONTRIBUTION ТО
ovule, being erect, is only slightly anatropous and hence can be regarded as exhibiting |
the very earliest stage in the evolution of the anatropous ovule.

'The stage in evolution evident in Jug/ans is followed by the complete inversion of the
nucellus, owing to the curvature of the raphe as in the Cactace and Saxifragaces, and
further by the congenital union of the raphe and integument as exemplified by the
Һовасеге. This phase of evolutionary advance is characterized by the separation of
several anatropal types. Two types obtain in the Rosacez: one “ascending,” with the
nucellus directed upwards as in Cerasus juliana; the other, “ descending,” with the

Text-fig. 13.

0
en
AAAS St eb
РТР
FELIX) QN >
ааа 5 4 as OF DS
wint ШЕТУ TEN >
SAK GEIS
Месна Ке
id : S
Д, Da

o]
EURO
Sar Ore
ce
à

KOA

Diagrams illustrating the structure of the ovule.

A, Pyrus Malus; В, Helwingia ruscifolia; С, Davidia involucrata; D-G, Aucuba japonica.

ad=adplacental side; ab=abplacental side; c—chalaza; e.sac=embryo-sac; int=integument ; int.r=cell-rows
of integument ; in.int=inner integument; out.int= outer integument ; up.int=upper margin of integument :
un.int = lower margin of integument ; /=limit of integument ; m=micropyle; n=nucellus; p=pivot cell.

nucellus directed downwards, as in Crategus Ozyacantha and Pyrus Malus (Text-
fig. 13, A). The former type is not far removed from the “dorsal” ovule which
prevails throughout whole families, such as the Araliacese, whilst the ventral ovule in
Согпасеге (Text-fig. 13, E, Aucuba; В, Helwingia) is conceivably an advance upon the
“descending” type. Specialized forms of the “dorsal” and “ventral” ovule are found
in Davidia and Helwingia respectively (Text-fig. 13, C and B), where the integument

(int) exists as a separate organ for only a short distance below the summit of the nucellus
—to the line marked / in the figures.

THE STUDY OF THE EVOLUTION OF THE FLOWER. 299

The line of anatropal advance and specialization is accompanied by the transition
from the bi- to the uni-tegumentary condition.

Benson *, in 1894, found all stages from two completely free integuments in Fagace:e.
to two more or less fused in Casuarinacee and Corylacez. ^ Péchoutre T, in 1902,
demonstrated a similar series in Rosacez and both the uni- and bi-tegumentary con-
dition in one genus, Spirea.

A third series occurs in the Ranunculacez, whilst one or two integuments characterizes
the Dilleniacee and Осһпасее. А single integument is found, for example, in the
Triuridales, Ceratophyllaceze, Ешреітасеге, Aquifoliacez, Cochlospermaceze, Гоаѕасеге,
Umbellifore, and prevails almost exclusively in nearly all the cohorts of the
Sympetale.

A third line of advance lies in the direction of the reduction in bulk of the nucellus.
The axile rows are numerous in the Rosacez, few in the АтаПасеге, and reach their
minimum development in Lilium, Helwingia, and Inula.

OvuLAR Connections.—Where each loculus contains several series of ovules, as in
the Saxifragaceze, the main placental bundles branch repeatedly throughout their course
on their outer side; branches of the second and third order arise from these and the
ultimate branchlets supply the ovules—one to each. Transitional vascular stages,
accompanying the change from the multiseriate to the biseriate arrangement, have not
been studied; but when the ovules are in two series, as in Leycesteria, the system is
reduced, each main bivalent axial bundle produces short branches on either side,
supplying ovules in the right- and left-hand loculus respectively—one branch to each
ovule. When the ovules in the double series are reduced to a terminal pair, rendering
each loculus biovulate, the branches are also reduced to two as in Aralia and the
Hamamelidacez. The further change from the biovulate to the uniovulate loculus
involves several instances. Thus the change from Aralia or Corylopsis, where one of a
pair of ovules is sterile, to Hamamelis, where one placenta of each pair is barren, is
accompanied by the abortion of the branch pertaining to the barren placenta and the
ovule receives a single bundle. But where the single ovule is axially borne, the ovular
trace is frequently double, receiving a vascular supply equivalent to that of a pair of
axially borne ovules. In Davidia, the branches uniting to form the ovular trace arise
from several small bundles extending in the right- and left-hand septa and pertain to the
right- and left-hand sides of the carpel; in Cornus, the branches arise from bivalent
peripheral bundles; and in Helwingia, where the median branch from an axial bundle
forms the ovular trace, the axial bundle represents two united marginal bundles of the
same carpel, so that the median branch may be considered bivalent.

In more advanced stages of reduction, where the septa disappear and the ovary
becomes uniovulate, various anomalies occur, such as have been deseribed in detail for
Viburnum, Corokia, Griselinia, and Nyssa.

It is highly probable that a parallel series of stages will be found leading towards

* M. Benson in Trans. Linn. Soc., Bot. 2nd ser. iii, (1894) 409-424, t. 67-72.
+ F. Péchoutre іп Ann. Sci, Nat., Bot, 8me вет, xvi. (1902) 88.

300 MR. А. S. HORNE—A CONTRIBUTION TO

genera where the ovule in the uniovulate loculus or ovary is basal. Іп the case of
Juglans regia, already cited, the ovule, when parietally borne, is supplied by the bundle
in the corresponding placenta, as in Aralia; but when basal (axially situated), bundles
from both placente are contributory as in Cornus.

OvuLAR VASCULAR SystemM.—Integumentary vascular tissue has been known since
1829, when Mirbel* described and figured the vascular systems of the ovules of Myrica
pensylvanica, Corylus Avellana, Juglans regia, and Quercus Robur, and states :—*“ On voit
fréquemment à la superficie de la Primine des nervures tantót simples tantót ramifiées
qui partent de la chalaze et se dirigent vers l'Exostome." А. Gris f, in 1861, figured
integumentary bundles in Ricinus communis; Van Tieghem T, in 1872, indicates several
examples selected from different families, and Le Monnier ў, in the same year, described
and figured the nervation in many seeds. Тһе vascular system in Myrica, Juglans, and
certain Amentiferz was rediscovered by E. M. Kershaw || in 1909.

Balfour | suggested that in non-vascular seeds the integument forms the water-
supply of the ovule.

L.-S. Gibbs** subsequently showed that the outer integument in the Alsinoides
actually performs this function.

The writer tt explained that in Davidia the ovular bundle branches and forms a
considerable system in the young seed, and the “integument” with its vascular
system is organized as an elaborate reservoir during the accumulation of an abundant
endosperm.

The water problem has to be solved in connection with seed-development from ovules
of numerous types, and it is clear that a similar device is adopted time and again in
phylogeny, early or late in ovular development according to special needs. Hence the
possession of a vascular supply is scarcely of considerable phylogenetic importance, as
suggested by E. M. Kershaw.

PHYLOGENY OF ANGIOSPERMS.—The author of each great system has in turn realized
that his classification is far from perfect and contains speculative groups difficult to
define. Hence the finding that two families of Angiosperms are polyphyletic, in itself,
merely corrects or modifies а fragment of Englers system, and, indeed, Engler
has already recognized the polyphyletic nature of his original Cornaces by adopting
Wangerin's not satisfactorily established arrangement in ‘Das Pflanzenreich.’ But the
faets that have influenced the writer to deduce polyphyletic origin for these families
are derived from an investigation in detail as yet unattempted, except perhaps in the

* Mirbel in Mém. de l'Acad. des Sciences, ix. (1830) 632.

T A. Gris in Ann. Sci. Nat., Bot. 4me sér. xv. (1861) 5-9, t. 2.

t Ph. van Tieghem in Ann. Sci. Nat., Bot. 5me sér. xvi, (1872) 228-232.

$ G. Le Monnier in Ann. Sci. Nat., Bot. 5e sér, xvi. (1872) 233-305, t. 9-12.

|| E. M. Kershaw in Ann. Bot. xviii. (1909) 353, 692.

4 І. B. Balfour in Proc. Brit, Assoc. Adv. Sci., Bot. Sect. (1901).
** L. 8. Gibbs in Ann. Bot. xxi. (1907) 25-55, t, 5, 6.
tt A. S. Horne in Trans. Linn. Soe., 2nd ser. Bot, vii. (1909) 315.

THE STUDY ОЕ THE EVOLUTION ОЕ THE FLOWER. 301

case of very few families, and he claims polyphyly for at least three genera of the
Caprifoliaceze and eight of the Cornaces, although in the latter family some genera
have remained closely associated for a considerable period. Now very many families
comprise a certain number of obscure forms, and subsequent analysis may determine
not only that these cannot be included in any existing family, being the sole repre-
sentatives of extinct lines, but even the apparently stable members of the family may
show evidence of parallel development and convergence.

Allowance must also be made in phylogenetic systems for cases where differently
organized forms may possess probable affinity, such as, for instance, Polyosma and
Diervilla, but the suggested origin of one from the other must be compatible in
all respects with morphological structure and evolutionary methods established by
experiment; thus, relatives of Polyosma should possess the power of advance towards
the condition in Diervilla, both with respect to gamopetaly and the number and nature
of the ovules.

It is а long-established custom to regard certain characters as more constant and
dependable than others, conceiving that the ovule, for example, would be less liable to
change than more superficial characters. According to De Vries's view of indiscriminate
mutation, however, the ovule would be equally liaole to change, but since mutation
is a rare phenomenon, difficult to perceive. Even then, however, the ovule would pass
through relatively fewer mutations than the floral organization, so that a certain form
of ovule might remain constant for a considerable range of floral and vegetative change.
Hallier reverts to an indiscriminate use of characters without due regard to relative
value; thus Garrya is retained in the Согпасег owing to the discovery of aucubin by
Hérissey and Lebas* in this genus, although a criterion of this kind is of doubtful
value unless supported by considerable research. Hallier + has attempted that which
cannot be accomplished from the evidence yet accumulated, and the inevitable result
of this finds expression in the extreme instability of his system.

Knowledge of the phylogeny of Angiosperms can only be truly advanced by the
detailed morphological and experimental investigation of many more families, and then,
. but not till then, can Engler's system be replaced by a greater scheme more nearly
approximating to natural relationships. In such a scheme the barrier between the
Polypetale and Sympetale will doubtless be removed, and the various families, upon
reconstruction, will be arranged in plexuses according to their probable origin and
relationship. It is possible to forecast the lines upon which this will be attempted, but
the effort lies beyond the scope of this present work.

I desire to acknowledge my indebtedness to Mr. C. E. Jones, B.Sc., F.L.S., for many
friendly suggestions; to Dr. A. B. Rendle, F.R.S., for valuable help and criticism ; and,
lastly, to Professor J. B. Farmer, F.R.S., for his continuous advice and encouragement
throughout the investigation.

* Hérissey and Lebas in Bot. Centralbl. exvii. (1911) 176.

+ H. Hallier in Beihefte Bot. Centralbl. xxiii. 2 (1908) 81, and Arch. Néerl, d. Sci. Ex. et Nat. sér. 3, Ва. i.
146 (1912),

802 MR. А. S. HORNE—A CONTRIBUTION ТО

BIBLIOGRAPHY.

. Arser, Е. A. NzwzLL.—Synanthy in Lonicera. Journ. Linn. Soc., Bot. xxxv. (1903) 463-474.

Асвект, M. Р. L.—Organogénie de la fleur dans le genre Salix. Adansonia, xi. (1873-76) 183-186,
t. 10

. Ваилом, H.— Observations sur les Saxifragacées. Adansonia, v. (1864-65) 282-304; vi. (1865-66)

1-15.

—— Recherches sur l'Aucuba. Tb., v. (1864) 179-203.

— — Traité du développement de la fleur et du fruit. УП. Corylées. Tò., xi. (1874) 163-174.
X. Castaneacées. Iób., xii. (1876-79) 1-20. XII. Garryacées. Tb., 264—269.

6. —— Sur deux nouveaux genres apétales. 10., x. (1871-73) 112-117.

7. —— Sur les affinités des Helwingia. Bull. Soc. Linn. Paris, i. (Nov. 1877) 137-189.

8. —— Sur l'organogénie florale et la graine des Garrya. Г. с., i. (Nov. 1877) 139-141.

9. —— Histoire des Plantes. Paris. Hamamelidez, iii. 456-461 ( 1871-2); Combretacez, vi.

977-988 (1877); Cornacex et Umbelliferze, vii. 79-83 (1879).

о м

о

т

10. —— Natural History of Plants. (Trans. by Hartog.) 1871 etc.
11. —— Nouvelles notes sur les Hamamélidées. Adansonia, x. 120-137 (1871-73).
12. Plantes de Madagascar, Atlas, iii. (1896).

13. Bager, J. G—Further Contributions to the Flora of Central Madagascar. Journ. Linn. Soc., Bot.
xxi. (1884-85) 317-353, 407-455.
Melanophylia crenata. Hook. Ic. Pl. xxv. (1896) +. 2499.
15. Bateson, W.—Mendel’s Principles of Heredity. London, 1913. 8уо.
16. Bateson, W., Miss E. В. Saunpers, and В. C. Punnetr.—Experimental Studies in the Physiology
of Heredity. Reports to the Evolution Com. Roy. Soc. ii. (1905) ; iii. (1908).
.17. Benson, Marearet J.— Contributions to the Embryology of the Amentiferz. Part I. "Trans. Linn.
Soc., 2nd ser. Bot. iii. (1894) 409-424, t. 67-72.
18. Benson, Mancanzr J., ELIZABETH Sanpay, and Emiry BennrpoE.—Contributions to the Embryology
of the Amentifere. Part II. Trans. Linn. Soc., 2nd ser. Bot. vii. (1906) 37-44, t. 6.
19. Benson, MARGARET J., and Evetyn J. WELsrogp.—The Morphology of the Ovule and Female
Flower of Juglans regia and of a few allied genera. Ann. Bot. xxiii. (Oct. 1909) 623-633. _
20. Вехтнам, G.—Plante Hartwegianz. Londini, 1839-57. 8уо.
21. Вехтнам, G., et J. D. Hooxer.—Genera Plantarum. i. (Hamamelidacex), 664-669 ; (Cornacez)
1. c., 947-959; ii. (Caprifoliacee) 1-7.
22. Bessey, Ernst A.— The Comparative Morphology of the Pistils of the Ranunculacez, Alismacee,
and Rosacem. Bot. Gazette, xxvi. (1898) 297-312, t. 25.
23. Виллмез, Е. H.—Beitrige zur Kenntniss der Samenentwicklung. Flora, lxxxix. (1901) (Capri-
foliace:z, p. 303).
24. Вісме, С. L.—Museum Botanicum Lugduno-Batavum, i. (1849-1851).
25. Воквгловв, J. G.—Flora van Nederlandsch Indie, i. Polypetalze (1890).
26. Brown, Rosert.—Observations on the Natural Family of the Plants called Composite. ‘Trans.
Linn. Soc. xii. (1817) 76-142.
27. —— Characters and Descriptions of three new species of Plants (1818). Miscell. Bot. Works of
R. B. vol. ii. (1867) 319—328.
28. Campsett, D. H.— Entwicklung des Embryosackes von Peperomia pellucida. Ber. Deutsch. Bot.
Gesell. xvii. (1899) 452—456, t. 31.
29. —— А peculiar Embryo-sac in Peperomia pellucida. Ann. of Bot, xiii. (1899) 626.
зо. CANDOLLE, А. ре.— Prodromus. Helwingiacez, xvi. и. 680-681 (1868).

THE STUDY OF THE EVOLUTION OF THE FLOWER. 303

. Сакківвк, Е. A.— Revue Horticole, 1869, 17-18.

$ 1879, 154-155

. CLARKE, BENJAMIN. — On the Structure of the Оты of Marlea and Affinities of Alangiez.
Hooker’s Kew Journal of Botany, vol. 11. (1850) 129-135, t. 5.

. CLARKE, C. В.-Оі1гасее. | Hooker's Flora of British India, iii. 590-618.

. CLEMENTS, Freperick EÉpwaAnp.—Research Methods in Ecology. Lincoln, Nebr., 1905.

. Courter, J. M., and W. Н. Evans.—A Revision of the N. American Cornacez. Bot. Gazette, xv.
(1920) 30-38, 86-97.

. Coutrer, J. M.—Contribution to the Life History of Ranunculus. Bot. Gazette, xxv. (1898)
73-86, t. 4-7.

. Darwin, Cuartes Rosert.—Animals and Plants under Domestication. London, 1868. 2 vols.,
8vo. 2nd ed. revised (1893).

. Decatsne, J.— Remarque sur les affinités du genre Helwingia, et établissement de la famille des
Helwingiacées. Ann. Sci. Nat. Bot. 2™° sér. vi. (1836) 65-76.

. Decaisne, J., et E. Le Maovr.—Traité Générale de Botanique. Paris, 1868. 4%.

A General System of Botany (trans. by Mrs. Hooker). London, 1873. 4to.

. Dicsy, L.—The Chromosomes of the Hybrid Primula kewensis. Proc. Brit. Assoc. Adv. Sci.
Portsmouth, 1911 (1912), 585-586.

. Ducamp, L.— Recherches sur l'embryozénie des Araliacées. Ann. Sci. Nat., Bot. 8% sér. xv. (1902)
311-402, t. 6-13.

. Еленев, А. W.—Blütendiagramme. Leipzig, 1875-78. 2 vols.,

. Enpuicner, S.--Cornez, Genera Plantarum (1839) 798-799; dede v. (1850) 17: Hamamelidez,
Gen. РІ. (1839) 803-805 : Lonicerez, Gen. РІ. (1839) 565-570.

. Емегев, A.—Syllabus der Pflanzenfamilien. Ed. vır. Berlin (1912).

. Encrer, A., and К. Prantt.—Die Natiirlichen Pflanzenfamilien. Araliacex, see Harms; Саргі-
foliaceæ, see Fritsch ; Cornaceae, see Harms; Hamamelidacez, see Niedenzu.

. Farmer, 7. B.—On tiie Development of the Endocarp of Sambucus nigra. Ann. of Bot. 11. (1889)
389—392.

. Гвітзсн, Canr.—Caprifoliaceze. Engl. u. Prantl, Nat. PRanzenf., Teil iv. Abt. 4, 156-169 (1891).
Verhandl. d. Zool.-bot. Ges. in Wien, xlii. (1892) 7-10.

. бівве, ІлілАм Suzetre.—Notes on the Development and Structure of the Seed in the Alsinoides.
Ann. Bot. xxi. (1907) 25-55, t. 5, 6.

. Соввег, CAnL.—Organography of Plants. Transl. by Balfour: Oxford, 1904-1905. 2 vols. 8yo.
. Свйгот, Pacr.—Hecherches sur la nervation carpellaire chez les Gamopétales bicarpellées de
Bentham and Hooker. Compt. rendus, exxii. p. 1144 (1896); exxv. p. 380 (1897).

. —— Recherches sur le systéme libéroligneux floral des Gamopétales bicarpellées. Aun. Sci. Nat.,
Bot. 8° sér. v. (1897) 1-154, t. 1-8.

. Gris, A.— Note sur le développement de la graine du Ricin. Ann. Sci. Nat. Bot. 4"* sér. xv.
(1881) 5-9, t. 2.

. Gurcxanp, Léox.—Recherches sur le вас embryonnaire des Phanérogames Angiospermes. Ann.
Sci. Nat., Bot. 6"* sér. xiv. (1832) 26-42, t. 3-7.

. Намлев, Hans.—Ueber den Umfang, die Gliederung und die Verwandtschaft de Familie des
Hamamelidaceen. Beihefte 2, Botanisches Centralblatt, xiv. (1903) 247—260.

58. —— Provisional Scheme of the Natural (Phylogenetic) System of Flowering Plants. New

Phyt. iv. No. 7 (1902) 151-162.

59. —— Neue е zur Stammesgeschichte der Dicotyledonen. Beihefte, Botanisches Central-

blatt, xxiii. 2. (1908) 81-265.

бо. —— Arch. Néerl. d. Sci. Ex. & Nat., Sér. r. і. (1912) 146.
SECOND SERIES.—BOTANY, VOL. VIII. дү

304 МВ. А. S. НОВХЕ--А CONTRIBUTION ТО

eo
л

. Harms, Hermann.— Ueber die Bliithenverhaltnisse der Gattung Garrya. Ber. Deutsch. Bot. Gesell.

Band xv. (1897) 19-21.

. —— Die Gattungen der Cornaceen. Ber. Deutsch. Bot. Gesell. Band ху. (1897) 21-29.
. — Araliacez, Cornacee. Engl. u. Prantl, Nat. Pflanzenf., Teil ii. Abt. 8, 1-69 (1894)

250-270 (1898).

. Hock, F.—Zur systematischen Stellung von Sambucus. Botanisches Centralbl. li. (1892)

288-284. `

. Hormetster, W.—Embryobildung des Phanerogamen :—Caprifoliaceen. Pringsh. Jahrb. і. (1858)

120-121.

. —— Neue Beiträge zur Kenutniss der Embryobildung der Phanerogamen. Abhandl. Kónigl.-

Süchs. Gesell. Wiss. (1859) 533—673, t. 1-97.

. Hooker, J. D.—Flora Nove Zelandiæ, i. 1853—55

Illustrations of Himalayan Plants, t. 19 (1855).

. Ноокев, J. D., et T. Taomson.—Precursores ad Floram Indicam : Caprifoliacez. Journ. Linn.

Soc., Bot. ii. (1858) 163-180.

79. Horne, A. S.—The Structure and Affinities of Davidia involucrata, Baill, Trans. Linn. Soc.
2nd ser., Bot. vii. (1909) 303—326, t. 31-33.
271. — The Polyphyli Origin of —€— Proc. Brit. Assoc. Adv. Sci., ` Portsmouth, 1911

(1912) 58

72. Jounson, D. up the Endosperm and Embryo of Peperomia pellucida. Bot. Gazette, xxx.
(1900) 1-11, t. 1.

7% On the Development of certain Рірегасев. Bot. Gazette, xxxiv. (1902) 821-340.

74. Кввзнау, Ерітн May.—Note on the Relationship of the Julianacez. Ann. of Bot. xxiii. (1909)
336-337.

75. — Тһе Structure and Development of the Ovule of Myrica Gale. Ann. of Bot. xxiii. (1909)
353—362, t. 24.

76. —— ` Further observations on the Structure of the Ovules of Мугісасев and allied groups.
Ann. of Bot. xxiii. (1909) 692.

77. Kunz, Surriz. Journ. Asiat. Soc. xliv. и. (1875) 201, pl. 15 (Nyssa).

. Le Monnier, Grorces.—Recherches sur le nervation de la Graine. Ann. Sci. Nat., Bot. 5%е sér.

xvi. (1872) 233-305, t. 9-12.

. Linviey, Joun.—The Vegetable Kingdom. 2nd ed. (1847); 3rd ed. (1858).

Garrya elliptica. Edwards’s Botanical Register, vii. (1835) t. 1686.

. Ілоур, Е. E.—The Comparative Embryology of the Rubiaceæ. Mem. Torrey Bot. Club, viii.

(1902) 1-112.

. Lyon, FroneNcE May.—A Contribution to the Life History of Euphorbia corollata. Bot. Gazette,

xxv. (1898) 418—426, t. 22-94.

. Maximowicz, Cart Jonann.—Diagnoses de Nouvelles Plantes Asiatiques. Bulletin de l'Académie

Impériale des Sciences de St. Pétersbourg. Lonicera, xxiv. (1878) 85; Viburnum, xxvi.
(1880) 474 ; Triosteum, xxvii. (1881) 476; Abelia, Diervilla, xxxi. (1887) 54—69.

. Merz, M 4... über die бше өен der Utricularieen. Flora, Ixxxiii. (1897)

77-87, t. 8.

. Маввег, С. F. Brisseav.—Nouvelles recherches sur la structure et le "D de l'Ovule

végétal (1828). Мет. de l'Acad. des Sciences, ix. (1880) 609—629.
--- Addition aux nouvelles recherches sur la structure et les sinas enden на l'Ovale
(1829). Mém. de l'Acad. des Sciences, ix. (1830) 629-684.

: аннат, Franz. oe + ч. Prantl, Nat. Pflanzenf., Teil iii, Abt. 2А. (891)
22 115-130.

Wk aa

THE STUDY ОЕ THE EVOLUTION OF THE FLOWER. 305

. ОввзтЕр, Ахрекѕ S.—Til Belysning af Slaegten Viburnum. Vidensk. Meddel. Naturb. Foren.
Kiobenh. 1860 (1861) 267-305, t. 7, 8.

со
oo

89. Отлуек, DawrEL.—Syeopsis, a new genus of Hamamelidacee. Trans. Linn. Soc., xxiii. (1860)
83-89, t. 8.

90. —— Note on Hamamelis and Loropetalum ; with a description of a new Anisophyllea from
Malacca. ` L. с. (1862) 457—161, t. 48.

gt. —— Nyssa sinensis. Hooker’s Ic. РІ. t. 1964 (1891).

92. —— Davidia involucrata, L. с. t. 1961 (1891).

93. Oriven, Frank Warr.—On the Structure, Development, and Affinities of Trapella, а new genus
of Рейа пез. Ann. of Bot. п. (1888) 15-115.

94. Patacny, J.—Plantarum Madagascariensium. Fase. iv. (1906).

95. Payer, J. B.—Lecons sur les Familles naturelles des Plantes (1860-1862) 124.

96. Organogénie de la Famille des [Ombelliféres]. Ann. Sci. Nat., Bot. 3° sér. xx. (1853)
111-118, t. 14.
97 Traité d'Organogénie comparée de la Fleur. Paris, 1857. 2 vols., 8vo.

98. PÉcnournz, F.—Développement de l'ovule et de la graine des Rosacées. Ann. Sci. Nat., Bot.
87° sér. xvi. (1902) 1-158. |
99. Завскмт, С. S.—Silva of N. America: v. (Hamamelidacee to Sapotacez). Boston & New
York, 1893. 8vo.
тоо. ӛснхква, H.— Beitrüge zur Kenntnis der Gattung Gunnera. Flora, xe, (1962) 161-208.
тот. Зевтовтоз, ApoLr.— Beiträge zur Kenntnis der Anatomie der Сотпасее. Diss., München, 1895 ;
in Bull. de P Herbier Boissier, i. (1893) 469-184, 496-512, 551-570, 614-689. `
102. Ѕноємакек, D. N.—On the Development of Hamamelis virginiana. Bot. Gazette, xxxix. (1905)
248—260.
тоз. Тахғахт, C.—Morfologia ed istologia del frutto e del seme delle Apiacez. Nuove Giorn. Bot.
Ital. xxiii. (1891) 451-469, t. 4-7.
104. Тнохвеко, C. P.— Flora Japonica, pp. 4 & 64, tab. 12 & 13 (1784).
105. Тоамгмт, F.— Ricerche di morfologie ed anatomia sul flore feminille e sul frutto del Castagno. |
Atti del В. Ist. Bot. dell’ Univ. di Pavia (1892) (Estr. pp. 35, tt. 3).
106. Van Тткенкм. Рн. E. L.—Recherches sur la structure du Pistil. Ann. Sei. Nat., Bot. 5"* sér. ix.
(1868) 127-226, t. 9-12.

107. —— Note sur les divers modes de nervation de l'ovule et de la graine. Ann. Sci. Nat., Bot.
5me sér, xvi. (1872) 228-292. . U

108. —— Structure de quelques ovules et parti qu'on en peut tirer pour améliorer la Classification.
Journ. de Bot. xii. (1898) 198-220.

109. —— Traité de Botanique. Paris, 1884, 2 vols., 8vo. ; 2™° éd., ib., 1886-1891, 2 vols., 8vo.

110. Үкваов, J.—Développement du sae embryonnaire des Phanérogames Angiospermes. Ann. Sci.
Nat., Bot. 6™¢ sér. vi. (1878) 237-285, t. 11-16.
111, —— Nouvelles recherches sur le développement du вас embryonnaire des Phanérogames
Angiospermes. L. с. viii. (1879) 261—890, t. 12-21.
112. Vipar, Lovis.—Sur la structure et le développement du pistil et du fruit des Caprifoliacées,
Bull. Soc. de Statistique, Grenoble, 4"* sér. t. iv. (1899) 157.
113. —— La course des faisceaux dans le receptacle floral des Labiées. Journ. de Bot. xii.
(1898) 46-52. ' '
114. Vries, Ново рк.—Тһе Mutation. Theory ; transl. by J. В. Farmer and A. D. Darbishire.
London, 1910-11. 2 vols., 8vo.
115. Warrrns, W. G.—Annales Botanices Systematice. (Hamamelidacee) п. 272-276 (1851-52);
_ (Lonicerz) ib. 789-788; (Согпеюж) ib. 795; (Hamamelidacee) iii. 842 (1853).
| 2:2

806 MR. А. 8. HORNE—A CONTRIBUTION ТО

116. WANGERIN, W.—Garryacez, Муззасее, Alangiaceæ, and Cornacez, in Engler’s Das Pflanzen-
reich, Heft 41 (Leipzig, 1910).

117. — Die Umgrenzung und Gliederung der Согпасее. Engl. Bot. Jahrb. xxxviii. (1906),
Beibl. n. 86.

118. Warmine, Е. J. B.—De l'Orule. Ann. Sci. Nat., Bot. 6™° sér. v. (1877) 177-266, t. 7-18.

119. Wess, J. E.—Morphological study of Flowers of Spirea. Pot. Gazette, xxxiii. (1902) 451-460.

120. WrrrsTEIN, В. von.—Handbuch der systematischen Botanik, i. & ii. (1901-08).

121. Wirrrock, V. В. Om Linnea borealis: En jemnfórande biologisk, morfologisk och anatomisk
undersókning. Botaniska Noti-er, Lund, (1878) 17-32, 49-54, 83-96, 122-197; (1879) 9-20,
137-150.

122. —— Linnea borealis, L. : Species polymorpha et polychroma. Act. Hort. Bergiani, iv. n. уп.
(1907) pp. 187, tt. 13.

EXPLANATION OF THE PLATES.

PLATE 28.

Fig. 1. Cornus Атотит. Young flower.

Fig. 2. С. Amomum. Young flower with two petals removed.

Fig, 3. C. alba. Fruit.

Fig. 4. Marlea platanifolia. . Portion of a very young inflorescence with flowers in a stage prior to
the development of the ovarian cavity.

Fig. 5. M. platanifolia. Single flower older than 4, ovary still Padriga

Fig. 6. M. platanifolia. Portion оЁ a flower slightiy older than 5.

Fig. 7. M. platanifolia. Flower slightly older than 6, with some of the petals and stamens removed.

Fig. 8. Aucuba himalaica. Normal flower.

Fig. 9. A. himalaica. Flower with six petals.

Fig. 10. А. himalaica. Flower with five petals and one stamen.

Fig. 11. A. himalaica. Fiowers with petals removed showing bilobed style.

Fig. 12. А. himalaica. Flower with petals removed showing three styles.

Fig. 13. Garrya elliptica. Female catkin, nat. size.

Fig. 14. С. elliptica. Female flower with one style bilobed.

Fig. 15. G. elliptica. Young fruit slightly enlarged.

Fig. 16. С. elliptica. Portion of male inflorescence.

Fig. 17. G. elliptica. Portion of male inflorescence, bracts removed.

Fig. 18. С. elliptica. Single flower.

Fig. 19. С. Thuretü. Portion of male inflorescence, showing one “flower” with four perianth-
segments and nine stamens.

Fig. 20. G. Thuretii. Single “ flower” with seven perianth-segments and seven stamens.

Fig. 21. С. Thuretii. Another “ flower” with five perianth-segments and ten stamens

Fig. 22. С. laurifolia. A terminal male flower (6r=bract).

Fig. 23. G. laurifolia. The same flower, different view.

.. Fig.24. G. laurifolia. Terminal flower.

. Fig. 25. G. laurifolia. Lateral flower.

‘Fig. 26. Helwingia ruscifolia. 'Їгїшегсиз male flower.

= 27. н. ы” "Tetramerous male flower.

THE STUDY OF THE EVOLUTION OF THE FLOWER. | 307

Fig. 28. H. ruscifolia. Young female inflorescence.

Fig. 29. H. ruscifolia. Young female flower.

Fig. 50. Н. ruscifolia. Mature flower, pentamerous.

Fig. 31. H. ruscifolia. Flower with four perianth-leaves and five carpels..
Fig. 82. Н. ruscifulia. "Tetramerous flower.

Fig. 88. Corokia Cotoneaster. Young flower.

Fig. 34. C. Cotoneaster. Mature flower.

Fig. 85. Griselinia lucida. Young male flower.

Fig. 36. С. lucida. Hexamerous male flower, side view.

Fig. 87. С. lucida. Hexamerous male flower, viewed from above.
Fig. 38. G. litloralis. Male flower, closed.

Fig. 39. С. littoralis. Male flower, open.

Fig. 40. G. lucida. Very young female flower, highly magnified.
Fig. 41. G. lucida. An older stage.

Fig. 49, С. lucida. Mature flower.

Fig. 48. G. lucida. Fruit.

Fig. 44. G. littoralis. Young female flower.

Fig. 45. G. littoralis. Mature flower.

Fig. 46. G. littoralis. Young fruit.

Fig. 47. G. littoralis. Older fruit.

Fig. 48. Nyssa sylvatica. Portion of a young inflorescence.
Fig. 49. №. sylvatica. Young flower.
Fig. 50. N. sylvatica. Young flower with petals removed to show the single stamen present in this

specimen.

PLATE 29.

[Abbreviations :—arch. = archesporium ; ant. = antipodals; az. = axial row; е. ғас. = embryo-sac ;
/.=loculus; int.=integument; т. = micropyle; m. р. = megaspore ; n. ep.= epidermis of nucellus;
par.= parietal tissue. ]

Figs. 51-54. Symphoricarpus racemosus. Development of sporogenous tissue in sterile ovules. 51,
primary archesporial cell; 52, two sporogenous cells ; 53, four sporogenous cells; 54, one
of four megaspores has germinated to form an incomplete embryo-sac. х 600 approx.

Figs. 55-57. Viburnum Lantana. Development of the megaspores in the incomplete ovules. 55,
primary archesporial cell; 56, binucleate sporogenous cell ; 57, quadrinucleate sporogenous
сей. х 600 approx.

Figs. 58-60. V. Lantana. Further development of the sporogenous tissue in the incomplete ovules.
58, portion of multinucleate protoplasmic sac showing e'ght of the twelve nuclei ; 59, nuclei
approximated in pairs ; 60, segmented protoplasmic sac. х 600 approx.

Fig. 61. Sambucus migra. Vertical section through a young ovule, showing single integument and
nucellus with primary archesporial сей. х 290 approx.

Fig. 62. S. nigra. Section through an ovular rudiment showing the enlarged subepidermal cells.
x 590 approx.

Figs. 63-64. S. nijra. The nucleus in the enlarged cells has
nuclei. х 590 approx.

Fig. 65. Benthamia fragifera. Vertical section of a young ovule showing an early stage in the
upgrowth of the single integument. x 290 approx.

Fig. 66. В. fragifera. Vertical section of a young ovule, showing the nucellus with axial rows.

x 290 approx.

divided to form two or four daughter

308 MR. A. 8. HORNE—A CONTRIBUTION ТО

Fig. 67. Cornus sanguinea. Showing the single integument and axial rows. х 290 approx.

Fig. 68. Aucuba japonica. Vertical section showing the earliest stage in the development of the
ovular mamelon. х 600 approx.

Fig. 69. A. japonica. Ап early stage in the formation of parietal tissue. x590 approx.

Fig. 70. А. japonica. Vertical section through the nucellus showing the enlarged embryo-sac-mother-
cell. x 290 approx. |

Fig.71. А. japonica. Transverse section of the nucellus showing two megaspores belonging to
different axial rows. x 290 approx.

Vig. 72. A. himalaica. Reconstruction from three vertical sections of the nucellus showing a complete
embryo-sac and an enlarged megaspore. х 290 approx. :

Fig. 73. Helwingia ruscifolia. Vertical section of a young ovule showing an early stage in the
development of the single integument and the nucellus. х 580 approx.

Fig. 74. H. ruscifolia. Vertical section of the nucellus showing the embryo-sac. x 290 approx.

Fig. 75. Griselinia lucida. Vertical section of the nucellus at a very early stage. х 580 approx.

Fig. 76. С. lucida. А slightly older stage showing axial rows. х 580 approx.

Fig. 77. Nyssa sylvatica. Transverse section of a very young ovule showing the nucellus invested
by a single integument. х 580 approx.

N.B.—Figs. 51-77 are semi-diagrammatic.

PLATE 80.
| Abbreviations :— а, b, с, d, e-- principal bundles; az.=inner peripheral bundle (corresponds to axial
bundle); іп. ет. = dorsal and lateral carpellary branches; ov. 1, ov. 2, ov. З = placental branches
from az.; ре. = petal-bundle; s¢.=stemonal trace; 2, y=auxiliary bundles; z’, у’ = branches
from т, y respectively. |

Fig.78. Photograph of a reconstruction in wax of the vascular system of the flower of Viburnum
Lantana. х 60 approx.

ABBREVIATIONS OCCURRING IN TEXT-FIGURES 1-19.

a. Fig. 3, ovular rudiment; Fig. 4, principal bundle.
а’. Fig. 3, ovular rudiment.
ат. Figs. 2, 3, 9 and LO, axial bundles.
5. Fig. 3, ovular rudiment; Fig. 4, principal bundle.
b'. Fig. 3, ovular rudiment.
бт. Fig. 4, bracteole ; Fig. 11, branches from outer series.
c. Fig. 3, ovular rudiment ; Fig. 4, principal bundle.
с’. Fig. 3, ovular rudiment.
d. Figs. 1, 8-10, dorsal carpellary bundle; Fig. 4, principal bundle; Fig. 5, dorsal bundle.
e. Fig. 4, principal bundle.
emb. Fig. 8, embryo. |
end. Fig. 8, endosperm.
f. Fig. 8, epigynous foliole (bracteole).
in. ser. Figs. 1-12, inner series.
| inf. Figs. 9 and 10, integument.
_ L Fig. 1, lateral сагреПағу bundle ; Fig. 9, branches from axial bundles.

out. ser.
ov., ov. 1, ov. 2, etc.

THE STUDY OF THE EVOLUTION OF THE FLOWER. 309

. Figs. 1, 5-12, loculus.

. Fig. 1, septal or placental bundles; Fig. 8, marginal carpellary bundles.
. Figs. 5, 8, 9 and 12, nucellus.

. Figs. 5 and 6, branches concerned with ovular supply.

Fig. 5, branch from o.

. Fig. б, branch from о.
. Fig. 5, inner branches ; Fig. 6, dorsal and lateral carpellary bundles.

Figs. (ete.), outer series.
Figs. 2 and 3, series of ovules ; Fig. 4, placental branches from az ; Fig. 10, 11,
and 12, vascular tissue supplying the ovules,

. Fig. 1, perianth traces; Figs. 2, 6, 7 and 10, petal bundles; Fig. 11, petal,
. Figs. 1, 3, 4 and 10, placenta.

. Fig. 11, rudiments of petals.

. Fig. 11, rudiments of stamens.

. Fig. 7, rudimentary axial bundles.

. Fig. 9, sepal bundle; Fig. 11, sepal.

Figs. 1, 2, 4, 10 and 12, stemonal bundles.

. Figs. б, 9-12, stylar canal.
. Fig. 6, toral cylinder.

. Fig. 4, auxiliary bundles.

. Fig. 4, branches from zy.

ОНҒА =.

Ногпе. x Trans. Linn. Soc. SER. 2. Bor. Vor, VIII. P1.28.

; West,Newman imp.
CORNACEAE.

Trans. Linn. бос.; Ser. 2. Bor. Vor. VIII. Рі. 29. .

Новме.

3⁄2

Grout, sc. & imp.
CAPRIFOLIACE/E AND CORNACEÆ.

данда

Новме. TRANS. Linn. Soc. SER. 2. Вот. Vor. VIII. Pr. 30.

=" сы poen a st nm

E

Grout, sc. & imp

A. S.H. phot. .
VASCULAR SYSTEM OF THE FLOWER OF VIBURNUM LANTANA

LINNEAN SOCIETY OF LONDON.

——n -—
— T mag

MEMORANDA CONCERNING TRANSACTIONS. - a

The First Series of the Transactions, containing both Botanical and Zoological contributions, has been
in 30 Vols., and a few entire sets are still for sale. Only eertain single volumes, or parts to complete sets, ME
E obtained at the original prices. The me of the Index to Vols. 1-25 is 8s. to the publie, and 6s. to Fellows; to
ls, 26-30, do: to the fabli Ж гаа 3. to Fellows.
The Second Series of the Transactions is divided into Zoological and Botanical sections. The prices of the
— n of these which have been published are as undermentioned. (For the Zoological parts see see Zoological | К
‘wrapper,

SgcoNp Sextes.—Boranx (continued).
Volume, Phi. T Price to the
УИ. Part. I 1904.....
(IIS UA ше .8 yh 22 Раб ПІ. 1905.....-
а, Parts ХҮІ. (complete). колеи
| 1138, 6 610 .... 5 ° : с a
: ae Рай У, 1907...
ш Parts 1.-ХІ. (completo). i ооо Part У
о сә... 1 1210 .... ART s

еп = TRANSACTIONS. 4
THE LINNEAN SOCIETY OF LONDON.

Е CERTAIN STRUCTURES CONCERN
| IN THE GENUS GNETUM.

[ Stra

IX. Notes on the Morphology of certain Structures concerned in Reproduction in the
Genus Gnetum *. By Н. Н. W. Pearson, Sc.D., F.L.S., Henry Bolus Professor
of Botany in the South African College.

(Plates 31, 32.)

Read 4th June, 1914.

THE following is an account of an investigation of material already described +,
together with further supplies $, for which I am indebted to

1. Mr. J. Gossweiler (Gnetum africanum).

A. Collected at Montobello, Angola, in February 1912.
B. From the same locality, collected in December 1912.

2. Dr. В. H. Lock (Gnetum Gnemon).

C. Collected in Ceylon.
D. Collected in Ceylon on various dates between August 5 and October 27, 1912.

3. Mr. I. H. Burkill.
E. Collected in Singapore, August 1913.

In all these cases the ovules and male spikes were carefully fixed in various
fixatives—weak Flemming’s solution, alcohol formalin, mercuric chloride solution, and
chromo-acetic acid. While the fixation was usually successful, delicate structures have
frequently suffered to some extent in transmission. Consequently the free-nuclear stages
of the embryo-sac are somewhat contracted, and no trustworthy information regarding
the structures in the pollen-tube and fertilization has yet been obtained. Stages in the
development of the pro-embryo and embryo have been figured, but their publication is
postponed until the account can be made more complete.

A. THE SPIKE AND THE FEMALE Fiower. (G. Gnemon.)
Most writers have hitherto referred to only two forms of spike in Gnetum, viz. :--
1. The pure female spike bearing ovules, each with three envelopes. Many of
these are normally arrested ; the rest are fertile.
2. The pseudo-androgynous (a male) spike, bearing functional male flowers and
infertile ovules with two envelopes only.
* Percy Sladen Memorial Expeditions in South-West Africa (assisted by grants from the Royal Society),
Report No. 22.
T Pearson, 1912.
5 Material of G. scandens, recently sent from Darjeeling by Mr. G. H. Cave, awaits investigation.
SECOND SERIES.— BOTANY, VOL. VIII. 97

819 DR. Н. Н, W. PEARSON ON THE MORPHOLOGY OF

A third form has recently been described *, viz. :—

3. A pure male, usually without any trace of infertile ovules or their envelopes.
This has so far been described only in the two African species, G. africanum
and G. Bucholzianum. In G. africanum an incomplete female flower is very
rarely present.

For many years there has been reason to suspect that still another form of spike
sometimes occurred in certain species, particularly іп 6. Gnemon. Lotsy T refers to “a
plate in Rumphius picturing a branch of a male tree of Gnetum Gnemon with some
apparently normal adult fruits." Strasburger{ found in an androgynous inflorescence
of G. Brunonianum (=G. Gnemon) received from Beecari that “ waren die männlichen
Blüthen bereits abgeblüht, theilweise abgefallen, die weiblichen noch erhalten."
Commenting upon this and upon statements made by Blume 6 and Griffith | that the
2-enveloped ovules of the male spike may later develop a middle envelope, Strasburger
remarks €J :—“ Könnte vermuthet werden, dass diese Blüthen (i. e. incomplete female
flowers of the ‘male’ spike) bei Gnetum Brunonianum sich auch noch weiter entwickeln,
ja dass es auch noch andere Gnetwm-Arten giebt, bei welchen die weiblichen Blüthen der
androgynen Blüthenstinde noch keimfühige Samen erzeugen; unmöglich ist dieses
nicht, mir persónlich aber kaum wahrscheinlich."

On a later page Strasburger | writes:—“ Die einzige positive Angabe über drei
Hüllen an der weiblichen Blüthe der pseudo-androgynen Blüthenstánde, wo sicher nur
solche vorlagen, wird von Tulasne** für Gnetum Leyboldi gemacht, allein aus den
eigenen Figuren Tulasne’s {+ ist zu ersehen, dass diese Angabe nur auf einer Tauschung
beruht." Lotsy ff obtained from a collector in the neighbourhood of Buitenzorg “a
male branch with а young fruit very much larger than the female flowers in the andro-
gynous infloresences.” This inflorescence was broken, and bore no male flowers, but it
came from a male tree. The “ fruit" which it bore possessed three normal integuments,
but was not fertilized. ‘Consequently, we have found here on a male tree a female
flower corresponding to those which occur normally on the female trees, 16 possessing
like these three integuments and not two like the female flowers of the androgynous
inflorescences.”

In the material of Gnetum Gnemon now under consideration, ovules with three,
occasionally with four (see figs. 2, 3) envelopes occur, the former not uncommonly, in the
spikes which bear the male flowers. These complete female flowers when present are
usually found in the lowest node. They are larger than the incomplete flowers, and
in the cases observed it was not certain that they originated in the uppermost ring of
the node; the male flowers of the nodes which bear them appear always to be fewer in
number than at ordinary nodes. In connection with the occurrence of these complete
flowers in the lower nodes of the spike, the following observation made by Mr. I. H.
Burkill may be of interest. He writes from Singapore (20. viii.13) :—“ After flowering,

* Pearson, 1912. + Lotsy, 1899, р. 53. + Strasburger, 1872, р. 105 et seq.
§ Blume, 1848, | Griffith, 1859, p. 301. € Strasburger, 1. с. p. 110.
** Tulasne, 1858, p. 116. tt Tulasne in Mart. Fl. Bras. iv. 1. Tab. 103. ++ Lotsy, l с.

CERTAIN STRUCTURES IN THE GENUS GNETUM. 313

the upper part of the spike (2. e. androgynous of G. Gnemon) breaks off and falls to the
ground, leaving the flowers of the lowest axils.” In the first case found there were two
such flowers at the lowest node; one had three envelopes, the other four (figs. 2, 8); each
contained more than one embryo-sac, all in an early free-nucleate condition. The
former contained many pollen-grains in the micropyle; two of these are shown in
detail in fig. 9. This node bore no incomplete female flowers; these were, however,
present in their usual position at the upper nodes.

This observation is of interest, in that it undoubtedly explains the statements of
previous authors referred to above. Since these ovules do become pollinated, it may be
presumed that fertilization ensues in some cases at least and that fertile seeds may be
borne upon the spikes which produced the male flowers. In other words, in Guetwm
Gnemon, as probably also іп б. Leyboldi and other species, there is found a truly
androgynous as well as a pseudo-androgynous spike.

The four types of spike now described establish a sequence ranging from the strietly
unisexual to the bisexual, or conversely, which leaves no doubt that the two commoner
types are members of the same evolutionary series. Whether this series progressed in
two directions from a primitive bisexual form or from two unisexual inflorescences
towards a combined type, has been a fruitful subject for discussion. Wettstein *
holds the latter view for all the genera of the Gnetales. The principal points in his
argument are the following :—

1. The primitive flower of the Gymnosperm was unisexual. Bisexual inflorescences
are unknown in the Cycads, but are relatively common in the higher forms.

2. Ephedra campylopoda (the only member of the genus in which a bisexual
inflorescence occurs at all commonly) does not, in its general characters, give
the impression of being a primitive form. |

3. The archegonia of the female flower of the androgynous spike of G. campylopoda
show no signs of reduction, although they apparently do not become
fertilized.

4. If the occurrence of 9 flowers in this androgynous inflorescence is really due
to reversion, then a similar reversion would be expected to occur in the
9 inflorescence. с

5. If, оп the other hand, the appearance of bisexuality is due to “ progressive
mutation" (as Wettstein believes), then it would be expected to appear in
the ¢ inflorescence rather than in the %, on account of the normally
unlimited growth of the former.

Wettstein quotes the evidence of Blume, Griffith, and Lotsy as to the occasional
appearance of a bisexual inflorescence in Gnetum, and is inclined to think that “auch
morphologisch diese ‘ zweigeschlechtigen’ männlichen Blüten von Welwitschia analog
erklärt werden können, wie die zweigeschlechtigen Inflorescenzen bei Ephedra.”

These arguments are not convincing. №. 1 would have greater force if the living

Gymnosperms formed a monophyletic group. The history of certain sections of the

* Wettstein, 1907, p. 28.

272

814 DR. H. H. W. PEARSON ON THE MORPHOLOGY ОЕ

group is obscure, and it is highly improbable that its complete elucidation will establish
the view that they represent a simple series. The statement contained in No. 2 is no
doubt true, but has it necessarily any bearing upon the point at issue? The conditions
which influence the development of the vegetative organs do not necessarily affect the
reproductive parts. An angiospermic parasite, for example, frequently shows extreme
reduction in all parts of the plant-body save in those concerned with reproduction,
which may, indeed, be of a highly specialized character. The fact that the archegonia
of the imperfect female flowers of Ephedra campylopoda show no signs of reduction is
not conclusive, if true. An argument based upon Ephedra must almost certainly apply
to Gnetum and Welwitschia. In the latter, sporogenous tissue is not organized in the
so-called hermaphrodite flower; in the former, gametes do not appear in the incomplete
female flower. The female flower of Ephedra campylopoda is less incomplete than that
of either of the other two genera, but if reduction can account for the one it must
explain all. The last two points are of less importance. If an Angiospermic species
with polygamous flowers may be held to be descended from a more primitive form with
hermaphrodite flowers, a similar explanation of the occurrence of female, male, and
more or less hermaphrodite inflorescences or spikes in the Gnetales is not ruled out.
Since the date of Wettstein’s paper it has been shown that Z. campylopoda is ento-
mophilous *; the same is true for Welwitschia t, and not improbable for Gnetum +
In all cases it is probable that the ovule of the “ hermaphrodite” spike or flower has
become specialized in connection with insect-pollination. This specialization is much
more likely to have occurred through the reduction in some or all of the parts of
a primitively functional hermaphrodite flower than by “ progressive mutation.” In the
latter case the fact that a new and previously functionless structure adapted to attract
insects bears so curious a resemblance to a functional female flower would be difficult
to explain.

Strasburger $ discussed the question fully and, at least for Welwitschia and Ephedra,
arrived at the conclusion that the female flower in each case represented an advanced
stage in the reduction of a primitive hermaphrodite flower, and that the pseudo-
hermaphrodite flower of Welwitschia represents an intermediate stage passed in the
course of this reduction. While it is not impossible that the actual course of events has
been more complicated than is here suggested, Strasburger’s view that the existing
structures are the result of a process of reduction seems to be in closer accord with
the facts than the contrary opinion of Wettstein ||.

Both the female and the pseudo-androgynous spikes of Gnetum Gnemon in the
specimens under notice very frequently bear a single terminal ovule; this was also
recorded by Strasburger{. This fact, together with the occasional reduction of the
male spikes to two nodes separated by an internode (@. Bucholzianum**) and the
presence of “ a ring of complex groups of vascular strands arising from the branches іп

ж Porsch, O., 1910. + Pearson, 1909.
+ Karsten, 1893; Lotsy, 1899, &e. $ Strasburger, 1872, р. 150 et seg. ; see also Goebel, 1905.
| Ligmer et Tison, 1912, р. 178. Ч Strasburger, 1872, р. 158. ** Реагзоп, 1912.

CERTAIN STRUCTURES IN THE GENUS GNETUM. 815

the base of the female flower of Gnetum Gnemon” *, afford some support to the view
that the primitive inflorescence from which the present complex spikes have arisen was
of the nature of an axis bearing a terminal ovule surrounded by a single ring of male
flowers * (or perhaps anthers).

In the ovule with four integuments, referred to above (fig. 2), there is no trace of the
“ring of complex groups of vascular strands ” described by Miss Berridge for the normal
ovule with three envelopes, but the three lower of the envelopes possess an apparently
equivalent vascular supply. This might, perhaps, be regarded as indicating that this
reduced ring of vascular tissue is a vestige, not of a ring of male elements, as Miss Ber-
ridge suggests T, but оға missing fourth envelope. On the other hand, it will be seen
on reference to the figure that the second and third envelopes appear to arise from
à common base—indicating, perhaps, a doubling of the original middle integument.
This suggestion is of interest when compared with the accounts given of the development
of the incomplete female flowers. Lotsy $ finds that a middle envelope is present in
a rudimentary condition in the younger stages, but lateris unrecognizable. It therefore
seems probable that the middle envelope is less stable in character than either of the
other two; that it is liable to be suppressed or duplicated aecording to circumstances.
It is at least improbable that the outermost envelope is supernumerary. Before it was
cut, the ovule enclosed in its envelopes had a perfectly normal appearance, and there
was nothing to suggest that its outer covering was in any way different from that of an
ordinary complete female flower, and as far as the outer envelope itself is concerned,
this resemblance is seen also in the sections. In any case, it is clear that in the present
state of our knowledge some degree of caution should be exercised in assigning any of
the three normal envelopes to a particular morphological category.

The same ovule shows another feature of interest which appears not to have been
previously described. А tangential section (fig. 3) shows а bud arising between the
outer and the succeeding envelopes (5). As this is the only instance of the kind
Observed, it is not certain that the bud stands opposite the median line of one of the
primaries from which the outer envelope is derived. Also its vascular supply cannot be
determined at this stage. It may, however, be presumed that it is truly axillary.
If so, the foliar character of this envelope is beyond doubt. If the view suggested
above, that the middle integument in this ovule is duplicated, is correct, and the outer-
most of the four envelopes is therefore homologous with the outermost of the three of
а normal flower, an interesting conclusion is suggested by the presence of this axillary
bud. For it shows that there exists a tendency to the development of axillary structures
by which the spike of Gnetum might arise by “proliferation om a pnimiuve
inflorescence (or flower) bearing a single terminal ovule. A similar origin for the male
spike of Welwitschia has already been suggested by Wieland §. Also, if this is the true
outer envelope, the presence of а bud in its axil is in favour of the conclusion that
the origin of the flower-envelope of the Gnetales is to be sought, not in the sterilization

+ Berridge, 1912.

* Pearson, 1912.
§ Wieland, 1908.

+ Lotsy, 1899.

316 DR. H. H. W. PEARSON ON THE MORPHOLOGY OF

of sporophylls, but in the specialization of barren leaf-structures *; and it might also be
regarded as in opposition to the current view that the outer envelope is a perianth +.
On the other hand, if a perianth homologous with that of the Angiosperms is present in
the Gnetales, it is not probable that in this group it will have attained the high degree
of specialization which is characteristic of the perianth of the former. But the
unqualified use of the term “ perianth " implies a homology which may be correct, but
of which there is at present no direct proof.

The nature of the axillary bud is uncertain (fig. 3); but an examination of the
whole series of sections passing through it gives the impression that it is a female
flower, complete or incomplete, in an early stage of development.

The material at present available does not contain early stages of development of com-
plete female flowers. Very young incomplete female flowers are shown in РІ. 81. figs. 4
and 5. Slight swellings on the axis (figs. 4 and 5, m), between the insertions of the
outer and inner envelopes, are doubtless identical with those considered by Lotsy i to
represent a rudimentary middle envelope. If so, these figures may be regarded as
confirming the statement that these envelopes arise in acropetal succession, which has
been made by previous writers (Beccari, Strasburger, Karsten, and Lotsy $). They
therefore correspond in development to the envelopes of the complete female flower as
described by the same authors. The same order of succession in development is found
in the ovular envelopes of Ephedra || and Welwitschia J. While all recent investigators
are in agreement on this point, Coulter and Chamberlain ** state, without discussion,
that in all three genera * there are two integuments which arise іп basipetal succession.”

B. Tug MALE FLOWER Місвовровв AND PoLLEN-Grain. (©. Gnemon.)

A single young pseudoandrogynous spike of G. Gnemon has been available for study.
It has yielded certain stages in the development of the male flower which have not been
previously described for this species. Karsten ff figures three stages ranging from the
early condition of the envelope to the periclinal division of the primary tapetal
layer. Lotsy's observations on the germination of the microspore have not yet been
published 11.

The internodes of the spike are somewhat longer than those of G. scandens $$, but
- much shorter than in the African species ||. As in G. scandens, the flowers at the
nodes of G. Gnemon receive their vascular supply from a single series of strands arising
from the leaf-traces M, and the capsule is without the large mucilage cavity which is
so marked a feature in the African species ***, x

Тһе male flower of б. Gnemon usually bears two anthers. These are always placed
transversely to the axis of the spike, so that a longitudinal median section of the latter
either passes between them or through one of them. Occasionally four anthers are

* Goebel, 1905. % Lotsy, 1899, pp. 63, 64. + Lotsy, 1899, pl. 3. fig. 14.
$ Вессагі, 1897; Strasburger, 1879; Karsten, 1893; Lotsy, 1899. || Jaccard, 1894.
€ Hooker, 1863; Pearson, 1906. ** Coulter and Chamberlain, 1910. Tt Karsten, 1893, figs. 12-14.
tt Lotsy, 1899, р. 82. $$ Pearson, 1912, fig. 1a. 1] Pearson, 1912, figs. 1 b & c.

44 Pearson, 1912, fig. 2; Thoday, 1912. _ *** Pearson, l. c. figs. 2, 9, & 4.

CERTAIN STRUCTURES IN THE GENUS GNETUM. 317

present (РІ. 31. fig. 6). When this is the case the two supernumerary anthers decussate
with the others. The vascular structure of the axis bearing more than two anthers has
not been investigated.

The single envelope at the stage shown in fig. 7 has the form of an approximately
equal collar arising from the base of the axis. Later on, growth is confined to two
points at right angles to the plane in which the anthers lie, the posterior lobe so
produced overlapping the lower or anterior*. Тһе origin of the sporogenous tissue and
tapetal layers is as described both for this species} and for 6. africanum}. Тһе
elongation of the floral axis commences at about the same time as the periclinal divisions
arise in the primary tapetum layer. Up to this time the anther primordia are sessile
in the minute axis, which bears the perianth-collar (fig. 7). The axis thus formed by
intercalary growth finally attains such a length that it may project as much as 3 mm.
beyond the edge of the subtending bract-capsule. Its elongation is probably not
completed when the dehiscence of the anthers commences,

There has been some discussion as to the morphological character of this apparent
axis supporting the anthers. Strasburger regarded it and its homologue in the male
flower of Ephedra as being of the nature of a true ахіѕ ў. More recently, however,
evidence has been adduced in support of the view that the antherophore is not an axis
but a foliar structure formed by the fusion of what must be regarded as the filaments
of the anthers ||. This opinion is to some extent supported for Gnetum by the late
origin of this structure by intercalary growth. As pointed out by the authors cited,
this view has the advantage of offering a rational basis for the comparison of the male
flower of Welwitschia with those of Ephedra and Gnetum. If the vascular knots which
Miss Berridge J describes as existing between the origins of the outer and middle
envelopes of the female flower are the vestiges of a whorl of microsporophylls, the male
flower of Gnetum may well be derived from a hermaphrodite structure not essentially
different from that which persists in a reduced form in Welwitschia. ae |

The development of the sporogenous tissue has not been traced in detail in this
species. Several stages, however, have been seen which correspond exactly with those
described for G. africanum **. As in the latter species, the microspore mother-cell is
chambered (fig. 8), the limiting wall and the septa being 1n both eases very thick and
mueilaginous Tf. Coulter determined the reduced number of chromosomes in the
homotypie division of the mother-cell of this species to be 12 tt —

The germination of the pollen-grain has not been tollowed in detail ; its results xp
apparently the same as in 6. africanum $$. The exine of the adult pollen-gram -
thick, eovered externally by minute protuberances and without any obvious pores эң
their areas ||. The two pollen-grains shown in fig. 9 were found, with others, in the
micropylar tube of the ovule of fig. 2. Тһе nuclei were easily visible in most of these

* Pearson, 1912, pl. 60. fig. 2. + Karsten, 1894, t. 8. fig. 14. + Pearson, = figs. 5, 6.
$ Strasburger, 1872, р. 157. | Thoday and Berridge, 1912. 1 л, ^ |
** Pearson, 1912. ++ Of, Pearson, 1912, fig. 15. 22 Coulter, 1908.
$$ намә; 1912, fig. 20. iil) Of, Karsten, 1893, р. 358 ; Strasburger, 1872, р. 156.

818 DR. Н. Н, W. PEARSON ON THE MORPHOLOGY OF

grains; some contained two, others three. АП appeared to be quite free, and there
was no clear indication of the presence of a generative “ cell ” such as Karsten records
for later stages of various species and Lotsy* for G.moluccanum. The adult pollen-
grains of various Indo-Malayan species + and of G. africanum t possess three nuclei.
In all cases described, one of these (presumably prothallial in character) disappears
before the pollen-grain reaches the micropyle. Іп 6. Guemon it would appear that all
three may persist until pollination has occurred, or one of the three may disappear as in
other species. Whether the third always disappears before the pollen-grain germinates
is not known.

C. THE MacnosPORE AND ENDOSPERM. (G. africanum.)

The structure of the ovule of ©. africanum has been described in detail by
Mrs. Thoday $, that of б. Gnemon and of other species by various authors |. Much
has also been written upon the embryo-sac, embryo, and endosperm, especially of
G. Gnemon |; there are, nevertheless, still very important gaps in our knowledge
of these structures. Тһе origin and morphology of the endosperm is a subject
regarding which further information is very desirable. Two types of endosperm are
known to occur in the Gnetales. One of these, a prothallus bearing archegonia,. is
characteristic of Ephedra. The second, arising from cells uniformly constituted by the
fusion of numerous nuclei, occurs in Welwitschia. When the latter was first described **,
it was suggested that the morphological differences between these two types of endo-
sperm were of а more or less fundamental character; that Ephedra retains the
intrasporic prothallus of the lower spermaphytes ; that the endosperm of Welwitschia
is morphologically distinct from anything hitherto described for the gymnosperms,
representing, perhaps, an early term in an evolutionary series which has ended in the
highly specialized endosperm of the Angiosperms—a structure whose morphological
value is doubtful. However they are to be explained, the facts themselves are
important. Apart, then, from any morphological interpretation which may be placed
upon it, it is clearly a matter of some interest to determine whether in respect to its
endosperm Gnetum is to be classed with Ephedra or with Welwitschia.

Since previous investigators have left no room for doubt that the female gametes of
Gnetum are free nuclei as in Welwitschia, and that in both cases the accessory structures
of the archegonium are absent tt, it would appear probable, a priori, that the endosperm
of Gnetum also resembled that of Welwitschia. Although the details of endosperm
formation were not followed either by Karsten or by Lotsy, their accounts, read in the
light of what is now known for Welwitschia, contain a little internal evidence in favour
of a homology with the latter 11.

* Lotsy, 1899, p. 94. Т Karsten, 1893, p. 859; Lotsy, 1899, р. 94.
$ Pearson, 1912. $ Thoday, 1911.

|| Strasburger, 1872, 1879; Karsten, 1892, 1893; Lotsy, 1899; Berridge, 1911, 1912; &c.
ЧТ Bower, 1882; Karsten, 1. с.; Lotsy, 1. c.; Coulter, 1908. ** Pearson, 1909.
tt G. Lignier et Tison, 1912, p. 171. 33 Pearson, 1909, pp. 356-358.

CERTAIN STRUCTURES IN THE GENUS GNETUM. 319

Karsten records that in the species investigated by him *, fertilization occurs while
the sac is still unseptate and all the nuclei are free. After fertilization the lower part
of the sac (in G. Rumphianum and б. ovalifolium) becomes cellular by septation, the
“сей” frequently containing two or more nuclei. А similar condition is found also
in б. edule}. At length, apparently, all the cells of the endosperm become
uninucleate $.

Lotsy found ў that in С. Gnemon the lower (sterile) part of the sac becomes septate
before fertilization, and that at first some at least of the compartments into which it
is divided are multinucleate. All eventually become multinucleate |.

In G. Ula fertilization has not been described, and Lotsy was at first of opinion
that this species is parthenogenetic. More recently, however, he is inclined to interpret
differently the structures originally supposed to be apogametic embryos, and to regard
them as the homologues of the embryo-sac tubes of Welwitschia **, However this
may be, in this, as in the other species described formerly by Lotsy and by Karsten,
the lower part of the sac becomes septate to form a solid endosperm tissue.

In 1908, Coulter ++ reinvestigated the embryo-sac of Gnetum Gnemon. He encountered
the difficulty which has so far confronted every one who has attempted to study
Gnetum, viz. the absence of * many intervening stages" from his material. Coulter's
conclusions differ in some important respects from those of Lotsy. He describes
a clearly differentiated region of nucellar tissue at the base of the sac under the name
of “ pavement tissue.” At the time of the appearance of this tissue the embryo-sac
contains free nuclei throughout its length, and Coulter concludes that Lotsy’s “ antipodal
tissue is clearly this glandular pavement tissue developed in the chalaza”: in other
words, Lotsy’s primary endosperm is a part of the nucellus which, as Coulter correctly
states, is later destroyed by the advance of the true endosperm $$.

Lotsy reaffirms his previous statements §§ without, however, being able to throw any
new light upon the origin of the endosperm.

In the related genus Welwitschia ||| the whole вас becomes septate, before pollination,
into multinucleate compartments. In those compartments which contain many nuclei,
nuclear fusion occurs. ТЬе uninucleate cells so formed are primary endosperm cells,
all of which arise in this way. These occupy almost the whole (frequently the whole)
of the chalazal two-thirds or three-quarters of the sac. In those compartments which
enclose fewer nuclei, the latter remain free and form the functional gametes ; the walls
of the compartments grow out into the nucellar cap forming the embryo-sac tubes.

The youngest ovules of б. africanum available are just at the stage at which the
upper half of the ovule emerges beyond the edge of the cupule (РІ. 31. fig. 1). In the
youngest of these the embryo-sac contained four free nuclei qq. Ав the sac elongates
and the nuclei increase in number, they depart from the regular peripheral arrangement

* Karsten, 1893 (2). % Karsten, 1892, fig. 9 A. + Karsten, 1892, fig. 21.

$ Lotsy, 1899. - | Lotsy, 1899, р. 98. 4 Lotsy, 1903.
** Lotsy, 1911, pp. 350-352. tt Coulter, 1908. £t L. c; Coulter and Chamberlain, 1910, p. 383.
$8 Lotsy, 1911, p. 355. ||] Pearson, 1909. 9 Cf. Lotsy, 1899, fig. 25 (after Strasburger).

SECOND SERIES.—BOTANY, VOL. VIII. ЗА

320 DR. H. H. W. PEARSON ON THE MORPHOLOGY OF

which characterizes the earlier stages, and tend to become aggregated more or less in
the chalazal end of the sac (fig. 9). At this stage the embryo-sac does not show the
“biscuit-like” shape described and figured for a slightly later stage of б. Gnemon*.
Nor, in fact, does it eventually assume this form. As the nuclei further increase in
number, the aggregation at the chalazal end becomes more marked, and occupies
approximately the lower half of the sac; the cytoplasm of this region also becomes
much less vacuolated than in the micropylar half. It is clear, therefore, that the
differentiation of the sac into micropylar (fertile) and chalazal (sterile) parts is
recognizable at least as early in Gnetum as in Welwitschia. Meantime the nucellar
tissue immediately beneath the chalazal end of the вас commences to form what Coulter
has called the “ pavement tissue” +, a fairly regular'tissue in which the cells are more
or less strictly arranged in longitudinal rows (РІ, 31. fig. 13).

The next important change in the embryo-sac is the blocking-out of the cytoplasm of
the lower half into compartments. At the time of its occurrence the sac seems to be
in a particularly delicate state, and more liable to contraction than at any other period
of its history—a characteristic of the Welwitschia sac at a corresponding stage in its
development f. The first stage of septation—snch as that figured for Welwitschia $—
has not been seen. Тһе youngest stage found (fig. 12), probably slightly contracted,
shows that each compartment possesses a definite limiting membrane which, however,
does not yet react to cellulose stains |. бо far, then, the septation of the lower end of
the sac presents many similarities with that of Welwitschia. Тһе outlines of each
compartment are irregular, and the sectional area varies within wide limits. Many sacs
in this condition have been examined. Every compartment contains more than one
nucleus; most can be seen to possess five or more. The upper part of the micropylar
end of the sae of fig. 12 is shown in fig. 11. The cytoplasm is irregularly vacuolated
and shows no sign of septation; the nuclei are fairly uniform in size and appearance,
both among themselves and with the free nuclei enclosed in the lower compartments.
Their distribution in the cytoplasm is irregular. Area for area, the number of nuclei
in the upper unseptate part of the sae is smaller than in the lower half. According
to Lotsy T, this micropylar end of the sac remains unseptate aud the nuclei unchanged
until fertilization occurs.

In the septate chalazal region of the sac the nuclei in each compartment become
massed together, and the constituents of each mass fuse to form a single nucleus. The
process of fusion, as in Welwitschia, appears to be of а simple character. The contiguous
portions of the nuclear membranes cease to be visible (Pl. 32. fig. 15), and gradually
the outline of the whole mass assumes а more regular form, in which the individuality
of the fusing nuclei is lost. For a time each fusion-nucleus contains several nucleoli
(fig. 16), which, however, are soon reduced to a single one, doubtless by fusion **
(figs. 19, 22). The fusions are by no means simultaneous, for in a single sac will

* Lotsy, 1899, p. 92, fig. 98, t Coulter, 1908. + Pearson, 1909, p. 347.
$ Pearson, 1912, figs. 22, 93, || Cf. Pearson, 1. c. fig. 32. «| Lotsy, 1899, p. 93.
** Cf. Pearson, 2. c. figs. 36, 38 A; Tischler, 1900, figs. 23-36.

CERTAIN STRUCTURES IN THE GENUS GNETUM. 321

be found compartments still containing free nuclei side by side with those which by
the completion of the nuclear fusion have become uninucleate cells (Pl. 32. figs. 19, 22).
Fig. 22 shows the completion of what may be described as the primary endosperm.
Pollination usually occurs between the stages of figs. 18-22, while the nuclei of
the micropylar end of the вас appear to be in exactly the same condition as they were
when the chalazal end became septate preparatory to nuclear fusion in that region
(cf. Pl. 31. figs. 11, 14; Pl. 32. figs. 16, 20, 21).

It appears then, that while the nucellar ** pavement tissue " described by Coulter in
G. Gnemon is present in G. africanum, this species also shows the fertilization antipodal
tissue of the embryo-sac, such as was found by Lotsy in Œ. Gnemon, the existence of
which has been denied by Coulter *. This *antipodal tissue" in the embryo-sac is the
primary endosperm, consisting of cells whose nuclei are formed by fusion. The material
used in this investigation has not permitted the development of the endosperm of
С. Gnemon to be followed throughout. It does, however, show stages which leave
little doubt that in this respect G. Gnemon and G. africanum are exactly alike, and
that Lotsy's account in respect of the endosperm is perfectly accurate.

Lotsy has followed the growth of the pollen-tubes down to the embryo-sac in
G. Gnemon, and although his account of fertilization is far from complete, there can be
little doubt that any of the free nuclei of the upper unseptate part of the sac are
capable of being fertilized—in other words, are gametes.

For the present it may be safely assumed that the same is true for 6. africanum.
If so, we are in a position to establish a comparison between the unfertilized embryo-sacs
of Gnetum and Welwitschia. In the latter case we have a вас which, before septation
and nuclear fusion, contains many nuclei apparently equal in origin and potentially
equivalent in function. So far as is known, those of the upper or micropylar region of
the sac are all of them capable of uniting with sperm nuclei to produce pro-embryos.
Since in the lower part of the sac some may escape fusion, and in this case travel up
into the nucellus in embryo-sac tubes t, it appears that all the nuclei of this region are
of the same quality, and lose their sexuality only when they fuse together to form the
single nuclei of the primary endosperm cells. So far as can be seen, Gnetum africanum
agrees entirely with Welwitschia, save only that in Gnetum there is no septation in the
micropylar half of the sac. In Welwitschia this septation is, perhaps, merely of the
nature of an adaptation to the peculiar form of siphonogamy which prevails in this
genus. It appears, then, that in Gnetum, as in Welwitschia, the sac before septation
contains free, potentially fertile nuclei; that after the septation of the lower end of the
sac the nuclei here lose their sexuality by fusion, while those of the upper end, which
remain unfused, retain it. Therefore the primary endosperm of Gnetum is in all
respects homologous with the primary endosperm of Welwitschia.

The cells of the primary endosperm of Gnetum soon undergo cell-division, some of
them immediately after the completion of the fusion, with the result that the endosperm

ж Coulter, 1908; Lotsy, 1899. + Pearson, 1909, text-fig. р. 365; see also р. 373.
x i ЗА 2

822 DR. Н.Н. W. PEARSON ON THE MORPHOLOGY ОЕ

becomes a tissue composed of very unequal cells (РІ. 32. fig. 28) as in Welwitschia *. Later,
the micropylar portion of the sac becomes filled with an endosperm tissue resembling
that of the lower end. According to Lotsy, this occurs after fertilization. Тһе origin
of the upper part of the endosperm is at present not known with any certainty. Lotsy
states that *a greater or smaller number of the (unfertilized) nuclei of the fertile part
of the embryo-sae surround themselves with a denser protoplasm, a membrane, and thus
form cells" *, and that these cells later become obliterated by the upward growth of
the primary endosperm tissue below it. There is reason to think that this account шау
require some amendment. Іп any case, all but the apex of the sac—which usually
becomes disorganized in G. africanum after fertilization—is filled with an endosperm
tissue of а uniform character, and in the upper part of this the pro-embryos are
embedded $,

D. Summary oF RESULTS.

1. Gnetum Gnemon.

A. A fourth type of spike bearing male flowers and both incomplete and complete
female flowers is described. The latter are pollinated. This is apparently a
spike which has been obscurely noticed by previous writers.

B. The complete female flower of the androgynous spike occasionally possesses
four envelopes. The fourth envelope is probably due to the duplication of the
middle envelope of the normal complete flower. u

In an ovule with four envelopes, the outermost bears an apparently axillary
bud.

Both the female and the pseudo-androgynous spikes frequently bear a single
terminal ovule.

Ав described by nearly all recent investigators, the envelopes of the incomplete
female flower appear in acropetal succession. A swelling on the axis between
the insertions of the two integuments, considered by Lotsy to represent the
aborted middle integument, is recognized.

The vascular supply of the pseudo-androgynous spike resembles that already
described for б. scandens. Тһе capsule also resembles that of 6. scandens in
that it contains no mucilage cavity.

The male flower usually bears two anthers; two supernumerary anthers are
occasionally present.

Тһе antherophore arises by intercalary growth; like that of Ephedra, it is
probably a foliar structure representing two fused filaments.

I. The microspore mother-cell is four-chambered.
K. The pollen-grains in the micropyle contain either two or three free nuclei.

ы °

Ен

5

BE я

* Cf. Pearson, 1909, pl. 30. fig. 88. T Lotsy, L c. pp. 97, 98.

i Coulter, 1908, fig. 7. Mrs. Thoday (1911) states that she examined “ large fertilized ovules (of G. africanum)
containing well-developed endosperm but no embryos. This is no doubt correct. But ovules of the same
gathering, and probably in the same stage of development as those to which she refers, contain numerous

pro-embryos which are easily mistaken for endosperm cells.

CERTAIN STRUCTURES IN THE GENUS GNETUM. 323

2. Gnetum africanum.
A. The youngest embryo-sac found contained four free nuclei.
B. In a later stage the free nuclei become somewhat crowded in the chalazal end
of the sac. |
About this time the “pavement tissue” appears in the nucellus below the
embryo-sac.

. When the last nuclear division in the free-nucleate sac is completed, the lower
end of the sac becomes septate by the formation of division-walls of irregular
form ; each compartment so enclosed is multinucleate.

. Meanwhile the micropylar half of the sac remains unseptate and its nuclei free
and unchanged in appearance. So far as is known, the nuclei throughout the
sac at this stage are of the same generation; they are alike in all visible
characters.

. All the nuclei іп the multinucleate compartments enter into fusion; the fusion
is of precisely the same character as that described for Welwitschia. Each
multinucleate compartment thus becomes a uninucleate cell—a cell of the
primary endosperm.

The fusion is not simultaneous in all the compartments.

. The primary endosperm so constituted undergoes cell-division and growth.
As in Welwitschia, the cells of the later stages of the endosperm are of various
sizes. The number of chromosomes appearing in the division of the primary
endosperm cell is larger than that characteristic of the sporophyte (fig. 24).

. According to Lotsy, fertilization occurs soon after the constitution of the
primary endosperm is complete (cf. Lotsy [1899] fig. 34).

Q

g

Eo ke} ЕЗ

егі

Е. DISCUSSION.

Although there are some important gaps іп the preceding account of the origin of the
endosperm of Gnetum africanum, it is probably safe to assume that the endosperm in this
species at least is homologous with that of Welwitschia. Taking into account the previous
statements of Karsten and Lotsy regarding the development of the endosperm of the
species studied by them, there can be little doubt that, except perhaps in the time of
its formation, the endosperm of Gnetum аз a whole resembles that of Welwitschia, and
that both differ in the particulars described from that of Ephedra. Is any morphological
importance to be attached to these differences, or not ?

In the account previously given of the Welwitschia endosperm *, its morphological
character was discussed. The conclusion arrived at was that, inasmuch as the nuclei of
the primary endosperm cells were formed by the fusion of nuclei of the z-generation
which were potentially sexual, and lost their sexuality only as a result of fusion, the
endosperm was not a prothallus, but a new structure recognized as a definite morphological
entity under the name of * Trophopbyte." It was further pointed out that the endosperm
of the Angiosperm might perhaps prove to be a more highly specialized form of the

* Pearson, 1909.

324 DR. Н.Н. W. PEARSON ON THE MORPHOLOGY OF

trophophyte found in Welwitschia. There are therefore two questions concerned here,
of which the more comprehensive is that of the relation of the endosperm of Welwitschia
to that of the Angiosperm. ‘The force of any argument based upon Welwitschia alone
could not but be diminished by the aberrant character of the genus. It might be urged
that the manner in which the endosperm was formed was but a link in a chain of
adaptations to peculiar conditions. The occurrence of the same type of endosperm in
Gnetum is a fact which must be accorded great weight in any view that may be taken.

This interpretation of the Welwitschia endosperm has been discussed by Lotsy*.
After giving a brief outline of the hypothesis referred to in the last paragraph, the
author says :—“ Was ist nun der Grund, dass Pearson bei Welwitschia die Kerne des
Embryosackes, trotzdem sie, ebenso wie bei allen anderen Gymnospermen, aus wieder-
holter Teilung des Macrosporenkernes entstehen, nicht als Prothalliumkerne sondern
als Gametenkerne auffassen will? Der Grund ist

* (а) dass die des oberen Teiles befruchtet werden ;
* (6) die des unteren Teiles miteinander fusionieren kónnen."

With regard to the latter (5), Lotsy very truly remarks that it is “kein triftiger
Grund.” It may be admitted, without further question, that the fact of fusion does not

in any way help us to determine the nature of the nuclei which enter into fusion; nor,
in fact, was it ever put forward as having any bearing upon the question. These nuclei
would have been regarded as gamete-nuclei if they never fused.

The other reason given by Lotsy (a), which he describes as “ eigentlich nur ein Spiel
mit Worten,” is more to the point, though, standing by itself, its meaning and force
are not clear.

On the previous paget Lotsy admits that “alle Kerne im Embryosack von Welwitschia
homolog sind.” Therefore, if the nuclei of the micropylar region of the sac are gametes,
as they undoubtedly are, the free nuclei of the lower region, which normally fuse to form
the nuclei of the primary endosperm cells, muy be sexual also. The reasons for regarding
them as sexual, which were given in detail}, шау be repeated here, omitting the first
three, since the homology is admitted :— |

4. When septation is completed some compartments contain few nuclei (six or
less); others a larger number. Тһе latter are normally confined to the lower
three-fourths of the sac; the former may occur anywhere, though they usually
fill the upper fourth of the sac and are either absent from or much less
abundantly developed in the lower three-fourths.

5. In compartments containing less than six, wherever they may be situated, the
nuclei do not fuse, but remain free; these are functional gametes. Fusion
occurs only in those compartments in which the nuclei are in greater number ;
a sterile tissue is thus formed.

6. In an abnormal case in which the upward growth of the sac was impeded by а
second one above it, compartments which normally would have contained

* Lotsy, 1911, р. 340. + Lotsy, 1911, p. 339. 3 Pearson, 1909, р. 373.

CERTAIN STRUCTURES IN THE GENUS GNETUM. 325

gametes and produced embryo-sac tubes became uninucleate endosperm cells
by nuclear fusion.

7. Under certain circumstances, no endosperm is formed at all, but, with the
exception of a few which succumb to the greater activity of their neighbours,
all the nuclei remain free and every compartment produces an embryo-sac tube
of the normal type.

8. In two cases (РІ. 32. figs. 16, 17) about six gametes іп an advanced embryo-sac
tube have fused to form a nucleus in all respects like that of a primary
endosperm cell.

9. A small tissue found in an embryo-sac tube and consisting of seven cells is
interpreted as an endosperm formed by such a nucleus as those referred to
in § 8.

It is submitted that, when these grounds are substituted for those given by Dr. Lotsy,
it is difficult to avoid the conclusion that the unfused nuclei of the eventually sterile
part of the sac are not only homologous with, but also potentially the physiological
equivalents of, the functional gametes. They are potential gametes; they cease to be
these when they fuse. |

The same conclusion for the fusing nuclei of Gnetum is founded partly on fact, partly
on analogy. But until complete proof is forthcoming, it may reasonably be adopted.
It has already been adopted by Coulter and Chamberlain, who write :—“ The free nuclei
(of the Gnetum sac) are potential egg-nuclei, although a group near the antipodal end of
the sac may be as distinctly vegetative as are the antipodal cells of Angiosperms ” *.

When this nuclear fusion in Welwitschia was first described, reference was made to
the occurrence of multinucleate cells in the prothallia of other gymnosperms (including
Ephedra) and to the nuclear fusion which occurs in these cells in Tarus 7. Ав Lotsy |
appears to regard this fusion as in all respects similar to that which occurs in Wel-
witschia—and now in Gnetwm—further reference may be made to it.

Multinucleate cells occur in the prothallus in many Conifers $. These шау be due
either to the formation of multinucleate alveoli or to the occurrence of free nuclear
division in uninucleate endosperm cells. In most cases this condition, however attained,
is not followed by nuclear fusion. If a multinucleate cell becomes uninucleate, it does
so, not by the fusion of its nuclei, but by the formation of new dividing walls; otherwise
it remains multinucleate. Saxton, discussing these cases, concludes that “there is no
evidence that the (multinucleate) character is of any importance, except as a nutritive
adaptation.”

There are two conspicuous cases in which the multinucleate condition in the prothallial
cells of the lower Gymnosperms is followed by fusion, and they are worthy of attention.
In Ephedra, Strasburger || states, free nuclear division occurs in the neighbourhood
of the archegonia, not only in the endosperm but also in the nucellus. The division is

* Coulter and Chamberlain, 1910, p. 388. + Pearson, 1909, p. 356. + Lotsy, 1911, р. 340.
§ Lawson, 1904, 1907, 1909; Saxton, 1909, 1910, 1913; Sinnott, 1913.
|| Strasburger, 1880, pp. 106, 107.

326 DR. Н. И. W. PEARSON ON THE MORPHOLOGY OF

accompanied by the formation of cell-plates, which, however, do not stretch completely
across the cell and are evanescent. Usually there is only one such division in the cell ;
sometimes a second one follows, producing four free nuclei*. In the nucellar tissue
adjacent to the pollen-chamber the cells become disorganized, the nuclei of those which
are multinucleate fusing to a “formless mass." The endosperm cells, especially those
of the jacket-layer, mostly become binucleate ; shortly before the cells are destroyed the
nuclei fuse +. In the endosperm cells lying further from the archegonia, an incomplete
nuclear fusion occurs not rarely, but here the nuclei entering into fusion do not lose
their individuality. The fact that the cells of the nucellus in the region most liable to
exhaustion undergo changes similar to those described for similarly situated endosperm
cells, leaves little doubt that we are here dealing with degenerating cells. The multipli-
cation of nuclei and their subsequent fusion are nutritive phenomena preceding the
destruction of the cell.

Jaeger's i account of the multinucleate cells of the prothallus of Tagus is of especial
interest, since this is the case particularly referred to by Lotsy. Until June or early
July the prothallus is composed of uninucleate cells. Early in July, a month or more
after fertilization, single cells become binucleate by indirect division; cell-plates are
formed, but are of short duration. About the middle of July some of the endosperm
cells contain as many as eight free nuclei—these having possibly arisen by direct division.
At the end of July cells with sixteen free nuclei are found §. At this stage the nuclei
are apparently in perfect condition. About the end of August they show signs of
degeneration; they have decreased in size, have lost their spherical form, and have
become “ unregelmassig, zerrissen, ausgefranst." At the beginning of September the
nuclei in the endosperm cells beneath the growing embryo are usually reduced to one,
the rest to six or less. These have become further reduced in size and are of indefinite
form. Finally, the nuclei in each ceil fuse to form 4, 3, or 1 “ Kernfleck” |. These
processes occur first of all in those cells of the endosperm which are contiguous to the
archegonia or the embryo. “Das dürfte wohl auf Embryowirkung zurückzuführen
sein."

Here also we have a case of tissue-degeneration in which nuclear fusion represents
almost the final stage. It will also be noted that both in Ephedra and in Tagus the
multinucleate condition is attained by a cell which was at first uninucleate.

The nuclear fusion which constitutes the primary endosperm cells of Gnetum and
Welwitschia is of a very different character. The multinucleate character of the com-
partments in which fusion occurs is produced in the primary septation of the sac.
Neither it nor the subsequent fusion can be held to be in any way influenced by the
activities of a parasitic organ or organism. The fusion is not one of a series of changes
leading to degeneration and death; on the contrary, it inaugurates a new phase of growth
and subdivision, in the course of which a degree of organization appears in the tissue
produced ; for not only does it show a well-marked apical growth and regional distribution
of nutritive cells, but, іп Gnetum, a differentiation into “а central region of smaller,

* Jaccard, 1894, p. 34, fig. 48. Т Strasburger, 1880, pp. 106, 107. t Jaeger, 1899.

$ Jaeger, 1899, t. 17. fig. 29. | Jaeger, 1899, t. 17. fig. 30. @ Pearson, 1909, p. 370.

CERTAIN STRUCTURES IN THE GENUS GNETUM. 327

more compact cells, and a more extensive peripheral region of larger looser cells " *.

In its physiological results it is to be compared with the fusion of the polar nuclei of the
Angiosperms, not with the fusions which characterize the degenerating of the endosperm
of Taxus and of the endosperm and nucellus of Ephedra. The former is a fusion which
precedes organization ; the latter, disorganization. |

Those who favour the latter comparison, must at least accept the onus of proving that
it is areal опе. Lotsy has passed over the difficulties without any reference to them.

Lotsy} would regard each multinucleate compartment of the Welwitschia sac—whether
its nuclei are destined to function as gametes or, by fusion, to constitute a primary endo-
sperm cell—as homologous with a Coniferous multinucleate alveolus of the “ Sokolowa-
type." If the free nuclei of the latter are potential gametes, this may well be. But the
difficulty is that, so far, they are not known to fuse; those of Welwitschia and Gnetum
do. And if the nuclei of the * Sokolowa " alveolus did tuse and constitute a cell which
divided to form a tissue, would this tissue be gametophytic? In Taxus and Ephedra
the fused nuclei lose one of the most distinctive characters of the generation in whose
disorganization they appear, and they leave no descendants.

If, then, the endosperm of Welwitschia and Gnetum is not merely an adaptation of the
fusion of degeneration occurring in a few of the lower Gymnosperms, what is it? It
must be a modification of the gametophyte, or of the sporophyte, or something different
from both. In discussing this question it is well to bear in mind that no criterion dis-
tinguishing the gametophyte from the sporophyte, applicable to all cases, is yet known.
But when the life-history includes fertilization, there must be a compensating reduction
of ehromosomes. Consequently, in such cases the number of the chromosomes is at least
а character of importance. Тһе endosperm nuclei in these two genera contain a larger
number of chromosomes than even those of the sporophyte. Another character which
is always regarded as decisive when fertilization occurs, is that the gametophyte
bears the gametes. The endosperm in these genera does not in any sense bear the
gametes. Much has been written about the differences between the prefertilization-
endosperm of С. Gnemon (such as also occurs in С. africanum апа Welwitschia) and the
endosperm formed after fertilization in other species of Gnetum described by Karsten.
But, so far as is known, all species of Gnetum resemble Welwitschia in this, namely, that
the endosperm is not formed until after the constitution of the functional gametes. That
being so, no great stress can be laid upon the occurrence of fertilization before or after
the formation of the primary endosperm. Іп the cases investigated the endosperm does
not bear the gametes, but is itself produced by them i. For these reasons it is not
regarded as a gametophyte.

Some writers on the morphology of the endosperm of the Angiosperms have argued
that this tissue, arising normally from the fusion of two or three (sometimes more) nuclei
of the z-generation, is a monstrous sporophyte $. Some of the reasoning applied to
these cases applies also to Gnetum and Welwitschia. It is true that no male nucleus
participates in the fusion of these genera, but it is well known that a sporophyte is

* Coulter, 1908, p. 45. + Lotsy, 1911.
$ 6. Brown and Sharp, 1911, р. 449. § Le Monnier, 1887; Sargant, 1900; East, 1911.
(D SERIES.— BOTANY, VOL. VIII. 3B

828 DR. Н. Н. W. PEARSON ON THE MORPHOLOGY OF

frequently produced by the fusion of gametophyte nuclei of female origin only. The
monstrous form of the sporophyte might well be accounted for by the large number of
nuclei entering into fusion *.

The third alternative, namely, that the endosperm of Welwitschia (and therefore of
Gnetum) is neither a gametophyte nor a sporophyte, has been suggested +. It was
regarded as a distinct morphological entity, appearing for the first time in the higher
Gymnosperms and persisting in a highly modified form in the Angiosperms, “a by-
product resulting from the fusion of potentially sexual nuclei” and functionally
replacing the prothallus of the lower spermophytes.

It is admitted that the evidence for this last view is small, and may prove to be quite
inadequate. Except that it seems impossible to refer this endosperm to the gametophyte,
it is probably wiser at this stage not to give it a definite morphological label. But the
fact that it occurs in Gnetum, as well as in Welwitschia, still further emphasizes the
necessity of recognizing that its morphological importance is not negligible. These are
the only genera among living Gymnosperms in which an endosperm constituted іп this
manner is known to exist. The only other known case of the origin of an endosperm
from nuclear fusion is found in the Angiosperms. Since, also, there is a general consensus
of opinion that Welwitschia and Gnetum are nearer to the Angiosperms than are any
other living Gymnosperms, the possibility that the peculiar endosperm of the Angio-
sperms had its origin in a nutritive tissue not essentially different from that of these
two genera cannot be overlooked. Іп the light of present knowlege it is hardly possible
to discuss adequately the morphology of the endosperm of the higher group without any
reference to that of the Gnetoidee ў.

The relationship of the endosperm of Welwitschia to that of the Angiosperm has
recently been discussed by Samuels §. The author accepts the view that the free nuclei
of the Angiosperm sac are phylogenetically related to the free nuclei of the Welwitschia
sac :—“ La formation libre des noyaux dans les Gnétacées est assurément identique a
celle qui a lieu chez les Angiospermes.” “ТІ n'existe, à mon avis, aucune différence
essentielle entres les produits de fusion du Welwitschia et ceux des Angiospermes " ||.
He considers that the prothallus has disappeared both from the upper Gymnosperms
and from the Angiosperms €, but he does not otherwise pronounce a definite opinion
upon the morphology of the endosperm.

Samuels quotes the following expression used in comparing the saes of Welwitschia
and Peperomia ** :—“There is no evidence that the fusing nuclei of Peperomia are
potential gametes, but they are probably to be regarded as the arrested representatives
of nuclei which in earlier forms were potentially sexual" Commenting upon this

* It may be noted here that pro-embryos of the ordinary type are stated to be formed in EpAedra distachya by
the fusion of nuclei derived from the jacket-cells (Berridge and Sanday, 1907).

+ Pearson, 1909. i Of. Coulter, 1911. 5 Samuels, 1912,

|| Samuels, 1912, р. 85. А misinterpretation of the use of the word “ trophophyte ” may here be corrected.
The author says (p. 85) that the endosperm of Welwitschia, “ pour le distinguer des noyaux de l’albumen des Angio-
spermes, a été appelé par Pearson trophophyte." On the contrary, the term trophophyte was applied to the fusion-
product in both Welwitschia and the Angiosperm, to distinguish it from the prothallus of the lower Gymnosperm.

Ч Samuels, 1912, рр. 86, 97. v ** Pearson, 1909, рр. 378, 379 ; Samuels, 1912, p. 97. |

CERTAIN STRUCTURES IN THE GENUS GNETUM. | 829

he writes :--“ Comme Pearson n'argumente pas cette assertion, nous пе pouvons com-
prendre pourquoi il considére les noyaux du sac du Welwitschia entrant en fusion,
comme noyaux potentiels, ou noyaux en repos, её pourquoi il ne considére pas comme
tels les noyaux des Angiospermes." In the case of Welwitschia there is some evidence
that the nuclei which fuse are potential gametes, for if by chance fusion does not occur
in a compartment embedded in the endosperm, its walls may grow out into the nucellus
as an embryo-sac tube *. In Huphorbia virgata t, in which sixteen free nuclei appear
in the sae (megaspore), it is not known that more than one actually retains the power of
functioning as a gamete. If the remaining fifteen have lost that power, then they may
be described as “the arrested representatives of nuclei which in earlier forms were
potentially sexual.” Therefore, Euphorbia virgata being substituted for Peperomia, the
opinion expressed in the sentence above quoted is almost identical with that advocated
by Dr. Samuels.

Coulter £, regarding the endosperm of the Angiosperm as a gametophyte, considers
that “conditions in the embryo-sac favour fusions of any free nuclei, in any number and
of any origin.” ‘This conclusion will probably not be questioned. And that * there is
no necessary phylogeny of such a performance” may also be true for this case, but
it seems to demand more evidence in support of it than the author has been able to bring
together. Before it can be accepted as the whole explanation of the phenomena, it will
be necessary to show why a set of different conditions leads also to nuclear fusion in
Gnetum and in Welwitschia, and why this fusion produces physiological results so closely
comparable to those achieved in the Angiosperm. It would appear that the intracellular
conditions in the multinucleate alveolus of Cryptomeria japonica §, for example, are as
favourable to nuclear fusion as those which exist in the Angiosperm sac and in the multi-
nucleate compartment of Gnetum. And yet fusion does not occur. And if the fusion
itself possesses no morphological significance, the phylogeny of the nuclei which fuse is
still to be determined. Among the very diverse views that have been entertained, the
possibility that the polar nuclei of the Angiosperm may be morphologically (as well as
functionally) the representatives of the fusing nuclei of Welwitschia and Gnetum, has
perhaps received too little attention.

While the whole question is involved in much obscurity, it is surely not desirable at
the present stage of the enquiry that the search for a primitive type of endosperm, from
which that of the Angiosperm may be derived, should cease—which would tend to follow
from the assumption that the fusion of the polar nuclei, with or without a sperm nucleus,
means nothing more than that they could not do otherwise in the conditions prevailing
in the embryo-sac. Widely different as are the views adopted by the many authors who
have considered the question, nearly all agree that the phenomenon is one the expla-
nation of which is to be found in phylogeny |. There is at least a prima facie case for

* Pearson, 1909, p. 365, text-figure

Т When the comparison between Welwitschia and Peperomia was instituted, Brown’s (1908) results were

. unknown to me. The fact that the Peperomia sac represents four spores robs the argument of its cogency. "The

same applies to Peneca (Stephens, 1909) and Gunnera (Samuels, 1912). Euphorbia virgata (Dessiatoff, 2-2 is
free from this objection,
$ Coulter,1911, ` ` 8 Lawson, 1904, fig. 23. || Cf. Davis, 1904,
өзі i a _ во

330 Ў DR. Н.Н. W. PEARSON ON THE MORPHOLOGY ОЕ

attempting to trace the origin of the fusion to a more primitive form, such as that which
persists in what are admittedly the highest of the Gymnosperms. Those who think
otherwise have not yet stated precisely the grounds upon which their objections are based.
National Botanic Gardens,
Kirstenbosch.

BIBLIOGRAPHY.

Вкссаві, О. 1877.—Della organogenia dei fiori feminei del Gnetum Gnemon, L. Nota. Nuovo
Giornale Bot. Ital. vii. 91; Bot. Zeit. 1877, 342.

Веквюірве, Е. М. 1909.—Fertilization in Ephedra altissima. Annals of Botany, xxiii. 509.

1911.—On some Points of Resemblance between Gnetalean and Bennettitean Seeds. The New

Phytologist, x. 140.

1912.—The Structure of the Female Strobilus in Gnetum Gnemon. Annals of Botany, xxvi. 987.

Веватосе, Е. M., and E. Sanpay. 1907.—Oogenesis and Embryogeny in Ephedra distachya. The
New Phytologist, vi. 127.

Втоме, C. J. 1848.—Rumphia, iv. 2-10, tabb. 174-176.

Bower, Е. О. 1882.— The Germination and Embryology of Gnetum Gnemon. Q.J.M.S. xxii. 278.

Brown, W. Н. 1908.—The Nature of the Embryo-sac of Peperomia. Bot. Gaz. xlvi. 445.

Brown, W. H., and L. W. Swarr. 1911.--Тһе Embryo-sac of Epipactis. Bot. Gaz. Ш. 439.

Coutrer, J. М. 1908.—The Embryo-sac and Embryo of Gnetum Gnemon. Bot. Gaz. xlvi. 43.

1911.—The Endosperm of the Angiosperms. Bot. Gaz. lii. 380.

Courter, J. M., and C. J. CHAMBERLAIN. 1910.—Morphology of Gymnosperms. Chicago.

Davis, В. М. 1904.—Studies on the Plant Cell. V.sect.iv. Amer. Nat. xxxviii.

Desstatorr, М. 1911.—Zur Entwickelung des Embryosackes von Euphorbia virgata, W. К. Ber. d.
Deutsch. Bot. Ges. xxix. 33-39.

East, Е. М. 1913. Xenia and the Endosperm of the Angiosperms. Bot. Gaz. lvi. 217.

Gorset, К. 1905.—Organography of Plants, vol. ii. Oxford.

Скіггітн, №. 1859.— Remarks оп Gnetum. Trans. Linn. Soc. xxii. 299-312, tt. 55, 56.

Ноокек, J. D. 1863.—0n Welwitschia, а new Genus of Gnetaces. Trans. Linn. Soc. xxiv. 1-48,
tt. 1-14.

Jaccanp, P. 1894.— Recherches embryologiques sur l' Zphedra helvetica. Lausanne.

Јлевек, L. 1899.—Beitrage zur Kenntniss der Endospermbildung und zur Embryologie von Tazus
baccata, L. Flora, 1хххуі. 241.

Karsten, (+. 1892— Beitrag zur Entwickelungsgeschichte einiger Gnetum-Arten. Bot. Zeit. 1. 205.

1893.— Untersuchungen über die Gattung Gnetum. I. Beitrag zur systematischen Kenntniss

der Gnetum-Arten im Sunda Archipel. Ann. Jard. Bot. Buitenzorg, xi. 195.

—— 1893(2).—Zur Entwickelungs-Geschichte der Gattung Gnetum. Cohn’s Beitr. z. Biol. d.
Pflanzen, vi. 337.

Lawson, A. A. 1904.—The Gametophytes, Fertilization, and Embryo of Cryptomeria japonica.
Annals of Botany, xviii. 417.

— 1907,—The Gametophytes and Embryo of the Саргезвіпеге, with Special Reference to Libocedrus
decurrens. L. c. xxi. 281.

--- 1909.—The Gametophytes and Embryo of Pseudotsuga Douglasii. L. c. xxiii. 163.

Le Мохмікв, б. 1887.—Sur la valeur morphologique de l'albumen chez les Angiospermes. Journ. de
Botanique, i. 140.

` LIGNIER, O., et A. Тіѕох. 1912.—Les Gnétales, leur Fleurs et leur Position systématiques. Ann. & .
Sci. Nat., Bot. 9™° sér., viii. 55. `

Lorsy, J. P. 1899.—Contributions to the Life-history of the Genus Gnetum. I. The Grosser
a of Production of Gnetum Gnemon, L. Ann. Jard. Bot. Buit. (2) i. 46.

TRANS LINN бос ӛвнв.Ә.Вот.УоГҮШІРІ.31.

Pearson.

С.Ноддез, lith. & imp
ІМ GNETUM.

REPRODUCTION

Т.В. & HHWE del.

TRANS.LINN. Ѕос. SER 2.Вот Мо VIILPL32.

Реагзоп.

› y^ "A
ЖАТР,

n
v

‚т

БЕ та

ғ
„=“

5

ГРА,
*
Pe,
vi
К

JN
»

LB. HRWP del

EH ‚К & imp.
IN GNETUM. ipe сир ины

J REPRODUCTION

CERTAIN STRUCTURES IN THE GENUS GNETUM. 391

Lorsv, J. P. 1908.—Parthenogenesis bei Gnetum Ula, Brongn. Flora, xcii. 397.
1911.—Vortráge über Botanische Stammesgeschichte gehalten an der Reichsuniversitat zu
Leiden. Ein Lehrbuch der Pflanzensystematik. III. Teili. Jena.

Pearson, Н. Н. W. 1906. Some Observations on Welwitschia mirabilis, Hook. t Phi Trans; B.
excviii. 265-304, pls. 18-22, figs. 1, 2. :

— 1909.—Further Observations on Welwitschia. L. с. сс. 881-102, pls. 22-30.

—— 1912.— On the Microsporangium and Microspore of Gnetum, with some Notes on the Structure
of the Inflorescence. Annals of Botany, xxvi. 603-620, pl. 60, 6 figs.

Porscu, О. 1910.—Ephedra campylopoda, C. А. Mey., eine entomophile Gymnosperme. Ber. d.
Deutsch. Bot. Ges. xxviii. 40.

SAMUELS, J. А. 1912.—Etudes sur le développement du sac embryonnaire et sur la fécondation
du Gunnera macrophylla, Bl. Arch. für Zellforschung, viii. 53.

Sarcant, Е. 1900.—Recent Work on the Results of Fertilization in Angiosperms. Annals of Botany,
xiv. 689. -

Saxton, W. T. 1909.— Preliminary Account of the Ovule, Gametophytes, and Embryo of Widdring-
tonia cupressoides. Bot. Gaz. xlviii. 161.

— 1910. Contributions to the Life-history of Widdrinytonia cupressoides. L. с. 1. 81.

---- 1913.—Contributions to the Life-history of Tefraclinis articulata, Masters, with some Notes on

the Phylogeny of the Cupressoidez and Callitroides. Annals of Botany, xxvii. 577.

1913 (b). Contributions to the Life-history of Actinostrobus pyramidalis, Мід. Г. с. xxvii. 322.

бімхотт, E. W. 1913.—The Morphology of the Reproductive Structures іп the Podocarpinew.

Annals of Botany, xxvii. 39.

Srernens, Е. L. 1909.—The Embryo-sac and Embryo of certain Penæaceæ. Annals of Botany,
xxiii. 363.

Srraspurcer, E. 1872.— Die Coniferen und die Gnetaceen. Leipzig.

1879.—Die Angiospermen und die Gymnospermen. Jena.

1880.—Zellbildung und Zelltheilung. Jena.

Зукиз (afterwards Тнорлх), М. G. 1910.—The Anatomy and Morphology of the Leaves and
Inflorescences of Welwitschia mirabilis. Phil. Trans., B. cci. 179.

Тнорду (formerly Svxzs), М. б. 1911.--Тһе Female Inflorescence and Ovules of Gnetum africanum,

with Notes on Gnetum scandens. Annals of Botany, xxv. 1101.

1912. Note on the Inflorescence Axis іп Gnetum. L. с. xxvi. 621.

TiscHLER, Q. 1900.—Untersuchungen über die Entwicklung des Endosperms und der Samenschale
von Corydalis cava. Verh. d. Naturhist.-medizin. Vereins zu Heidelberg, vi. 351.

TuLAsNE, 1863, in Martius, Flora Brasiliensis, iv. pars 1. t. 103.

1858.—Gnetacez Americz australis. Ann. d. Sci. Nat., Bot. 4"* sér., x. 111.

Werrsrern, R. von. 1907.— Ueber das Vorkommen zweigeschlechtiger Infloreszenzen bei Ephedra.
Festschr. d. Naturwiss. Ver. an der Universitit Wien, 21-28, tab. 1.

WirLawp, G. В. 1908.--Accelerated Cone-growth in Pinus. Am. Journ. Sci. xxv. 102.

EXPLANATION OF THE PLATES.

PLATE 31.
Figs. 1, 10-23 — Gnetum africanum ; Figs. 2-9 — G. Gnemon.
Fig. 1. Female spike of б. africanum, showing aborted ovules and immature ovules and seeds.
Natural size.
Fig. 9. Median longitudinal section through an ovule from the lowest node of an androgynous spike of
С. Gnemon. v.b.— vascular system (diagrammatic); e.s.= embryo-sacs; 1,2, 3, 4—envelopes.

=
9%
©

i

к.
03

сл

Fig. 23.

ON THE MORPHOLOGY OF THE GENUS GNETUM.

. Tangential longitudinal section through the same ovule. М = nucellus; ó = bud in ахі] of

outer integument. x 300.

. Part of a longitudinal section through the node of the pseudo-androgynous spike, showing an
_ Incomplete female flower, two male flowers, and undifferentiated meristem (М). N=

nucellus; О = outer, I= inner; m= ? middle envelopes of ovule; 4 = male flower;
P= perianth of male flower. x 300.

. Longitudinal section of an incomplete female flower, slightly older than that of fig. 4.

О = outer, [= inner, m= ? middle envelopes. х 300

. Longitudinal section of a male flower bearing four anthers. P= perianth. х 70.
. Longitudinal section of a very young male flower. The primordia of the two anthers are

indicated. Р = ring-like perianth. х 300

. A four-chambered microspore mother-cell. x 1200.
. Two pollen-grains from the micropyle of a complete flower of an androgynous spike. х 1200.
. A longitudinal section through an embryo-sac of G. africanum in the free-nucleate condition.

The nuclei at the chalazal end of the sac are crowded. х 300.

. Top micropylar end of the same sac as fig. 12, showing free gamete-nuclei.
. The chalazal end of a later stage of the embryo-sac, showing septation into multinucleate

compartments. x 700.

. The chalazal end of an embryo-sac with the underlying nucellar tissue, showing an early stage

in the formation of “ pavement tissue ” ( p.t.).

. Gamete-nuclei from the micropylar half of ап embryo-sac.

PLATE 92.

. Three cells from the chalazal end of the same sac as in fig. 14, showing stages in nuclear

fusion. x 1200.

. Free gamete-nuclei from the ор half of a вас.
. Cells from the upper part of the lower half of the same sac, showing an advanced stage in

nuclear fusion, the nucleoli being still unfused. х 1200.

. Outline of a more advanced embryo-sac, showing approximately the division into sterile (0) and

fertile (с) regions. х 70.

The lower part of the endosperm-tissue of the sac of fig. 18, with a part of the underlying
nucellar pavement tissue. In most of the endosperm cells nuclear fusion is proceeding or
completed ; in a few the nuclei are still free.

- The upper part of the micropylar region of the same sac, showing the irregularly distributed

free gamete-nuclei.

. Longitudinal section through micropylar half of a more advanced embryo-sac, showing fertile

region with free gamete-nuclei and its junction with the top of the primary endosperm.

- Chalazal end of the same sac, showing almost the whole of the primary endosperm. Except in

a few cells, nuclear fusion is complete. х 300.
The pollen-tube has traversed the greater part of the length of the nucellar cap, but has
not yet reached the embryo-sac.
A portion of the endosperm of a more advanced (? fertilized) ovule after the division of some
of the primary endosperm cells, showing the very different sizes of the constituent cells
(cf. Welwitschia, Pearson, 1909, fig. 88). х 800.

[ 2958 |

X. Оп Isoëtes japonica, 4. Br.
By Cyr West, F.L.S., and Н. TAKEDA, D.I. C.

(Plates 33—40 and 20 Text-figures.)

Read 3rd December, 1914.

EXTERNAL MORPHOLOGY.

[SOERTES JAPONICA, А. Br., which has а fairly wide distribution in Japan, is the
largest in habit of any known species of this genus. The caudex * of an old plant often
attains a diameter of 4 cm., and, according to Makino (22, р. 146), a diameter of 8 cm.
may be reached in very large specimens. The caudex is tri-lobed. In young plants
the caudex is sometimes very flat (РІ. 33. fig. 6), but in older plants, in which it may
attain a height of over 3 cm., it appears roughly cylindrical or napiform with three deep
longitudinal grooves (РІ. 33. figs. 4, 7). Except in young specimens, each of the three
main lobes of the caudex becomes longitudinally subdivided by a number of very irregular
shallow fissures (about half as deep as the main grooves), so that, in transverse section,
each lobe appears irregularly incised (РІ. 38. figs. 56-59). The upper half of the caudex
is covered by the ‘rosette’ of leaves and by the very short leaf-bases. The cells of the
latter become slightly thickened and suberized, but they never form a phyllopodium as
defined by A. Braun (6, p. 581). The structure designated ‘ phyllopode’ by Makino
(1. с. pp. 181, 146) is actually the vagina of the leaf.

_ Along the grooves and fissures the peripheral cells of the cortex also become
suberized and form a thin covering of periderm. А similar periderm is formed on the
lower surface of the caudex. Numerous roots of a dirty brownish colour anchor the
plant firmly in the mud at the bottom of a lake or stream.

Leaves are produced in large numbers ; occasionally, more than 200 fully-developed
leaves may be counted on a single specimen. These, in plants growing in deep water,
attain a length of 1 m. or more. The terminal portion of the leaf is usually aerial and
beset with stomata, but leaves which happen to be totally submerged are devoid
of stomata. den

In transverse section the leaves appear semiterete-quadrangular with the adaxial
surface much flattened. Тһе leaf is dilated towards its base, as in all other species of
Isoëtes, but the wings are comparatively narrow, the width of the vagina at its base
never exceeding 2 сш. А few sterile leaves occur on the margin of the leaf-rosette, but
the remainder are fertile. In the centre of the rosette of leaves, where a funnel-shaped

* We prefer the term ‘caudex’ for the structure usually referred to as the ‘stem,’ ‘rhizome,’ ‘trunk,’ or
ç к?

934 MESSRS. CYRIL WEST AND Н. TAKEDA ON

depression is formed in the caudex, there is a cluster of short immature leaves, which
will develop in the following season (РІ. 33. figs. 7, 8; РІ. 34. fig. 10).

It is generally stated (3, р. 84; 6, р. 563; 27, pp. 7 57-8), that in Isoëtes the outer-
most leaves of the ‘rosette’ are megasporophylls *, and that these are followed by a
eycle of microsporophylls whilst the innermost zone is made up of sterile leaves. This
statement, however, is incorrect. In J. japonica the sterile leaves, which form the
transition from one year’s increment to the next, are found on the periphery of the
‘rosette.’ The mega- and micro-sporophylls are not arranged in any definite order, but
are irregularly distributed throughout the wide zone of fertile leaves. When a plant
is dug up late in the year, it often happens that the sterile leaves of the outermost whorl
of the rosette have by that time completely decayed away; under these circumstances
sporophylls are found on the margin. But if such а leaf-rosette be carefully dissected,
young sterile leaves will be found in the centre; these are generally regarded as
belonging to the present season's growth, but actually they represent the outermost
cycle of next year's rosette.

"The sporangia are elliptical or narrowly oblong and are embedded in the fovea, the
margin of which does not project as a velum.

A well-developed ligule is borne just above the fovea. Тһе ligule is ovate-lanceolate,
cordate at the base, and elongate-acuminate towards the apex (Pl. 36. figs. 25, 26, 27,
33). Ina small specimen the length of the ligule is about 8 mm., but in a large plant
a length of nearly 15 mm. is reached, whilst the width may exceed 5 mm.

MATERIAL AND METHOD.

The material used in this investigation was collected near Tókyó by one of us in
September, 1913. Plants of all ages (several weeks to many years old) were represented.
The larger specimens were collected from a fairly rapid stream some 3 feet deep, whilst
the smaller plants were gathered from narrow ditches where the water was very shallow.
All the specimens were preserved in alcohol (about 75 per cent.).

Sections were cut either by mierotome or by hand, and were mounted in Canada
Balsam or in Glycerine-jelly to which various stains had been added. For the general
work we employed the Safranin-Hsematoxylin combination, but for special purposes
several other stains and reagents, including Azo-blue, Corallin-soda, Gentian-violet,
Phloroglucin, Chlor-zinc-iodine, Iodine-sulphuric acid, and Sudan III were found useful.

GENERAL ANATOMY OF THE CAUDEX.

We have nothing to add to the existing accounts of the structure and development
of the very young sporophyte of Isoëtes, the anatomy of which has been carefully
studied by Hofmeister (16), Bruchmann (7), Farmer (14), and Campbell (10). The
" conclusion generally arrived at is that the axile stele of the young sporophyte is made

* Milde (25, p. 275), on the other hand, states that the outermost leaves are microsporoph y lls.

pw

ISOETES JAPONICA, А. BR. 335

up entirely of the united basal regions of the leaf- and root-traces. With this explanation
of the vascular anatomy of the very young sporophyte we quite agree, but our

Text-fig. 1.

`
'

\

'

`
4-

ч
3

a asue as

a
Diagram showing arrangement of primary and secondary stelar tissues of Zsoétes japonica. From a median
longitudinal section of the caudex cut in the plane of a main furrow. х 20.

e. = cambium. 1.4. = leaf-trace. mer. = primary meristem.
ph. = primary phloem. ph.” = secondary phloem. p.m. = parenchymatous mantle.
i ту. = xylem.

SECOND SERIES.—BOTANY, VOL VIII. 3c

336 MESSRS. CYRIL WEST AND Н. TAKEDA ON

investigations have convinced us that the view held by many botanists, who regard the
primary vascular cylinder of the adult stem as being made up of a sympodium of
leaf-traces, is untenable. :

For the reasons given in detail below (p. 338), we have adopted the view already
held by Bruchmann (7), Hegelmaier (15), and Scott and Hill (28), that part, at least, of
the stele of the adult plant is cauline.

Text-fig. 2.

> M e
Ñ > ot 32:

Diagram showing arrangement of primary and secondary stelar tissues of Isoëtes lacustris. From a
median longitudinal section of the caudex cut in the plane of the furrows. х 50.
(For lettering, see text-fig. 1, p. 335.)

Тһе stele of 1500468 is generally regarded as a single structure (7. e. a stem-stele), but
we are of the opinion that the stele of the adult sporophyte consists of two distinct
parts, namely, a vertical cylindrical portion to which the leaf-traces are confined and
which constitutes the stele of the stem proper, and a relatively flattened bi- or tri-lobed
basal portion to which all the root-traces are attached (text-figs. 1, 2). The latter we
propose to call the rhizophore-stele, since it belongs to a perfectly distinct root-bearing

pr.

ISOETES JAPONICA, А. BR. 337

organ of the plant [cf. Lang (20), р. 784], analogous to the swollen basal region of
the stem of Selaginella spinosa [Bruchmann (8)], or of Pleuromeia (Solms-Laubach,
(3121.

We assume, however, that owing to its extremely stunted growth, the plant has
completely lost all external morphological differentiation into stem and rhizophore.

APICAL MERISTEM AND DEVELOPMENT OF THE PRIMARY TISSUES OF THE STEM.

The question as to the exact nature of the apical meristem of the stem of Isoëtes has
given rise to much discussion in the past.

The first botanist to investigate this problem was Hofmeister (16), who concluded
that the tissues of the stem are derived from the segments of a single apical cell, the
septa of which are turned towards the furrows of the stem. This apical cell was found
at the apex of a blunt cone of cellular tissue situated at the base of the funnel-shaped
depression in the upper surface of the cortex (16, pl. li. fig. 4). In arriving at this
conclusion, Hofmeister was probably influenced by the prevailing conception of a single
apical cell at the apex of the stem of all Vascular Cryptogams.

In the year 1874, this question was discussed in papers by. Hegelmaier (15) and
Bruchmann (7) respectively. The former maintained that the apical meristem of the
stem of Г. velata and of I. Durieui consists of an actual apical-cell surface (Scheitel-
zellenflüche), and that the cells situated near the centre of this superficial layer of
meristem give rise to the mother-cells of the axile woody mass, hence they may be
regarded as plerome initials. Тһе more peripheral cells of this meristematic layer, by
repeated divisions in an oblique plane, give rise to the primary cortex and to the leaves.

Bruchmann, as a result of his investigations on J. lacustris, arrived at a somewhat
similar conclusion. He states (7, p. 576), that “ Der Stamm wachst durch Urmeristem-
Initialen, die anfangs nur aus wenigen Zellen bestehen.” In other words, the actual
apical meristem is confined to a group of a few cells, which in older plants forms a slight
elevation near the centre of the superficial cells of the stem apex. No definite apical
cell was observed, but these cells (Urmeristem-Initialen), which шау be distinguished
from the other superficial cells by their denser protoplasm and by a slight difference in
size, divide tangentially, giving rise to an underlying group of meristematic cells, from
which both the central mass of xylem and the tissue of the cortex are produced.

In the first edition of his * Traité de Botanique, Van Tieghem, referring to the growth
of the stem of Isoëtes, remarks (33, p. 1294) that “Та tige croit, comme celle des
Lycopodes, par un groupe de petites cellules méres; mais ici lon n'observe dans ce
groupe aucune differentiation d'initiales propres." But, in a later edition of the same
work (34, p. 1429), the following sentence bearing on this subject occurs, “ La tige croit,
comme celle des Lycopodes, par une seule petite cellule тёге,”

Campbell (то, p. 249) states, in reference to the very young sporophyte, that * the
growing-point of the stem .... is a nearly flat area, whose surface is nearly at right

angles to the axis of the leaf."
8c2

338 MESSRS. CYRIL WEST AND H. TAKEDA ON

Farmer (14) follows Hegelmaier in ascribing to the flat apex of the stem a layer of
columnar cells in which most of the divisions are anticlinal, periclinal divisions occurring
at rare intervals. The infrequent occurrence of periclinal divisions was correlated with
the slow growth in length of the stem in this genus.

A similar conclusion was arrived at by Wilson Smith (30, p. 228), who adds, however,
that in longitudinal sections of small plants the stem apex is distinguishable as a slight
elevation. |

Scott and НШ (28), оп the other hand, bring forward, with some reservation,
evidence in favour of Hofmeister’s original conception of a single apical cell. In our
opinion no importance should be attached to the presence of a single cell of slightly
larger dimensions than its neighbours, since several large cells can usually be found
irregularly distributed over the entire surface of the apical region of the stem. More-
over, these cells often possess a comparatively large nucleus, and may possibly represent
leaf-primordia. : :

Text-fig. 3.

Isoëtes lacustris : longitudinal section through the apex of the stem. Тһе plane of the section corresponds
to that of the main furrows. G.P.=growing-point of the stem. x 300.

A careful examination of many series of microtomed sections, both transverse and |
longitudinal, which were compared with thick hand-sections cleared with potash or
‘eau de Javelle, has revealed the true apex of the stem as a relatively large conical
protuberance at the base of the funnel-shaped depression in the centre of the leaf-rosette
(text-figs. 1, 2, 3; РІ. 39. figs. 65-72).

We have observed this conical mass of tissue at the apex of the stem of four distinct
species, namely, T. japonica, I. lacustris, I. Hystrix, and I. velata * ; it occurs even in
ihe very youngest specimens of J. japonica and of Г. lacustris which we examined.
This protuberance has probably been mistaken for one of the leaf-rudiments to which,

* We are indebted to Prof. J. B. Farmer, F.R.S., for the loan of several slides of the two last-named species.

ISOETES JAPONICA, A. ВВ. 339

however, it bears only a very superficial resemblance. Whereas the leaf-rudiments
possess a thin layer of cuticle, the conical protuberance which constitutes the apex of
the stem has no cuticle.

It is impossible to distinguish any definite apical cell amongst the superficial cells of
this protuberance, all of which, as Bruchmann (7, p. 568) correctly observed, are of
similar shape and size. These cells usually divide by anticlinal walls, although occasional
periclinal divisions are found.

In the underlying tissue (‘Urmeristem’ of Bruchmann), the cells of which are
irregularly arranged, anticlinal and periclinal divisions occur in equal numbers. The
relatively large parenchymatous cells lying immediately below the protuberance are
arranged in more or less regular series (Pl. 39. figs. 67, 69, 70) and constitute a poorly
defined plerome, the cells of which, probably owing to the peculiar habit of the plant,
are not vertically elongated.

Traced still further downwards, a certain number of these large cells are found in all
stages of lignification (РІ. 34. figs. 10, 12). These lignified elements form the tracheids
of the stem-stele. Scott and Hill (28, p. 420, Pl. 23. fig. 6) follow Van Tieghem in
ascribing a centripetal development to these lignified elements, but in every one of our
preparations of J. japonica the development of these tracheids is obviously centrifugal
(Pl. 34, fig. 12). It is, perhaps, inadvisable to lay great stress on this point, since a
certain amount of irregularity may be expected in such a slow-growing organ.

The remaining cells retain their meristematic condition for a considerable period and
ultimately give rise to (1) the numerous small parenchymatous cells (xylem parenchyma),
which with the tracheids constitute the xylem of the stem-stele, and (2) a zone of
parenchyma which completely surrounds the xylem except where interrupted by the
outgoing leaf-traces. This peripheral zone of parenchyma occurs as a definite structure
in every species examined ; we therefore propose to call it the * parenchymatous mantle’
(text-figs. 1, 2; Pl. 34. figs. 10,13; РІ. 38. fig. 62).

The sieve-tubes of the primary phloem are usually differentiated from the central
cells of this zone of parenchyma, in which case they are surrounded both internally
and externaliy by one or more layers of parenchyma (Pl. 84. figs. 10, 13). But
occasionally ihe more peripheral cells of the parenchymatous mantle become differ-
entiated as sieve-tubes, and since the cambium arises from the outermost layer of the
plerome (i. e. outermost layer of the ‘ parenchymatous mantie ' it follows that the
sieve-tubes of the primary phloem will then abut directly on the tissues which result
from the activity of this cambium (РІ. 35. fig. 21; Pl. 38. fig. 64).

We therefore agree with Scott and Hill (28, p. 428) and Stokey (32, p. 320) that
the tissue in immediate contact with the primary tracheids is always parenchymatous ;
as to whether this parenchyma belongs to the xylem rather than to the phloem we
prefer to leave an open question, but we usually find that near the apex of the
stem the xylem tracheids are differentiated before the sieve-tubes of the primary
phloem (text-fig. 1; РІ. 34. figs. 10, 12).

340 MESSRS. CYRIL WEST AND H. TAKEDA ON

Possibly this layer of parenchyma, which retains its meristematic activity for a
considerable period, was mistaken for a cambium by Scott and Hill (28, Pl. 23. fig. 8).

The primary cortex is formed by the active division of the cells of the apical meristem
which abut on the plerome. These cells (х x in Pl. 39. figs. 67-69) divide most frequently
by oblique anticlinal walls and give rise to parenchymatous tissue which is not sharply
marked off from that of the secondary cortex and of the leaf-bases.

STRUCTURE OF THE PRIMARY TISSUES OF THE STEM.

1. Primary Xylem.

Stokey (32, p. 314) maintains that “іп the short and compact stem there is no
indication of a procambial strand, and as there are never any tracheids present above
the last leaf-traces the evidences of a cauline portion are not satisfactory.” However,
neither of these objections will stand the test of a critical examination. Let us consider
the first statement that “there is no indication of a procambial strand." Та all four
species examined by us, a regular arrangement of the meristematic cells lying
immediately above the mass of primary xylem tissue was noticed; these meristematic
cells form а procaimbial strand, which is less distinct than that of Lycopodium, owing
to the fact that in Zsoëtes the cells are not vertically elongated; this difference may be
correlated with the very slow growth of the stem of Isoëtes as compared with that
ot Lycopodium, With reference to the second objection, we need only mention the
fact that both transverse and longitudinal sections through the apical region of the
stem clearly show that the tracheids lying immediately below the procambial strand are
fully differentiated before those of any leaf-traces which would connect up with the stem
stele at that level (Pl. 34. fig. 12),

Hofmeister (16, p. 356), referring to Isoëtes lacustris, states that “ тапу other cells
besides the primary cells of the vascular bundles take part in the wood-formation."

Lastly, as further evidence for regarding the stem-stele of the adult Isoëtes as a
cauline structure, we may mention the fact that the tracheids of the leaf-trace are
markedly narrower than, and altogether different in appearance from, those of the stem-
stele (Pl. 34. figs. 11,12; Pl. 35. fig. 21).

Hence it follows that at least part of the central xylem mass of the adult plant
represents the xylem belonging to the stem itself. In other words, the adult plant of
Isoétes japonica possesses a true cauline stele [af. Scott & Hill (28), p. 426]. Тһе
primary xylem-cylinder of the stem of 7. japonica is composed of numerous large, oval,
reticulate tracheids intermixed with abundant xylem parenchyma, the cells of which
retain their meristematic condition (Pl. 34. figs. 11, 12; РІ. 40. figs. 75, 78).

The tissues of the primary xylem-cylinder are obviously subjected to a very
considerable strain in the horizontal direction, which results in the almost complete

disorganization of the tracheidal elements. The cells of the xylem parenchyma, which
are capable of further division, are able to accommodate themselves to this pulling
strain, and appear healthy even in quite old plants (РІ. 35. fig. 14).

ISOETES JAPONICA, А. ВВ. 341

The factors which bring about this result were investigated by Farmer (14), and we
cannot do better than quote his explanation, which is entirely confirmed by our own
investigations. Farmer (14, р. 41) states that “This result is indirectly brought about
by the activity of the cambium. The traces of the roots and leaves which pass through
this zone to join the axile vascular cylinder are subjected to great strain by the faci
that, whilst the cells round them are growing, and, so to speak, travelling outwards, they
themselves are stationary, and can only increase in length by extension. This takes
place to а limited extent, and effects а great distortion of their constituent cells, but
finally the loose central tissue gives way, and they are partially relieved of the strain
at the expense of the cohesion of the inner tissue." Moreover, this explanation accounts
for the fact tbat the pull exerted is in the horizontal direction. In this connection the
‘parenchymatous mantle’ probably plays an important part. As mentioned above, the
primary xylem is completely enveloped by the ‘parenchymatous mantle,’ which is
uninterrupted except by the leaf-traces. The parenchyma of the leaf-traces is connected
up with that of the ‘ parenchymatous mantle,’ so that as the leaf-traces are pulled away
from the central axis of the stem they drag the ‘mantle’ with them, with the result
that the weaker internal tissues of the stele are compelled to give way. The presence
of this parenchymatous mantle also explains the regular outline of the primary tissues
of the stele of an adult plant as shown in а median longitudinal section of the stem;
if the leaf-traces were connected up directly with the loose internal xylem-parenchyma
or with the tracheids of the stem-stele, the outline of the latter would be very
irregular (cf. text-figs. 1, 2; РІ. 83. fig. 8; Pl. 34. fig. 10).

2. Primary Phloem.

The sieve-tubes of the primary phloem have thick cellulose walls, which, in well-
stained preparations, sbow typical sieve-areas. In many of these primary sieve-tubes
deposits of callus were observed, and in some cases the deposit was so abundant as to
completely fill the cavity of the cell (РІ. 34. fig. 13; РІ. 38. fig. 64).

Primary phloem of a similar character was also found іп Isoétes lacustris (Pl. 38.
fig. 62).

The sieve-tubes, which are continuous with those of the leaf-traces (Pl. 34. fig. 13;
РІ. 35. fig. 15), are associated with unmodified cells of the ‘mantle,’ several of which
contain a semi-transparent substance, which is remarkably resistant to stains. Possibly
these parenchymatous cells have a function similar to that of the companion cells of
the sieve-tubes of Phanerogams.

The cells of the primary phloem are seldom arranged in regular radial rows, and it
is obvious from their arrangement that they cannot have been produced by the activity
of the secondary cambium. They are often elongated tangentially (Pl. 34. fig. 13),
being subjected to the same horizontal pull as that mentioned above in connection with
the primary xylem. The primary phloem of the stem would serve to connect up the
phloem of the first-formed leaf-traces with the corresponding tissue of the earlier roots,
prior to the formation of the secondary phloem. |

342 MESSRS. CYRIL WEST AND H. TAKEDA ON

3. Primary Cortex.

Тһе primary cortex of the youngest specimen examined by us is composed of large
parenchymatous cells packed with starch and proteid granules. In the younger part
of the plant these cells are isodiametric and form a compact homogeneous tissue, but
in the central region of the cortex the cells become loosely arranged with large
intercellular spaces. Towards the outer limits of the cortex these cells become
collenchymatous, their walls meanwhile increasing in thickness. At this stage the
starch-grains suddenly disappear, leaving a loose alveolar mass of protoplasm which
has been ‘moulded, so to speak, by the starch-grains before their disappearance.
РІ. 88. fig. 60 shows a cortical cell in this condition; numerous proteid granules can
still be seen attached to the protoplasmatic framework.

On the margin of the caudex no definite epidermis was observed, but the outermost
layers of cortical cells form a rudimentary periderm, the walls of these cells giving the
recognized suberin reactions on treatment with Sudan 111., ete.

A point of especial interest in connection with J. japonica is the presence of an
intracellular mycorrhiza in the peripheral cells of the primary cortex. The very fine
freely branched non-septate mycelium of the fungus was found in many of these cells.
No reproductive bodies of any description were noticed, but the hyphee formed numerous
spherical swellings [vésicules of Janse (18, р. 148, РІ. 13. figs. 7, 8)], either terminally
or in an intercalary position (РІ. 38. fig. 61). These swollen portions of the hyphe
contain, at an early stage in their development, a dense granular substance very similar
to that found in the hyphz themselves. The chemical nature of this granular
substance was not investigated ; it is probably a proteid substance. As these “үбвісшіев”
increase in size they appear to lose their contents, the largest, and therefore, oldest
‘yésicules’ being quite empty. It is not known why these swellings are produced, but,
judging from their wide distribution in the mycorrhiza of several distantly separated
groups of plants *, they no doubt possess some unknown but nevertheless important
function connected with the nutrition of the fungus or of the host.

Owing to the complete absence of reproductive bodies we were unable to determine
the systematic position of this endophytic fungus.

STRUCTURE OF THE SECONDARY TISSUES OF THE STEM.

1. Cambium.

In Isoëtes japonica cambial activity starts very early; obvious periclinal divisions in
the outermost cells of the ‘parenchymatous mantle’ were observed in a very young
sporophyte, the greatest diameter of the caudex of which measured only 2 mm. The
cambium, once formed, persists throughout the life of the plant; more than one
cambium in the stem was never observed in our preparations.

* E. g, Marattiacere, Ophioglossaceze [Campbell (11), р. 18), Psilotacee [Janse (18), р. 143].

ISOETES JAPONICA, А. BR. 343

2. Secondary Phloem.

The cambium, as stated above, arises very early. It at once begins to cut off new
tissues both internally and externally.

Towards the interior the cambium cuts off a tissue the true morphological nature of
which has, in the past, provided the subject for much discussion. This ‘tissue consists
of compact angular cells more or less regularly arranged in radial rows, and completely
surrounds the primary stele of the stem with the exception of the apical region of the
latter, forming a zone which is thickest in the oldest (2. е. lowest) region of the stem
(text-figs. 1, 2).

Considerable interest attaches to this tissue, which is, in many respects, unique in the
Vegetable Kingdom.

Von Mohl (24), to whom we owe the first really scientific description of the structure
of the genus Jsoétes, suggests that the secondary tissue consists entirely of parenchyma.

According to Hofmeister (16) а few spiral cells become added to the wood of old
vigorous plants as a result of the cambial activity.

Russow (26), as a result of his anatomical investigations upon herbarium material
of several species of Isoëtes, arrived at the conclusion that the tissue cut off internally
by the cambium consists of parenchyma interspersed with tracheids and with cells
which he referred to the phloem.

Hegelmaier (15) refers to this tissue in Г. velata and in Г. Durieui, but leaves its
morphological nature an open question. He was apparently handicapped by the poor
microscopical technique of his day.

Farmer (14) gives a detailed description of the various elements comprising this tissue
in 7. lacustris, but leaves the question of their exact morphological nature undecided.

Wilson Smith (30) considers the adoption of the term ‘phloem’ inadvisable until
further evidence as to the physiological properties of the cells in question is available.
He remarks (7. c. р. 227) that “по clearly defined sieve-tubes, the essential elements of
the phloem, have been found either in the stem or in the leaf.” Also, “ the inner cells
of the prismatic zone are unknown to become secondarily thickened and transformed
into xylem tracheids.”

Scott апа Hill (28), whose investigations were confined to Г. Hystrix, a terrestrial
species, confirmed Russow’s original suggestion that this tissue was made up of both
xylem and phloem together with a certain amount of parenchymatous tissue.

The most recent description of the secondary tissue of Zsoéfes is given by Miss Stokey
(32), who was fortunate in having material of four different species at her disposal.
She maintains (2. с. р. 333) that “the so-called ‘ prismatic layer’ is secondary xylem.
The cambium does not form phloem.” Moreover, she adds that there is no primary
phloem in the stem. This conclusion is of special interest, since the four species
examined exhibited a very considerable range in habit. That undoubted phloem occurs
both in the leaf and in the root, is admitted; if this be so, it is difficult to under-
stand how the products of metabolism can be conveyed from one part of the plant

SECOND SERIES.—BOTANY, VOL. VIII. x 9D

244. MESSRS. CYRIL WEST AND Н. TAKEDA ON

to another. There is very little evidence in favour of the view that the parenchyma
of the secondary cortex takes over the function of phloem. Although it is usual for
the xylem of aquatic plants to undergo reduction to a greater or lesser degree, we know
of no other genus, either amongst Vascular Cryptogams or amongst Phanerogams, in
which the phloem of the stem is reduced to the point of extinction. On the other hand,
many aquatics possess a well-developed phloem *.

As a matter of fact, the tissue cut off internally by the cambium is largely composed
of sieve-tubes T which are very similar in form to the parenchymatous cells, from which,
however, they can easily be distinguished by the presence of definite sieve-areas and
frequent deposits of callus and by the absence of starch-grains.

The sieve-areas, which are of the typical cryptogamic type, occur on the radial,
tangential, and horizontal walls of the sieve-tubes (РІ. 34. fig. 13; Pl. 35. figs. 18, 19 ;
РІ. 38. fig. 64; РІ. 40. figs. 74, 76) ; hence it is strange that their presence should have
been overlooked or that they should have been mistaken for secondary xylem elements x
in process of lignification. Many of these sieve-tubes have comparatively thick cell-
walls divided by irregular reticulate thickenings into numerous sieve-areas; it is
these cells with irregularly thickened walls which appear to have been mistaken for
secondary xylem elements by Wilson Smith and by Miss Stokey.

As mentioned above, deposits of callus giving positive reactions to Corallin-soda
and other recognized ‘ callus’ stains are very frequently met with in these cells (Pl. 35.
figs. 18, 19). Occasionally this substance is deposited in such quantity as to
completely fill the lumen of the cell (РІ. 40. figs. 74, 76), especially in the sieve-tubes
first cut off by the cambium. A small nucleus is sometimes found in these elements,
but, when present, the nucleus is always much smaller than the nuclei of neighbouring
parenchymatous cells and is probably undergoing degeneration. Thus, it is evident that
the sieve-tubes of the secondary phloem of Isoëtes japonica are essentially similar in struc-
ture to those found in other Vascular Cryptogams, differing only in shape from the usual
type. It has already been shown by Seward and Ford (29) that the so-called * querge-
streckte Zellen' of Zenetti ( 37), which occur іп the stem of Osmunda апа Todea, are
actually sieve-tubes, which are considerably elongated in the horizontal (2. е. tangential)
direction. The description of these peculiar cells given by Seward and Ford (7. с. р. 245
and Pl. 29. fig. 40) would serve equally well for the sieve-tubes of Jsoétes. Moreover,
these authors (2. с. р. 242) remark that in the apical region of the stem of Todea these
cells contain large nuclei.

We have invariably found that on the phloem (4. е. lower or abaxial) side of the
leaf-trace the sieve-tubes of the secondary tissue are in direct organic connection with
those of the leaf-trace, as is clearly shown in РІ. 35. fig. 21. Moreover, а gradual
transition from the extremely short sieve-tubes of the secondary phloem to the

* Е. g., Potamogeton, Ceratophyllum, Elodea, etc.
+ Strictly speaking, these elements аге sieve-cells and differ from the sieve-tubes of the leaf- and root-traces,
which consist of rows of cells (= ‘members’ of De Bary, 12, p. 172).

ISOETES JAPONICA, А. BR. 945

more elongated sieve-tubes of the leaf-traces was observed. Тһе sieve-tubes of the
basal region of the leaf-traces are remarkably short considering the length of the
leaves themselves.

A possible explanation of the peculiar form of the sieve-tubes of the secondary phloem
of Isoétes may be deduced from the very slow stunted growth of this plant. It is
generally agreed that some sort of correlation exists between the actual rate of growth
of a plant-structure and the length of its individual conducting elements. Numerous
examples of this correlation might be given; for instance, the segments or members
of the sieve-tubes of certain tropical lianes attain a length of more than 2 mm. (e. g.
Calamus Rotang; Bignonia spp., etc.).

Isoétes lacustris was examined for comparison with 1. japonica. Short, box-like
sieve-tubes, similar in every respect to those described above, were found in the internal
secondary tissue (Pl. 35. fig. 20) associated with starch-containing parenchymatous cells
and with secondary xylem elements. Іп this species also, the sieve-tubes of the
secondary tissue are in organic connection with those of the leaf-traces (РІ. 35. fig. 16).

In 7. japonica the parenchymatous cells (2. е. phloem-parenchyma) cut off internally
by the cambium can easily be recognized by the presence of a relatively large
nucleus and numerous starch-grains. The cells labelled ph.p. in Pl. 88. fig. 64 are
of this nature, although at first sight they might easily be mistaken for the actual
cambial cells; the latter, however, as previously pointed out by Scott and Hill (28,
р. 424), do not contain starch-grains.

These parenchymatous cells are slightly less numerous diu the sieve-tubes; they are
always irregularly distributed through the secondary tissue, and, even in very old
specimens, are never arranged in definite zones, as described for several other species
of Isoëtes by Hegelmaier (15, p. 501), Farmer (14, p. 42), and Stokey (32, p. 323).
Since the secondary tissue completely envelopes the primary stele, it follows that the
cells of the secondary tissue come into contact with the tissues of the leaf-trace as
it passes out from the central mass of primary tissue. On the xylem (i. e. upper or
adaxial) side of the trace a number of parenchymatous cells occur, and these, as already

. described, connect up directly with those of the * parenchymatous mantle.’

3. Secondary Xylem.

No secondary xylem was found in I. japonica; apparently it does not occur in this
species, since plants of all ages were examined. Secondary xylem has been described
in several other species of Isoëtes, namely :—

I. lacustris (Hofmeister, 16, p. 361; Russow, 26, р. 139);
Т. Duriewi (Hegelmaier, 15, p. 504);
I. Hystriz (Scott and Hill, 28, p. 421);
1. echinospora and 1. Engelmanni (Wilson Smith, 30, p. 297);
Т, Nuttallii and Т. melanopoda (Stokey, 32, p. 926).
In Isoétes lacustris xylem elements occur spasmodically throughout the secondary
302

846 MESSRS. CYRIL WEST AND Н. TAKEDA ON

tissues of the stem-stele. These secondary tracheids are very similar to those of the
primary xylem and usually occur in groups (РІ. 38. fig. 62), although single elements
are occasionally found. The reticulate thickenings of the walls of these elements are
lignified; hence, in sections stained with the Safranin-Hematoxylin combination, they
stand out sharply from the surrounding secondary tissues, which consist of sieve-tubes
and parenchyma. Moreover, these secondary xylem elements have no contents.

4. Secondary Cortez.

It is generally agreed that the tissue cut off externally by the cambium consists,
in all the species of Jsoétes hitherto examined, of typical parenchymatous cells, which
form the relatively bulky secondary cortex. These cells, which are produced very
rapidly, are at first arranged in radial rows and form a compact homogeneous tissue.
But, sooner or later, as they move farther and farther away from the centre of the
stem, their arrangement becomes correspondingly less and less regular, whilst large
intercellular spaces appear between them. These cells contain numerous starch-grains
and proteid granules, but oil-drops were never observed.

The walls of the peripheral cells of the secondary cortex become suberized and form
a feebly developed periderm.

It is somewhat difficult, even in young specimens, to distinguish the cells of the |
primary cortex from the parenchymatous cells formed by cambial activity, since they
show a general agreement in size, shape, and contents.

APICAL MERISTEM AND DEVELOPMENT OF THE PRIMARY TISSUES OF THE RHIZOPHORE.

The results obtained by us from a careful study of the rhizophore region of the
caudex of 1. japonica and of Г. lacustris, cannot be reconciled with the conclusion
arrived at by Scott and Hill with reference to the development of the same region
of I. Hystrix. These botanists (28, p. 431) give the following explanation of the
mode of growth of the rhizophore: “the downward growth of the base of the stem
is entirely due to the activity of the cambium, and to the addition of new root-bases.”
And (l.c. р. 441), “Тһе wood, which is added at the bottom of the stele, and which
remains functional when the primary wood higher up is disorganised, is entirely built
up by the sympodium of root-bases.” `

It has already been shown that growth of the stem-stele of J. japonica is acropetal,
fresh xylem and phloem-tissue being added continuously by the slow and gradual
differentiation of large plerome cells derived from the apical meristem. The growth
of the rhizophore-stele takes place in much the same way ; that is to say, the growth
of the rhizophore-stele is also acropetal, fresh xylem and phloem-tissue being gradually
differentiated from the primary meristem *, which, however, differs from that of the

* This meristematic tissue may possibly be regarded as a part of the cambium ; but it is present in the embryonic
plant (с/. Hofmeister, 16, Taf. 48. fig. 3), persists throughout the life of the plant, and is quite distinct both in
origin and in appearance from the true cambium. It therefore seemed advisable to distinguish between the two
tissues ; the term ‘primary meristem’ is therefore employed with reference to the former.

ISOETES JAPONICA, A. BR. 347

Text-fig. 4,

Diagram showing arrangement of primary and secondary stelar tissues. From a transverse

Isoëtes japonica,
section of one of the lobes of the rhizophore-stele. х 20.
с. = cambium. тег. = primary meristem. „т. = * parenchymatous mantle.’
ct. = cortex. ph., ph.” = primary, secondary phloem. r.i, = root-trace.

Text-fig. 5.

Yt %
2 vt
Я A
io et
+ xX
“г ^5
gi
`
NS
X
і * `
А <
<
р
NS

Ditto: but from an oblique transverse section of a stelar lobe cut in the plane indicated
by the line a-b of text-üg. 1. х 20.

318 MESSRS. CYRIL WEST AND Н. TAKEDA ON

stem-stele in several important details. For instance, the apical meristem of the
rhizophore is distributed over a much larger area, since it extends along the whole
length of the curved lower edge of each of the three lobes of the stele as indicated in
text-figs. 1, 4, 5.

However, it must be remembered that this primary meristem is confined to a very
narrow strip of tissue and is completely absent from the lateral surfaces of the stelar
lobes. This is clearly shown in text-figs. 4 and 5, which indicate in a diagrammatic
manner the arrangement and distribution of the various structures met with in transverse
sections through one of the lobes of the rhizophore-stele. РІ. 40. fig. 77 is reproduced
from an untouched photo-micrograph of a transverse section through a similar lobe, cut
near its apex. In this figure, only those tissues which are found along the lower margin
of the lobe are shown. From the central region of this primary meristem the tracheids
and parenchyma of the xylem are slowly differentiated (Pl. 39. fig. 63; Pl. 40. figs. 75,
77, and text-figs. 4, 5; also cf. Hofmeister, 16, Taf. 52. fig. 6). This results in the
continual addition of xylem elements to the primary stele of the rhizophore. From
the more peripheral regions of this meristem the ‘parenchymatous mantle, primary
phloem, and cambium of the rhizophore-stele are produced (text-figs. 4, 5).

The primary. meristematic tissue extends to a point slightly beyond the apex of
the stelar lobe, so that the ‘parenchymatous mantle’ and primary phloem are found
along the upper surface as well as on both flanks of the lobe (text-figs. 1, 4, 5).
Consequently, the actual growth of the primary tissues of the rhizophore-stele is
confined to three radiating lines (а, 0, с, in text-fig. 10) separated from one another by
an angle of 120" and inclined at an angle of approximately 80* to the vertical axis of the
plant (text-figs. 1, 11). Hence, there сап be no lateral growth of the primary tissues
of the stelar lobes of the rhizophore, and consequently no lateral increase in dimensions
apart from that brought about by the extension of tissues already formed. However,
the primary tissues of the rhizophore-stele are subjected to a pulling strain by the
out-going root-traces (Pl. 40. fig. 80, and text-figs. 4, 5).

The primary meristem of the rhizophore is situated not at the actual periphery of the
caudex, as in the stem apex, but is separated from the exterior by several regularly
arranged layers of parenchymatous cells (Pl. 38. fig. 63; РІ. 40. figs. 75, 77), which
contain numerous large starch-grains. These cells (— primary cortex of the rhizophore)
form a structure in many respects analogous to a * root-cap.'

Jsoétes lacustris was examined for comparison with the above, from which it differs
only in the possession of two, instead of three, lobes to the rhizophore-stele.

STRUCTURE OF THE PRIMARY TISSUES OF THE RHIZOPHORE.
l. Primary Xylem.

The tracheids of the rhizophore-stele differ both in shape and size from those of the
stem-stele. "The former are not oval in shape but are considerably elongated (Pl. 38.
fig. 63; Pl. 40. figs. 75, 77); they are rather smaller than the tracheids of the stem-stele
and are spirally or reticulately thickened.

ISOETES JAPONICA, А. BR. 349

| 2. Primary Phloem.

The primary phloem of the rhizophore-stele consists of sieve-tubes which are essentially
similar to those of the primary phloem of the stem ; they are accompanied by unmodified
cells of the ‘parenchymatous mantle, which may possibly constitute the phloem-
pareuchyma.

3. Primary Cortex.

The primary cortex of the rhizophore region needs no further description.

STRUCTURE OF THE SECONDARY TISSUES OF THE RHIZOPHORE.

1. The Cambium.
In the rhizophore cambial activity starts very early, the cambium arising from the
peripheral cells of the ‘parenchymatous mantle’ just below the apical meristem
(text-figs. 4 % 5).

2. Secondary Phloem.
The tissue cut off internally by the cambium consists of parenchyma and sieve-tubes,
which are exactly similar to those of the secondary phloem of the stem.

3. Secondary Cortex.

The greater part of the caudex of Isoëtes consists of secondary cortex, derived partly
from the cambium of the stem and partly from the cambium of the rhizophore.

The secondary cortex of the rhizophore consists of large spherical parenchymatous
cells containing starch and proteid granules. These cells are cut off externally by the
cambium, which completely surrounds all three lobes of the rhizophore-stele apart from
the region of the apical meristem (РІ. 40. fig. 80, and text-figs. 4, 5). The lobing of
the caudex depends entirely upon the distribution of the cambium, since the three
main furrows always coincide with the three narrow zones of primary meristem, from
which, as stated above, relatively very few cortical cells are differentiated. This
accounts for the almost spherical outline of the caudex of the very young sporophyte,
in which the formation of secondary cortical tissue has not had time to proceed
very far.

Owing to the centripetal development of the secondary cortical tissues, the diameter
of the caudex rapidly increases; but nevertheless a very considerable increase in height
as compared with the increase in diameter is noticeable in old specimens (РІ. 33.
figs. 4, 7). This result is brought about by the constant decortication of the older
(i. e. most peripheral) region of the caudex. This process of decortication takes place
in the following manner. Certain irregular patches of cortical cells situated at or near
the periphery of the caudex are stretched by the mechanical pull exerted by the
rapidly increasing radial growth of the secondary cortex (text-fig. 6). The walls of the
cells bordering on these ‘islands’ or ‘ bays’ of stretched parenchyma gradually become

350 MESSRS. CYRIL WEST AND H. TAKEDA ON

suberized (text-fig. 7). These cells ultimately become torn apart, and in this way
irregular fissures are produced in the oldest parts of the cortex (РІ. 33. fig. 7; PL Ja
figs. 56-58).

It frequently happens that these fissures meet in the interior of the cortex, in which
case irregular fragments of cortical tissue, including portions of unmodified parenchyma,

Text-fig. 7.
Text-fig. 6.

Diagram showing later stage of
The dotted areas denote cells with
From a transverse section of a lobe

Isoëtes japonica. Diagram showing early stage of Isoétes japonica.
The stretched cortical cells are indi- decortication.
suberized walls.

of the eaudex. x 6

decortication.
eated by the areas included by the dotted lines.

From a transverse section of a lobe of the caudex.

x 6.
Text-fig. 8.

Isoëtes japonica. Diagram showing a late stage of decortication. The fragment А is completely
cut off from the cortex. From a transverse section of a lobe of the caudex. x 6.

become more or less isolated and finally break away (text-figs. 7, 8). Тһе suberized
cells bordering these: fissures form a thiu layer of periderm, which is continuous with
that of the primary cortex of the stem aud of the rhizophore. А somewhat similar

ISOETES JAPONICA, А. BR. 351

process takes place at the base of each of the three main furrows of the caudex. In
these regions the primary cortical cells are greatly stretched in a plane at right angles
to that of the furrow, whilst a few cells bordering on the stretched cells become feebly
meristematic, dividing by periclinal walls (РІ. 33. figs. 1, 2). Тһе walls of the stretched
cells gradually become suberized and are finally torn apart (Pl. 33. fig. 3). In this way
the main furrows increase in depth.

A series of transverse sections through the caudex of an old plant are represented in
РІ. 38. figs. 55-59. Тһе caudex immediately below the *leaf-rosette' has a roughly
circular outline, as shown in fig. 55. Immediately below this level the caudex
begins to assume its trilobed form, but the grooves appear relatively shallow (fig. 56).
Sections cut through the caudex at successively lower levels (figs. 57, 58, 59) show the
typical tri-lobed outline, the three lobes of the caudex being very prominent in the
rhizophore region of the caudex.

DEVELOPMENT, STRUCTURE, AND ARRANGEMENT OF THE ROOTS.

1. Development of the Roots.

We were unfortunately unable to investigate the development of the primary root,
which was fully developed in the youngest plant we examined. The young roots, which
are differentiated from the primary meristem of the rhizophore, penetrate the cortical
tissues of the rhizophore and ultimately emerge from its lower surface. In Isoëtes
Japonica the root never undergoes dichotomy whilst the apex is still embedded in the
cortex.

Our observations on the mode of growth at the apex of the root, so far as they go,
confirm the account given by Farmer (14, рр. 51, 52). Distinct histogenetic layers
are clearly shown, and the plerome appears to grow from a single initial cell. We also
agree with this author that the root in the genus Isoëtes is really an early type of.
adventitious root. |

2. Structure of the Adult Root.

The adult roots of Zsoétes japonica are essentially similar to those of other species of
the genus. They are long and fibrous, and dichotomize repeatedly at a short distance
from the caudex (РІ. 40. fig. 73). They have a dirty-brown colour, which is due to the
suberization of their peripheral cells. Іп a young root, only the outermost layer of cells
is suberized ; but in older roots several layers of cortical cells may become suberized.

The roots bear numerous long unicellular root-hairs, which are most abundant towards
the root-apex.

The general structure of the monarch vascular bundle of the root of Isoëtes is well
known. It is surrounded by a well-defined endodermis, and occupies an excentric
position in the root. Тһе protoxylem, аз a rule, is endarch; but a certain amount of
irregularity in the position of the first-formed elements of the xylem was observed. The
xylem consists of narrow tracheids with spiral or annular thickenings. True sieve-tubes

SECOND SERIES.— BOTANY, VOL. VIII. ЗЕ

852 MESSRS. CYRIL WEST AND Н. TAKEDA ON

occur in the phloem (Pl. 35. fig. 17); they are in direct organic connection with the
sieve-tubes of the rhizophore-stele.

3. Arrangement of the Roots.

The arrangement of the roots in Isoëtes has already been described by Von Mohl
(24, p. 183), Hofmeister (16, p. 358), Scott and Hill (28, pp. 428-430), Stokey
(32, p. 815), and Lang (20, p. 784).

However, a reinvestigation of this question, which was greatly facilitated by the large
dimensions of 7. japonica, has led us to put forward an entirely new interpretation.

Тһе corm-like caudex of Isoëtes is perennial, and may live for very many years. The
annual increment in the number of roots is very large indeed; Hofmeister (16 p. 359)
states that as many as six generations of roots may be produced during the vegetative
period of one year. Moreover, the duration of each root is far less strictly limited to any
definite period than that of the leaves, which persist for one season only. Thus it

Text-fig. 9. Text-fig. 10.

Diagram showing arrangement of the first six roots Diagrammatic transverse section of the caudex of
of a young plant of /soctes japonica. ‘The roots are Isoëtes japonica, showing the order in which the roots
numbered consecutively, | are differentiated. Compare text-fig. 11.

follows that a very considerable area is required for the attachment of these numerous
roots.

The roots are borne upon a specialized structure, the rhizophore, which in this genus
must be regarded as an organ sui generis.

In the very young sporophyte, the first three roots, when fully developed, lie
along the centre line of the three lobes of the caudex repectively (text-fig. 9). It
must, however, be remembered that the caudex of the very young plant is roughly
cylindrical ; the lobing of the caudex shown іп the diagram does not represent the
actual state of affairs. The fourth, fifth, and sixth roots lie alongside the first, second,
and third respectively (text-fig. 9). Even in the very young sporophyte, the primary
meristem of the rhizophore can easily be distinguished at the base of the group of |

ISOETES JAPONICA, А. BR. 353

tracheids produced at the junction of the first few root-traces (PI. 40. fig. 75). From
this primary meristem the young roots and the primary xylem, phloem, and cortex of
the rhizophore are differentiated.

Differentiation of the primary tissues proceeds along three radiating, but slightly
ascending, lines, which correspond to the three main furrows of the caudex as
explained above (р. 346 ей seg.) In this way, the stele of the rhizophore becomes

Text-fig. 11.

Isoëtes japonica. Diagram showing arrangement of roots on one flank of a lobe of the rhizophore-stele.
¢=cambium. (For explanation see text.) The consecutive numbers denote the succession of roots on one flank
of a stelar lobe, whereas the numbers 1,7, 13, 19, ete., denote the succession of roots on the plant. Compare

text-fig. 10.

distinetly three-lobed, each lobe appearing irregularly triangular in a radial longitudinal
section of the caudex cut in the plane of the corresponding main furrow.

The young roots are differentiated from the primary meristem of the rhizophore in
rows (hereafter referred to as *series") roughly parallel to the lower edge of the stelar
lobe; these series rise in acropetal succession (text-fig. 11). All the roots belonging to

| | 3E2

354 MESSRS. CYRIL WEST AND Н. TAKEDA ON
a series are differentiated simultaneously, that is to say, the roots belonging to a series

are always of the same age. These series of roots are regularly arranged on both flanks
of the stelar arms (text-figs. 4, 5).

Let us consider the roots which emerge from any one lobe of the caudex. Text-fig. 12
represents a drawing of the lower surface of one complete lobe of a fairly young plant

of I. japonica, from which the roots had been artificially removed. It will be noticed

Text-fig. 12.

Part of the caudex of a young plant of Isoëtes japonica (cf. Pl. 33. figs. 5 & 6) seen from below

. The roots have
been artificially removed.

that the roots, as they emerge from the cortex, are arranged in several pairs of more or
less curved series on the right and left halves of the lobe. This arrangement is to be |
explained by the fact that the series of roots which emerge from the right half of the -
lobe are derived from the flank of one arm of the rhizophore-stele, whereas those which | |

ISOETES JAPONICA, А. BR. 350

emerge from the left half of the lobe are derived from the flank of another arm of the
rhizophore-stele (text-fig. 10).

In text-fig. 18, which was prepared from the same specimen, the roots which cannot
be seen in a surface view of the actual specimen owing to the pronounced curvature of
the edge of the lobe, are also included (cf. Pl. 33. fig. 6; РІ. 34. fig. 9).

In the diagram represented in text-fig. 11, the series of roots bear numbers corre-
sponding to those of text-fig. 13; hence the oldest series, which always consists of a

Text-fig. 13.

+

` дег” PA

о
4000 е Doet g
06 m".

. Jsoétes japonica. Diagram showing the actual arrangement of the roots on one lobe of the caudex represented
in text-fig. 12. Тһе roots are numbered as in text-fig. 11.
single root, is number 1. This root occurs on the extreme edge of the lobe, and in
old specimens is invariably missing, since it is cut off together with the most peripheral
_ portion of the lobe as soon ав decortication takes place.

The theoretical arrangement of the roots, based upon the same specimen, is repre-
sented in text-fig. 14, in which all the roots of every series are shown. In this diagram
every root differentiated by the primary meristem is made to appear on the surface of
the caudex, whereas in text-fig. 13 only those roots which at that stage have actually
appeared on the surface are represented. It so happens that the nearer the long axis of

356 MESSRS. CYRIL WEST AND H. TAKEDA ON

the lobe a root is differentiated, the narrower the zone of cortical parenchyma it has to
traverse before reaching the surface; it follows, therefore, that in any given series, all
the roots of which are differentiated simultaneously, that root which has to travel the
shortest distance before reaching the surface, e.g., root 9 of series 9, represented in
text-fig. 18, will be the first of that series to emerge from the cortex. As a result of
the regular succession in the appearance of the roots of any series at the surface of the
lobe, together with the more rapid centripetal production of secondary cortical cells in
the centre of the lobe, the roots in this region are carried towards the outer margin of

Text-fig. 14.

Isoétes japonica. Diagram showing theoretical arrangement of the roots on one half of а lobe of the caudex.
The roots are numbered as in text-fig. 11.
the lobe (text-figs. 12, 13), the remaining roots of each series emerging at a point
nearer the angle of the furrow.

In this connection a brief description of the root-arrangement of those species, such
as I. lacustris, I. echinospora, etc., which generally possess a bi-lobed caudex, may not
be out of place. It is well known that the basal portion of the stele (i.e., rhizophore-
stele) of these bilobed forms has two arms which lie in the plane of the two main
furrows of the caudex. In a median longitudinal section of the caudex cut parallel
with the furrows, the rhizophore-stele appears crescent-shaped (text-figs. 2, 15; also
cf. Von Mohl, 24, p. 183, Taf. 3. figs. 4, 9).

ISOETES JAPONICA, А. BR. 357

Roots from the two arms of the rhizophore-stele are arranged on the right and left
halves of each lobe of the caudex respectively (text-figs. 15, 17).

In general, the arrangement of the roots on the flanks of the two stelar arms
is essentially the same as in the tri-lobed species (compare text-figs. 15, 16, 17 with
text-figs. 9, 10, 11; also of. Stokey, 32, Pl. 19. fig. 6%); further description is therefore
unnecessary.

Before leaving the subject of the root-arrangement іп 1500/68, a very simple method
of caleulating the approximate total number of roots produced by any given plant may
be mentioned. In text-figs. 11, 14, 15, 18, it will be noticed that the number of
roots produced іп any series always corresponds to the number of that series, e. g., in
series 1 there is one root, in series 2 there are two roots, in series 9 three roots, and so
forth.

Text-fig. 15.
Stele
TN o2/
°
^ i м. i] i
po Án]
239 е9 А else 1 vA ° РЕТ

- Isoëtes lacustris. Diagram showing arrangement of the roots on one half of the rhizophore-stele. The
numbers correspond with those of text-figs. 16 and 17.

Text-fig. 16. Text-fig. 17.
m до 3&4 » 11

^k 78 22 23 19 ‚у
Isoëtes lacustris- ` Diagram showing order in which the
roots are differentiated. From a transverse section

of the caudex.

Isoétes lacustris. Diagram showing the arrangement
of the first seven roots of a young plant.

* Stokey (1. с. p. 315), however, states that “the youngest sets of roots are those nearest the furrow ; and of those
_ in the furrow the youngest are at the ends.” This statement is incorrect, because all the roots belonging to any
. Series are of the same age (see text-fig. 18).

398 MESSRS. CYRIL WEST AND H. TAKEDA ON

"Therefore, if the number of series on a flank of either of the stelar lobes be
determined *, then the approximate total number of roots produced by that plant may
be calculated from the following formula, where

z = approximate total number of roots produced by the plant,
y — number of stelar lobes f,
z — number of series on the flank of a stelar lobe.
gua tu rd rm)
ur SË gz ж

Text-fig. 18.

"NOTER

© (9 (9 OG (Q2 (9 (9) (9) (2).
OOOO O@ 000;
OIGIGICIGIGIGION
0020900,
060606000,
9690090,
©©@®@..
99%,

Isoëtes lacustris. Diagram showing theoretical arrangement of the roots on one half of a lobe of the caudex. The
consecutive numbers denote the succession of roots on one flank of a stelar lobe, whereas the numbers 1, 5, 9,
13, etc., which correspond to those of text-figs. 15, 16, and 17, denote the succession of roots on the plant.

ARRANGEMENT OF THE LEAVES.

As stated above, many leaves are produced every year. A. Braun (3, p. 34), referring
to Isoétes setacea, another tri-lobed species, says that even in a young specimen the leaves
are arranged on the caudex іп а spiral of complicated phyllotaxis. In young plants of

* The number of series can easily be determined from thick tangential hand-sections of the caudex cut parallel
with the plane of a main furrow.

| + Two in bi-lobed species and three іп tri-lobed species, since the number of stelar arms is ERA equal to the -
number of lobes of the caudex.

ISOETES JAPONICA, А. BR. 359

Т. japonica the phyllotaxis 181%; but we were unable to determine the complicated
phyllotaxis of older plants with any degree of accuracy. On the other hand, ·
Hofmeister (16, p. 536) points out that the leaves of young plants of J. lacustris, a
bi-lobed species, are always at first arranged in a distichous manner, the distichous
arrangement gradually passing over into a spiral of increasing complexity, viz., 5, $, в,
$5, until finally an 38; phyllotaxis is attained.

DEVELOPMENT OF THE LEAVES.

The primordia of the leaves arise from certain superficial cells of the primary cortex
situated around the base of the apical protuberance (РІ. 39. figs. 65-70). Ав a result
of the luxuriant growth of these superficial cells of the primary cortical tissue, the
primordia can very soon be identified as smaller or larger, flattened, more or less ovate,
or crescent-shaped papilla. Subsequent growth of the leaf, as Farmer (14, p. 46) has
already described, is intercalary, the meristematic tissue being for the most part
restricted to a zone (of varying thickness) at or near the base of the young leaf.
Ultimately, however, this meristematic zone becomes localized in a region immediately
above the insertion of the ligule.

It must be remembered, however, that considerable elongation results from the
further growth, both in length and іп diameter, of the individual cells of the leaf
(ef. Wilson-Smith, 30, pp. 229—230).

That part of the procambial strand of the foliar bundle which traverses the primary
cortex, is differentiated at a very early stage in the development of the leaf. It
originates by the division of certain cells of the primary cortex, which retain their
meristematic character for a considerable period. A strand of small cells, easily dis-
tinguishable by their relatively large nuclei, is produced in an upward and outward
direction (РІ. 39. fig. 70). The upper extremity of the strand extends to the base of the
young leaf. Connection with the primary xylem and phloem tissues of the stem-stele
is established by the downward prolongation of the procambial strand, the tissues
of which are differentiated from the ‘ parenchymatous mantle.’

The differentiation of the procambial strand of the leaf immediately precedes the
differentiation of the cambium of the stem from the outermost cells of the ‘ parenchy-
matous mantle’; hence it is evident that the developing leaf-trace is not interrupted
either by the cambium itself or by the products of its activity. We agree with Farmer
(14, р. 40) that Hegelmaier's statement that the cells of the cambium contribute to the
formation of the leaf-trace is incorrect, for. as Farmer (/. с.) remarks, the particular
direction of the division of the cells which take part in the formation of the leaf-trace
is determined before the cells in their vicinity assume the freshly active meristematic
condition.

Further differentiation of the procambial strand of the leaf takes place centrifugally,
the cells of the basal part of the strand being the first to become converted into
tracheids, sieve-tubes, etc. This process gradually extends upwards and outwards
towards the base of the leaf until finally the lamina is reached.

SECOND SERIES.—BOTANY, VOL. VIII. 3 F

860 MESSRS. CYRIL WEST AND H. TAKEDA ОМ

ANATOMY AND HISTOLOGY OF THE LEAVES.
1. Epidermis.

The epidermis of the leaf of 780%ев japonica exhibits no special features, although the
outer wall is very highly differentiated for a water-plant. Тһе outer wall is compara-
tively thick (4-8 ш), and consists of an inner cellulose layer and an outer cuticularized
layer (РІ. 36. figs. 22, 23). Тһе outer surface of the epidermis is covered by a very
thin layer of cuticle. The primary membrane of the epidermal cells is very conspicuous,
particularly in the parting-wall between adjacent cells. On the inner side of the
primary membrane the cell-wall consists chiefly of cellulose, but on the outer side the
thickening is more or less cuticularized. This well-marked differentiation of the outer
_ eell-wall into two distinct layers appears to have a mechanical, but not a physiological,
function. Moreover, this type of epidermis is not peculiar to Isoëtes amongst water-
plants; it also occurs, for instance, in the stem of Ranunculus Jluitans. A comparison
of Pl. 36. fig. 23 with Pl. 36. fig. 22 will show that the outer wall of the epidermis is
somewhat thicker in the lamina than in the basal region of the leaf. Тһе cells of the
epidermis usually contain chlorophyll granules.

2. Stomata.

The apical portion of the leaf, unless it be completely submerged, is usually provided
with stomata. These stomata are of the gramineous type (Pl. 36. figs. 23, 24), and
possess no subsidiary cells. It appears, therefore, that а stoma-mother-cell divides but
once longitudinally and forms a stoma directly. Тһе guard-cells are provided with а
conspicuous outer ridge.

3. Mesophyll.

Large air-canals, which are always four in number, traverse the mesophyll. Тһе
development of these lacunze, which are formed by the lysigenous degeneration of
certain definite groups of cells in the very young leaf, has already been described in
detail by Farmer (14, p. 47) and by Wilson Smith (30, p. 230 e£ 8е4.). In these
lacunze, transverse septa or diaphragms occur at intervals of a few millimetres. Тһе
individual cells of these diaphragms have the stellate form commonly found in such
structures, especially amongst water-plants (РІ. 36. fig. 28).

No distinct palisade tissue is formed, most of the mesophyll cells containing abundant
chlorophyll (РІ. 36. figs. 23, 41). Тһе cells bordering the lacunze towards the centre
of the leaf are always empty; that portion of their wall which abuts on the lacuna is
slightly cuticularized (РІ. 36. fig. 41).

Towards the base of the leaf the mesophyll cells contain very little chlorophyll, but
occasionally the presence of reserve starch can be detected in these cells.

Whereas in some specimens of T. japonica the leaves possess hypodermal groups of
fibrous elements, in others no trace of these tissues сап be detected. Тһе presence or
absence of the hypodermal fibrous strands probably depends entirely upon the environ-
mental conditions to which the leaves are subjected at a very early stage in their

ISOETES JAPONICA, A. BR. 861

development. А somewhat similar conclusion was arrived at by Уарр (36) as a result
of his ecological studies on several morphological features exhibited by Spiræa Ulmaria.
These strands are differentiated from the mesophyll parenchyma, and normally form
four main and two accessory groups of fibre-like cells. Of the former, three are situated
on the adaxial and only one on the abaxial side of the leaf. The accessory strands occur
between the dorsal and corner strands. The main strands are very well developed, but,
even in the terrestrial species (e.g. 7. Hystrix), the individual fibres are only slightly
lignified. Possibly these strands have but little mechanical significance; lignification
of a cell-wall, whilst increasing the rigidity of the cell, decreases its strength. It was
found, even on the same plant, that some leaves possessed these strands whilst others
did not, and it sometimes happens that only three or four strands are developed.
Moreover, individual strands do not necessarily extend along the whole length of the
leaf, but usually show a tendency to die away or even completely disappear at certain
points, especially in the basal region of the leaf. This fact may be correlated with the
intercalary growth of the leaf, since the region of meristematic activity is situated near
the base of the leaf.

Towards the margin of the leaf, particularly near the base, the cells become tan-

gentially elongated, and their walls are thickened but not lignified.

4. Vascular Bundle.

A single collateral vascular bundle with normal orientation occupies an approximately
central position in the leaf. The vascular bundle in the adult leaf is surrounded by a
layer of large cells, which are fairly compact and regularly arranged. This row of
cells probably constitutes the true endodermis, for although the radial walls of these
cells are unthickened, yet they are highly refractive and present a more or less wavy
appearance.

Immediately within this layer of cells we find a varying amount of conjunctive
parenchyma, amongst which a few schizogenous canals are formed. These canals
(= eaneaux aériféres of Janczewski, 17) arise at a very early stage in the development
of the leaf, i. e., before the differentiation of metaxylem and metaphloem. The walls of
the cells bordering these canals are not cuticularized.

For our present knowledge of the structure and development of the phloem tissues of
the leaf of Isoëtes we owe much to the excellent work of Kruch (19), who made а
comparative study of the conducting tissues of the leaves of several distinct species.
Whereas the development of the xylem tissues varies in a striking manner in different
parts of the leaf, the phloem tissues, on the other hand, are well developed throughout
the whole length of the leaf.

In the young leaf, the phloem consists of sieve-tubes without companion-cells, and
parenchyma. The former are arranged in an are with its concave side facing the
xylem. Protophloem occurs in the central part of the phloem-arc, whilst the later-
formed sieve-tubes are usually collected together into two lateral groups separated from
one another by the few protophloem elements, which, at a later stage, appear as a very
narrow layer of crushed cells surrounded by ground-tissue. In the mature leaf,

3 F2

362 MESSRS, CYRIL WEST AND H. TAKEDA ON

the cells of the ground-tissue surrounding the protophloem undergo a considerable
change. They appear to undergo considerable elongation, and their walls, although
remaining cellulosic, become greatly thickened. These cells, according to Kruch, have
a mechanical function. The thick-walled cell figured by Janezewski (17, Pl. 4. fig. 9)
and designated a sieve-tube, evidently represents one of these modified conjunctive
cells, which do not in the least resemble the sieve-tubes.

The sieve-tubes are composed of elongated cells with more or less inclined terminal
walls (Pl. 37. figs. 43,44). Sieve-areas of the usual Cryptogamic type occur in abundance,
both on the lateral and terminal walls of these cells. They are often rendered very
conspicuous by a deposition of callus (РІ. 37. figs. 43, 44).

Sieve-tubes of a similar nature аге also found in the leaves of Isoëtes lacustris (РІ. 31.
fig. 45).

In both species the sieve-tubes of the foliar bundle connect up directly with similar
elements of the primary and of the secondary phloem of the stem-stele (РІ. 34. fig. 13;
РІ. 35. figs. 15, 21).

The xylem of the foliar bundle consists of tracheids and parenchyma. Towards the
apex of the leaf the tracheids are very feebly developed, only a few metaxylem elements
being present besides the canal formed by the disorganization of a single protoxylem
element. Неге the position of the protoxylem is exarch (Pl. 37. fig. 46). The transverse
or slightly oblique terminal walls of the protoxylem elements are reticulately thickened,
the thiekening being lignified (Pl. 37. figs. 49, 50). Тһе primary membrane does not
break down; hence the protoxylem canal consists of several superposed tracheids, with
almost horizontal terminal walls, and is not а ¢rachea in the strict sense, since only the
annular thickenings of the lateral wails become disorganized. "These protoxylem canals
are said to conduct water; they must, therefore, retain their original function even
after the disorganization of the secondary thickening.

In the adult leaf the protoxylem canal is always surrounded by a layer of regularly
arranged cells, the morphological nature of which has never been satisfactorily explained.
In J. japonica the radial walls of these regularly arranged cells, as well as that part of
the wall which abuts on the canal, are slightly thickened, highly refractive, do not
dissolve in strong sulphurie acid, and present most of the characters usually associated
with an endodermis. Scott and Hill (28), and Janezewski (17), have regarded this row
of cells as a true endodermis, but we prefer to apply the term ‘ pseudo-endodermis '
to these cells, since they differ from а typical endodermis both in origin and position.
In the lower region of the young leaf it is evident that the protoxylem elements are
surrounded by a distinct row of cells, some of which are typical metaxylem tracheids,
whilst others are parenchymatous (Pl. 37. fig. 47). Thus, in this part of the leaf some
of the * pseudo-endodermal’ cells are homologous with the water-conducting elements of
the foliar bundle. The fact that these cells occur in the centre of a vascular bundle and
generally surround only one of many tracheids, is sufficient evidence to show that they do
not constitute a true endodermis. Moreover, the slight thickening and cuticularization
of the radial walls of these cells are not features that occur only in endodermal cells: a
layer of cells exhibiting all the characters generally associated with an endodermis

X

ISOETES JAPONICA, А. ВВ. 868

completely surrounds the glossopodium of the ligule (Pl. 36. fig. 42). In Elodea, and in
some species of Potamogeton, the lysigenous protoxylem canal is surrounded by a row
of regularly arranged cells, the radial walls of which, however, do not appear to be
cuticularized.

As many as five protoxylem canals may be found in the supra-ligular region of the
leaf. They are invariably formed by the disorganization of protoxylem elements, and
are always surrounded by а row of ‘ pseudo-endodermal ' cells.

In the ligular region of the leaf a marked increase in the number of metaxylem
elements is found, but the position of the protoxylem remains exarch (РІ. 87. fig. 48).
In the sporangial region of the leaf, and lower down also, a few metaxylem elements
are differentiated on the phloem side of the protoxylem, while many are present on the
opposite side; hence, in this region of the leaf, the protoxylem is mesarch,

The metaxylem of the laminar portion consists of elongated spiral or subreticulate
tracheids with very oblique terminal walls (Pl. 37. fig. 51). In the sporangial region
the tracheids are somewhat wider, have less oblique terminal walls, and pursue a sinuous
course. In that part of the leaf-trace which traverses the cortex of the stem the xylem
is better developed and more compact than in the leaf itself. Sooner or later, however,
the conducting tissues become more or less stretched as a result of the formation of fresh
secondary tissues (2. е., phloem and cortex) by the cambium, and of the increase in size of
individual cells of the cortex of the stem. For this reason the tracheidal elements of an
old leaf-trace embedded in the cortex become greatly distorted. On the other hand, the
xylem parenchyma, which retains its meristematic condition for a considerable period,
is able to adapt itself to the changing proportions of the surrounding cortical tissue by
dividing both transversely and longitudinally (Pl. 37. fig. 54).

The phloem is subjected to a similar strain, but its complete destruction is somewhat
delayed by repeated divisions of the phloem-parenchyma. As a result of this cell-division
the leaf-trace traversing the cortex becomes much wider with age (РІ. 37. fig. 54).

5. Ligule.

The earlier stages in the development of the ligule of Г japonica agree with those
described and figured for Г. echinospora and I. Engelmanni by Wilson Smith (30, р. 232,
Pl. 14. figs. 12-19). "The ligule arises from a single epidermal cell of the very young leaf
(РІ. 36. fig. 35). This cell contains very dense protoplasm, and rapidly increases in size ;
it then divides by a periclinal wall into two daughter-cells (РІ. 36. fig. 36), each of which
immediately divides again by periclinal walls (РІ. 36. fig. 37).

The apical region of the young ligule divides most vigorously, and ultimately gives
rise to the laminar portion, whilst, by the subsequent growth of the basal region,
the glossopodium and ligular sheath are formed (РІ. 36. figs. 38-40; also df. РЕ 39.
figs. 66, 68).

The form of the mature ligule is ovate-lanceolate, cordate at the base, with a lacerate
margin (РІ. 36. figs. 25-27, 33). Тһе central region of the ligule is many cells іп
thickness, but towards the margin it is never more than one cell in thickness. The
laceration of the margin is brought about by a very peculiar method of sliding growth.

+

364 MESSRS. CYRIL WEST AND Н. TAKEDA ON

The individual cells of the marginal region exhibit unequal growth, as a result of which
they become very loosely attached to one another, and ultimately produce narrow,
irregular lobes (РІ. 36. fig. 34).

The cells of the thicker central region of the ligule form a very loose tissue, and
enclose large intercellular spaces (Pl. 36. figs. 30-32). Even the epidermal cells
are not compact; they are always covered with a very thin, but nevertheless quite
distinct cuticle, which does not follow the outline of the cells, but leaves small inter-
cellular spaces between them, thereby producing an effect which recalls that produced
in the leaves of a plum-tree when attacked by the ‘silver-leaf’ fungus (=Sterewm
purpureum).

'The cells of the lamina of the ligule have dense protoplasmic contents; these cells
secrete mucilage (РІ. 36. figs. 29-31). Тһе secretion of mucilage by the ligule has

Text-fig. 19.

Isoétes lacustris. Transverse section through base of leaf-rosette. The ligules are darkly shaded. x 20.

previously been described for Г. lacustris by Farmer (14, р. 44), and for Г. Hystrix by
Scott and Hill (28, p. 442).

The cells of the glossopodium are very large, and contain relatively little protoplasm ;
but in 7. japonica a band of large empty cells, such as Wilson Smith describes and
figures for I. echinospora (30, p. 233, Pl. 14, g in fig. 22), was never observed.

The glossopodium is surrounded by a sheath of more or less isodiametric cells, the
central portion of the radial walls of which appears to be lignified (РІ. 36. fig. 42). This
structure, which closely resembles an endodermis, has already been described and figured
by Bruchmann (7, p. 576, Taf. 24. figs. 22-25).

_ Certain cells of the leaf-base outside the glossopodium become converted into short
reticulate tracheids. These elements, which are invariably found in close proximity to

ISOETES JAPONICA, А, BR. 365

the ligule, frequently abutting directly upon the ligular sheath, occur both in the fertile
and in the sterile leaves, They are not continuous with the tracheids of the leaf-
bundle, so that their function is probably not water-conduction but water-storage.
However, since these cells are also present in J. lacustris, which is practically always
submerged, it is possible that their original function has been lost.

Bup-PROTECTION.

We have already mentioned that the apex of the stem together with the youngest
leaves and leaf-rudiments are situated at the base of a comparatively deep depression in
the cortex (РІ. 33. figs. 7,8). Thus they are surrounded on all sides by cortical tissue,
whilst above they are completely enclosed within a canopy formed by the overlapping

Text-fig. 20.

Asoétes japonica. Transverse section through leaf-rosette. Тһе ligules are darkly shaded. х 10.
т = mucilage.

vagine of the older leaves. In some species, e. g. I. lacustris, the wings of the leaves are
of sufficient width to completely enclose the next younger leaf (РІ. 37. figs. 52, 53, and
text-fig. 19). In this species the ligule is relatively very small. The wings of the leaves
of 1. japonica, on the other hand, are not sufficiently broad, except at the very base of
the leaf. Their protective function, however, is taken over by the ligule, which is
remarkably well developed in this species (text-fig. 20). Тһе ligule not only envelops,
and thus protects, the younger leaves, but provides an additional protective device in
the form of a copious secretion of mucilage. This secretion of mucilage is particularly
abundant in the youngest ligules, 2. e. those found in the very centre of the leaf-rosette,
and therefore it usually happens that the actual apex of the stem is covered with a film
_ of mucilage (m in text-fig. 20). -

866 MESSRS. CYRIL WEST AND Н. TAKEDA ОМ

SYSTEMATIC OBSERVATIONS.

A. Braun (4), describing the present species for the first time, expressed the opinion
(1. с. р. 460) that it belonged to his section “ palustres s. amphibize ” of the genus Zsoétes.
In a later and more detailed account of this species the same author (5, p. 8) retained
this classification *. Some years later, in 1882, Motelay and Vendryés (25, p. 318)
placed this species in their section **palustres," which, together with the section
* amphibize," they had established on the above-mentioned group of A. Braun. Baker
(r, p. 132), on the other hand, transferred it to the group “ Amphibize " f. Sadebeck
(27), in a more recent work on Isoëtes, placed Г. japonica in the section “ amphibiæ,”
following A. Braun in his classification of the genus.

We are perhaps justified in taking advantage of this opportunity of discussing the
classification of this genus.

A brief summary of the various classifications which have been put forward is given
below in order of priority :— |

1. ALEX. Braun (5, p. 8; 6, p. 598), followed by Milde (23) and by Sadebeck (27).
Section 1. Aquatic: (s. Submerse).

Leaves with wide lacunæ, the outer wall of which, including the E
consists of from 2-3 layers of cells; stomata, hypodermal fibres, phyllopodes,
and phyllades (** Niederblátter ") absent.

Section 2. Amphibiee (s. Palustres).

Leaves as in 1; but with stomata and groups of hypodermal fibres, usually 6
in number. Phyllopodes and phyllades absent.

Section 8. Terrestres. . |

Leaves with narrow lacunz, the outer wall of which consists of only one
layer of cells, 2%. e. the epidermis itself; stomata numerous; 4 large groups of
hypodermal fibres; phyllopodes and phyllades present.

2. MorELAY and VENDRYÉS (25, p. 317).
Group 1. Aquatice.
Section 1. Submersze.

Leaves with wide laeuns, the outer wall of which consists of from 2-3
layers of cells. Phyllopodes and phyllades absent.

Section 2. Palustres.

Leaves as in 1; but 6 groups of hypodermal fibres present without exception.

Phyllopodes absent. Phyllades absent, or, if present, very feebly developed.
Section 3. Amphibiz.

Leaves as before, with 6 groups of hypodermal fibres; stomata usually

present. Phyllopodes and phyllades absent.

* In an appendix to the paper, Braun refers to this section as “ Palustres s. Inundate " (l. с. p. 93).
T Baker established a fourth group, “ Subaquatice,” characterized by the presence of stomata and the siamo of
hypodermal strands.

ISOETES JAPONICA, A. BR. 867

Group 2. Terrestres.
Leaves with narrow lacun:e, the outer wall of which consists of only one
layer of cells. Stomata numerous.
Only 4 groups of hypodermal fibres present. Phyllopodes present, with or
without spines; the phyllades forming a scale.

3. BAKER (1, p. 124).
Group 1. Aquatice.

Plants permanently submerged. Leaves without stomata, groups of hypo-

dermal fibres, or phyllopodes.
Group 2. Subaquatieze.

Plants inhabiting shallow water. Leaves with few stomata, but without

groups of hypodermal fibres. Phyllopodes absent.
Group 3. Amphibiz.

Plants inhabiting shallow water, where they are subjected to desiecation
during certain periods of the year. Leaves with groups of hypodermal fibres ;
stomata abundant. Membranous leaf-bases occasionally persistent.

Group 4. Terrestres.

Plants growing in damp or water-logged soil. Leaves with groups of hypo-

dermal fibres; stomata abundant. Phyllopodes present.

4. EATON (13, pp. 58, 59) *.
$1. Aquatics, А. Br.
Plants submerged ; leaves cylindrical, fleshy, without groups of hypodermal
fibres or stomata. ;
$ 2. Amphibize, A. Br.
Plants of inundated shores or tidal flats; leaves with stomata, but without
groups of hypodermal fibres.
§ 3. Palustres, A. A. Eaton.
Plants inhabiting the extreme edge of ponds or streams; leaves with stomata
and groups of hypodermal fibres.
5 4. Terrestres, A. Br. б
Plants inhabiting dry situations; leaves setaceous with numerous stomata
and 4 groups of hypodermal fibres ; velum absent.
Using the above classifications it would be very difficult to determine the section in
which J. japonica should be included. For instance, this species may grow in situations
where the plant is always completely submerged, or it may grow on damp soil at least
for a short period. On the one hand, it produces leaves devoid of both stomata and
hypodermal fibres; on the other hand, the leaves are provided with both fibres and
stomata or with groups of fibres alone f. Thus one and the same species may belong to
* Aquatice ” of A. Braun and of Baker, іо “Submersze” of Motelay and Vendryés,
to * Amphibize ” of A. Braun and of Baker, and to “ Palustres " of Motelay and Vendryés
and of Eaton.

* Includes North-American species only.
+ It is possible that specimens may be found with stomata but without groups of fibres.

_ SECOND SERIES.—BOTANY, VOL. VIII. 3 G

368 MESSRS. CYRIL WEST AND H. TAKEDA ON

Let us take another example. Isoëtes echinospora is always regarded as belonging to
* Aquatics,” auct., or to “ Submersæ,” Mot. et Vendr. Nevertheless, the var. Braunii
would fall into the section “ Subaquaties," Bak., and has actually been placed in the
“ Amphibie," Eaton (13, р. 58)*. It is surely absurd that a species and its variety
should belong to different sections!

The following examples also demonstrate the weak points in these artificial classi-
fications :—

Т. saccharata, placed by Motelay and Vendryés in their section “ Palustres,” is trans-
ferred by Baker to his “ Subaquatieze."

Т. tegulensis is also regarded by Motelay and Vendryés as belonging to section .
* Palustres" ; Baker, however, considers it to be one of the “ Amphibize."

Т. Tuckermanii, according to Baker, belongs to the group '*Subaquatiez," but is
placed in “ Aquatic " by Eaton.

I. Butleri furnishes us with a more extreme case. This species is placed іп
“ Amphibize ” by Motelay and Vendryés and by Baker, whereas Eaton regards it as the
sole representative in North America of the most distinct section “Terrestres” How-
ever, an examination of the specimens collected by В.Е. Bush at Eagle Rock, Missouri,
оп May 22nd, 1898, has convinced us that this species cannot belong to “Terrestres,”
A. Br. In arriving at this conclusion we were influenced by the account given by
Geo. D. Butler (9, p. 2), the discoverer of the plant. This species is closely related to
I. melanopoda, a species belonging to “ Amphibize,” but prefers less humid situations.
Presumably Eaton considered the last-mentioned character of sufficient importance to
justify its inclusion in the section “Terrestres”! But according to its morphological
and anatomical characters, this species should be placed in the section “ Amphibize."
The definition which Eaton (13, р. 59) gives for section “ Terrestres” may fit Г. Butleri,
but does not agree with A. Braun’s definition of “ Terrestres.”

All the above-mentioned discrepancies point to the fact that the existing classifications |
of the genus Isoëtes are both unnatural and arbitrary, being based for the most part on
certain very unstable morphological characters, which have been shown to vary with the
environmental conditions of the plant.

Isoétes japonica furnishes us with a very clear case of this correlation of anatomical
characters with the habitat relations of the plant, since it may acquire the characters of
at least three different sections of the genus.

It is evident that species usually included in sections “ Aquatice " or “ Submersze "
may so alter their characters as a direct result of a change in their environment (e. g. by
adopting an amphibious mode of life) аз to justify their inclusion in “ Subaquatice.”
In the same way, there can be but little doubt that any species of ** Amphibiz " may live
entirely submerged or may even adopt a “terrestrial” habit. Possibly the species
belonging to “ Terrestres” are only extreme forms of “ Amphibie.” We are therefore
led to the conclusion that this artificial division of the genus Jsoétes into sections can
serve no useful purpose. |

We may, however, distinguish two sections, comparable to the primary groups of |

: y Eaton ` regards A. Braun as the author of this section, but his definition does not жеді ет? to that of :
E Mene Mes HB x | | x

ISOETES JAPONICA, А. BR. 369

Motelay and Vendryés, viz. “ Aquaticee” and “ Terrestres ” (25, p. 317). But we are of
the opinion that the use of these names is misleading and should therefore be avoided.
The chief characters of the two sections are as follows :—

Section 1. Eu-Isoirss, nob.

Leaves with 4 wide laeunse, the outer wall of which is composed of 2-3-5 layers of
cells (including the epidermis); sometimes provided with hypodermal fibrous strands,
normally 6 in number, 4 main (1 abaxial and 3 adaxial) and 2 accessory (lateral).
Sterile leaves very similar to the sporophylls. Leaf-bases seldom persist as membranous
scales, never forming phyllopodes.

Species growing either submerged, in shallow water (where they are subject to
temporary desiccation), or in damp soil.

Section 2. CEPHALOCERATON, Gennari (gen. ).

Leaves with 4 narrow lacunse, the outer wall of which consists of one layer of cells
(i. e. the epidermis alone); normally provided with only 4 hypodermal fibrous strands
(1 abaxial and 8 adaxial). Sterile leaves conspicuously smaller than the sporophylls,
eventually giving rise to phyllopodium-like scales. Leaf-bases always form persistent
phyllopodes.

Includes species inhabiting damp soil, subject to desiccation during a certain period
of the year.

DISCUSSION.

The systematic position of Zsoéfes still remains undecided. This genus has in turn
been compared with: 1. Certain Angiosperms (Hofmeister); 2. Eusporangiate Ferns
(Vines, Farmer, and Campbell, 10); 3. Lycopods (including Lepidodendree, Sigillarice,
and Pleuromeia) (Wilson Smith, Scott & Hill, Bower, Stokey, and Lang) *. |

The supposed resemblance between Isoëtes and the Angiosperms is both fanciful and
unscientific, and cannot, be accepted as any indication of affinity.

The Filicinean affinities of this genus are almost entirely based upon gametophytic
characters, the importance of which, however, is too frequently underestimated.

With regard to the morphology and anatomy of the sporophyte, 1800068 has many
characters in common with the Lycopodines (in the wider sense), but these are not of
sufficient importance to justify the view that Zsoéfes belongs to this group.

The isolated position of the genus 1800068 amongst recent Vascular m has
also been suggested, a separate Class, the Isoétales, co-ordinate with the Filicales,
Lycopodiales, Sphenophyllales, etc., being proposed for its reception (Campbell, 10 А;
followed by Lotsy, 21). 2

Тһе evidence derived from the present anatomical investigation 18 distinctly favourable
to this view. | U
Im the structure of the stem-apex and the development of the primary tissues of
. . the stem, Isoëtes differs in many respects from Lycopodium or Selaginella, while the
г 5 са In these papers the systematic position of /soctes is discussed in detail; recapitulation is ere unnecessary.

370 MESSRS. CYRIL WEST AND Н. TAKEDA ON

secondary tissues of this genus bear no resemblance whatever to those of any species of
Lepidodendron or Sigillaria hitherto examined.

The presence of a distinct root-bearing organ does not necessarily indicate any affinity
with the recent or fossil Lycopods which possess such a structure (е. 5. Selaginella,
Stigmaria, Pleuromeia, etc.).

On the other hand, the anatomy of these structures is quite unlike that of the rhizo-
phore of Isoëtes, to which even the four-lobed root-bearing basal region of the caudex of
Pleuromeia bears only a superficial resemblance. Whereas in the latter the lobes of the
stele are situated in the centre of the caudex lobes (31, Taf. 8. fig. 6), in the former they
alternate with the lobes of the caudex and correspond in position to the main furrows.

It is not unlikely that a distinct root-bearing organ has been developed along two or
three lines of descent, in which case the possession of a rhizophore is of little or no
yalue in determining the phylogenetic position of a plant.

Unfortunately, the geological history of Isoëtes is exceedingly meagre, a fact which
may be correlated with the almost complete absence of hard tissues (apart from the ripe
spores) from the sporophyte of this genus.

We are accordingly led to the conclusion that Isoëtes belongs to a separate Class of
Vascular Cryptogams, the Isoétales, co-ordinate with the phere Lycopodiales,
Psilotales, Equisetales, and Sphenophyllales.

SUMMARY.

The trilobed caudex of Isoëtes japonica consists of two distinct structures, viz. Stem
and Rhizophore, to which the leaves and roots are respectively attached ; but owing to
the stunted growth of the plant, all external morphological differentiation between the
two organs has been completely lost.

Stem.—The stem-apex has the form of a conical mass of tissue situated at the base of
the funnel-shaped depression in the cortex. In this protuberance no definite apical cell
can be distinguished.

Primary xylem, phloem, and cortex are differentiated from the primary meristem of
the stem.

The cauline primary vascular axis is a non-medullated monostele. Primary phloem,
in which true sieve-tubes occur, surrounds the central xylem-core.

An endophytic mycorrhiza is found in the peripheral cells of the primary cortex.

The cambium, which arises very early from the outermost layer of the plerome,
cuts off secondary cortex externally and secondary phloem internally.

Sieve-tubes with sieve-areas of the typical cryptogamic type occur both in the primary
and in the secondary phloem, and are continuous with those of the leaf-traces.

No secondary xylem is formed in Isoëtes japonica.

Rhizophore.—'The roots, the vascular bundles of which are collateral with iaa
endarch protoxylem, are arranged in acropetal series upon a distinct root-bearing organ,
the rhizophore, which in this genus must be regarded as an organ swi generis.

The primary growth of the rhizophore muet from а primary meristem situated
| sma three онар и lines which ее to the main fissures in the caudex.

ISOETES JAPONICA, A. BR. 371

The primary and secondary tissues of the rhizophore are essentially similar to the
corresponding tissues of the stem.

Leaf.—The protoxylem of the collateral vascular bundle is exarch in the lamina, but
becomes mesarch in the sporangial region of the leaf.

True sieve-tubes occur in the phloem of the leaf. Тһе ligule is very well developed in
Isoëtes japonica. It has a protective function; the young ligule envelops the younger
leaves and also secretes mucilage.

Systematic.—The species of Isoëtes can be grouped together under two sections,
Eu-Isoétes and Cephaloceraton.

Isoëtes occupies an isolated position amongst recent Vascular Cryptogams, and is
regarded as the sole living representative of the Class Isoétales.

In conclusion, we wish to express our thanks to Professor J. B. Farmer, F.R.S., for
his valuable advice and kindly criticism throughout the course of this investigation.

POSTSCRIPT.

Since the foregoing was written, the first two of a series of articles by Prof. Lang,
dealing with the organization of the ** stock " of 1500068 lacustris, have appeared %,

Although our interpretation of the general morphology of this plant agrees on the
whole with that of Prof. Lang, still, his account of the development of the stelar
tissues of the stem from the apical meristem cannot be reconciled with our observations.

We also differ from him in our interpretation of the mode of growth of the roots from
the rhizophoric region, since we consider that there is an actual penetration of the
cortex of the rhizophore by the developing roots.

Also, our conclusions as to the affinities of 1800068 differ from those of Prof. Lang,
who claims that this genus represents the last stage of reduction of the Lepidodendron-
type. x

BIBLIOGRAPHY.

т. Bager, J. G—Handbook of the Fern Allies. London, 1887.

2. Bower, Е. O.—Origin of a Land Flora. London, 1908.

3. Braun, A.—Weitere Bemerkungen über Isoëtes. Flora, xxx. 1847. : |

4. Ueber eine neue Art der Gattung Isoëtes. Monatsb. der kónigl. Akad. d. Wiss. zu Berlin,
1861. Í

5. —— Zwei deutsche Iso#tes-Arten nebst Winken zur Aufsuchung derselben (und ein Anhang über
einige auslándische Arten derselben Gattung). Verhandl. d. bot. Verein. d. Prov. Brandenb.
und die angrenzend. Lánd., Heft. iii., iv. Berlin, 1862; 299-333.

6. —— Bemerkungen über die Jsoéfes-Arten der Insel Sardinien. Monatsb. der kónigl. Akad.
d. Wiss. zu Berlin, 1863. x

7. Brucumann, H.—Ueber Anlage und Wachsthum der Wurzeln von Lycopodium und Isoétes.
Jenaische Zeitschrift f. Naturwiss. viii. 1874.

8. —— Untersuchungen über Selaginella spinulosa, A. Br. Gotha, 1897.

* Lang, W. H., “Studies in the Morphology of Isoëtes.” Part 1, Mem. Proc. Manchester Lit. Phil. Soc. lix.
,W.H.," f
` * No. 3, Feb. 1915; Part 2, l с. No. 8, May 1915. .

372 MESSRS. CYRIL WEST AND H. TAKEDA ON

9. Burer, G. D., in б. Ехвегмахм, “The Species of Isoëtes of the Indian Territory.” Bot. Gaz.
vol. 11. 1878.

IO. CAMPBELL, D. H.—Contributions to the Life-History of Isoëtes. Annals of Botany, vol. v. 1891.

IOA. Mosses and Ferns. 1905.

IH. Eusporangiate. Carnegie Institute of Washington Publication 140. 1911.

12. Рк Bary, A.—Comparative Anatomy of the Phanerogams and Ferns. Oxford, 1884.

13. Eaton, А. A., in Gray's ‘Manual,’ ed. 7. 1908. :

14. Farmer, J. В.--Оп Isoëtes lacustris. Annals of Botany, vol. у. 1890.

15. Нескгматек, F.—Zur Kenntniss einiger Lycopodium. Bot. Zeitung, xxxii. 1874.

16. Ноғмківтев, W.—Germination, Development, and Fructification of the Higher Cryptogamia.
Ray Soc. London, 1862.

17. Janczewski, E. nz.—fitudes comparées sur les Tubes cribreux. Ann. Sci. Nat. Bot. sér 6, xiv.
1882.

18. Janse, J. M.—Les Endophytes radicaux de quelques Plantes Javanaises. Ann. du Jard. Bot. de
Buitenzorg, xiv. 1897.

19. Kruca, O.—Istologia ed istogenia del fascio conduttore delle foglie di Isoëtes. Malpighia, iv.

20. Lane, W. H.—On the Stock of Isoëtes. Report of the British Association for the Advancement of
Science, Sheffield, 1910.

21. Lorsy, J. P.—Vortráge über botanische Stammesgeschichte, Bd. п. 1909.

22. Макіхо, T.—Observations оп the Flora of Japan. Tôkyô Bot. Mag. xviii. 1904.

23. Миле, J.—Filices Europe Atlantidis, Asi: Minoris et Sibirie. Leipzig, 1867.

24. Mont, H. von.—Ueber den Bau des Stammes von Jsoétes lacustris. Linnsza, xiv. 1845.

25. Moreray, L., et А. VexpRYís.—Monographie des Isoótez. Actes de la Soc. Linn. de Bordeaux,
xxxvi. 1882.

26. Russow, E.—Vergleichende Untersuchungen der Leitbiindel-Kryptogamen. Mém. del'Acad. Imp.
des Sciences de St. Pétersbourg, sér. 7, xix. 1872.

27. SADEBECK, R.—Isoétacez in Engler u. Prantl, Natürliche Pflanzenfamilien, i. Abt. iv. 1902. `

28. Scorr, D. H., and T. G. Нил.—ТЬе Structure of Isoëtes Hysiriv. Annals of Botany, vol. xiv.
1900.

29. Sewarp, А. C., and S. О. Forp.—The Anatomy of Todea, with Notes on the Geological History
and Affinities of the Osmundacee. Trans. Linn. Soc. Lond., Bot. vi., 1908, pp. 237-260,
tt. 27-30. : :

30. 8мітн, В. Wrrsos.—The Structure and Development of ће Sporophylls and Sporangia of Isoëtes.

Bot. Gaz. vol. xxix. 1900, pp. 225-258, 323-346, tt. 13-20.

31. Sorms-Lavsacu, Н. Скар zu.— Ueber das Genus Pleuromeia. Bot. Zeitung, lvii. 1899.

32. Stoxey, А. G.—The Anatomy of Isoëtes. Bot. Gaz. vol. xlvii. 1909.

33. Van Ттвенвм, Pu.—tTraité de Botanique, ed. 1. 1884.

Traité de Botanique, ed. 2. 1891.

35. Vines, S. H.—The Systematic Position of Isoëtes. Annals of Botany, vol. ii. 1889.

36. Yarr, В. H.—Spirea Ulmaria, L., and its Bearing on the Problem of Xeromorphy in Marsh
Plants. Annals of Botany, vol. xxvi. part 2. 1912.

37. ?кмхетті, P.—Das Leitungssystem im Stamm von Osmunda regalis, L. Bot. Zeitung, liii. 1895.

ISOETES JAPONICA, А. BR. : 878.

EXPLANATION ОҒ ABBREVIATIONS.

с. = cambium. ph., ph? = primary phloem, secondary phloem
cel.l. = cellulose layer. (ph.p. = phloem parenchyma).
ct. = cortex (ct. р. = cortical parenchyma). pler. = plerome.
cul. = cuticle. præ, = protoxylem.
cut.l. = cuticularized layer. ps.e. = pseudo-endodermis.
gr’, 977. = main furrow, secondary furrow. 7. = root.
1; = leaf. rh. = rhizophore.
14. = leaf-trace. r.t. = root-trace.
lig. = ligule. 8. = stem.
lig.sh. = ligular sheath. 84, = sieve-tube.
т. = mantle. tr. = tracheid.
mer. = meristem (primary). v. = vesicle.
п. = nucleus. 2. = xylem ( = primary xylem, 2° =
р. = parenchyma. secondary xylem).
per. = periderm. ж.р. = xylem parenchyma.

EXPLANATION OF PLATES 33-40.
PLATE 33.

Figs. 1-3. From transverse sections of the caudex, showing successive stages in the formation of a
main furrow. )
а = cells without starch, probably phellogen ;
& = cells with suberized walls (= periderm).
In all three figures the left-hand side is towards the centre of the stem. Figs. 1 & 2 x 220;
fig. 3 x 100.
Fig. 4. Lateral view of a fairly old caudex. Nat. size.
Fig. 5. Caudex of a young plant (about two years old) viewed from below. Some of the roots have
been removed in order to show their arrangement on the caudex. Nat. size.
Fig. 6. Lateral view of the same specimen. The leaves and roots have been completely removed.
Nat. size. |
Fig. 7. A median longitudinal section of a plant about four years eve x 15.
a = present season’s growth ;
b = groups of stretched cortical cells.
Fig. 8. Central portion of the same section magnified five diameters.

PLATE 34.
Fig. 9. Longitudinal section through the caudex represented in figs. 5 & 6. The plane of the section
is shown by the dotted line «..... 8 in fig.5. х 23.

. Fig. 10. Apical portion of the same section (diagrammatic). Since this section is not strictly median,
| it does not pass through the actual apex of the stem. Note at z the early differentiation
of the stelar tracheids. х 45.

374 | MESSRS. CYRIL WEST AND Н. TAKEDA ON

Fig. 11. Part of the same section, including the basal portion of the uppermost leaf-trace (on the left)
and the adjacent stelar tissues. X 250.

Compare the size and shape of tracheids belonging to the leaf-trace and stele respectively.

Fig. 12. Slightly oblique transverse section of a stem passing through the youngest part of the stele.
x 250. Note the leaf-traces surrounding the stelar region, many of which are not yet
differentiated.

Fig. 13. Part of a transverse section through the oldest part of the каш showing connection of leaf-
trace with stem-stele. x 200.

PLATE 35.

Fig. 14. Ditto, showing the horizontally stretched xylem parenchyma of the stem-stele. х 220.
Fig. 15. Part of a tangential section of the stem (cut in a plane immediately outside the stele) showing
several leaf-traces cut transversely. х 220.
The phloem of the leaf-traces is shown in organic connection with that of the иаа
Note the deposits of callus.
Fig. 16. A similar section of the stem of Tsoëtes lacustris. x 285.
Fig. 17. Part of a sieve-tube from the root, showing sieve-areas on the lateral walls. х 700.
Figs. 18 & 19. Sieve-tubes of the secondary phloem; from a transverse section of the stem. — x 700.
Note the callus deposited in some of the cells.
Fig. 20. Sieve-tubes of the secondary phloem of Jsoétes lacustris ; from a longitudinal section of the
stem. х 700.
Fig. 21. Part of a transverse section of the stem, passing through the junction of a leaf-trace with the
stele. x 220. Note the phloem of the leaf-trace in direct organic connection with that of
the stem.

PLATE 36.

Fig. 22. Transverse section through epidermis of the submerged portion of a leaf. x 450.

Fig. 23. Transverse section through epidermis of the aerial portion of a leaf. Two stomata are
shown. х 450.

Fig. 24. Stoma in surface view. x 220.

Figs. 25 & 26. Basal portion of two sporophylls from the intermediate region of the ‘ rosette.’

Nat. size.

. Basal portion of one of the innermost leaves of the ‘ rosette.’ Nat. size.

. Stellate cells from leaf diaphragm. x 220.

. A cell of the ligule. Хх 1000. Note the mucilaginous appearance of the cytoplasm.

Fig. 30. Transverse section of the central region of a ligule. х 100. Note the irregularly arranged
cells and intercellular spaces, both of which may contain mucilage.

Fig. 31. Part of the above more highly magnified. x 220.

Fig. 32. Cells of the central region of the ligule in surface view. х 220. Note the characteristic
intercellular spaces in the epidermis.

Fig. 33. Fully developed ligule. x 6.

_ Fig. 34. Marginal portion of the above more highly magnified. х 100.

_ Figs. 35-40, Successive stages in the development of the ligule. х 285.

E

9%
w
“4

E
йз 93
t2 t5
© o

ISOETES JAPONICA, А. BR. 7678

Fig. 41. Part of a transverse section of a leaf showing the large clear cells abutting on а lacuna.
X 285.
Fig. 42, Endodermis-like cells surrounding the glossopodium. х 220.

PLATE 37.

Fig. 43. Sieve-tubes from a leaf-trace embedded iu the cortex. х 350. Note deposition of callus
on the walls.
Fig. 44. Sieve-tubes from the vascular bundle in a leaf. x 450.
Fig. 45. Sieve-tube from the vascular bundle in a leaf of Isoëtes lacustris. х 450.
Figs. 46—48. Series of transverse sections of the vascular bundle of a young leaf. All x 285.
Fig. 46 is drawn from a section from the middle region of the leaf ;
Fig. 47 is drawn from a section at some distance from the glossopodium ;
Fig. 48 is drawn from a section just above the glossopodium.
Fig. 49. Longitudinal section of the central canal of а leaf. х 450. Note the remains of the lignified
bars of the end wall.
Fig. 50. А complete end wall of a central canal in surface view. х 1000. Note the lignified reticulum
and the middle lamella seen through the pits.
(From a very young leaf.) е
Fig. 51. Tracheids of а leaf-bundle, showing the very oblique terminal walls. х 225.
Figs. 52 & 53. Young leaves of 1306405 lacustris from the centre of a leaf-rosette, showing vaginee of the
older leaf, completely enveloping the younger leaf. x 1:5.
Fig. 54. Longitudinal section of an old leaf-trace embedded in the cortex. х 100.
Between the tracheids notice the parenchymatous cells, several of which have recently under-
gone division.

PLATE 38.

Figs. 55-59. Successive series of transverse sections of an old caudex :
Fig. 55 is cut just above the apical meristem, while
Fig. 59 is cut very near the base of the caudex. Slightly magnified.
Fig. 60. Cell from near the periphery of the primary cortex. X 500. Тһе starch-grains have dis-
appeared, leaving alveolar protoplasm to which numerous small proteid granules are attached.
Notice the collenchymatous cell-wall.
Fig. 61. Cell of the periderm with mycorrhiza. х 590. Note intercalary and terminal vesicles.
Fig. 62. Part of a transverse section of the stem of Isoëtes lacustris. х 200.
Fig. 63. Part of a longitudinal section of the rhizophore showing apical region of rhizophore stele.
x 160
Fig. 64. Part of a transverse section of the stem showing the peripheral tissues of the stem-stele.
x 280.
Puate 39.
Figs. 65-70. Median longitudinal sections through the growing-point of the stem. All x 320.
Fig. 65. An old plant of Г. lacustris.
Fig. 66. A very young plant of 1. lacustris.
Fig. 67. An old plant of 7. Hystriz.
Fig. 68. An old plant of 7. velata.
Fig. 69. An old plant of J. japonica.
5 Fig. 70. А very young plant of T. japonica.
2 SECOND SERIES.—BOTANY, VOL. VIII. 3H

. 78.
274.

2 709,
. 79.
ig. 80.

ON ISOKTES JAPONICA, А. BR.

Surface view of the growing-point of the stem of Isoëtes japonica. х 920.
Drawn from a thick hand-section cleared with * Eau de Javelle.’

Surface view of the growing-point of the stem of Isoëtes lacustris. х 320.
Drawn from a microtomed section.

PLATE 40.

Longitudinal section through the apex of a dichotomizing root. x 50.

Secondary phloem from a transverse section of an old stem. x 200. Note sieve-fields and
deposits of callus.

. Median longitudinal section through the stele ofa very young plant. Тһе plane of this

section is at right angles to that of a main furrow. x 165.

. Secondary phloem from a transverse section of an old stem. х 190. Note deposits of callus.
. Transverse section through the lower half of one of the lobes of the rhizophore-stele. "This

section is cut near the apex of the lobe. х 165.
Note how the cambium (seen above and below) is interrupted by the apical meristem (mer.)
of the rhizophore. Тһе arrow points towards the lower surface of the stelar lobe.
Transverse section of the stele of an old stem. x 23. -
Transverse section of the stele in the transitional region between stem and rhizophore. x 45.
Transverse section of the stele of the rhizophore. x 40.

West & Takeda TRANS:LINN.SOC.SER.2. Вот. Vou. VIL. Pl. 33.
/ I AZ 3 А | АЕ Я.
| З «я Бы a | One ne a
— JSU. ү

8

” d a T SN ў
D In
Huth lth. et imp.

ISOETES JAPONICA, A.Br.

TRANS.LINN.Soc, SER.Z.Bor. VoL VIII, P1.34.

West & T&keda.

NN

Huth lith.et imp.

\
PASE : |
— ® =
xe

' : и à dier
NN fij LA E

K | 4 Met 2 5
AUR

a é ' | | E e
ee SW 2) Ұй SN. | š i 8/ A 5 {з ж
De SEU Wah ку я E D бои * 3
: | | 7er ss HA ON j Y ү ру Š z A 3y б) 59 AS
WY ces EU се aise
B ы ресс DOO ce т
k. d X ` C o ФА ср
j Hn 0) ONKOL |96 ONG,
| x < ооч Се этесе үр»

©

ISOETES JAPONICA, A Вг.

NS one o) ОА DOE
Soe Bearers С) 9 Jalsa (де ox 010/9)
а ж Шо 3i С ЧЫ:

ENON CN 20227
3 DX у су сз = Сыз lo) ie uw»
OS S TENES De с ру

er

)
2
Z
Н.Т. et C.W.del.&d nat

West & Takeda TRANS.LiNN.Soc.SER 2.Bor Vor VIILPL135.

су, vete

Ns nd
w... j 4
Ye
yay а

K

21. 20. к.
НТ. del. ad nat. : uth lith.etimp.

IBORTES JAPONICA, А Br.

West & Takeda. TRANS.LINN. Soc. Зев.2. Вот. Vou.VII. P1. 36.

З"
°

атауда

OAT EET

Di

> 7906
А Оту Te ы
4. % (
ZO 9 ges ру Є
Clo “өзі. =
`

Huth hth etimp.

TRANS.LINN.SOG.SER.2.BoT. VoL. VII. P1.37

West & Takeda.

51.

54.

7

Huth lith. et imp.

|» .HT.del.ad nat.

ISOETES JAPONICA, A. Вг.

Trans.Linn. Soc. SER. 2. Вот. VoL VIII. P1 38.

West & Takeda

/

Š de go

S SN 27)
elle

N ЯТ
N S

N

Huth lith et imp.

Ji

1

Sa S

||

с GO Ao
oe ere
E39 y 2 o

©

*.
м
т.

“1 :
«o efe

^

r

SÁ

HINS

Ma
=

YL

AN
223

%

'
I
ч!

A. Вт.

Deno aD
sae ae

it cna

Хо 96)
p

776162
T
299;
5-2
(«4

o 19
SPINE
AN

к
ЛАМА»
(4

XO
э o kel |

JAPONICA,

Lele)
(955
ТИ

УЗ КЛ!
ТО»
ng

а)
г?
o2

©
Жы
e

Sih
Shirase)
Mo I
lof 102,170 Н
ЖОК Ыз"

[>

%„ ©З
6,
rel

($)

5225
72 | X
SL)
mess

EA
[o7

со;
ee)

ISOETES

ж

у

CW et НТ. del. ad nat.

West & Takeda TRANS. LINN SOC. SER.2.Bor. Vox VIIL PL 39.

D С REN
DD 9 QS —

5 | 56 © CONO
RS SORES т
О б», 6
OR o> D
ENS OE

ое YN (OKA ӘӘ 95509
САО О ONES y Ota

ре
FAN

52.2.

! (9/2) es COR put
о Я 033 9 Í
SOAS. ^

O
Bg

Е:

Q

SE
аз

хо
s

v
NS

(o

5
00
59
x
À

| р) oe |
о езе та у а -—
LA 0 A ers 65 OL Ker 22 te

2-5 И % 70. —

{о [oe

Huth lith.et imp. i `

ISOETES JAPONICA, A.Br.

о
ч
=
4
O
E
-
о
m
dr
m
fj
ds
O
©
oe
Z
Z
Ж
Ф
m
E

West & Takeda.

Bre

2

Ul

ax ш»

UM ы mea ТЕ

Huth сой

IBOETES JAPONICA. A. Вр

` s . MEMORANDA CONCERNING TRANSACTIONS.

First РА of the Transactions, containing both Botanical and Zoological contributions, has я nae gaa E.
à few entire sets are still on sale. Only certain single volumes, or parts to com may
T oe f the Index to Vols. 1-25 is 8s. to the public, aud бз. to Fellow»; + >

2nd Ser. BOTANY.) (VOL. VIII. PART 9.

THE

| "TRANSACTIONS
THE LINNE

IN DEX.

(Хотв,--бупопутав and Native Names are printed in italics. А star * is added to names which appear
to be used for the first time ; a + for fossils. | :

Abelia, Br., mentioned, 951;
vascular tissue, 252, 253, 260.
Acanthopanax, Decne. & Planch.,

mentioned, 248, 259, 260.
Acanthophora, Датх., 128.
orientalis, J. Ag., 107, 123;
mentioned, 109.
specifera, Borges., mentioned,
1238.

Thierii, Lamx., mentioned, 123.
Adoxa, Linn., mentioned, 240, 295.
Alangiacee, mentioned, 241, 242,

243, 268.
Alangies:, mentioned, 267, 268.
Alangium, Linn., 265-268; men-
tioned, 242, 289, 291, 296.
Alsinoidex, mentioned, 295.
Altingia, Nor., mentioned, 295,
Amentiferz, mentioned, 241, 243.
Amphibiæ (Jsoétes), mentioned, 366,
367, 368.
Amphiroa, Lamz., 141.
foliacea, Гатл., 107, 141;
mentioned, 110,
fragilissima, Lama., 107, 141;
mentioned, 110.
Amphisbetema *, Weber-van Bosse,
133; mentioned, 106; ftnote,
135

indica *, Weber-van Bosse, 108;
mentioned, 110, 133, 134,
135, 136.

Anacardium, Rottb., mentioned, 152.
Angiosperms, mentioned, 300,
313, 316, 318, 323, 324, 325,

327, 328, 329, 369.
Anomozamites Nilssoni *t (Phill),
Sew., n. comb., 187-188; men-
tioned,189, 190,195, 196,198,199.

Anomozamites Wiellandiella*+, Sew.,
mentioned, 188.
Antithamnion, Маедей, 136.
adnatum, J. Ag. 107, 136;
mentioned, 110,
Aporosa, Blume, mentioned, 278.
Aquatics, mentioned, 366, 367, 368,

Aquifoliacez, mentioned, 299.

Aralia, Linn., mentioned, 244, 248,
956, 959, 260, 273, 280, 289,
990, 993, 296, 299, 800.

АтаПасеге, mentioned, 240, 241, 242,
244, 259, 273, 278, 280, 281,
296, 299.

Ата Неге, mentioned, 242.

Araucarian stems, mentioned, 94.

Araucariee, mentioned, 95, 96, 97,
99, 100.

Araucarines, mentioned, 71.

Arber, À., An Anatomical Study of
the Palseozoie Cone-genus Lepi-
dostrobus, 205—232.

Arthrotaxites indicus f,
Morris, mentioned, 71.

Aucuba, Linn., 269-273 ; mentioned,
241, 242, 244, 261, 271, 277, 286,
287, 290, 291, 293, 296, 297, 298.

himalaica, Hook. f. $ Thoms.,
mentioned, 269, 270, 289.

japonica, Linn., mentioned, 269,
270, 289, 298.

Aucubes, mentioned, 272.

Avrainvillea amadelpha, 4.4 E. Gepp,
mentioned, 106.

Oldh. 4

Balanophora, Forst., mentioned, 145,
146, 147.

SECOND SERIES.-—BOTANY, VOL. VIII.

Balanophoraces, 143; mentioned,
146.

Bancroft, N., On some Indian Ju-

rassic Gymnosperms, 69-81.
Rhexowylon africanum, A new
Medullosean Stem, 87-101.
Bennetittalean frouds, mentioned,

— group, mentioned, 196.
Bennettitales T, 179-189;
tioned, 198, 199.
Bennettites Gibsonianus t, Carruth.,
mentioned, 76.
Morierei 1, Lign., 182, 183;
mentioned, 197.

Benstedtia T, Sew., mentioned, 72.
Benstedti T, Anowlton, 73.
Benthamia, Lindi., mentioned, 242,

261, 264

Betula, Linn., mentioned, 268, 977,
278.

men-

Betulaceæ, mentioned, 295.
Bignonia spp., mentioned, 345,
Bowenia, Hook., mentioned, 157, 160,
171, 172, 173, 174, 177, 181, 183,
197.
serrulata, Chamberlain, 172.
spectabilis, Hook., var.serrulata,

Brachyoxylon notabile T, Holl. $
Jeffr., mentioned, 72.
Brachyphyllum t, Brongn., 71.
macrocarpum T, Newberry, men-
tioned, 71.
mamillare T, Brongn., 71-72.
Brugmansia, Pers., mentioned, 147.
Bruniacee, mentioned, 296.
Bucklandia, В. Br., mentioned, 247.

9I

378

Bucklandia Millerianat, Carruth.,
mentioned, 69.

Вис апа!014е», mentioned, 246,
2

Buxus balearica, Гат., mentioned,

Сасбасеге, mentioned, 298.
Calamus Rotang, Linn., mentioned,

345.
Calliblepharis, Kitz., 117.
prolifera, J. ш: 196. НЕ:
mentioned, 106.
Campanulacez, mentioned, 297.
Camptotheca, Decne., mentioned,
2
Caprifoliacez, 251—260; morphology
of ovary, 251-255; ovule, 255-
256; vascular tissue, 256-258;
summary, 258-960; mentioned,
240, 243, 215, 264, 273, 231,287,

294, 295, 301.

tioned, 156. us
Carpinus Betulus, Linn., mentioned,
250

i : |
Castanea, Linn., mentioned, 243 |
|

Савпагіпасезв, mentioned, 299.
Сегатіасет, 135; mentioned, 110.
Ceramium, Lyngb., 186.
cinnabarinum, Hauck, 107;
mentioned, 110, 136.
вр., 107, 108, 137; mentioned, |
` 10. |
Cerasus juliana, hort., mentioned, |
298. ` |
СегаторһуПасеге, mentioned, 299.
Ceratophyllum, Linn., mentioned,
ftnote 844.
Ceratozamia, Brongn., 157, 166, 167;
mentioned, 174, 175, 177.
fusco-viridis, D. Moore, 166,
wo
Küsteriana,
166, 167
latifolia, Mi, 166, 167.
mexieana, Brongn., 166, 167;
mentioned, 176.

Regel, mentioned

Chetangiacee, 117; mentioned,
08.
Champia, Desvaux, 119. |
compressa, Harv., 107, 119;

mentioned, 109, 120.
parvula, J. Ag., 120.

| Cladhymenia, Hook. 4 Harv., 124.
Carboniferous Pteridosperms, men- |

| Сіздоху1еге, mentioned, 98, 99.

| Cochlospermacex, mentioned, 299.
| Codium, Ag., mentioned, 132.
‚ Colarthrum, Вогдез., 120.

, бот егосашоп t, F'liche, 72.

| Conifers, mentioned, 199.

INDEX.

Champia sp., 107, 119-120; men-

tioned, 109.
Chantransia, Schmitz, 111.
barbadensis, Vickers, men-
tioned, 11

efflorescens, Kjellm., mentioned,
r.

Liagore *, Weber-van Bosse,n.sp.,
107,111; mentioned, 108.
Nemalionis, Vickers, mentioned,

Chlorophyces, mentioned, 105.
Chondria, С. Agardh, 125.
pumila, Vick., 107, 125; men-
tioned, 109.
pusilla, Hook., mentioned, 106.
simpliciuscula *, Weber - van
Bosse, 125; mentioned, 107;
108, 110.
Chylocladia, G'rev., 120.
perpusilla *, Weber-van Bosse,
120; қанай; 108, 109

oblongifolia, Hook. $ Hurv., 108,
124; mentioned, 106, 109.
ladoxylon +, Unger, mentioned,
8. 2
Kidstoni t, Solms,
98, 99, 101

mentioned,

Albertisii, Bórges., 107,
121; mentioned, 109.
Ceelochondria, Falk., mentioned,

120-

Colpoxylon $, Brongn., mentioned,
‚ 98.

ceduense +. Brongn., mentioned,

Combretaceze, mentioned, 242, 267, |
268, 296, 297. |
Composite, 294; mentioned, 297. |
Сот етге, mentioned, 71, 73

Benstedtii t, Stopes, 73.

colymbezforme t, Sew., 72.

.Т, 72, 73; internal strut-
6

Corallinacew, 141; mentioned, 110.
Corallopsis, Grev., 117.
`СасаНа, J. Ag., 107, 117; men-
tioned,
minor, J. d. dou: f 127.

Согалфег t, mentioned, 95, 96, 97,
99, 100

Cordaitean stems, mentioned, 94,

Cornacee, 261-291; mentioned, 239,
240, 242, 243, 244, 245, 248.

Сотпеге, mentioned, 241, 242, 261,

Cornoidex, mentioned, 239.

Cornus, Linn., 261-265 ; mentioned,
241, 242, 248, 268, 272, 278,276
277, 281, 283, 290, 291, 295,
296.

alba, Linn.,

969, 965

Amomum,
61

mentioned, 261,

Mill., mentioned,
mas, Linn., mentioned, 261, 262.
sanguinea, Linn., mentioned.

261, 262, 263.

Corokia, A. Cunn., mentioned, 242,
244, 261, 282, 283, 286, 287,
290, 291, 296, 999.

Cotoneaster, Raoul, mentioned,
282, 290.

Corylace:, mentioned, 299.

Corylopsis, Sieb. $ Zucc., mentioned,
245, 948, 950, 951, 999.

- spicata, Sieb. $ Zucc., vascular
tissue, 249; mentioned, 247,
‚248, 249.

Corylus Avellana, Zinn.., mentioned,

30

Crategus Oxyacantha, Linn., men-
tioned, 298.

| Cruoriella, Crowan, нь 138.

Cruoriopsis, Dufour, 1
cruciata, Dufour, о 140-141;
mentioned, 110.
Cryptomeria japonica, D. Don, men-
tioned, 329.
Cryptonemia, J. Ag., 138.
seminervis, J. Ag., 108, 138 ;
mentioned, 110.
sp., 108, 188; mentioned, 110.
Ctenis T, Lindl. $ Hutt.. mentioned,
190, 196, 199. -
falcata +, Lindl. $ Hutt., 193-
194; mentioned, 190.
sp. T, mentioned, 178.
Ctenozamites +, Wath., mentioned,
178.

саса mentioned, 71, 72.
Curtisia, Ait., mentioned, 241,

` Сусадасеж, mentioned, 100, 161.

Cycadales, mentioned, 196, 199.

Сусайеап cuticles, mentioned, 157,

fronds, mentioned, 74, 156, 188,
197, 190

stems, mentioned, 188.

Cycadeoidea t, Buckl., mentioned,
16. 18i
gigantea T, Sew., mentioned,

ingens T, Ward, mentioned, 7

Сусадео14еге, mentioned, 78.

Cycadeoidean rhachis, mentioned,
75.

Cyeadophyta, mentioned, 79, 156,
195.

Cyeadopteris Brauniana t, Zeill., 178.
Cycads, mentioned, 75, 78, 80, 155-
200, 3138.
Cycas, Linn., 160-162; mentioned,
90, 157, 174, 198.
circinalis, Linn., 160, 161;
mentioned, 162, 173, 177.
media, R. Br., 160,
Micholitzii, Dyer, 160, 161.
revoluta, Zhunb., 160, 161;
mentioned, 162, 175, 177,

91.
undulata, Desf., 160, 161.

Dasya, Ág., mentioned, 135.
elegans, Ag., mentioned, 128.
indica, J. Ag., 105; mentioned,

106 :

Dasyes, mentioned, 128.
Dasyopsis, Zanard., 128 ; mentioned,
127 U

aperta *, Weber-van Bosse, 128—
129; mentioned, 110, 126,
127

cervicornis, Falkenbd., mentioned,
29

Geppii *, Weber-van Bosse, 130 ;
mentioned, 107, 108, 110,
126, 131.

palmatifida *, Weber-van Bosse,
130-131; mentioned, 107,

110.
spinella, Zanard., mentioned,
129.

Stanleyi*, Weber-van Bosse,
108, 128; mentioned, 110,
129, 130.

INDEX

Davidia, Baill., 239, 940; men-
tioned, 242, 243, 244, 248, 267,
273, 280, 988, 289, 290, 291,
292, 293, 296, 299.

involucrata, Baill.,
239, 298.

Delesseria hypoglossoides, Harv.,121.

Delesseriacez, 121; mentioned, 109.

Dermonema, G'rev., 112.

dichotomum, Harv., 107, 112;
mentioned, 108.
Dicoryphe, Thou., mentioned, 248.
Dicotyledons, sterility in, mentioned,

mentioned,

Dictyopteris falcata +, Morris, men-
tioned, 74.

Dictyozamites Т, Oldh., 74-76; men-
tioned, 178, 185, 189, 196.

falcatus Т, Oldh., nob., 74,
76; сона, 80.
Hawellit, Sew., 186; men-

indicus t, Feist., mentioned, 74.
Johnstrupi T, Nath., mentioned,
178, 186; 187.
Dictyurus, Bory, 135.
purpurascens, Bory, 107, 108,
135; mentioned, 110.
Diervilla, Linn., mentioned, 256,
260, 295, 301.
Lonicera, Mill., floral structure,
254; mentioned, 255.
Dilleniacez, mentioned, 299.
Dioon, Lindl., 157, 165; mentioned,
166, 174, 175, 177, 191, 198
angustifolium, Mig., mentioned,

175.
edule, Lindl., 165, 166; men-
tioned, 175, 177.
spinulosum, Dyer, 165, 166.
Dioonites Buchianus +, Ettingsh., 195.
Dunkerianus f, Schenk, 178,

Dipelta, Maxim., mentioned, 253.
Disanthus, Maxim., mentioned, 251.
cercidifolia, Maxim., mentioned,

247.
Distylium, Sieb. $ Zucc., 245-247;
mentioned, 250, 296.
Dracena Benstedtii Т, König, men-
tioned, 72.

Echinostrobus expansus 7, Feistm.,
71

879

Elodea, Michx.,
ftnote, 344.

Empetracez, mentioned, 299.
neephalartos, Lehm., 162-164;
mentioned, 157, 174, 177, 197,
198

mentioned, 363;

Altensteinii, Lehm., 174.

caffer, Miq., 162; mentioned,
164; ftnote, 166.

Ghellinckii,
mentioned, 17

horridus, Lehm., 162, 163.

lanuginosus, Lehm., 162.

Laurentianus, Wilden., 162.

Lehmanni, Eckl., 162, 163, 164.

villosus, Lemaire, 162, 163, 164.

Woodii, hort., 162, 163.

Endosiphonia, Zanard., mentioned,
1

162, 164;

Lem.,

6, 123, 1
clavigera, F«lkenb., a

124
Ephedra, Tn. , mentioned, 313,314,
316, 317, 318, 322, 393, 325, 326,
327.
campylopoda, DC., 313, 314.
distachya, Forsk., mentioned,
ftnote 328.
Equisetales, mentioned, 370.
Ethelia *, Weber-van Bosse, n. sub-
gen., 138; mentioned, 140. `
Eucheuma, J. Ag., 115.
Cottonii %, Weber-van Bosse, 108,
115; mention 109.
Schrammi, Crouwan, жекке.
15

spinosum, J. Ag., mentioned,
115.

Eu-Isoétes* ($), С. West & Takeda,
369, 371

ісе онаа, Weber-van Bosse,
п, subgen., 139,

Euphorbia virgata, Waldst. $ Kit.,
mentioned, 329,
Euryomma, Schmitz, mentioned,

115.
Eustigma, Gardn. 4 Champ., men-
tioned, 246, 250.

Fagacer, mentioned, 299.
Fauchea, Mont. $ Bory, 118.

Filieales, mentioned, 369, 370.
Flora of Frankonia, mentioned, 178.

312

380

Forstiera, А. 8. Horne, error for
Forestiera, Poir., mentioned, 278.
Fothergilla, Murr., mentioned, 248,
Gardeni, Murr., 248 ; vascular
tissue of, 249.
major, Lodd., vascular tissue of,
249; mentioned, 247, 248,
249, 250.
Frenelopsis Hoheneggeri T, Schenk,
mentioned, 178.
Fucus obtusus, Huds., 121.
Valentie, Turn., 118.

Galaxaura, Lamz., 112.
collabens, J. Ag., mentioned, 118.
Decne., 107, 108,

hawaiiana, Butt., 108, 113; таеп-
tioned, 108.

Liebmanni, Kjellm., 107, 112,
113; mentioned, 108.
obtusata, Гатх., 108, 113;

mentioned, 108.
veprecula, Ajellm., 107, 113;
mentioned, 108.

Garrya, Dougl., 274-278; men-
tioned, 241, 242, 261, 272, 273,
280, 290, 291, 296, 301.

elliptica, Doug/., mentioned, 241,
945, 274, 975, 277, 290;
vascular tissue, 276.

Fadyenii, Hook., mentioned, 274,

Fremontii, Zorr., mentioned,

79.

gracilis, Wang., mentioned, 275.

laurifolia, Benth., mentioned,
274, 275, 277.

longifolia, Hon mentioned, 276.

Macfaydiana, Carrière, sphalm.,
275

ovata, Benth., mentioned, 27

Thuretii x , Carriére, mentioned,
241, 275, 276, 289

Veatchii, Kellogg (err. Veitchii),
mentioned, 275, 276

Wrightii, Torr., mentioned, 274,
975

Garryacee, mentioned, 241,
243, 277.

Garryales, mentioned, 277.

Gates, R. R., A Contribution to a
Knowledge of the Mutating (Eno-
theras, 1-62,

Gelidiaceze, mentioned, 109, 114,

242,

INDEX.
Gelidiopsis, Schmitz, 116; mentioned,
137.

rigida, Weber-van Bosse, 107,
variabilis, Schmitz,
mentioned, 109, 119, 126.
Gelidium, Lamz., 114; mentioned,

crinale, Lamæ., 107, 114; men-
tioned, 109
pannosum, Grun.,
mentioned, 109.
rigidum, Okamura, 116.
Gelinaria, Sond., mentioned, 137.
Geranium longipes, DC., mentioned,
297.

107, 114;

Gigartinaceæ, 114; mentioned, 109.
Ginkgo, Linn., mentioned, 199,
Ginkgoales, mentioned, 179.
Gloioderma, J. A
expansa*, Weber-van Bosse, 108,
119; mentioned, 108.
Gloiophlea, J. Ag., 112.
articulata*, Weber-van Bosse,
108, 112; mentioned, 108.
scinaioides, J. Ag., mentioned
2.

Gloiothamnion Schmitzianum, Rhein-
bold, 137.
Gnetales, mentioned, 199, 316.
Gnetoidew, mentioned, 328,
Gnetum, Linn., mentioned, 175, 325,
326, 327, 329
africanum, Welw., 311; macro-
spore and endosperm, 318--
322; described, 323; men-
tioned, 312, 317, 327.
Brunonianum, Griff., 312.
Bucholzianum, Zngl., mentioned,
312, 314
campylopoda, lapsu= Ephedra
campylopoda, 313.
edule, Blume, mentioned, 319,
Gnemon, Linn., 311-318; male
flower microspore and pollen-
grain, 316-318; mentioned,
119, 319, 320, 321, 322, 327.
Leyboldi, Tul., mentioned, 312,
313.

moluccanum, Lotsy, mentioned,
ovalifolium, Poir, mentioned,
319

Rumphianum, Весс., mentioned,
319

scandens, Roxb., mentioned, 316,

| Gnetum Ula, Brongn., mentioned, 319.
| Gracilaria, Grev., 116; mentioned,
115.
compressa, Grev., 107, 116, 117;
mentioned, 109.

5. Ag.,
108, er mentioned, 110.
ае orst. f. 284—287, 299;
mentioned, 242, 261, 290, 291,
293 ч
littoralis, Raoul, vascular tissue,
285; mentioned, 245, 284,
286, 290.
lucida, Forst. f., vascular tissue,
285; mentioned, 245,
285, 286, 287, 291

Gunnera, Linn., mentioned, ftnote

|
|
; |
| Grateloupiacess, ы: ; mentioned, 110,

?

Gymnosperms, mentioned, 156, 157,
175, 179, 197, 199, 200, 313, 325,

327, 328, 330.
|
| Haloplegma, Mont., 136.

Preissii, Sond., 108, 136 ; men-

s | tioned, 1 10.

Halymenia, J. Ag., 137.
Durvillei, Bory, mentioned,
137
formosa, Harv., mentioned, 137.
polyclada var. aldabradensis *,
4 E. Gepp, 108, 137;
mentioned, 110. -
Hamamelidacee, 240-951;
tioned, 259, 260, 264, 265,
977, 278, 293, 295, 296, watts 299.
Hamamelis, | Gronov mentioned,
240, 245, 247, 248, 950, "os 264,
273, 277, 296, 299.
arborea, Ottol., vascular tissue,
249; mentioned, 247, 248,

ciiam R. Br., mentioned,
2

virginica,
247, 248

Zuccariniana, Ottol., vascular
tissue, 249; mentioned, 247,
250.

Linn., mentioned,

Harvey-Gibson, R. J., Observations
on the Morphology and Anatomy
of the Genus Mystropetalon, Harv.,
143-153.

Hedera, Linn., mentioned, 259, 260,

Helix, Linn., mentioned, 292.

Helminthocladiacee, 111; men-
tioned, 108.

Helwingia, Willd., 278-283; men-
tioned, 241, 242, 261, 273, 290,
291, 293, 295, 296, 298.

ruscifolia, Willd., mentioned,
9,

Hoelwingiacere, mentioned, 241, 242,
267, 2

ыш, Хаедей, 126; men-

tioned, 136.
prorepens, Schmitz, 107, 126;
mentioned, 11
Heterosiphonia, Mont., 132; men-
tioned, 1
Rendlei *, Weber-van Bosse, 108,
182, 137: mentioned, 110,

DL

sp., 107, 108, 132; mentioned,

10.
subsecundata, Falk., mentioned,
3

Horne, A. S., A Contribution to the

Study of the Evolution of the |

INDEX.

‚ Isoëtes Butleri, Engelm., mentioned,
368.

Durieui, Bory, mentioned, 337,
343, 345.

echinospora, Durieu, mentioned,
345, 356, 363, 364, 368

var. Braunii, mentioned,
368.

Engelmanni, А. Br., mentioned,
345, 868

Hystrix, Bory, mentioned, 338,
343, 345, 346, 361, 364.

| japonica, А. Br., 333-372;

stem, 340-346 ; rhizophore,

roots, 851-858.

858-365; bud-pro-
tection, 365. ;

lacustris, Linn., mentioned, 336,
387, 338, 340, 341, 343, 345,
346, 348, 356, 357, 359, 362,
364, 365, 371.

melanopoda, J. бау, mentioned,

345,
2. d Br., mentioned,

Flower, with special reference to |

the Hamamelidacee, Caprifoliacez, |

and Согпасе», 239—306
Hydrangea, Gronov., mentioned, 295.
Hyoscyamus virginianus, Alpino, 16,

35d
Hypnea, Lama., 117.

divaricata, Grev.,
mentioned, 109.

pannosa, J. Ag., 107, 108, 118,
125; mentioned, 109, 125.

spathulata, J. Ag., mentioned,
12

108, 117;

sp., 107, 118; mentioned,
109.
Valentie, Mont., 107,
mentioned, 109, 117.
Hypoglossum, Kiitz., 121.
вр., 107, 121; mentioned, 109.

118;

Пех Pernyii, Franch. (err. Perneyi),
mentioned, 296,

Inula, Linn., mentioned, 244,. 290,
299.

Iridea, Bory, 11
вр., 106, = 114; mentioned,
1

Isoëtes, Linn., mentioned, 383, 334,
336, 337, 340, 343, 344, 345, 346,
351, 352, 357, 360, 361, 366, 368,
370, 371.

a Engelm., mentioned,
68.

setacea, Bosc, mentioned, 358,
tegulensis, Genn., mentioned,

Tuckermanii,
tioned, 368.
velata, 4. Br., mentioned, 337,

,

A, Br., men-

Jania, Lamæ., 141.
tenella, Kütz.,
tioned, 110.
Juglans, Linn., mentioned, 277,

107, 141; men-

regia, Linn., mentioned, 297,
298, 300.

Kaliphora, Hook. f., mentioned,
242.

Lathyrus odoratus, Linn., mentioned,
292.

. Laurencia, Lam»x., 121.

ceylaniea, J. Ag., 107, 108, 123;
mentioned, 109.

indica, Hauck, mentioned,

381

| Laureneia nidifica, J. Ag., 108, 121;
mentioned, р

obtusa, Гатх., 107, 108, 121-
122; mentioned, 109.

var. racemosa, Kütz,, men-
tioned, 122.

paunosa, Zanard., mentioned,

papillosa, Grev., var. austra-
lica, Kiitz., 107, 108, 120;
mentioned, 109,

Weber-van Bosse,

spinulifera, Kitz., 107,
mentioned, 109.
Lepidodendreæ, mentioned, 369.
Lepidodendron t, лс 207; шеп-
tioned, 370, 3
Harcourtü T, p 206,
vasculare Т, Binney, 206.
Veltheimianus Т, Sternb., 206,
207.
em oc жы T, Sternb., mentioned,
e AM T. Brongn., 205-
Bertrandi t, Zalessky, 207;
mentioned, 215.
Binneyanus *Т, А. Arber, 212-
217; form of cone and
Sporophylls, structure of
cone axis, 212, 213; sporo-
phyll traces in axis, 213;
structure of sporophyll, 213-
214; ligule, 214; sporan-
gium, 214-215; spores, 215;
mentioned, 211.
brevifolium t, Lesg., mentioned,
218.
Brownii T, Schimp., 206; men-
tioned, 207, 211, 224, 231,
232

foliaceus T, Maslen, 287, 208-
211; mentioned, 216,

gracilis *Т, А. Arber, 217-220;
cone апа sporophylls, 217—

; Structure of cone axis,

sporophyll traces in
axis,
219;
gium, spores,
mentioned, 211, 221,

lamiaatus f, A. Arber, men-
tioned, 210,

majus T, Brongn., mentioned,
ftnote 208,

882

Lepidostrobus oldhamius T, Will.,
208-211; sterile plate in sporan-
gium, 208-211; periderm in the
sporophylls, 211; mentioned, 206,
215, 223.
= ——— f. minor *{, A. Arber, 220-

223; form of cone and
sporophylls, 221; sporophyll
traces in axis, 221; structure
of вроторһуПв, 221-222;
врогев, 222.

—— f. pilosus*t, A. Arber,

223—232; form of cone and
sporophylls, 224; structure
of cone axis, 224-295;
sporophyll traces in axis,

225; structure of sporo-
phyll, 226-227; ligule, 227;
sporangium, 228;
228 ; mentioned, 211.

vasiabilis T, Lindl. 4 Hutt.,
208.

spores,

Veltheimianus f, Scott, 206;
mentioned, 207, 209, 211.
Wünschianus +, Binney, 206,

207.

Leyeesteria, Wall., mentioned, 243,
251, 253, 255, 256, 258, 259, 260,
273, 296, 297, 299.

formosa, Wall., vascular tissue,
272; mentioned, 256.
Liagora, Lama., 111.
elongata, Zanard., mentioned,

hawaiiana, Butt., 111, 112;
mentioned, 108, 111

Lilium, Linn., mentioned, 244, 278,
290, 299.
Linnea, Gronov., ad, 251,
253, 256, 260.
borealis, Linn., vascular tissue,
2

Liquidambar, Linn., mentioned, 247,
251, 293, 295
styraciflua, Linn.,
247, 248.
Tithothamnia, Fosl., mentioned, 105.
Loasaceæ, mentioned, 299.
Lomentaria, Lyngbye, mentioned,
20.

mentioned,

Albertisii, Pice., 120. ,
intertexta, Kütz , 120.
Lonicera, Linn., жетіні 241, 248,
251, 255, 256, 260, 293, 297.

INDEX.
Lonicera Caprifolium, Zinn., men-
tioned, 251
Perielymenum, Linn., vascular
tissue, 252.
Lonicerez, mentioned, 260.
Loropetalum, R. Br., 250; men-

tioned, 240, 246, 296.
chinense, Oliver, mentioned, 246.
Lycopodiales, mentioned, 369, 370.
Lycopodinex, mentioned, 269.
Lycopodium, Linn., mentioned, 230,
340, 369
Lycopods, mentioned, 337, 369, 370.
Lysimachia americana, Columna, 16,
IT
angustifolia, spicata lutea Lusi-
tanica, Barrelier, 17.

corniculta minor lutea Cana-
densis, Morison, 17. |
latifolia, spicata, lutea Lusi-

tanica, Barrelier,
lutea corniculata, Bauhin, 16,
17

latifolia Lusitanica,
Barrelier, 17.

non papposa Vir-
giniana major, Morison, 17.
minor,

Morison, 17.
Лоте globoso, Park., Ger.,
Morison, 17.

siliquosa Virginiana,

Parkinson, 17.

Virginiana, Ray, 17.

Virgimiana altera, foliis lati-
oribus. и. ege majori-
bus, Ray, 17, 1

— cor niculata,

Morison, 17.

latifolia lutea corniculata,

Morison, 17.

Macrozamia, Mig., 157, 167, 172,
167, 168;

cylindrica, С. Moore, 167, 168,
169, 170, 177.
Denisonnii, F. Muell., 167, 168,

Fraseri, Miq., 167, 168.

Hopei, 7. Hill, 167, 168, 169.

Macleayi, Miq., 167, 168.

spiralis, Mig., 167, 168, 170;
mentioned,

|

|

Marattiacee, mentioned, ftnote,
Marlea, /tox)., 265—268 ; mentioned,
241, 242, 244, 261, 277, 290,
297.
barbata, А. Br., mentioned,
begonifolia, Hoxb., mentioned,
265, 267, 291
platanifolia, Sieb. $ Zucc., men-
tioned, 265, 266, 267, 272.
Mastixia, Blume, mentioned, 241.
Medullosa t, Cotta, mentioned, 89,
96, 97, 98, 100, 101
anglica t, Scott, mentioned, 91,

Leuckarti ў, Gépp. 4 Stenz.,
mentioned, 96, 97.

Ludwigiit, Gopp. $ Stenz.,
mentioned, 98.

porosa t, Cotte, mentioned, 90,
96, 97

Solmsiit, Schenk, mentioned,
91, 96, 97, 98

var. incrassata *+, Bancr.,

mentioned, 96.

var. lignosa*t, Baner.,

mentioned, 96, 97.

var. typica *+, Tano. ,men-

tioned, 96,
stellata f, Cotta, mentioned, 90,
96, 97.
Medullosex, mentioned, 90, 94, 96,
98, 99, 100, 101.
Melanophylla, Baker, 284; men-
tioned, 242, 261, 287.
. alnifolia, Baker, mentioned,

aucubzifolia, Baker, mentioned,

284.
Humbloti, Baill. (non descr.),
mentioned, 284.
Melanothamnus, Born. $ Falk., men-
tioned, 134, 135

Melobesia, Lamæ., 141.
callithamniodes, Falk., 107,
141; mentioned, 110.
oeg Forst., mentioned, 280,
и Cycadean fronds, 178;
Cycadophyta, 80; flora, шеп-

tioned, 80; fronds, mentioned,

156.

Місгосусаз, A. DC., 157.

Monocotyledons, mentioned, 294.
usa, Linn., mentioned, 81.

Myeloxylon T, Brongn., mentioned,
93.

Myriea, Zinn., mentioned, 277.
panay isantes, hort., mentioned,

Myricales, mentioned, 277.

Myrtifloree, mentioned, 268.
Mystropetalon, Harv., 143-153;
rhizome and its relation to host
plant, 145-147; inflorescence,
147; male flower, 148-149;
female flower, 149-151; fruit,
151-152; identification, 152-

153.

Polemanni, Harv., 143, 144,
145; 5-5 149, 150,

Thomii, Harv., 143, 144, 145;
mentioned, 148, 149, 150,
151, 152; described, 153.
Мухорһусеге, mentioned, 105.

Nilssoniat, Brongn., 190; men-
tioned, 178, 179.
brevis Т, Brongn., mentioned,

compta 7, Bronn, 190-191;
mentioned, 192, 193, 196.
mediana t, Foa-Strangw., 191—

192; mentioned, 190, 191,
192, 196.
orientalis t, Heer, 192-193;

pterophylloides +, Nath., men-
tioned, 190, 196
sp., 192.
tenuinervis T, Nath., 192, 193.
Nilssoniales, 190—196.
Nilssoniopteris t, Nath., 192; men-
tioned, 193.
m Grev., 121.
., 107, 108, 121; mentioned,

109.

Ny Gronov., 987-991; men-
боры; 241, 242, 243, 944, 261,
980, 290, 291, 296, 299.

capitata, Walk., mentioned,
264

64.
sessiliflora, Hook. f. & Thoms.
mentioned, 281.
sylvatica, Marsh., 287, 288,
ОСА

INDEX.

Хуввасете, mentioned, 241, 242, 243,
268, 289

Ochnaces, mentioned, 299.
(Enothera, Zinn., 1-62; mentioned,
292, 293, 294.
albida, De Vries, mentioned,

ammophila, Focke, 11; men-

42.
argillicola, МасК., 11, 12, 13.
biennis, Linn., T 12, 39-40;
mentioned, 3, 4, ftnote 13,
14, 15, 16, 17, 18, 19, 20,
21, 27, 28, 41, 42, 43, 44,
45, 46, 47, 48, 51, 52, 53,
54, 55, 56, 57,59, 60, 61.
-—— x О. gigas, mentioned,
51.
— x 0. levifolia, mentioned,
53, 56.
x —— —— læta, men-
tioned, 56.
x O. Lamarckiana, men-
L 53, 55, 50.

—— — læta, men-
бйз. 56.
—— X — velutina,
mentioned, 56.
х ——-, . velutina
x O. Lamarckiana, men-
tioned, 56.
x O. muricata, mentioned,
55.
— x 0. nanella, mentioned,
54, 56.
x О. rubriealyx, men-
tioned, 56.
——- [var.] grandiflora, men-
tioned, 47.

Малда, Schout., mentioned, 30.

brevistylis, De Vries, 4, 27-28;
mentioned, 16, 21, 22, 23,
40, 47, 51, 55, 61.

canovirens, zum 11, 13; men-

tioned, 14, 3

eheradophila, Fw tlett, ix, 13;
mentioned, 14.

cognata, Hort. Panorm., men-
tioned, 11, 33, 34, 30.

Drummondii, Hook., 11, 12,

elliptica, De Vries, mentioned,

45.
erythrosepala, Barb., 47.

383

(Enothera foliis lanceolatis, dentatis,
caule hispido, Miller, 17.

ovato-lanceolatis planis,
Miller, 1

fruticosa, pee mentioned, 17.

gigas, De Vries, 2, 4, 5, 6, 31-
34; mentioned, 34, 35, 46,
49, 50, 51, 52, 55, 58, 59,
60, 61, 292. °

--- x 0. biennis, mentioned,
51.

—— x О. brevistylis, шеп-
tioned, 51,

x О. Lamarckiana, men-
tioned, 51,

—— x О. muricata, mentioned,
51.

—— x О. rubrinervis, шеп-
tioned, 51.

[var.] nanella, mentioned,
32, 33, 34.

grandiflora, Solander, 11, 12,
38-39; mentioned, ftnote 13,
14, 15, 16, 17; 18, 19, 20,
24, 25, 27, 35, 40, 42, 44,
45, 46, 47, 59, 60, 61,

Heribaudi, Lévl., 11, 13; men-
tioned, 14,

Hookeri, Torr. $ Gray, 11, 12;
mentioned, 14, 19, 20, 27.

Jamesii, Torr. $ Gray, 11, 12,13.

læta, De Vries, mentioned, 28,
52, 53, 54, 55, 56, 57, 61.

levifolia, De Vries, 4, 26-27;
mentioned, 21, 22, 23, 29,

[var.] nanella, Schout. (2),

mentioned, 27.

[var.] salicifolia, De Vries,
mentioned, 27.

Lamarckiana, Ser., variations
and mutations, 20-26; men-
tioned, 2-61, 67.

— x 0. biennis, mentioned,
53, 54, 55, 56.

x (О. Меп-
nis х О. Lamarckiana), шеп-
tioned, 57.

x x O. Lamarckiana,
mentioned, 55, 56.

x О. brevistylis, men-
tioned, 35. |
x О. gigas, mentioned,
49, 50, 51..

х О. rubricalyx, men-
tioned, 49, 50

384

(Enothera Lamarckiana x О. rubri-
nervis, mentioned, 49,

Lamarckiana, Ser., var. cruciata,
mentioned, 11, 25.

lata, De Vries, 5, 35-38, 293;
mentioned, 21, 26, 28, 32, 35,
87, 44, 45, 47, 48, 49, 50, 51,
52

— — X O. gigas, mentioned, 51,
52
——— x O. Lamarckiana, men-

tioned, 35, 37, 48, 49, 50.
х О. nanella, mentioned,

50.
— x 0.
tioned, 49,
x O. semilata, mentioned,

rubrinervis, men-

37.

nanella (?), mentioned, 34,

MacBridew, Nels., 11, 12, 13.
macrosceles, А. Gray, 11, 12,
18.

muricata, Linn., 40-42; men-
tioned, 2, 3, 11, 14, 15, 16,
17, 18, 19, 20, 27, 28, 41, 48,

55, 57, 60.
—— x 0. biennis, mentioned,
55.

— x О. gigas, mentioned, 51.

'— var. canescens, Robinson,

11,13; mentioned, 41, 42,43,

mut. albida, De Vries, mentioned,
1

brevistylis, De Vries, men-

tioned,

otia, De Vries, men-
‚11.

tion
fatua, De V ries, mentioned,

——— gigas, De Vries, mentioned,
ТЕ

---- levifolia, De Vries, men-
tioned, 11.

—— -—— nanella, mentioned,
11.

—— —— salicifolia, mentioned,
и.

lata, De Vries, mentioned,
1

leptocarpa, De Vries, men-

tioned, 11.

nanella, De Vries, men-
tioned, 11.

— — ——— elliptica, mentioned,
11

.

INDEX.

(Enothera mut. nanella lata, men-
tioned, 11.
oblonga, mentioned,

EE:

—— scintillans, men-
tioned, 11.
oblonga, De Vries, men-
tioned, 11.
—— rubrinervis,
mentioned, 11.
scintillans, De Vries, men-

De Vries,

tioned, 11.
semilata, De Vries, men-

tioned, 11
spathulata, De Vries, men-

tioned, 11.
sublinearis, De Vries,

mentioned, 11.

subovata, De Vries, men-
tioned, 11.

nanella, De Vries, 5, 84-35;
mentioned, 21, 24, 32, 33, 37,
38, 46, 48, 50, 53, 54, 56, 58,
59

-albida, mentioned, 84,
x О. biennis, mentioned,
53, 54, 56.

— x (0. La-
marckiana x О. Мепшіз),
mentioned, 56.

elliptica, mentioned,34,38.
-oblonga, mentioned, 34,38.
[var.] rubrinervis, men-
tioned, 50.

scintillans, men-

Oakesiana, Robbins,
mentioned, 14.

oblonga, De Vries, mentioned,
37, 44, 49, 56

ornata, Nels., 11, 12; mentioned,
14, 19, 40.

parviflora, Linn., 11; mentioned,

II, 12;

pusillus *,

Gates, 107; men-
tioned, 110.
rhombipetala, Мий., 11, 13;

mentioned, 14,

rubricalyx, Gates, 5,6, 11, 28- 81;
mentioned, 21, 26, 28, 29, 49,
53, 55, 56, 57, 58, 61.

--- x О. biennis, mentioned,
56.

— x О. Lamarckiana, шеп- |

tioned, 49.
x О. nanella, mentioned,49,

CEnothera rubrinervis, De Vries, 6,
2

, 29, 35, 41, 44, 47, 49,
50, 51, 53, 57, 58, 61.
— х О. gigas, mentioned,
51.
--- x О. Lamarckiana, men-
tioned, 29, 30, 35, 49, 57.
x d. F
marckiana x O. rubrinervis),
mentioned, 57.
x О. nanella, mentioned,
35, 49, 50.
nanella, mentioned, 50.
scintillans, De Vries, mentioned,
27, 293
nanella, 11.
semilata, De Vries, 35-88.
Simsiana, Ser., 11, 12, 13; men-
tioned, 16, 40.
sinuata, Linn., mentioned, 14.
strigosa, МасК. $ Bush, 11, 13;
mentioned, 14, 39.
suaveolens, Desf.; 47.
Tracyi, Bartlett, 11, 12; men-
tioned, 14, 40.
velutina, De Vries, mentioned,
28, 52, 53, 54, 55, 56, 57, 61.
Oleaces, mentioned, 278.
Oleandra, Cav., mentioned, 188. ~
Oleandridium *, Schimp., mentioned,

188.
Oligocladus, Weber-van Bosse, 126;
Boldinghii, Weber-van Bosse,
mentioned, 126, 127.
Prainii *, Weber-van Bosse, 107,
126-127; mentioned, 110.
pusillus *, Weber-van Bosse, 127.
Onagra angustifolia caule rubro, flore
minori, Tourn., 1
angustifolia, Топтп., 17.
latifolia, Tourn., 17.
Jloribus ampli, Tourn., 17.
Ophioglossacez, mentioned, ftnote,
2.

є,

Osmunda, Linn., mentioned, 344.
pterist, Lindl. & Hutt., men-
tioned, 178.

Otozamitest, Braun, mentioned, 75,
178, 178, 187, 188, 195, 196, 197,
198,

acuminatus +, Schenk, тпеп-
tioned, 178,

Beani +, Sew., mentioned, 178.

Bucklandi +, Schenk, 178.

Otozamites Feistmanteli t, Zigno, 186.
graphicus T, Leck., 178, 186, 188.

Paramecium, Duj., mentioned, 29.

Parrotia, C. A. Mey., mentioned, 245,
248, 250, 251

Jacquemontiana, Decne., vascular
tissue, 249; mentioned, 247,
248, 249

Paulownia imperialis, Sieb. д Zuce.,
227; mentioned, 226.

Pearson, Н. H. W., Notes on the
Morphology of certain Struetures
concerned in Reproduction in the
Genus Gnetum, 311-330.

Penea, Pearson, егг. = Репа, Linn.,

mentioned, ftnote 329.
Peperomia, Ruiz 4 Pav., mentioned,
328; ftnote, 329.
Peyssonnelia, Decne.,
tioned, 138,
biradiata *, Weber-van Bosse,
140; mentioned, 107, 110.

calcea, Heydr., 107, 139; men-
tioned, 110.

coccinea, J. Ag., 107, 108, 139;
mentioned, 110.

conchicola, Русс. $ Gran., men-
tioned,

Harveyana, Crouan, 108, 139 ;
mentioned, 106, 110.

polymorpha, Schmitz, 108, 139-
140

139; men-

f. Gardineri *, Weber-van
Bosse, 108, 139-140; men-
tioned, 110.

Ро ойуна, Griff., mentioned, 145.

Pheophycese, mentioned, 105.

Phyllanthez, mentioned, 278.

Pisum sativum; Linn., mentioned,

292.

Pleuromeia t, Corda, mentioned,
‚887,8
rests, қ Ag., 114.
van Hoevellii*, Weber-van Bosse,
114; mentioned, 107, 109.
Polyosma, Blume, mentioned, 260,
293, 295, 296, 301.
Polypetale, mentioned, 301.
Polysiphonia, Grev., 125 ; mentioned,
1

вр., 107, 125; mentioned, 110.
Potamogeton, Linn., mentioned,
ftnote 344, 868. | |(

INDEX.

Primula, Linn., mentioned, 51, 294.
floribunda isabellina x

Р. kewensis, mentioned, 51.
sinensis, Sabine, mentioned, 292,

var.

Prunus, Linn., mentioned, 296.
Pseudendosiphonia, Weber-van Bosse,
123; mentioned, 106, 108, 109.

Gardineri*, Weber-van Bosse,
123-124.
Pseudoeycas t, Nath., mentioned, 178,
179, 194, 195.
Psilotacese, mentioned, ftnote, 342.
Psilotales, mentioned, 370.
Pteridosperms, mentioned, 98, 199.
Pterophyllum t, Brongn., mentioned,
178, 179, 195.
spp., mentioned, 178.
Ptilophyllum T, Morris, 73-74, 75,
76, 180, 182, 195, 196, 198, 199;
mentioned, 75; fronds, mentioned,

aeutifolium T, Oldh. $ Morris,
78; mentioned, 75.
cutchense T, Ғейвіт., 73; men-
tioned, 70, 74, 75, 79.
hirsutum t, Я. Н. Thomas, men-
tioned, ftnote 183.
pecten *T, Morris, n. comb., 179,
180, 181, 182, 183; men-
tioned, 185, 187, 189.
sp. t, 183, 184.
Ptilozamites T, Nath., mentioned, 178,
179, 190, 196, 198, 199.
Leckenbyi t, Nath., 194; men-
tioned, 190.
Pyrus, Linn., mentioned, 296.
Malus, Linn., mentioned, 298.

Quercus, Linn., mentioned, 243, 244,
Robur, Zinn., mentioned, 300.
Quisqualis, Zinn., mentioned, 293.

Rafflesia, R. Br., mentioned, 147.
Ralfsia expansa, J. Ag., mentioned,
1

Ralfsiacew, mentioned, 140.
Ranuneulaces, mentioned, 299.
Ranunculus аавв,
tioned, 360.
Reinboldiella, De Toni, 137.
Schmitziana, De Toni, 107,137;
mentioned, 110.

Lam., men-

. SECOND SERIES.— BOTANY, VOL. VIII.

885

Rhexoxylon *+, N. Baner., 100.
africanum *, М. Baner., 87-101;

affinities and reference of

96-100; external
characters, 87-88; internal
structure, 88-89; structure
of ground-tissue, 89; structure
of steles, 91-93,

Rhizanthew, 143.

Rhizophoraceae, mentioned, 268.

Rhodoglossum, J. Ag., mentioned,

114.

specimen,

Rhodoleia, Champ., mentioned, 244,

6, 247, 250, 260.

Championi, Hook., mentioned,

Teysmanni, Miq., mentioned,
246.

Rhodomelaceæ, 121; mentioned, 109,
110, 128, 135
Rhadophyceæ, 105; mentioned, 106.

Rhodophyllidaceæ, 114; mentioned,
109.

Rhodophyllis peltata, Grun.,
tioned, 119.

Rhodymenia prolifera, Harv., 117.

Rhodymeniacew, 118; mentioned,
109, 119. |

Ricinus communis, Linn., mentioned,

men-

Rosaces, mentioned, 244, 248, 290,
295, 296, 298, 299.
Roschera, Sond., 125.
africana, Sond., 126.
glomerulata *, Weber-van Bosse,
125 ; mentioned, 107, 110.

БаПсасев, mentioned, 295.
Salicales, mentioned, 277.
Salix, Zinn., mentioned, 277, 278.
Sambucacew, 240, 260, 264, 265.
Sambucus, Linn., mentioned, 240,
248, 244, 953, 255, 256, 258, 959,
264, 265, 273, 291, 292, 295, 296.
canadensis, Linn., mentioned,
264, 265. |
nigra, Linn., floral structure,
954; mentioned, 265.
Santalaces, mentioned, 289,
Баріпдасеге, mentioned, 296, 297.
Sarconema, Zanard., 114.
furcellatum, Zanard., 107, 114,
115; mentioned, 109.

Saxifraga, ons mentioned, 250,251.
3K

Бах тасіп dis ig

SR

1 Steloxylon +, Solms, | mentioned, 96,

986

rei md s c
295, 298,

260, 264,

, 246, 248.
Scepa, Lindl., 278
Scinaia, Bivona, mentioned, 112.

Scolopendrium, Sm., mentioned, 188,

Scrophulariaceew, mentioned, 226.

‘Sealark’ Expedition, mentioned,106.
Selaginella, Spring, nientioned, 369,
370 ; >

lævigata, Baker, var.
Spring, mentioned, 232.

spinosa, Beauv., mentioned, 337.
‘Siboga’ Expedition, mentioned, 105.
Sigillaria T, Brongn., mentioned, 370.

Sigillariew, mentioned, 369.
* Bokolowa," type of nucleus, 327.
Әрімегососсасеге, 116; mentioned,

Sphenolepidium t, Heer, mentioned,

Sphenophy llales,

its Linn., mentioned, 296, 299.
Ulmaria, Linn.,

361.
Spyridia, Harv., 136

filamentosa, Harv., 108, 136;

mentioned, 110.

Squamariaces, 138 ; mentioned, 110.
Stangeria, Т. Moore, 157, 158-160,
173; mentioned, 162, 171, 174,

175, 177, 181, 188, 197, 198.

paradoxa, T. Moore, mentioned,

158, 159, 198.

Stangerites t, Bornem., mentioned,

LI

_ 98, 100, 101,
Ludwigii +, Solms, mentioned,
98,

сец purpureum, Fr., mentioned,

: Sigmar T

Pronto

e кылы; Stromf., 140.

ook Streblocladia, Schmitz, mentioned,

: Sutcliffia +, Жылына, 89, du
b

Lyallii,

mentioned, 369,

mentioned,

- var. schizodon, mentioned,

mentioned,

INDEX.

Sycopsis, Oliver,
246, 250, 296.

mentioned, 240,

Sympetale, mentioned, 297, 299,
301
Symphoricarpus, Juss., mentioned,

253, 255, 256, 258, 259, 260, 293.

racemosus, Miche., mentioned,
252, 253.
Teniopteris Т, Brongn., 188, 195,

major T, Lindl. $ Hutt.,
mentioned, 190,
stenoneura Т, Schenk, 178.
vittata T, Brongn., 188, 189;
mentioned, 190, 192, 193.
Tapeinodasya, Weber-van Bosse, 131;
mentioned, 106.
Bornetii, Weber-van Bosse, men-
tioned, 131, 139,
Ее] *, Weber-van Bosse, 131—
132; mentioned, 110. `
| Taxus, Linn., mentioned, 325, 326,

189;

327.
Tetrathyrium subcordatum, -Benth.,
‚ mentioned, 248.
Thinnfeldia f, Ettingsh., mentioned,
178, 195.
spp. T, mentioned, 178.
Thomas, H. Hamshaw, and Nellie
Bancroft, On the Cuticles of some
Recent апа Fossil Cycadean
Thuites expansus t, Phill., 71.
Todea, Willd., mentioned, 344.
Tolypiocladia glomerulata, Schmitz,
125; mentioned, 126, |
Trichocladus, Pers., нао, 945,
250.
cane Linn., mentioned, 960,
uridales, mentioned, 299.

| (ағы te of Gnetum, 323.

Udotea, Lama., mentioned, 120, 140.
Umbelliferze, mentioned, 242.
Umbelliflore, mentioned, 299.

Valerianacew, mentioned, 997.
Viburnum, Linn., mentioned, 241,
243, 253, 256, 258, 259, 260, 272,
273, 286, 296, 297, 299,
Lantana, Linn., floral struc-

ture, 254; mentioned, 253,

Viburnum Tinus, Zinn., mentioned,
255.
Volkelia +, Solms, mentioned, 98,

Weber-van Bosse, A., Marine Algæ,
Rhodophycexw, of the ‘Sealark
Expedition, collected by Mr. J.
Stanley Gardiner (communicated
by Gardiner), 105-141

Welwitschia, Hook. f., mentioned,
175, 199, 313-329,

West, Cyril, and Н. Takeda, On
Isoétes japonica, A. Br., 333- 372.

Wielandiella t, Nath., pe
188, 1

angustifolia T, Nath., 187,
Wikstroemia, Ændl., mentioned, 31.
Williamsonia T, Carruth., mentioned,

9, 76, 180, 183, 184.

gigas T, Carruth., mentioned,
79, 76, 180.

pecten T, Carruth., mentioned,
79, 179, 180.

scotica T, Sew., mentioned, 184.
Williamsonian type, fructification,
76; yep 78,:.80,. 196;
stems, 77, 78, 7
Wilson:a indica, s Toni, 133.

Yatesia T, Carruth., mentioned, 69.

Zamia, Linn., 157, 170 ; mentioned,
72, 75, 80, 159, 172; 174, 175,
Тут, 182, 197.

cylindriea, Liebm., 170.

Fischeri, Miq., 170, 171; men-
tioned, 172.

gigas T, Lindl. $ Hutt.,
tioned, 180. Bs

integrifolia, .4it.,. 170, 172;
mentioned, 173, 177.

Leiboldii, Miq., 170.

— Regel, 170, 171, 17

177.

теп-

Loddigesii, Miq., 10; aem

tioned, 72.
muricata, Willd., 170, 172.
var. picta, 174.

Ottonis, Mig., 170, 171, ne
kinneri, Warsz.,170, 171, 172
mentioned, 72, 85. D

Zamites t, Brongn., mentioned, 186, a

195, 196, 197. a
gigas Т, Lindl. uen 184186. —
тенир +, 185 x

NEAN SOCIETY

MEMORANDA CONCERNING ашы. pe о

still for sale. On ly certain s eem volumes, or parts ek mplete sets, may E —
The price of the Index to Vols. 1-22 is 8s. to es public, and ба; Fellows; i ко
lows. _ Е

е Transactions is divided i nto dnd and Botanical sections. The prices a the
? which Ms been pt мы are as andermentioned. (For the женыл раг rts see Zoolo gical /