THE

VOYAGE OF H.M.S. CHALLENGER.

BOTANY-VOL. II.

REPORT

(2.

ON THE

SCIENTIFIC RESULTS

%e,

OF THE

VOYAGE OF H.M.S. CHALLENGER

DURING THE YEARS 1 873-76

UNDER THE COMMAND OF

Captain GEORGE S. NARES, R.N., F.R.S.

AND THE LATE

Captain FRANK TOURLE THOMSON, R.N.

PREPARED UNDER THE SUPERINTENDENCE OF

THE LATE

Sir C. WYVILLE THOMSON, Knt, F.R.S., &c.

REGIUS PROFESSOR OF NATURAL HISTORY IN THE UNIVERSITY OF EDINBURGH
DIRECTOR OF THE CIVILIAN SCIENTIFIC STAFF ON BOARD

AND NOW OF

JOHN MURRAY

ONE OF THE NATURALISTS OF THE EXPEDITION

WO
H(
MAS

Botany — Vol. II.

Publishes bp <£rUer of i£>et Jflajestp's ^obernment

PRINTED FOR HER MAJESTY'S STATIONERY OFFICE

AND SOLD BY

LONDON :— LONGMANS & CO.; JOHN MURRAY; MACMILLAN & CO.; SIMPKIN, MARSHALL, & CO.

TRUBNER & CO.; E. STANFORD; J. D. POTTER; AND KEGAN PAUL, TRENCH, & CO.

EDINBURGH:— ADAM & CHARLES BLACK AND DOUGLAS & FOULIS.

DUBLIN :— A THOM & CO. and HODGES, FIGGIS, & CO.

1886

Price Fifteen Shillings.

PRINTED BY

BALLANTYNE, HANSON AND CO., EDINBURGH AND LONDON,

FOR HER MAJESTY'S STATIONERY OFFICE.

CONTENTS.

I. — Report on the Diatomace^; collected by H.M.S. Challenger during the years

1873-1876.

By Conte Abate Francesco Castracane degli Antelminelli, Eome.

(The Manuscript was received in December 1884.)

79073

EDITORIAL NOTE.

This Volume contains the fourth and concluding Part of the Botanical
Series of Eeports on the Scientific Results of the Expedition, and
is a Memoir by Conte Abate Francesco Castracane degli Antelminelli
on the Marine Diatomace^e collected during the Expedition. The
Report was originally written in the Italian language, and was trans-
lated into English under the direct supervision of Conte Castracane ; it
has, however, been found necessary to submit this translation to consider-
able modifications of form before sending it to press. In some respects
it would have been desirable had the Author included in the Report
lists of the species found in the deposits and in the surface waters
of different regions of the ocean ; such lists will, however, be given in
the Report on Deep Sea Deposits.

It was at one time hoped that an additional Report would be in-
cluded in this Volume upon Pelagic Alce other than the Diatomace^e,
such as Ocillatorise, Coccospheres, Rhabdospheres, and some other
organisms which are regarded as Pelagic Algse. The material brought
home was, however, not extensive enough, nor in a sufficiently good
state of preservation, to admit of the preparation of a satisfactory
Report ; indeed, it would not have been possible to add much to what
has been said about these doubtful organisms in the Narrative of the
Cruise, or to the remarks it will be necessary to make concerning them
in the Report on the Deep Sea Deposits. An interesting account of
these Pelagic Alg^ may be looked for from the first Naturalist, who
has the time and opportunity to examine [them in the living state on
board ship, immediately after having been taken from the waters of
the open ocean.

John Murray.

Challenger Expedition Office, 32 Queen Street,
Edinburgh, November 14, 1886.

EKEATUM.
On p. 123, line 10 from top, for "trocos" read "trochus."

THE

VOYAGE OF H.M.S. CHALLENGER.

BOTANY.

REPORT on the Diatomace^e collected by H.M.S. Challenger during the
Years 1873-76. By Conte Abate Francesco Castracane degli
Antelminelli.

PREFACE.

Although the group of the Diatomaceae was almost unknown until within the last half
century, the activity that has been exhibited in this very attractive department of
research has of recent years been so great that many hundreds of different species have
been already recorded. Two branches of the study, however, possess perhaps a more than
ordinary amount of interest. Among the numerous types that occur in the more super-
ficial waters of the ocean, or that have sunk to the bottom and now form part of the
oozes of the ocean bed, many unusual forms which are the representatives of several of
the most elegant of known genera have been met with, while many geological deposits of
marine or lacustrine origin, and occurring in various parts of Italy and in Sicily, as well
as several other regions of the globe, not unfrequently present rare and superb frustules.

Having for several years occupied myself with the careful study of many marine
soundings and with the examination of various geological deposits, and having taken
pains to procure frustules from various parts of the world for comparison with those
found in such deposits by communicating with other well-known micrographers, I lost no
time on the return of H.M.S. Challenger in expressing to the naturalists at the head of
the Commission the desire that I might procure from them spare materials for the
purpose of facilitating my work. The courteous assurance that this request would in due
time be attended to afforded me great satisfaction, and this was still further increased
when, after writing to Dr Radford of the importance of instituting a comparison between
the Italian deposits and the Antarctic diatomaceous banks first made known by Sir J. D.

(bot. CHALL. EXP. — PART IV. — 1886.) a

ii THE VOYAGE OF ELKS. CHALLENGER

Jlooker, I was informed by the late Sir "Wyville Thomson that everything would be done
to aid this investigation. The announcement which I received somewhat later from Mr
John Murray that Sir Wyville Thomson proposed that I should undertake a systematic-
examination both of the littoral and pelagic Diatoms brought home by the Expedition
caused me no little delight or surprise. The proposal, however, I gladly accepted, and
although conscious of the arduous nature of the work, I resolved to spare no pains in
making it as complete as the time at my disposal would permit.

The materials as intrusted to me consisted of about a hundred small bottles con-
taining muds or oozes from different regions, together with several surface collections
obtained by the use of the tow-net. The whole was in excellent condition, and the
depth and position of the gathering, as well as the date when it was procured, were
carefully recorded on every sample.

The time necessary for the preparation of illustrative slides from so many gatherings
and for the determination of new genera and species has necessarily been somewhat long,
but I am indebted to Mr Murray for an extension, more than once, of the period at first
"ranted to me, and this has enabled me to make the present Report somewhat more
complete than it otherwise could have been. It cannot be doubted, however, that
.among the great amount of material at my command many new or rare forms have
escaped observation, and these may be published at a subsequent time as a supplement
to the present work.

In several cases criticisms have been offered or emendations given of the definitions
of previously known genera and species. As was to be expected from the examination
of gatherings made in so many hitherto unexplored regions, the list of new species that
have been established — some provisionally — is considerable, but care has been taken to
limit these as far as possible. A few new genera have been founded, but on the other
hand several old ones have been abolished, our recent advances in the knowledge of the
Diatomaceaa being such as to reveal the inadequacy of the characters that have been
hitherto regarded as of generic value. In all cases where the differences between any
form and the already known typical species have been of a subordinate kind and quite
insufficient to justify the introduction of new specific names, varieties have been consti-
tuted, but these instances are not very numerous.

Although in the earlier stages of my work it was my endeavour to record the various
species that occurred in each gathering with a view of perfecting somewhat more fully
our knowledge of the environments and geographical distribution of the various species, I
soon abandoned this method, and have been satisfied by merely recording in most cases
the locality from which the various species described have been obtained.

The plates have been prepared under my direct supervision by Signor Cesare Cerri,
and to him my best thanks are due for the faithful manner in which he has executed the
work. It is perhaps somewhat unfortunate that a natural classification has not been

PREFACE. in

followed in these plates, but this was impracticable, as the earlier figures had to be
engraved before the entire collection was examined.

To the late lamented Professor Sir Wyville Thomson 1 lie under a permanent debt of
gratitude for the opportunity afforded me of examining and reporting on this interesting
group of organisms, and my best thanks are also due and rendered to Mr John Murray
for the valuable aid of all kinds which he has given me in my work ; as well as to Mr.
John Rattray of the Challenger Editorial Staff for the care and pains he has taken with
the proof sheets as they were passing through the press, and for many valuable suggestions.
To Dr Radford, Mr J. A. Tulk, Dr Stolterfoth, and Mr Weissflog of Dresden, who
have manifested the greatest interest in my work, and have often furnished me with
valuable advice, I also lie under the deepest obligation, and last, but not least, to Dr
James Rae, R.N., who has with the greatest liberality furnished me with many unique
preparations of new and rare frustules.

Although no one can be more conscious of the many imperfections of the work than
the author himself, it is earnestly hoped that these are not of a kind to derogate too
much from its value as a work of reference on this exceedingly interesting branch of
marine investigation.

CONTENTS.

INTRODUCTION .....

Limited Condition of our Knowledge of Diatoms

The Biology of Diatoms

The Geographical Distribution of Diatoms .

The Bathymetrical Limit of Diatoms

Hypothesis on the Formation of Banks and Deposits of Diatoms

1
1
3

9
11
14

DESCRIPTIONS OF GENERA AND SPECIES
Raphidie^e ....

PsEUDORAPHIDIE/E

CryptoraphidiezE

17
17
41
69

INDEX

EXPLANATION OF THE PLATES

171

INTRODUCTION.

LIMITED CONDITION OF OUR KNOWLEDGE OF DIATOMS.

The extreme difficulty of acquiring an adequate knowledge of Diatoms, owing to their
small size and the impossibility of following attentively and registering every stage of
their development under the microscope, renders the greatest care necessary in the obser-
vation of every phenomenon presented by the living species. As, however, the making
of such useful observations is entirely fortuitous, the majority of the earlier investigators
are found almost exclusively to have been satisfied with making a mere record of the
different types occurring in their preparations, the most of which were purchased from
traders in this department of microscopy, who vied with each other in procuring new
or rare forms.

The eagerness of these observers to possess such novelties, with the view, in many
cases, of designating new species by their own names, along with the too hasty work of
the first discoverers of Diatoms, and especially of Ehrenberg. had the effect of introducing
a nomenclature that was misleading, and of causing a confused synonymy that has been
justly designated the opprobrium of the science. Against the continuance of this state
of things Professor Walker Arnott raised an authoritative voice, going so far even as to
declare that any researches relating to types found in the stomachs of molluscs and fishes,
or obtained from marine soundings, were altogether useless. By thus acting it cannot be
doubted that this naturalist exceeded all reasonable limits, but it is to be borne in mind
that he did so with the view of putting a restraint on those who were too careless in
conducting their investigations.

In the present state of our knowledge, one of the most important applications of an
acquaintance with Diatoms lies in the field of Geology, where it serves as a guide in
tracing the history of many formations which teem with their remains or impressions.
Thus, although a siliceous bed may be found on the side of a mountain several hundred
feet above the level of the sea, and although it may be in a position of notable inclination
to the horizon, an accurate examination of the contained frustules may prove whether the
deposit was laid down in fresh or salt water, and whether, in the latter case, it occurred
in the open sea or in greater proximity to the land. Moreover, as the obliquity of
the bed demonstrates the existence of subterranean forces of upheaval subsequent to
its formation, so the presence in it of layers of chalk or clay point to a removal of

(bot. chall. exp. — paet iv. — 1886.) D 1

2 THE VOYAGE OF H.M.S. CHALLENGER

the diatomaceous growth from the area of deposition while these intrusive layers were
being formed.

That the comparison of recent and fossil or semi-fossil species is very useful as a means
of elucidating structural peculiarities cannot be doubted, yet the determination of species
is a matter of great difficulty. Thus although two given forms may be readily seen to
belong to the same genus, the points of difference between them may be merely varietal
and not specific, hence it becomes necessary to give careful consideration to the real
significance of these points. To those who hold that all organic forms are but modifica-
tions or evolutions of a monad or primitive cell, the existence of species is enigmatical and
such differential features are of but secondary moment ; while, on the other hand, to those
who regard the slightest deviation from a typical form as sufficient justification for the
creation of species, these characters become of great value. In any case, however, although
it is often impossible to decide whether two closely allied organisms differ in a varietal or
specific manner, the omission to record the existence of distinct structures, which tend to
make the history of the given type more complete, is unjustifiable. For this reason the
iudication of every notable form will frequently be made, but modifications will very often
be treated as varietal, and new species will never be founded save on the presence of
two or more interesting differences existing side by side. Where these are established, or
new varieties recorded, the arguments that lead to such a course will always be given.

With regard to the geographical and bathymetrical distribution of Diatoms, no
definite results can yet be laid down on account of the very limited condition of our
knowledge of marine forms. As to the former, however, the desirability of registering the
locality from which every organism has been obtained as a means of facilitating future
work is apparent. The bathymetrical range is a question of even greater difficulty, as
only living species, and the depth at which they are found, can be taken into account. It
may here be noted that the advantages arising from a careful record of the depth from
which Crustacea, corals, madrepores, &c, are obtained, are very great, as frustules recog-
nised to be living by the presence of the endochrome are often found adhering to their
surfaces, and thus facts may often be gathered that go to aid in the determination of the
limits in depth of vegetable life.

The importance of prosecuting investigations in this direction is so great that a means

of conducting experiments was suggested by me in March 1871. ' I had repeatedly

observed that, after exposing a vessel of water covered with a glass plate to the light

for some weeks, small Diatoms and other unicellular Algae appeared in it, and the

same result was obtained whether the water were fresh or marine. As this growth could

only have been due to germs pre-existing in the water, since carefully prepared artificial

sea-water led to entirely negative conclusions, it is very probable that if, with a suitable

apparatus, sea-water could be drawn from various depths, for example, at intervals of

1 Esame microscopico c note critiche su un campione di fango Atlantico ottenuto nella spedizione del
"Porcupine" nell' anno 1869, Atti. Accad. Pontif. d. nuov. Lincei, vol. xxiv. pp. 16-29, 1871.

KErORT ON THE DIATOMACE.E. 3

50 metres (27 fathoms), the depths being in all cases carefully recorded, and if this sea-
water could be placed in glass vessels covered with suitable glass plates and exposed to
diffuse light, so as to avoid the great heat of the direct rays of the sun, after a month
Diatoms might appear in some of the vessels. In this manner it is not unlikely that the
bathymetrical limit of plant life might be fixed with approximate certainty. Moreover,
the great facility with which this method of determination may be practised recommends
it as one likely to aid in solving this important problem.

THE BIOLOGY OF DIATOMS.

This interesting group of organisms, the knowledge of which we owe entirely to the
perfection of the modern microscope, has for upwards of half a century occupied the
serious attention of naturalists.

To the entire series of forms the designation Diatomophycese was given by Eabenhorst
and Bacillariece by Kutzing, but the name of Diatomacese is now generally adopted.

Diatoms may be defined as unicellular alga? having a siliceous coat. Each organism
forms a small box, the siliceous walls of which completely enclose a space ; these walls in
many, if not in all species, are formed by two distinct plates or valves, each possessing its
own hoop, one of which embraces and slides over the other like the tube of a telescope
or the lid of a box. This hoop, connecting zone or belt, may be single, double, or of
complex structure, as in Rhabdonema, Striatella, and some other species, where it is
formed of several hoops. Such a complete individual Diatom is called a Frustule, and it
may be viewed from two aspects, to which, however, confusing terms have been applied.
Thus Kutzing and W. Smith, followed by the English micrographers, speak of front view
when the organism presents its belt to the observer, and side view when its valve is next
the eye ; but, as the frustule is formed by two plates joined together, it seems reasonable
to speak of the front of the box when the valve is seen, and of the side when the union
of the two plates (i.e., the belt) is presented. This latter opinion has been adopted by
the German naturalists, who call Kutzing' s side view the secondary view and his front
view the primary. In order, however, to avoid confusion, the terms valval and zonal
might be applied according as the valve or connecting zone is next the observer.1

1 The following arrangement will serve to show the relationship hetween these various terms : —

C Lateral surface (Ehrenberg).

(a) Front view (Ralfs/W. Smith, &c), i.e., hoop next observer = i ^^ side (Kutzing).

I Secondary side (Rabenhorst).

v Zonal view.

C Dorsum and venter (Ehrenberg).
(6) Side view (Ralfs, W. Smith, fee), i.e., surface of valve next observer = \ Sefondary side (Kutzing).

J Primary side (Rabenhorst).
v Valval view.

4 TILE VOYAGE OF LI. M.S. CHALLENGER.

The siliceous cell of the Diatom is filled with finely granular protoplasm, as may be
seen after treatment with water slightly acidulated with hydrochloric acid. This causes
the protoplasm to contract so that a vacant space appears along the sides of the frustule.
The protoplasm encloses a coloured substance called endochrome, as well as a few
round granules of an oily nature which dissolve under the action of ether or carbon
bisulphide.

The endochrome is composed of chlorophyll and diatomine or phycoxanthine,1 but of
these the former is the more interesting, as it is capable, under the influence of light, of
decomposing carbonic acid and evolving oxygen. The endochrome is of a greenish-
yellow colour, and occurs in the form of bands, granules, or rounded masses which may
be arranged irregularly or in radiating lines.

The appearance which this endochrome assumes in the same species is not always
identical, and frequently, when it occurs as an amorphous mass, it becomes subsequently
divided into numerous granules of equal size and definite outline. That these are to be
regarded as so many perfect embryonal cells which are destined to reproduce the parent
form cannot be doubted (see pp. 7, 8).

In addition to the endochrome and oleaginous substance, there is, in the centre of the
cells, a protoplasmic mass distinguishable by its density from the protoplasm which occupies
the rest of the interior of the Diatom. This is called the Cytoblast or Nucleus, and
in its turn encloses the Nucleolus in its inactive condition, as may be demonstrated by
the action of a very weak solution of chloride of gold.

The cytoblast and the nucleolus certainly exist in all Diatoms, but it is not very easy
to distinguish them in small forms. In large species, however, such as Striatella unvpunc-
tata and a few Amphora, these structures may be readfly seen, and in these the cytoblast
may be observed to possess a stellate form having filaments resembling the pseudo-
podia of Diffluc/ia or other Infusoria which sometimes extend to the extremities of the
frustule.

Diatoms are solitary or gregarious, free or fixed. The fixed forms are attached by
means of a peduncle which may be simple or branched, or they may be enclosed in a cyst,
fastened on a small cushion, or embedded in a gelatinous substance which forms a frond
either in the form of a simple or compound tube, of a flattened plate, or globular mass.
In seriate species the frustules may be arranged either in a lineal or lateral manner. They
are also often met with in the form of zigzag chains, adhering to each other by means of
small angular cushions, as in Tabellaria.

The substance which exists in the form of a definite frond resembling a higher alga,
being either ramified, ulvaceous, tubular, saccate, or mucilaginous and amorphous, and

1 Sachs' Text-Book of Botany ; Millardet and Kraus, Compt. Rend., vol. Ixvi. p. 505 ; Askenasy, Botan.
Zeit., 18G9, p. 799; Hitzer in Heft, ii. Botan. Abhandl., edited by Ilanstein, Bonn, 1871 ; Quart. Jour. Micr.
Sex, 1872-73.

REPOET ON THE DIATOMACE^E. 5

which encloses a greater or less number of regularly or irregularly disposed diatomaceous
frustules, was designated by Brebisson by the name of " Coleoderma." It is to be noted,
moreover, that solitary and free Diatoms are not completely devoid of external
mucus, but, on the contrary, are surrounded by a thin stratum of this substance. This
becomes manifest when their movements in water coloured with indigo are observed
under the microscope, inasmuch as the granules of coloured matter do not come
into immediate contact with the Diatom, but leave a thin translucent area surrounding
the frustule.

But the most singular and curious phenomenon observable in some species, and one
which misled their first observers so far that they considered these organisms to be of an
animal nature, is the rectilinear motion with which, among others, the Naviculw, the Cym-
J>cUcB, &c, are seen to advance in the direction of their long axis and to return along the
same line. The strongest magnifying power, the most accurate and careful application of
the most perfect microscope, and the most skilful methods of observation, have failed to
discover the existence of any special organs of locomotion. This being so, the most
probable and generally accepted hypothesis suggested in explanation of the movement is
that the Diatom, in order to obtain the silex which constitutes its walls, must continually
absorb and reject the water which has in solution imperceptible traces of the siliceous
substance, and that the consequent action and reaction thus exerted upon the li^ht
frustule suspended in the watery medium, determines its alternate forward and backward
movement.

Besides this phenomenon of motion another circumstance connected with these interest-
ing organisms early attracted the attention of naturalists, namely, the extreme delicacy
of the details with which their valves are ornamented. There is probably not a single
Diatom which, when examined by the best microscopes, does not show the surface of
its valves to be adorned by exceedingly minute granules, generally arranged in lines
which are usually spoken of as "striae," although properly they are lines or rows of
points.

That these stria?, however, are sometimes perfectly continuous cannot be denied, although
some naturalists assert that the appearance of a continuous streak is purely illusory,
the so-called streak being but the expression of confluent granules. In this connection it
is to be noted that the phenomenon of diffraction produced by the obliquity and intensity
of the illumination has occasioned a belief in the confluence of the granules of pinnuloe,
&c, for example in Pinnidaria major, Raben., as such a result can be obtained at pleasure
in that and other frustules. This consideration, moreover, has led some to deny that there
is any value in the division of the Naviculaceos made by Ehrenberg when he instituted
the genus Pinnularia for navicular forms having pinnulae or coarse continuous strke, and
limited the idea of Navicula exclusively to navicular forms adorned by rows of granules.
Like the quality and delicacy of the ornamental details, the forms of the valves of

•

6 THE VOYAGE OF H.M.S. CHALLENGER.

Diatoms vary greatly. These are generally more or less convex on the outside and
concave on the inside, increasing in this manner the capacity of the cell and at the same
time enabling it more readily to resist the external pressure which would otherwise tend
to crush it. Not only are the forms of the two valves in most cases symmetrical with
each other, but each valve is divided in many species into two similar portions by a
central rib, called a " raphe," which serves to strengthen the valve.

In the centre of this raphe, and sometimes at its two extremities, swellings called
" noduli " often occur. The central nodulus, instead of being round, is sometimes extended
so as to form a kind of cross which is named a " stauros."

Although in a great number of species a central raphe occurs, in many others no
such structure is to be found. Hence, upon the presence or absence of this raphe,
Professor H. L. Smith has proposed a new classification of Diatoms, distinguishing them
into Raphidiece, Pscudoraphidiew, and Cryptoraphidiew. It is to be remarked that this
classification has the defect of not being natural, and, moreover, it would seem that the
Cvyptorapliidicae should be called Araphidiem or Anaraphidiece, inasmuch as none of
the types included exhibit traces of the existence of a raphe or of any other analogous
arrangement. Nevertheless, since this system has the merit of simplicity, and has
already been adopted by the Austrian naturalist Albert Grunow, and by Dr Henri van
Heurck in his Synopsis des Diatome'es de Belgique, it will be followed in the present
Report.

Diatoms, like all other living organisms, multiply and reproduce themselves. Multi-
plication, however, is not to be confounded with reproduction, but is to be understood as
an extension of individual life. Multiplication by duplication or division occurs in many
members of the vegetable kingdom ; on the other hand the process of reproduction is
common to every living thing. That multiplication by duplication is of frequent occur-
rence in Diatoms is proved by the fact that it has been already observed in forty or
more different species, yet it would appear that it ought to be regarded not as normal
but as exceptional. This process, which is also called fissiparous division or ternno-
genesis, takes place as follows: (1.) the nucleus or cytoblast becomes bisected ; (2.) a
central contraction of the internal protoplasmic mass takes place, and proceeds so far
that the latter becomes completely divided ; and (3.) a double wall or diaphragm, which
contemporaneously extends to the circumference in a direction normal to the ring or
cingulum, is formed. In this manner the primitive mother frustule gives origin to two
perfect daughter frustules, each of which is provided with an old and a new valve, the
latter being somewhat smaller in diameter than the former as its hoop is embraced by that
of the old valve. The process of duplication, as it progresses, must in this way lead to
a marked diminution of the diameter of the new frustules, and in a short time these would
become so small that they would be no longer characteristic of the species. In order,
however, to counteract this progressive diminution it is to be noted that the siliceous walls

REPORT ON THE DIATOM ACE. E. 7

of the frustules increase by reason of the process of fission. Dr Pfitzer ' of Bonn, by
denying the possibility of such growth, was induced to imagine the theory of the existence
of auxospores, and to believe that these constituted the only real reproduction of Diatoms
by a sexual process. According to this observer, after the frustules had, in the succession of
divisions, reached the smallest dimensions compatible with the species, the contents of the
small frustules escaped and united with each other so as to form one or two sporangia.
Within these one or two sporangial frustules called auxospores appeared, and by their
larger dimensions brought the Diatom to the beginning of another series of new graduated
forms. But such a theory, taken in a general sense, and assumed as the process of repro-
duction common to all genera of Diatoms, is fundamentally false, because it rests on the
gratuitous supposition that the diatomaceous walls are incapable of any increase in size.
That such an increase, however, does take place has been proved by the Rev. Professor
W. Smith in his classic Synopsis of the British Diatomacese (plate lii. fig. 335), where
some sporangial frustules of Orthosira diclcieii are represented in which it is evident
that the siliceous walls increase with the growth of the contents. It is also to be noted
in this connection that the distinguished Hugo von Mohl maintains that the cytoderm of
a Diatom is not entirely inorganic, but only an organic membrane which is impregnated
with silex, it having been already shown that silica is sometimes substituted for carbon in
the formation of cellulose.

Again, if during the process of duplication an expansion of the cell wall did not take
place, a hundred frustules of a Fragilarian filament would exhibit some difference in their
longitudinal diameter. But no such difference is observable. Moreover, as it is impossible
to understand the formation of the two new dividing walls in the centre of the parent cell
in all their minute details unless it be admitted that the new frustules are stereotyped
upon the old ones, it follows that such a process cannot be verified except in the genera in
which the two valves of the frustule are perfectly identical in a symmetrical position.
It is to be remarked that fissiparous division has not yet been observed in a single case
to form an exception to the above rule.

Finally, against the theory which regards the sporangial frustule as destined to initiate
a new descending series, Dr Wallich'2 remarks that that frustule, instead of being, as
heretofore assumed, the primary or parent frustule of a new and vigorous generation,
constitutes in reality the expiring phase in the life of an old one. Professor H. L. Smith
seems to be of the same opinion, for, considering the possibility of Stauroneis phcenicen-
teron, Ehrenb., being a sporangial form of Stauroneis gracilis, Ehrenb., he points out that
it is only an abnormal and transient form from which the Stauroneis gracilis, differing
from it so much as it does, could never originate.

1 Untersuchungen iiber den Bau und Entwickelung der Bacillarien (Diatornaceen), Bonn, 187E

2 On the Relation between the Development, Reproduction, and Markings of the Diatomaceae, Month.
Micr. Journ., vol. svii. p. 61, 1877.

8 THE VOYAGE OF H.M.S. CHALLENGER.

In considering the process of reproduction, it is to be noted that the phenomena
recorded during a fortunate observation in one case only cannot be regarded as typical for
the entire group. It has, however, been perfectly established that reproduction can take
place by means of two conjugated frustules, or by the two valves of the same frustule
separating themselves so that the contents form a globular mass or sporangium, within
which the formation of one or more young frustules takes place. These young forms
sometimes exhibit duplication or fission, but this does not always happen, nor does it occur
in all frustules. On the contrary, some species give rise to very minute frustules which
in their turn form a new progeny. Thus when the endochrome occurs not in a more or
less amorphous condition but as numerous rounded masses of equal size and definite out-
line, the frustule is to be regarded as sporiferous, the rounded masses representing so many
sporules that are destined to reproduce the species. The escape of these sporules has been
observed by Eabenhorst in a species of Melosira, by O'Meara in a Pleurosigma, and by
myself in a Podosphenia, while Dr Cohn, as noted by Dr Pfitzer, saw a winged Amphi-
prora emitting small frustules from which the species could be recognised.

It is further to be borne in mind that diatomaceous sporules will gradually develop
into the typical frustule either internally in the parent frustule, or externally in the spo-
rangium. Sometimes the parent frustule takes the place of the sporangium, in other
cases the sporangium is the production of the Diatom, destined for the incubation and
maturation of the sporules.

These sporules are emitted in a free and solitary condition, but they are sometimes seen
in couples, as in Mastogloia. Here each couple is enclosed in an oval cyst, and all stages
of development, from two minute oblong greenish corpuscules up to large fully formed
frustules, can readily be traced. All the couples are immersed in a dense gelatinous
mass with which no extraneous bodies are mixed. Numerous round sporules have also
been observed to form a mass, and transitional stages have been traced from the round
green granular condition to that of spherical hyaline cysts which included a number
of small but distinct Naviculce. Among the cysts some were observed to move by
means of two very slender flagelliform threads, and these must be regarded as real
zygozoospores.

It seems evident that the process of reproduction by means of sporules is the only one
that can explain the enormous rate of multiplication of the majority of the Diatomaceae,
while the other processes of duplication and reproduction will account for the greater
rarity of some forms. Viewed from a physiological standpoint, therefore, sporular repro-
duction is of the highest importance, as by this means countless individuals are formed,
all of which co-operate in supporting animal life by giving off the oxygen which is essen-
tial for the existence of the latter.

REPORT ON THE DIATOMACE/E.

THE GEOGRAPHICAL DISTRIBUTION OF DIATOMS.

Diatoms are to be found at all seasons covering with a mucilaginous coating of an
olivaceous green colour the watery surface of fountains and other damp places, and they
have been seen by me to live in a small aquarium without the least sign of inconveni-
ence for a space of three days, although all the water remained in a frozen state.

Marine Diatoms flourish in all latitudes and at all seasons of the year in the warmest
and coldest seas ; it is well known that they are so abundant in the Frigid Zones as
sometimes to colour the surface of the sea and to tinge with a peculiar hue the blocks
of floating ice. Seeing that they are capable of surviving in conditions so diverse it is
difficult to believe that any fixed laws of geographical distribution can be discovered with
respect to them ; on the contrary, it might rather be supposed that the continuity of
adjacent seas, the surface and submarine currents, the movements of tidal waves, the
existence of periodical and other winds, the traffic of ships and the movements of fishes,
would all tend to facilitate or bring about the mingling of local floras. It is true, indeed,
that, since the temperature of the sea is nearly constant at great depths, an actively
moving Diatom has only to sink deeper or approach more closely to the surface in order
to find the temperature best suited to its nature and most favourable for its development,
its vertical range of motion being limited, among other things, by the depth to which
sunlight penetrates, this being assumed to be necessary to its existence. Since, moreover,
the quality possessed by organic beings of adapting themselves to conditions of life
different from those in which they came into existence becomes greater as the position
occupied by the organism in the scale of life becomes lower, it is by no means easy to
understand how, during the long course of centuries, the different types have not been
distributed far and wide and rendered common inhabitants of all seas.

That several distinct floras nevertheless do exist may be confidently affirmed, although
it may be premature to determine finally the question of distribution according to the
genera and species that inhabit different areas. The distinction, for example, between
pelagic and littoral Diatoms has already been clearly recognised, and the importance of
the application of this knowledge to geological problems has been fully pointed out.1

Again, as regards the occurrence of distinct chorological floras it may be noted that,
by those conversant with Diatoms, collections from the Antarctic Ocean would not be con-
founded with those from the Sea of Japan or from the Arctic Sea, nor would the flora
of the Adriatic or of the Mediterranean as a whole be mistaken for that of the Indian
Ocean, although some of the same genera and species might exist in all.2

1 See Atti. Accad. Pontif. d. nuov. Lincei, Anno xxxii., Sezzion lc, December 15, 1878.

2 See Narrative of the Cruise, vol. i. p. 933.

(bot. chall. exp. — part iv. — 1886.) D 2

10 THE VOYAGE OF H.M.S. CHALLENGER.

It must, indeed, be admitted that the presence of a single common form in two widely
distant seas — such as the Arctic and Antarctic Oceans — throws a difficulty in the way of
admitting different local floras. Thus, if only a single species could, by currents or
from any other cause, have migrated in one or more stages from one point to another, it
must follow that the general mingling of the various marine genera and species is reduced
to a mere question of time. It remains, however, to be proved that in the polar seas the
Diatoms that occur, though admittedly belonging to the same genus, are also positively
of the same species. On account of our imperfect knowledge and of our limited means of
observation, it cannot be asserted that two such frustules do not form two distinct species,
though their specific characters may appear to be identical.

But allowing that some common species do exist in widely separated oceans, it may be
safely concluded that, since the complete fusion of all marine floras has not taken place in
the course of the thousands of centuries during which Diatoms have existed, it will not
happen for many ages to come.

What, then, is the meaning of setting a limit of any kind to the habitat of marine
species 1 In the present state of knowledge no definite information can be offered, and
it may be long before the circumstances under which the development of one diatoma-
ceous form takes place in a given locality, to the exclusion of another, are fully known.
Remembering, however, the great facility with which Diatoms adapt themselves to the
most varied conditions of life, the numerous opportunities afforded for the distribution of
microscopical marine organisms, and the occurrence of different temperatures at different
depths in the sea in the same latitude, it may with some probability be asserted that the
extension of species in the sea is determined to a great extent by the temperature of the
water. If such were not the case it would not be easy to explain why the Diatoms of
temperate seas should not be found in company with glacial forms, while, on the other
hand, it must be expected that, apart from considerations of pressure and other physical
conditions of environment, the species of the polar seas should be met with in great depths
in temperate waters.

Another circumstance which may have tended to prevent the fusion of local micro-
florae into a universal one may here be referred to, having been recently pointed out by
Mr John Murray in a preliminary report of observations made on board the Challenger,
read before the Royal Society of London.1

" In the Southern Ocean," says Mr Murray, " south of Kerguelen, in the Arafura Sea,
off the coasts of Japan, New Guinea, North America, and in enclosed bays and river
deltas, — in short, wherever the specific gravity of the sea is low from an admixture of fresh
water, we have met with very many Diatoms on the surface."

After examining an abundant supply of materials from the Red Sea I have been led to

1 On Surface Organisms and their Relation to Ocean Deposits, Proc. Roy. Soc. Lond., vol. xxiv. p.
533, 187G.

EEPORT ON THE DIATOMACE.E. 1 1

the same conclusion. Here Diatoms of different species do occur, but they are relatively-
scarce — a circumstance which may be explained by the fact that no large rivers flow into
this area to compensate for the enormous evaporation to which it is subjected on account
of its vicinity to the equator, so that its waters are more dense and more heavily charged
with saline constituents.

If this explanation be correct, it follows that the warmest, and, by their great
evaporation, the most saline parts of the ocean, serve as all but insuperable barriers
to the mixture of different floras, so that each may retain almost indefinitely its own
special characteristics.

Silica, of which the walls of Diatoms are composed, exists in all kinds of water, but
generally in so small quantities that its proportion cannot be determined. Hence
Professor Bischoff has lately asserted that siliceous Infusoria — meaning thereby Diatoms —
must take into their interior, in order to form their valves, the same quantity of water, in
proportion to their mass, as would be swallowed by a man who drank a cubic foot per
second.

If water were absolutely devoid of silica, Diatoms could not exist, but this and other
important substances are being continually carried into the sea by rivers. The calcareous
material present in the ocean goes to form the tests of Foraminifera and Crustacea and the
coral formations of tropical zones, whfle the silica is removed by Polycistinse, Radiolaria,
certain Foraminifera, Spongidse, and Diatoms.

The secretion of lime takes place, according to Mr. Murray, most actively in the
warmest and saltest seas where solar evaporation is most rapid, and it is a fact no less
striking that the greatest abundance of Diatoms is found in localities where the water is
of relatively low specific gravity and the temperature in some instances likewise low.

THE BATHYMETRICAL LIMIT OF DIATOMS.

Problems of distribution are, from the very nature of the conditions, much more
difficult to solve in the case of marine than of terrestrial organisms, and with regard to
the Diatomaceae it has hitherto been impossible to define the greatest depths at which they
are capable of surviving in the various ocean basins. A method has been already referred
to by the application of which it may be proved whether they do vegetate at any given
depth (pp. 2, 3), but in addition to this, careful records should be made of the occurrence
of living frustules on fragments of corals, shells, &c, which are brought up from a known
depth, as a check is thereby placed on the experimental working out of this most
important question.

From a knowledge of their physiological functions of decomposing carbonic acid and
evolving oxygen under the influence of sunlight, it may be inferred that Diatoms cannot
live where light is unable to penetrate. "With this theoretical conclusion daily experience

12 THE VOYAGE OF H.M.S. CHALLENGER.

is in perfect harmony, for these minute organisms multiply with extraordinary rapidity
in positions illuminated by and exposed to the sun, but abound less in shady places and are
never found to vegetate in darkness. Hence in the great depths of the sea, to -which light,
because of absorption, cannot penetrate, only dead valves of Diatoms will be found.

I have mven at length in another communication ' arguments based on the existence
of Diatoms in the stomachs of abyssal Echini, which seem to me to show that these
organisms might possibly live at depths of 1340 fathoms, and that sunlight might
penetrate to greater depths than at present supposed. Mr. Murray, however, informs me
that it was observed during the Expedition that the great majority of deep-sea animals
lived by eating the surface layers of mud or ooze forming the bed of the ocean, which
usually contained a large number of the remains of surface animals and Diatoms ; his
opinion is that a small amount of organic matter is carried down with the dead shells and
frustules from the surface waters where the organisms lived, and that the deep-sea animals
obtain their nutriment from it. Had Diatoms been observed in the alimentary canal of
fishes, they might have been obtained at less depths. As, however, Echinodermata live
only on the bed of the ocean itself, it must be regarded as very probable that these
frustules formed a portion of the food of these animals.

The significance of such observations in connection with deep-sea life is of the highest
value, as tending to explain the existence of animals in a normal state at the greatest
depths — a fact that has already been recognised by many marine scientific expeditions
organised by Great Britain, the United States, and other nations. The existence of life
in darkness must be regarded as exceptional, although there are numerous and well-known
instances of rare and blind animals (such as Proteus) which inhabit caves into which no
light can penetrate.

On the other hand, the facts that the eyes are fully developed in some of the animals
that inhabit the deep sea, and that they possess beautiful and brilliant colours, seem
necessarily to imply the presence of light in these abysses, since, in its absence, the organs
of vision should become atrophied, and the colouration be more or less dark.

These conclusions have been universally accepted, but the belief that light could not
penetrate to such depths on account of absorption in its passage through the water afforded
ground for advancing hypotheses to explain the facts, the reality of darkness at the
bottom not being doubted.

Among these hypotheses it has been held that the phosphorescent light given off by the
bodies of abyssal animals is sufficient to make up for the absence of sunlight. Although
the weight of this suggestion is very great, it may be asked — Can marine animals
emit light when alive and in their normal state ? Fishes, Crustacea, and other marine
creatures are phosphorescent when they are dead or at least out of their proper medium.

' Nuove Osservazione sulla Profundita cui giunge la Vegetazione delle Diatomee ncl Mare. — Mem. Pont.
Accad. d. Nuov. Lined, voL i. 1885. ;

REPORT ON THE DIATOMACEiE. 13

The sea on dark nights is lighted up under the strokes of oars, in the track of boats and
ships, along the shores, and among the breakers, and it is known that this attractive
phenomenon is chiefly due to myriads of Noctiluca?, Peridinia, and other small pelagic
forms ; yet these organisms emit a momentary splendour only when they are disturbed by
an extraneous body or by the mechanical action of the waves, although the highly
specialised apparatus in Nyctvphanes norvegica1 is apparently an evidence that in some
animals this power is under the influence of the will. Mr. Murray also informs me that
he has observed very many animals emit phosphorescent light in the stillest waters without
any apparent external stimulant.

Granting that the phosphorescence of abyssal animals is a normal characteristic, it
must still be proved that it is not only sufficient to dispel the darkness so effectively
as to explain the presence and development of visual organs, but that it is capable of
so great actinic action as to determine the vivid colouration above referred to.

Just as the bathymetrical limit of marine life, which was laid down at 300 fathoms
by Professor Edward Forbes, is now no longer accepted, so the distance to which light can
penetrate may prove to have been misunderstated hitherto, and should, in the interest of
truth, be re-examined by physicists.

It is stated that in the passage of light through sea water the first rays to be absorbed
are the calorific, then the luminous, and lastly the actinic or chemical, which are precisely
those that have most influence on colours and on chlorophyll and consequently on
vegetation in general, and attempts have been made to fix the limit of the penetration of
light by observing the gradual descent of a white object, and by noting the moment when
that object could no longer be discerned. But even although the distance thus obtained be
more than doubled, and the observer placed in a condition to be in no way influenced by the
external light, such observations are not very reliable.

Since several indications of the influence of light in marine abysses now exist, an
explanation of the manner in which it can penetrate to even very great depths should be
sought after.

An interesting observation may here be noticed. When two French aeronauts were
recently crossing the English Channel, and when at a great height above its surface, they
were struck by the circumstance that its bed could be distinctly seen, and that all the details
of the irregularities of its bottom could be traced. It may be said that the depth of this
narrow belt of water is not to be compared to that of the great oceans, but assuredly even
its depth could not be seen by an observer near the surface, and perhaps it is not less than
the limit hitherto placed on the distance to which light can penetrate.

1 Narr. Chall. Exp., vol. i. p. 743.

14 THE VOYAGE OF H.M.S. CHALLENGER.

HYPOTHESIS ON THE FOEMATION OF BANKS AND DEPOSITS OF

DIATOMS.

In 1876 Mr Murray read before the Eoyal Society of Loudon a Preliminary Report '
on Oceanic Deposits, appended to 'which is a map in which the existence of a bank
formed almost exclusively of the siliceous remains of Diatoms is indicated ; and further
details have been more recently given by Mr Murray in the Narrative of the Cruise.2 This
bank extended between lat. 60° 52' S., long. 80° 20' E., and lat. 53° 55' S., long. 108° 35' E.
It therefore measured not less than 1700 miles in length.

Such accumulations lead one to inquire how it is that Diatoms which vegetate in all
seas should accumulate in this manner in particular localities which are in all probability
few in number, since similar formations were not recorded by the naturalists on board the
Challenger in any other region.

A Diatom as soon as it ceases to live is no longer supported by the globules of oxygen
which it evolves and which adhere to it, but is left to the mercy of the waves, and in
virtue of its own weight, it sinks to the bottom, which it will reach in a relatively short
space of time, notwithstanding the disturbing tendencies exercised by oceanic currents, as
the density of the water is but little affected by the constantly increasing pressure. It is,
moreover, not improbable that the cooling and consequent descent of the surface waters in
the neighbourhood of the Antarctic ice, facilitate to some extent the sinking of the dead
frustules and their accumulation on the bottom in these regions.3

It is a matter of very considerable importance from a geological point of view to
determine exactly the conditions under which diatomaceous deposits like those discovered
by the Challenger are laid down, inasmuch as what is now taking place in the Antarctic
supplies a clue to the elucidation of the phenomena which took place in remote epochs
when similar marine diatomaceous banks were laid down.

Italy affords a very interesting example of such formations. For several years it has
been known to the scientific world that, in the celebrated sulphur mines of Sicily, siliceous
schists occur which not unfrequently contain specimens of fish. These schists have been
found to consist very largely of Diatoms and Radiolaria, so that their marine origin cannot
be doubted.

More recently the curiosity of palaeontologists has been attracted to a locality in

Central Italy called Mondatino, in Montefeltro, where, among the strata of siliceous marl,

a number of fishes were found in schists, which, like the former, were extensively formed

of the remains of marine Diatoms and of the skeletons of Eadiolaria and Polythalamia.

This formation lies along the spurs of the Apennines, extending on the one side as

1 Proc. Roy. Soc. Land., vol. xxiv. p. 471, 1876. " Narr. Chall. Exp., vol. i. p. 432.

3 Exploration of the Antarctic Regions, Scottish Geographical Magazine, vol. ii. p. 537.

REPORT ON THE DIATOMACE/E. 1 5

far as the right bank of the Savio above Ccsena, and on the other as far as the hills which
flank the valley of the Misa above Sinigaglia. It also occurs at the extremity of the
peninsula in the province of Catanzaro.

The microscopic examination of the Diatoms contained in these deposits warrants the
conclusion that they belong to the same formation and constitute so many tracts of the
same bank, which, though as yet imperfectly known in all its parts, probably extends over
the whole of Italy from Sicily to the vicinity of the most ancient Jurassic formation of
the Alps.

It may perhaps prove to be an interesting problem for the geologist who possesses a
knowledge of the palaBontology and stratigraphy of the lands which surround the basins of
the Mediterranean and Adriatic Seas, to compare this marine tripoli with the diatomaceous
deposits of the Antarctic, and to endeavour to determine the directions of tertiary oceanic
movements, and compare them with those now prevailing in the Southern hemisphere.

In opposition to the hypothesis that has just been advanced, it is to be noted that
Mr Murray is of opinion that the great Antarctic diatomaceous bank is but a portion of
a much more extensive formation which girds the Antarctic Pole.1 I believe, however,
that sufficient proof of the existence of this belt has not yet been obtained, and that
should it exist the circumstances that bring about such a formation would remain to be
determined. It is true, indeed, that Sir Joseph D. Hooker, during the expedition of the
"Erebus" and "Terror," discovered, between the parallels of 76° and 78° S. lat., a
bank of Diatoms 400 miles in length and 120 miles in breadth, but such a discovery
cannot be regarded as sufficient to prove the uninterrupted extension of the deposit
around the Antarctic Pole. The bank found by the Challenger naturalists did not extend
beyond 60° 26' S. lat., and was succeeded by a deposit of blue mud, which, however, con-
tained many Diatoms, and may be continuous with that observed by Sir J. D. Hooker.
Finally, it may be noted that should such a singular diatomaceous annulus gird the South
Pole its counterpart might be expected in the Arctic Sea. Yet, although this area has
been much more explored than the Antarctic, and although it is much less deep, so that
dead frustules might much more readily reach its bottom, no such ring-like deposit has
been discovered. Mr Murray points out that this is probably to be accounted for by
the Diatoms being much less numerous in the surface waters, and by the fact that their
presence would be almost completely masked by the great abundance of terrestrial debris
in all the deposits of the Arctic Ocean.

1 Loc. cit., p 541.

DESCRIPTION OF GENERA AND SPECIES.

TRIBE I.— RAPHIDIE^.
Amphora, Ehrenb.

The essential characters of the frustules of this genus are so difficult and complicated
as to be but little understood. So much is this the case that Professor Arnott severely
writes l that Dr Gregory having, in his pamphlet on the Diatoms of the Clyde, failed to
indicate their distinguishing features, has thereby deprived his work of its claims to be
considered scientific. These observations adequately prove the magnitude of the dif-
ficulties that exist, and indulgence may be claimed for any errors that occur in the
treatment of so obscure and intricate a genus.

The general form of an Amphoran frustule may be taken to resemble that of a grain
of coffee, but Dr Gregory's conception that the so-called frustule should be regarded as
two perfect frustules that have become united apart from any process of conjugation
cannot be entertained. If some species occur as very delicate cymbelloid forms, such as
Amphora cqffcceformis, Kg., this is attributable to the character of the connecting zone,
which is extremely slender, and which may be destroyed by the slightest possible action of
acids, even when diluted, or by fire.

Amphora speciosa, n. sp. (Plate XXVII. fig. 1.)

Elliptica sublinearis ; apicibus rotundatis ; valva introrsum infiata ; striis punctulatis
subradiantibus. Ad portum Thaiti.

This species would perhaps be identical with Amphora oblonga, Greg.,2 were it not for
the profile of the frustule, which is described by Gregory as " elliptic, rather broad," with
" ends obtusely acuminate." The present form, however, is narrow, its margins are almost
linear, and its extremities are rounded less sharply. The striation also differs in the two
cases, being subradiate and evidently moniliform in Amphora speciosa, while in Amphora
oblonga it is continuous. This new species, which is figured in its zonal aspects, was
obtained at the port of Tahiti.

Amphora meneghiniana, n. sp. (Plate XXVII. fig. 16.)

Sublineariter elongata ; apicibus rotundato-truncatis ; medio late subinflata ; striis
transversis conspicuis subradiantibus. Ad portum Thaiti.

This magnificent new species was also first obtained at the port of Tahiti. It is of an

1 Quart. Journ. Microsc. Sci., vol. vi. p. 184, 1858. 2 Diatoms of the Clyde, p. 43, pi. v. fig. 78.

(bot. chall. exp. — part iv. — 1886.) D 3

18 THE VOYAGE OF H.M.S. CHALLENGER.

elongated sublinear form, its centre is slightly inflated, and its extremities are roundly
truncated. The continuous stria? form subradiating bars, and are easily discerned. The
valves are slightly depressed at the centre.

It has been named in honour of the Italian naturalist Professor Giuseppe Meneo-hini,
one of the earliest cultivators of the study of the Diatornaceae.

Amphora scalaris, n. sp. (Plate XXVII. fig. 19.)

Oblonga, truncata, lineis dorsali et ventrali late biarcuatis, centrali medio angulariter
inflexa ; striis continuis raris transverse parallelis. Ad portum Jackson in Australia.

Of this remarkable form, which was collected at Port Jackson, Australia, only a single
valve was observed. In this the dorsal line, like the ventral profile, is singularly biarcuate,
while the raphe is angularly folded at the centre. The widely separated, transverse, equi-
distant, and continuous striae are especially characteristic. The extremities are roundly
truncated.

The specific name has been suggested by the appearance of the transverse striae.

Amphora polyzonata, n. sp. (Plate XXVII. fig. 18.)

Forma lata, medio angulariter constricta ; apicibus late rotundatis, truncatis : valvis
introrsum inflexis ; zona connectiva in plures fascias divisa ; zona et valvis quadrato
ordine punctulatis. Ad portum Thaiti.

In this species from Tahiti, just as in some Naviculce, well-marked constrictions occur.
Thus the valval involution is so strong and acute as to simulate the staurotic dilatation
of a central nodule. A strongly biarcuated line forms a raphe, and the frustule, which
is twice as long as broad, is widely and roundly truncated at the extremities. The con-
necting zone bears several longitudinal bands which, like the valves, have the granules
arranged in a quadrate manner.

Plate XXVII. fig. 8, represents a two-lobed form which is angularly constricted, but
far less deeply than Amphora polyzonata. The raphe is also strongly biarcuate, but the
extremities are round and non-truncated. The length of the frustule is more than twice
its breadth. The connecting zone bears longitudinal bands which, like the valves, are
punctated. It may therefore be regarded as a mere variety of the former, the distinctions
indicated being such as might result from differences in development. It was obtained
at the port of Tahiti.

Amphora decora, n. sp. (Plate XXVII. fig. 14.)

Late inflata, elongata ; apicibus obtusis subproductis ; valvis medio inflexis ; zona
connectenti in fascias divisa, binis granulorum ordinibus et striis alternis insignita. Ad
mare Philippinarum.

This beautiful little Amphora, from the mud of the Philippine Seas, has a widely
inflated profile, and obtuse, slightly prolonged extremities. Each of the two valves,

REPORT ON THE DIATOMACE^. 19

which are very transparent, presents a central inflection which seems to indicate a depres-
sion in the raphe and in the plane of the valve. The connecting zone is distinguished
by bearing linear bands, which are decorated in a special manner, being marked by stria?
alternating with double files of Granules.

The specific name has reference to this zonal sculpturing.

Amphora philippinica, n. sp. (Plate XXVII. fig. 2.)

Valvis ad dorsum arete sinuato-constrictis, apicibus productis acutis ; linea media
recta ad polos vix deorsum inflata ; striis transversis perspicuis punctulatis. In mari
Philippinarum.

An acute constriction also occurs in this bi-lobed species from the Philippine Sea. Its
apices are very prolonged, sharp, and slightly bent, and its transverse striae are well defined,
granulated, and subradiating.

By referring to A. Schmidt's Atlas 1 (Plate xxv. fig. 62) it may be observed that the
present valve resembles the Amphora coarctata of Grunow, of which no description has
been given. The latter, however, possesses capitated apices, and the striae are thin, non-
sranulated, and not subradiatins;.

The specific name has reference to the locality in which it was found.

Amphora thaitiana, n. sp. (Plate XXVII. fig. 15.)

Elliptico-oblonga, polis rotundatis ; valvis cymbiformibus, dorso late convexo, ventre
inflato, apicibus obtusis ; striis transversis validissimis per sulcos longitudinales interruptis.
Ad portum Thaiti.

This type seems to be akin to the form represented in Plate xxviii. fig. 1 5 of the Atlas
of A. Schmidt, which is a novelty if it be not the Amphora egregia of Ehrenberg. The
present form, however, may be distinguished by its rounded extremities — in the form
represented by Schmidt the ends are roundly truncated — and by the ornamentation of the
valves. This consists of continuous costae interrupted by longitudinal lines, while in
Schmidt's frustule, and in Amphora crassa of Gregory,2 which it resembles, the valves are
distinguished by continuous costae, and the dorsal areas by numerous longitudinal lines
separated by rows of spots or small lines.

The distinction between the appearance of the extremities, apart from other structural
differences, might have been explained by supposing that the frustules had attained
different stages of development.

This new form is very abundant in the rich flora of Tahiti, and its specific name is
intended to indicate this fact.

1 Atlas der Diatomaceenkunde, in Verbindung mit der Herren Grundler, Grunow, Janiseh, Weissflog imd
"Witt, herausgegeben von Adolf Schmidt, 1875. See also a reproduction of the above about one-half the
linear dimensions of the original, by C. Henry Kain, Camden, N.J., 1884.

2 Op. cit. p. 52, pi. vi. figs. 94, 94Z>, 94c, and 94tf.

20 THE VOYAGE OF H.M.S. CHALLENGER.

Amphora staurophora, n. sp. (Plate XXVII. fig. 6.)

Valvis cymbiformibus, apicibus acutis ; linea ventrali subrecta, dorsali convexa ;
nodulo medio in staurum ad margineni dilatato. Ad portum Thaiti.

The narrow stauros of this little cymbiform valve extends to the margin. The dorsal
line is convex and presents a slight constriction at its junction with the stauros. The
ventral line is straight, and the extremities are slightly inflexed. The apices of the
valve are acutely rounded. The striation is very delicate and transverse.

The specific name has been derived from its long and well-defined stauros.

Amphora oceanica, n. sp. (Plate XXVII. fig. 20.)

Valvis deorsum subinflatis, introrsum late concavis; apicibus obtusisrotundatis; nodulo
centrali lineam marginalem subattingente ; striis tenuissimis transversis. Prope Sydney.

This new form, from the waters near Sydney, appears at first sight to be a variety of
Amphora obtusa,1 Greg., but to regard it thus would widen the limit of that species to
too great an extent, so that the two forms cannot be included in the same category.

In Amphora oceanica the terminal nodules are not distinct, and the central nodule is
in close proximity to the internal profile of the valve. The striation is singularly delicate
and, as in Amphora obtusa, Greg., transverse.

Cymbella, Ag. Kg.

Two of the frustules described under this genus, and represented on Plate XXVII. figs.
5 and 13, namely, Cymbella criophila from the south of Heard Island, and Cymbella
marina from the neighbourhood of Yedo, Japan, seem, at first sight, to belong to the
genus Amphora. Since, however, the characteristic of that genus is to have a central
marginal nodule, they cannot be included in that group, but must be classed as Cymbella,
although no members belonging to the latter genus have been hitherto recorded as marine.

That an admixture of fresh-water and salt-water Diatoms should occur in marine
gatherings is to be expected, however, when it is borne in mind that frustules of the
former must often be carried into the sea by rivers. Thus Asterionella formosa,2 Hass.,
and Eunotia arcus,3 Ehrenb., which, like all other Eunotice, vegetates at an elevation of
several hundred feet above the level of the sea, have been detected in sea water ; yet it
cannot be doubted that Cymbella marina is a true pelagic form, and it is probable that
Cymbella criophila has the same habitat. That representatives of this genus do live
normally in sea water has indeed been clearly proved by me while working in the

1 Micr. Journ., vol. v. pi. i. fig. 34.

2 Micr. Journ., vol. viii. pi. vii. fig. 8 ; Microscopical Examination of the Water supplied to the Inhabitants
of London, by Mr. Hassall, p. 10; Smith, Synopsis of the British Liatomacese, vol. ii. p. 81.

3 Abhandl. d. k. Akad. d. Wiss. Berlin, 1840, p. 17 ; Rabenhorst, Siissw. Diat., fig. 6; Wigand in Hedwigia,
vol. ii. p. 43, pi. vii. figs. 13 and 14.

REPOKT ON THE DIATOMACE^E. 21

canal of Trail, Dalmatia, for the purpose of studying the flora of the Adriatic. In
this locality a perfect Cymbella was observed, in a preparation fresh from the sea, in
a moving condition. Species have also been recorded in sand obtained by Gwyn Jeffreys,
the well-known conchologist, in the vicinity of the Shetland Islands, although Professor
Dickie, who first studied these gatherings, believed that the cymbelloid frustules were
deposited on the bottom, after having been carried into the sea by rivers.

Cymbella criophila, n. sp. (Plate XXVII. fig. 5.)

Porma marina (?) elongata, sensim ad acutissimos apices declinans ; linea dorsali con-
vexa, ventrali vix convexiuscula ; striis perspicuis punctulatis subradiantibus, ad centrum
lineani mediam bine inde insequaliter non attingentibus. Ad meridiem insulae Heard.

This interesting species, which was obtained to the south of Heard Island, has an
extremely convex dorsal line, the ventral being only slightly convex. The frustule tapers
at each end to a very sharp point. The striae are well marked, and subradiating in
disposition, each being formed by a series of very minute points. Those towards the
middle of the frustule do not meet in the centre, where a smooth area which is notably
larger on the more convex than on the less convex side of the valve is to be found.

This Antarctic Diatom may have been transported by icebergs into the sea, although
the probability that it is truly marine must be borne in mind.

Cymbella marina, n. sp. (Plate XXVII. fig. 13.)

Forma marina, elongata ; linea ventrali recta, dorsali late arcuata ; apicibus rotundato-
acutis ; striis transversis parallelis, lineam mediam non attingentibus. Ad mare Japonicum,
prope Yedo.

This species was obtained near Yedo in the Sea of Japan. The ventral line is straight,
and the apices are rotundately acute. The valval striae are parallel and terminate near
the raphe, a non-striated nodule being left in the centre and being especially prominent on
the side next the dorsal line.

That this form is truly marine cannot be questioned.

Cymbella pelagica, n. sp. (Plate XXVII. fig. 4.)

Mediocris ; dorso turgido, ventre leniter convexo ; linea media subcurvata ; apicibus
obtuso-truncatis, productis ; striis moniliformibus subradiantibus.

The dorsal line, in this form, is greatly arcuate, the ventral being slightly convex.
The apices are obtuse and somewhat prolonged, and the raphe is slightly curved. The
striae are subradiating and moniliform.

This Diatom, like the preceding, must be looked upon as a truly marine form, inas-
much as it has been observed in a moving condition under the microscope, and has been
met with in more than one collection from the sea.

22 THE VOYAGE OF H.M.S. CHALLENGER.

Mastogloia, Thwaites.

This genus was established by Thwaites in 1848, and is thus referred to by the late
Dr L. Eabenhorst, who contributed so much to our knowledge of Cryptogams, in his
Flora Europsea Algarum aquse dulcis et submarinae : —

" Frustula rectangulo-oblonga, aggregata vel solitaria, valvis naviculaceis transverse
costatis, costis plerumque abbreviates, nodulo centrali distincto."

Pritchard, in his History of Infusoria, defines the genus in the following words : —

" Frustules oblong, naviculoid, annulate, in a gelatinous mammillate cushion or frond ;
annuli loculated ; loculi opening by foramina along the line of suture."

The chief difference between these two definitions consists in the fact that, in the
former, the characteristic circumstance that the frustules are annulate and that the
annulus is loculate, has not been referred to. It should, however, be noted that I have
several times observed valves of Mastogloia entirely or partially devoid of such an annulus,
so that this structure must be regarded as one of unessential significance.

Mastogloia thaitiana, n. sp. (Plate XXVI. fig. 11.)

Valvis elliptico-lanceolatis ; apicibus productis, obtusis ; striis decussatis lineam
mediam simplicem attingentibus ; annulo costato totum valvse circuitum ambiente.
Ad oras Thaiti.

This small form was obtained in an extraordinarily rich sounding made in the waters
around Tahiti. The profile is precisely similar to that of Mastogloia apiculata,1 W. Sm.,
but in the latter the striae are transverse, while in the former they are decussate and
oblique to the direction of the raphe.

The specific name of this Diatom has reference to the locality in which it was found.

Mastogloia kerguelenensis, n. sp. (Plate XV. fig. 11.)

Forma elliptica ; apicibus cuneato-rotundatis ; annulo loculato, medio latiori, hinc ad
apices decrescente ; valvse area centralis ordine subquadrato granulata in sex fascias longi-
tudinales dividitur ; nodulus centralis in area lasvi inscribitur. Ad insulam Kerguelen.

This singular and interesting species has an elliptical valve with cuneato-rotundate
extremities. The loculate annulus is wide in the centre, but diminishes regularly and is
very narrow at the ends of the frustule. The striation is transverse, subradiate, and
formed by lines of granules disposed in a subquadrate manner. The central area, circum-
scribed by the loculate annulus, is divided by a raphe and by four symmetrically disposed
longitudinal lines. The central nodule is represented by a smooth areola, which has an
indistinct outline.

The specific name has reference to the locality in which the frustule was found.

1 Op. cit. vol. ii. p. 65, pi. lxii. fig. 387.

KErOItT ON THE DIATOMACEiE. 23

Stauroneis, Ehrenb., Kg.

Althougli some authors, impelled by a desire of reducing the number of generic
distinctions among Diatoms, have united into a single genus not only the Pinnularice and
Naviculm but also the Stauroneides and Stauropterm of Ehrenberg, such a course cannot be
regarded otherwise than as prejudicial to science. It is to be remarked, however, that the
genera Stauroneis and Stauroptera, which do not contain a large number of species, may
be amalgamated, so that, following the example of W. Smith, Pritchard, and others, any
free navicular frustule with a middle line, terminal nodules, and a central stauros or
transversely enlarged nodule, may be called a Stauroneis.

Stauroneis pacifica, n. sp. (Plate XX. fig. 9.)

Lanceolato-oblonga ; apicibus obtusis ; nodulo centrali in aream linearem ad marginem
attingente ; striis moniliformibus tenuissimis radiantibus, lineam mediam non attingentibus.
In oceano Pacifico.

This frustule possesses a navicular lanceolate form, and is provided with obtuse pro-
longed extremities. The stauros extends to both margins. The striae are very delicate
granular, radiating and approximately parallel, they are obliquely disposed, and terminate
before reaching the median line.

The species is named after the ocean in which it was first observed.

Stauroneis thaitiana, n. sp. (Plate XX. fig. 16.)

Parva, elliptico-lanceolata ; apicibus cuneato-rotundatis ; area transversa lineari brevi ;
striis ten uissimis transversis, parallelis. Ad portum Thaiti.

This small and elegant form from the rich collection obtained at the port of Tahiti
possesses an elliptico-lanceolate outline, and is provided with roundly cuneate extremities.
The central nodule expands transversely as a narrow linear area which stops before
reaching the margin. The striae are very fine, tranverse and parallel. The raphe is
enclosed in a smooth narrow linear area.

Stauroneis salina, W. Sm., var. c, nov. (Plate XX. fig. 13.)

Probably this frustule must be regarded as a mere variety of the Stauroneis salina of
W. Smith.1 The following points of distinction between the typical species and the variety
may however be noted: — (1.) The extremities of the latter are considerably more acute
than those of the former; and (2.) the stria?, although erroneously represented coarser in
the present species, are approximately of the same degree of fineness in both.

1 Op. cit. vol. i. p. 60, pL xix. fig. 188.

24 THE VOYAGE OF H.M.S. CHALLENGER.

Since Rabenhorst 1 constituted a variety b for a form with obtuse apices, the present
frustule has been indicated as variety c on account of the acuteness of its extremities.

Stauroneis brebissonii, n. sp. (Plate XV. fig. 4.)

Oblonga, medio late constricta ; apicibus cuneato-obtusis, subproductis ; striis ad
nodulos radiantibus et medio aream laavem linquentibus.

This is one of the most curious forms of the genus Stauroneis. The valve is of oblong
form, being narrower in the middle than at its extremities, and the two longitudinal lines
are slightly sigmoid. The extremities are cuneato-obtuse and exhibit a slight constriction.
The striae are strongly radiating towards the centre and towards the nodules, and they
terminate as in Pinnularia divergens,2 W. Sm. They are absent in the centre of the
valve, so that this area is very wide and is bounded only by the nearest striae. The zonal
side is sub-rectangular.

This species is named in honour of one of the earliest naturalists who prosecuted the
study of Diatoms, Alphonse de Brebisson.

Stauroneis oblonga, Bail. (Plate XX. figs. 7 and 11.)

Different sizes of this species are here represented. On comparing the figures with
Plate xlviii., fig. 16, of A. Schmidt's Atlas, one cannot fail to recognise the justice of the
determination here arrived at, although the definition of the species, which is given by
Pritchard,3 does not at all agree with the present figures, which are of the form called
" linear " by Bailey.

In this species it may often be observed that the striae on one side touch the raphe,
whilst on the other they stop short of it. This, however, cannot be compared with what
takes place in Alloioneis, but is to be attributed to the fact that the raphe is raised and
forms an oblique projection on one-half of the valve.

Stauroneis glacialis, n. sp. (Plate XXVII. fig. 11.)

Magna ; valvis longitudinaliter late convexis, transverse arete convexis ; stauro lineari ;
striis dense transversis parallelis. In mari Antarctico.

This curious navicular frustule was collected in the Antarctic Ocean. It is of large
size, and its valve is very convex in a longitudinal direction but much more so in a
transverse, so that it cannot be entirely brought into focus at the same time. A median
stauros stretches in a linear manner between the raphe and the margins. The striae, which
are very fine, are directed transversely.

1 Eabenhorst, Flora Europsea Algarum Aquse dulcis et submarina?, p. 251. Hero variety b is characterised as
follows : "Forma sub polis leniter constricta, apicibus obtusis."

2 Synopsis of the British Diatomaceae, vol. i. p. 57, pi. xviii. fig. 177.

3 This definition runs as follows : " Linear, with acute, cuneate ends, and oblique punctato-asperate striae ;
stauros abbreviated, dilated outwards. . . . The size and markings of Staurojitera aspera, Ehrenb., but having
its valves oblong, with parallel sides, and acute angular ends." — History of Infusoria, 4th edit. p. 914.

KEPORT ON THE DIATOMACE^E. 25

The specific name has reference to the occurrence of this frustule in the frozen
Antarctic Ocean.

Plate XXVII. fig. 9, probably represents a variety of this species differing from
the typical form in the following respects — (1.) the dimensions are much greater, (2.)
the appearance is more swollen, and (3.) the striae are relatively much finer. The arrange-
ment of these striae and the character of the stauros show its close affinity with Stauroneis
glacialis, n. sp. ; nevertheless, apart from its sublanceolate profile, its true varietal or
specific character must at present remain somewhat doubtful.

Stauroneis pygmea, n. sp. (Plate XXIX. fig. 7.)

Parva, elongata-elliptica ; apicibus rotundatis ; area transversa sublineari marginem
non attingente ; striis moniliformibus subradiantibus. In mari Arafura.

This small form from the Arafura Sea possesses an elliptical outline and rounded ex-
tremities. The raphe is somewhat bent, and the stauros terminates in an irregular manner
at a notable distance from the edge of the valve. The striae are moniliform and radiat-
ing, and near the middle line they form a well-marked double somewhat irregular area.

Pinnularia, Ehrenb.

Nomenclature being nothing more than a means employed to facilitate the work of
the naturalist in the description of the innumerable organic types with which he has to
deal, Ehrenberg wisely divided the genus Navicula, Bory, which was already extremely
rich in species, into two, and included under the generic name of Pinnularia those
naviculoid forms whose valves were ornamented with pinnulse or costae and not by rows
of granules. It is true, indeed, that some refuse to admit this distinction on the ground
that the sculpturing of all naviculoid forms is ultimately reducible to granules, but,
although some species, such as Pinnularia peregrina, Ehrenb., have been attributed to
the genus Navicula because of the recognition of longitudinal lines which are made up
of a dense series of oblong granules, all similar forms are not really granulated, there being
many Pinnularice, which, when observed with the best object-glasses, with the most
accurate manipulation of the light, and by the use of monochromatic illumination, have
exhibited no granular structure.

Pinnularia raeana, n. sp. (Plate XV. fig. 3.)

Lanceolata ; valvis spiraliter inflexis ; linea media sigmoidea in area laevi lanceolata ;
costis circum radiantibus. Ad Zebu, ex insulis Philippinis.

This very singular and elegant species was discovered by Dr Eae in a collection
brought from Zebu in the Philippine Islands. It possesses a very graceful lanceolate
form. The central line or raphe is slightly sigmoid. The radiating continuous costse do

(bot. chall. exp. — part iv. — 1886.) D 4

26 THE VOYAGE OF H.M.S. CHALLENGER.

not reach the raphe, but leave a smooth somewhat lanceolate area around the centre of
the valve.

The specific name is given out of respect to its discoverer.

Pinnularia criophila, n. sp. (Plate XV. fig. 2.)

Anguste rhomboidea, transverse costata ; costis lineam mediam attingentibus ;
apicibus rotundatis ; valvis transverse convexis. Ad mare Antarcticum.

This form was brought from the glaciers of the Antarctic Ocean. The valve, as may
be noted by comparing the figures representing the valval and zonal aspects, is not flat
but transversely convex, so that its form somewhat resembles that of the diatom * shown
in Plate xlvii. fig. 1 of Dr. A Schmidt's Atlas, although it is at once longer and narrower
than the latter, which, moreover, is not transversely convex. It is rhomboidal in general
outline, transversely costate — the costse reaching the median line — and its apices are
rounded.

Pinnularia sp. ? (Plate XV. fig. 1.)

We have here represented a bilobate form in which the stria? are continuous. The
frustule is elegant in general appearance, elongated in shape, and provided with a wide
central contraction. The two lobes into which it is thus divided are elliptical. The
radiately disposed lateral costae are interrupted by two wide longitudinal bands. Between
these and the median raphe there is a long sublinear and elliptical area on which, towards
the middle line, the terminations of the costse may be traced. The central nodule is
oval, its long axis being directed transversely. The fact that navicular bilobate frustules
are apt to present a difference of structure according to the stage of development at wThich
they have arrived renders it difficult to decide on the true taxonomic value of this type,
hence it has been provisionally recorded as an undetermined species of Pinnularia.

Navicula, Bory.

The examination of the navicular forms in the Challenger collections has been the
means of still further augmenting this already extensive genus. In Pritchard's History
of the Infusoria 303 species have been described, in A.. Schmidt's Atlas of the Diato-
macese — a publication which unfortunately was interrupted — more than 200 forms are
figured, while in Habirshaw's Catalogue of the Diatom acete a still greater number
of specific names have been registered. It is true, indeed, that in both of these
lists forms possessing continuous striaj, upon which characteristic Ehrenberg con-

1 "With respect to this frustule Schmidt says : " 1. G. v. Mexico, 2. Spitzbergen, 3. St. George's River,
4. Yokohama, Formen, welche weder mit N. (i.e., Navicula) directa noch mit A. loiuja verbunden werden kbn-
nen. "—Loc. cit., Explanation of Plate xlvii.

REPORT ON THE DIATOMACE.E. 27

stitnted the genus Pinnularia, are included, but apart from this the species already
recorded as belonging to the genus Navicula, properly so called, are so numerous as to
tend to bring about confusion in the science, and to render difficult if not distasteful the
work of one who introduces new types. Yet such a course is inevitable when a minute
examination of marine micro-organisms is made, and it becomes of the greatest importance
to subdivide the larger genera, and above all the genus Navicula, distinguishing in them
groups based on some common and more salient characteristics, in order to make it possible
to remember the different forms. Hence, although recognising the weight of the authority
of Kiitzing and Brebisson, who join together the genera Navicula and Pinnularia because
some of the latter are found to be granulated, and because there is perhaps no specimen
the costules of which, according to these distinguished naturalists, may not be resolved
into points, still, as this assertion has not yet been proved, I am unable, resting on the
authority of "W. Smith and Eabenhorst, to adopt their arrangement. Still less can I
agree with those who join the genus Stauroneis to that of Navicula, although both are
naviculoid, because, in that case, many other genera which are likewise constituted on
navicular forms would have to be abandoned.

One of the most extensive group of the genus Navicula is that in which the valves
are more or less contracted in the middle and which Ehrenberg designated by the name of
Diploneis.1 In this section the species, on account of their number, are the most difficult
to define.

Navicula abnormis (?), n. sp. (Plate XXVIII. fig. 19.)

Panduriformis ; apicibus cuneato-rotundatis ; striis subradiantibus moniliformibus ;
nodulo centrali quadrato ; area centrali medio subconstricta.

This singular frustule is remarkable on account of the different magnitudes of the
two halves of its valves, and should probably be regarded as an anomalous and monstrous
form were it not that several specimens have been found in the same collection. If it
should ultimately prove to be a monstrous frustule it will be especially liable to sterility,
and although a few generations may come and go its' long survival in the struggle for
existence cannot be expected ; it may, however, for the present receive a specific name.
Its shape is panduriform, its apices cuneately rounded, and its moniliform striae are
subradiating. It possesses a quadrate central nodule, and the central area is slightly
constricted at the middle.

Navicula thaitiana, n. sp. (Plate XX. fig. 5.)

Bilobata, arete constricta ; lobis subcordatis ; apicibus cuneato acutis ; striis monili-
formibus radiantibus, medio evanescentibus ; linea media tenui, nodulis parvis. Ad
portum Thaiti.

1 Pritchard, loc. cit., pp. 892 et seq.

28 THE VOYAGE OF H.M.S. CHALLENGER

This elegant naviculoid frustule, found in the interesting collection from the port
of Tahiti, recalls, when viewed in profile, the Navicula proserpinae of Ehrenberg,1
which is identical with the Navicula kutzingii 2 of Grunow. Its two lobes are sub-
cordate and the extremities are cuneato-acute. The nodule, which in Ehrenberg and
Grunow's type is large and quadrate, is here very small, nor does the frustule present such
longitudinal bands as separate the strise of the Navicula proserpince and Navicula
kutzingii.

In Navicula thaitiana the radiating moniliform strise gradually become less visible
towards the centre.

Navicula entomon, Ehrenb., var. thaitiana, nov. (Plate XX. fig. 17.)

We have here represented a frustule which is probably only a variety of Navicula
entomon,3 Ehrenb. Its centre is more contracted than that of the typical species, and its
two lobes present an elegant elliptical profile. Both forms possess lines of slightly sub-
radiating but manifest granules, which differ from those of all other bi-lobed navicular
frustules. The present diatom has a very strong round central nodule, but this charac-
teristic, taken by itself, cannot be regarded as sufficient to constitute a species.

Navicula entomon, Ehrenb., var. (?) (Plate XX. fig. 10.)

This frustule can scarcely be said to be sinuately constricted, but its raphe, which is
slightly bent when viewed in profile, really possesses a double curvature and is flanked
on both sides by a band of granules.

With reference to these peculiarities it is to be remarked that they are probably merely
developmental stages in the life-history of the species. Thus in the younger frustule the
raphe may be straight instead of bi-curved, while in embryonic phases the granules are
not well defined but become more and more marked as age advances.

Navicula grunowii, Rabenh. (?) (Plate XV. fig. 9.)

The present figure represents an elegant Navicula possessing a valve which is almost
linear in profile. On comparison with Plate lxx. of Schmidt's Atlas it is found to coin-
cide almost exactly in its characteristics with Navicula pristiophora, Janisch, and with
Navicula grunowii, Eabenh.4 It may accordingly be regarded as intermediate between

1 = Diploneis proserpince, Ehrenb., Monatsber. d. k. Akad. d. Wiss. Berlin, 1858, p. 14; see Pritcbard,
loc. cit., p. 893.

2 This form is shown in Plate xiii. fig. 24 of Schmidt's Atlas ; see also Verhandl. d. k. k. zool.-bot. Gesellsch.
in Wien, 18G0, p. 532, T. i. fig. 15.

3 Established by Ehrenberg in 1840; = Pinmdaria entomon, Ehrenb. (Verb. T. i. 1, P. 3 et 4 ; Kiitz., Bac.,
p. 100, T. xxviii. fig. 74); and = Diploneis entomon, Ehrenb. Mikrogeol., T. xix. f. 30; see also Pritchard, loc.
cit., p. 893.

4 See Schmidt's Atlas, pi. lxx. figs. 72 and 73; and Rabenhorst's Flora Europaaa Akjarum, sect. 1, p. 203,
Lipsise, 1864.

REPORT ON THE DIATOMACE.E. 29

the forms just named, being less sinuately incurved than the former and somewhat more
so than the latter. All three possess transversely disposed continuous stria) which are
parallel to each other in the middle, but are directed in a radiating manner at the
extremities, and in all the stria? are interrupted by two longitudinal furrows, while their
central extremities are placed in close proximity to the raphe. Such considerations point
to the conclusion that we have here a form which is very probably identical with the
Navicula grunowii of Eabenhorst.

Navicula janischii, n. sp. (Plate XXX. fig. 5.)

Biloba, profunde constricta, lobis cuneato-subcordatis ; structura dense et inordinate
granulata ; plicis a margine radiantibus, hinc evanescentibus. Ad insulas Bermudas.

This frustule, from the neighbourhood of the Bermudas, possesses an exceedingly elegant,
bilobed form. The lobes are subcordate and cuneiform, and the margins are marked by a
row of well-defined dots, while the surface is irregularly granulated and carries radiating
lines which proceed inwards from the periphery but become indistinct and disappear before
reaching the raphe.

It bears a considerable resemblance to a frustule from Samoa which occurs in
a superb preparation by Moller, but in the latter the middle plica is absent and the
granulations of the surface are much less marked. It is also somewhat similar to the
Navicula jamaicensis described by Greville, but in the Grevillean species the points or
granules are disposed regularly instead of being agglomerated without order.

This new species is named in honour of Mr Charles Janisch the Diatomist.

Navicula bullata, Norman, var. carinata, nov. (Plate XXVIII. fig. 7.)

The specimen here represented is evidently closely related to the Navicula bullata1
of Norman. This is indicated by its elliptical form, its obtuse apices, and by the radiately
disposed rows of granules, which are interrupted by two distinct bands upon which a small
number of large rounded dots occur.

The apices of the frustule are clearly hollowed and the valves are longitudin-
ally plicated ; the longitudinal bands do not present any constriction in the centre, and
the granules are much less numerous than those of Navicula bullata.

Navicula bullata, Norman, var. obtusa, nov. (Plate XXVIII. fig. 10.)

This frustule is elliptico-rhomboidal and not simply elliptical like that established by
Norman, nor, and this is a matter of greater importance, are the apices of the former in
any way elongated or sub-produced, but are merely rounded. The significance of this fact
becomes more apparent when it is borne in mind that, among navicular Diatoms in

1 Micr. Joum., 1861, p. 8, pL ii. fig. 7.

30 THE VOYAGE OF H.M.S. CHALLENGER

particular, the shape of the apex is constant in each type. As in the variety last
mentioned there are two smooth longitudinal areas, but these are slightly constricted in
the middle, where they are ornamented by a few prominent rounded dots.
The specific name is taken from the character of the apex.

Navicula bullata, Norman, var. rhomboidea, nov. (Plate XXX. fig. 7.)

This bullate navicular frustule is subrhomboidal in shape, and the prolonged apices of
the valves are rounded. The valve is plicated by several longitudinal furrows, but, although
this characteristic is exceedingly distinct, the form can be regarded only as an interesting
variety of Navicula bullata. As in the previous varieties there are a few well-marked
dots at the centre of the smooth longitudinal areas which are here plano-convex in outline
and situated around the central area of the valves. This specimen was gathered at Zebu.

Navicula brasiliensis, Grun. (Plate XX. figs. 1 and 3.)

We have here represented the typical form of the Navicula brasiliensis of Grunow,1
but the valves cannot be said to be "oval," nor the apices "very acute," as that author,
whose figure agrees entirely with the present one, has declared. In Grunow's frustule,
moreover, the central nodule is "large and subquadrate," but in the form now being
considered the raphe is interrupted by a subrotund areola.

This form was obtained at the port of Tahiti.

Plate XX. fig. 3, represents a very small navicular frustule from the same locality
in which the central areola is round. It accordingly corresponds more exactly with the
frustule described by Grunow.

Navicula mammalis, n. sp. (Plate XX. fig. 2.)

Elliptico-lanceolata ; apicibus mammiformibus ; lineis granulatis radiantibus, ad cen-
trum cessantibus. Ad mare Philippinarum.

This form was obtained from, a sounding made in the neighbourhood of the Philippine
Islands. Its shape is elliptico-lanceolate, and it possesses mammiform apices. The
radiating granular strise extend from the periphery towards the median raphe, which,
however, they do not reach. The central nodule is surrounded by a smooth areola, which
is marked only by two arched shadows.

The name of this species is derived from the character of its extremities.

Navicula subrhomboidea, n. sp. (Plate XX. fig. 4.)

Parva, lanceolato-rhomboidea ; apicibus cuneatis ; striis punctulatis, transversis, lineam
mediam attino-entibus. In mari Pacifico.

1 Verhandl. d. k. li. zool.-bot. Gesellsch. in Wien% 1863, T. xiv. fig. 10,

REPORT ON THE DIATOMACEiE. 31

This species, which was obtained from the bottom of the Pacific Ocean, possesses a
lanceolate-subrhomboidal outline with wedge-shaped extremities. The surface is covered
with dense transverse and subradiating rows of granules that touch the middle line which
is interrupted at the centre.

The specific name has reference to its subrhomboidal outline.

Navicula oxeia, n. sp. (Plate XX. fig. 8.)

Parva, lanceolata ; apicibus acutis ; striis transversis, lineam mediam attingentibus.
Ad mare Japonicum.

This small navicular lanceolate frustule was gathered in the Sea of Japan. Its
extremities are notably acute, and its fine transverse striae touch the median raphe.

These characters correspond with those of the type represented in Schmidt's Atlas,
Plate xlix. fig. 17, which was obtained from Samoa, and of which the accurate deter-
mination was left in doubt. On the other hand, the frustules delineated in figs. 15, 16
and 18 of the same plate evidently belong to a different species, and it is by no means
easy to understand the grounds upon which Schmidt grouped the four forms together.1

The Samoan frustule, as well as that now recorded from the Sea of Japan, may be
designated Navicula oxeia on account of the singularly acute character of the extremities.

Navicula zanzibarica, Grev., var. zebuana, nov. (Plate XXVIII. fig. 8.)

The present specimen does not fully coincide with the Navicula zanzibarica of Greville.
On making an accurate comparison with the figure given by Greville,2 as well as with
that found in Schmidt's Atlas (Plate ii. fig. 3), it may be observed that, while in Greville's
species the striae are divided into two series by a smooth longitudinal band, in the present
frustule an indication of the striae can be recognised between the band and the raphe by
the presence of some minute granules. Moreover, in the latter the two spots, one of
which occurs on each side of the central nodule, are tolerably regular, and show in the
centre a lineal band formed by stronger granules, but in the former the bands are curved
in order to adapt themselves to the form of the granulated side of the valve. The present
frustule may accordingly be regarded as a variety, and named, from its origin, zebuana.

Navicula parallela, n. sp. (Plate XXVIII. fig. 12.)

Valvis lineariter elongatis ; apicibus cuneato-rotundatis ; striis tenuissimis, parallelis,
transversis ; areola laevi ad centrum. In portu Thaiti.

1 In his explanation of the plate Schmidt explains his figures as follows : —

" 15. Pudasjarri, 17. Norrland, N. (i.e., Navicula) bisulcata, Lagerstedt.
16. Sing-Sing, Hudson R., N.firma var. subundulata, Grun.
18. Sodanskylae, nach Grunow Mittelform zwischen N. firma u. N. bisulcata."

2 Trans. Micros. Soc, vol. xiv., 1866, pi. xii. fig. 22.

32 THE VOYAGE OF H.M.S. CHALLENGER.

This very delicate form, from the port of Tahiti, possesses extremely delicate striae.
The valve is linear in form, and its extremities are cuneately rounded and sometimes
slightly protracted. The specific name has reference to the parallel disposition of the striae.

Navicula, sp. (?) (Plate XXVIII. fig. 13.)

We have here represented another naviculoid very finely fluted frustule from the same
collection, and remarkable from the fact that its raphe is bent. This appearance may have
resulted during the final stages of development, when growth in the centre of the valve
may have continued after that at the circumference had ceased, thereby producing the
undulating raphe referred to.

Plate XXVIII. fig. 17, represents a curious lanceolate Navicula, provided with a
broad rounded median region which tapers towards the narrow obtuse extremities. The
central nodule is larcre and round, but the striae are almost invisible in Canada-balsam
preparations. Since all its characteristics could not be determined, its specific value must
at present remain undecided.

In Plate XXVIII. fig. 15, a naviculoid frustule from the collection made at the port of
Tahiti is delineated in its zonal aspect. From this point of view it possesses a biconvex
outline, the raphe, however, is low at the centre of the valves.

The magnificent frustule represented on Plate XXVIII. fig. 9, differs from Navicula
spectabilis, Grev., in the elliptico-rhomboidal form of its valve — a distinction which,
however, cannot be regarded as of sufficient importance to justify the establishment of a
new species for this form.

Navicula decipiens, n. sp. (Plate XXVII. fig. 17.)

Magna, late lanceolata, apicibns rotundatis ; striis transversis subtilissimis ; nodulo
centrali minimo. In mari Philippinarum.

This frustule may at first sight be regarded as belonging to the genus Amphora,
since, like the latter, it presents in the middle region two small noduli. It is to be noted,
however, that the two noduli are not at the same level, while the two peripheral lines
belong to different valves which are closely united by a connecting zone.

The form of the frustule and the very minute transverse striation may also lead the
observer to regard it as being identical with Navicula ostrearia,1 Kg., but the size of the
different specimens that may be observed, and the extreme smallness of the central nodule,
oppose this view. Moreover, the proximity of the two valves argue its distinct specific
character. From the apparent agreement with the genus Amphora the specific name
decipiens has been chosen for this form.

1 Kiitzing, Species Algarum, p. 77.

EEPOKT ON TIIE DIATOMACE^E. 33

Navicula (?) jejuna, A. S. (Plate XX. fig. 12.)

"We have here represented a frustule whose characteristics almost perfectly coincide
with those of the Java form figured in Schmidt's Atlas, Plate xlvi. fig. 7G. In the latter
the extremities, which are somewhat prolonged, are cuneately rotundate, while the ribs are
wide and transversely disposed in the middle but subradiating at the ends. In the present
case the outline is that of an ellipse of very great eccentricity, yet both forms must be
regarded as being specifically identical.

Whether the generic name of Pinnularia or Navicula should be adopted must for the
present remain uncertain.

Navicula (?) jejuna, A. S., var. nov. (Plate XXVIII. fig. 11.)

The present figure represents a variety in which the shape of the valve and the
character of the striation for the most part agree with the typical species. The difference
observable in the central nodule may be explained by the fact that the nodule and raphe
are somewhat more raised than usual, and are accordingly seen in a lateral position. It
may also be distinguished from the typical frustule figured by A. Schmidt when viewed
in profile, being somewhat broader in proportion to its length.

Navicula kerguelenensis, n. sp. (Plate XXVIII. fig. 16.)

Parva, elliptica ; apicibus cuneato-rotundatis ; striis punctulatis subradiantibus ; areola
media subelliptica transversa. Ad insulam Kerguelen.

This small naviculoid frustule was found in a collection of Diatoms from a soundino-
made near the island of Kerguelen. Its form is elliptical, and its extremities are cuneately
rotundate. The striae are fine and subradiating, while a subelliptical transverse areola
occupies the centre. No type hitherto recorded agrees with this form in its leading
characteristics, which has accordingly received a specific name indicative of the locality in
which it was obtained.

Navicula lyra, Ehrenb., var. signata, A. S. (Plate XXX. fig. 13.)

This figure represents an elliptical and elongated naviculoid frustule, in which the form
of the two bands that separate the striated regions indicates an agreement with the Navicula
lyra1 of Ehrenberg. Its greatly elongated and hollowed valves, however, render it necessary
to regard it as a variety of Ehrenberg's typical species. This view has already been
adopted in the case of an identical form by A. Schmidt,2 who has named it Navicula lyra,
Ehrenb., var. signata, A. S. The present frustule was obtained in the neighbourhood of
the Philippine Islands.

1 Ehrenberg, Verb. p. 131, pL i. 1, fig. 9; Gregory, op. cit., p. 13, pi. i. fig. 13; Janisch et Kabenhorst,
Hondur., pi. iii. fig. 7, &c.

2 Schmidt's Atlas, pL ii. fig. 4. i

(bot. chall. exp. — fart iv. — 18S6.) D 5

34 THE VOYAGE OF H.M.S. CHALLENGER.

Navicula mirabilis, n. sp. (Plate XXX. fig. 10.)

Rhomboidea-elliptica ; striis binatis moniliformibus radiantibus ter utrinque inter-
ruptis ; nodulo centrali subrotundo. Ad Zebu, ad insulas Philippinas.

This magnificent elliptical form seems at first sight to be identical with Navicula
sniithii,1 Breb., which, on the whole, resembles it in its general characteristics. In the
latter, however, each half of the valve is divided into two unequal parts by a single longi-
tudinal line, whilst in the present case there are three longitudinal lines on each side of
the raphe and cutting the transverse and subradiating granulated striae. Moreover, the
shape of the present valve is rhomboido-elliptical rather than elliptical, so that it must be
viewed as specifically distinct from Navicula smithii. Navicula mirabilis was obtained
at Zebu near the Philippine Islands.

Navicula cyclophora, n. sp. (Plate XXVIII. fig. 18.)

Minima, elliptico-subinflata ; apicibus obtuso-rotundatis ; striis radiantibus ; cyclo
laterali ad nodulum centralem. In mari Antarctico.

This form was gathered between Kerguelen and Heard Islands. It is remarkable in
having on one side of the central nodule a very small circle, which is not due to any
casual superposition, but is a part of the valve itself. A similar phenomenon occurs in
Cocconeis cyclophora as made known by Grunow.2

In a most interesting collection from Bahia given to me by Mr Weissflog I have
observed another species of Cocconeis with a lateral mark, and two years ago in the Gulf
of Naples a third form was recorded as possessing the same peculiarity. Hence, instead
of being viewed as of specific value, this character may be regarded as being analogous to
the craticular condition of some Navicula. This view is confirmed by a micro-photograph
of a Navicula lyra, Ehrenb., which has been furnished by Mr Weissflog, inasmuch as the
valve, which has its centre marked by small circles, is distinguished by possessing a lateral
hyaline area.

Navicula, sp. (?) (Plate XX fig. 6.)

In this figure there is delineated a small naviculoid frustule which differs slightly from
the preceding in the form, direction, and perhaps also in the fineness of the strise. More-
over, instead of the small lateral circle found in Navicula cyclophora, n. sp., it possesses a
non-striated space on one side of the central nodule. Both forms, however, agree in
having the two sides of the valve unsymmetrical. Its specific value must at present
remain uncertain.

1 See note on this species in Smith's Synopsis of the British Diatomacese, vol. ii. p. 92 ; Pritchard, loc. cit.
p/898.

2 Dr. H. van Heurck, Synopsis des Diatomees de Belgiquc, pi. xxx. figs.^24 and 25.

EEPOET ON THE DIATOMACE.E. 35

Navicula, sp. (?) (Plate XXVII. figs. 3 and 7.)

We have here delineated two other navicular forms, both being represented in their
zonal aspects, probably owing to the fact that they possess very narrow valves. Both
possess a similar striation, but while the profile in fig. 3 is curvilinear, that in fig. 7
is straight. This difference may be explained by saying that in fig. 3 the zonal side is
seen somewhat obliquely.

The perfect resemblance of the striation and the identity of their locality — both being
found to the south of Heard Island — render it very probable that both forms represent
the same species, or at most that one is a variety of the other.

Although the minute transverse striation might point to an identity between these
forms and Navicula liber,1 Sm., such an opinion, on closer examination, cannot be enter-
tained, the striation, which is characteristic of the connecting zones in the present frustules,
differing from the corresponding sculpturing in Smith's species.

Navicula, sp. (?) (Plate XXVII. fig. 10.)

This figure represents a single valve which must belong to another species of Navicula.
Its general outline is plano-convex, and its extremities well defined and bluntly recurved.
The longitudinal axis exceeds two and a half millimetres in length. It is transversely
granulated, save in the narrow areola which surrounds the central nodule. The probability
that this is a sporangial frustule may be borne in mind.

No navicular form hitherto recorded possesses so large a size as the present one, whose
specific designation must, however, remain undecided till further observations have been
made.

Alloioneis, Schumann.

The late Professor Schumann, recognising the advisability of separating from the
extensive genus Navicula all those forms that possess distinctive characters in common,
established the genus Alloioneis for frustules which, while they were symmetrical in their
external profile, had long strise on one side of the raphe and short on the other.

This new genus, which has been accepted, amongst others, by the Swedish micrographer
Professor P. T. Cleve, has been defined by Schumann 2 as follows : — " Navicula, ab altera
linese longitudinalis parte pinnulis longis, ab altera parte pinnulis brevibus prsedita, sectione
transversa rhomboidea."

Alloioneis antillarum, CI. et Grun., var. no v. (Plate XV. fig. 5 ; Plate XX. fig. 1 4 ;
Plate XXVIII. fig. 14.)

Plate XX. fig. 14, represents a diatom which resembles the species established by
Cleve under the name of Alloioneis antillarum, CI. et Grun.3 It may be noted, however,

1 Smith, op. cit, vol. i. p. 4S, pi. xvi. fig. 133. 2 Die Diat. d. H. Tatra, p. 73, 1867.

3 Diatoms from the West Indian Archipelago, hy P. T. Cleve, p. 8, pi. i. fig. 11, 1878, published in Bihang
k. Svenska Vet. Akad. Handl, Band V., No. 8.

36 THE VOYAGE OF H.M.S. CHALLENGEK.

that the present frustule differs from Cleve's species in the fact that the curvature of the
middle line is much less — a distinction which is probably due to the fact that we have to
deal with different stages of development in the two cases, and one which is not of great
importance.

This form occurs pretty frequently in the sea around the Philippine Islands as well
as in the Sea of Japan.

The form represented in Plate XV. fig. 5, is also closely related to the same species. It
possesses, however, a somewhat rhomboidal profile and the extremities are more obtuse,
the raphe being at the same time more excentric. Notwithstanding these differences, it
must be regarded as merely a varietal frustule.

Plate XXVIII. fig. 14, represents another variety of Alloioneis antillarum, CI. et Grun.,
which differs from the latter only in the much greater dimensions of the frustule and in
the proportion between the longitudinal and transverse axes.

Alloioneis japonica, n. sp. (Plate XX. fig. 12, bis.)

Linearis elongata ; apicibus cuneato-rotundatis ; striis laxis, transversis, ad apices
subradiantibus, hue lineam mediam attingentibus, illuc brevioribus et lsevem aream
linquentibus. Ad mare Japonicum.

This curious navicular form was found in mud procured from a sounding of 335
fathoms in the Sea of Japan. It possesses along narrow elliptical form and terminates in
two cuneato-acute extremities. The striae, which on one side touch the raphe, do not reach
it on the other side, and near the central nodule they leave a well-marked smooth area.

Pleurosigma, W. Smith.

The genus Pleurosigma is perhaps the most remarkable among the navicular Diatoms.
The surfaces of the valves are adorned with very minute granules, and this fact, as is well
known, is often made use of in testing the quality of objectives for microscopes.

The genus includes all those navicular forms which possess a granulated surface, and
whose valves, or at least central raphes, are curved in a sigmoid manner. The name
Gyrosigma was that first given by Hassall 1 to such forms, but the word Pleurosigma,
which was introduced by W. Smith, has now been universally adopted, being more in
conformity with the rules of nomenclature, and also because of the accuracy maintained
by that observer in his revision of the species which belong to this group.

The valves of all species of the genus are covered with very minute equal granules
arranged in a quadrate or in a decussate quincuncial order, and upon these character-
istics Smith based his general system.

1 Pritchard, op. cit., pp. 915, 916.

REPORT ON THE DIATOMACEiE. 37

The zonal side of the frustules is so much narrower than the valval as to render it
almost impossihle to meet with an individual that shows anything more than the plane
of the valve. This is generally elevated, and, when seen with the binocular stereoscopic
microscope of Nachet, shows itself in the form of an extremely depressed pyramid with a
very extended basis, so as to cause some doubt in the case of species in which the raphe
divides the extremities of the valve unequally, as in Pleurosigma formosum, W. Sin.,
Pleurosigma decorum, W. Sm., Pleurosigma balticum, W. Sm., &c, whether this
appearance may not be the effect of a difference in obliquity at that point of the extreme
sides of the pyramid.

Pleurosigma elegantissimum, n. sp. (Plate XXVIII. fig. 1.)

Valvis gracillime lanceolatis, rotundato-acutis ; linea centrali excentrica medio iuflexa ;
striis obliquis delicatissimis. Ad mare Japonicum.

This frustule was found in a sounding made near Yedo in the Sea of Japan. The
striae, which are very delicate, are arranged in a decussate manner. The form of the
valve is slightly lanceolate, and the extremities are acute. The profile reminds the observer
of the Pleurosigma delicatulum of W. Smith,1 but it may be at once distinguished from
the latter by the fact that the raphe divides unequally not only the extremities but also
the general surface of the valve. Moreover, towards the central nodule the raphe of
Pleurosigma elegantissimum is depressed so that the sigmoid line appears broken.

From such considerations the specific value of this form cannot be doubted.

Pleurosigma naviculaceum, Breb., var. nov. (Plate XXVIII. fig. 3.)

AVe have here represented a very minute Pleurosigma which was found in the
Arafura Sea, where many other very interesting forms have been obtained. It is remark-
able on account of its lanceolate profile and its greatly sigmoid raphe, which divides the
two halves of the valve unequally. The central nodule is oval and very large, and in
the closest proximity to this on both sides there is a well-defined smooth areola. The
extremities are somewhat lengthened and rounded, and slightly recurved in a sigmoid
manner.

Similar characteristics are found in Pleurosigma naviculaceum,2 Breb., with the
exception of the double indistinct areola near the central nodule.

Pleurosigma speciosum, W. Sm., var. nov. (Plate XXVIII. fig. 2.)

This figure shows another linear Pleurosigma, which was obtained in a sounding
taken at the port of Tahiti. The striae, which are arranged in a decussate manner, are

1 W. Smith, op. cit., vol. i. p. 64, pi. xxi. fig. 202.

2 Br^bisson, Diat. de Cherbourg, p. 17, fig. 7 j = (1.) Gyrosigma transversale, Microg. Dictionary, 1854,
pi. xl figs. 37 and 38; (2.) Pleurosigma transversale, W. Smith, op. cit., vol. ii. p. 96.

38 THE VOYAGE OF H.M.S. CHALLENGER.

not very minute. The central line or raphe is throughout the greater part of its course
straight but curves near the extremities of the valve, so that the surface of the latter
is divided into two equal parts while the raphe runs obliquely across it.

The Pleurosigma speciosum of W. Smith1 agrees with the present form in every
respect except with regard to the profile of the valve, which, instead of being linear, is in
the former linear-lanceolate. Hence the Tahitian frustule has been regarded as a variety
of Smith's species.

Pleurosigma arafurense, n. sp. (Plate XXVIII. fig. 5.)

E minimis, lanceolata, vix sigmoidea ; apicibus obtusis ; linea media excentrica ; nodulo
centrali grandiusculo, quadrato ; striis decussatis tenuissimis. In mari Arafura.

This very minute species, from the Arafura Sea, is provided with excessively delicate
strioe. It. is remarkable on account of the size and subquadrate form of the central nodule
as well as the very slight sigmoid curve of the profile of the valve. The median raphe is
central and the extremities are obtusely rounded.

The specific name has reference to the locality from which it has first been recorded.

Pleurosigma thaitiense, n. sp. (Plate XXVIII. fig. 4.)

Valvis lanceolatis, gradatim ad apices acuto-rotundatos terminantibus, et moderate
curvatis ; raphe centrali ad apices submarginali ; striis decussatis vix perspicuis. In
portu Thaiti.

This very elegant frustule possesses a lanceolate sigmoidal form and acutely rounded
extremities. The flexion is. moderate, and the central raphe becomes decidedly excentric
towards the ends. The granulation is very delicate and decussate.

Pleurosigma smithianum, n. sp. (Plate XXVIII. fig. 6.)

Valvis lanceolatis, acutis, aduncis ; raphe sigmoideo, ad centrum oblique transverso,
dein submarginali. Ad portum Thaiti.

In the collection which contained Pleurosigma thaitiense, n. sp., from the port of
Tahiti, the present frustule was often observed. The disposition of the raphe is particu-
larly noteworthy, remaining submarginal at the two extremities, then closely following
the outline of the valve, and finally abruptly crossing it near the centre in an elegant curve.
The extremities of the valve are acute and somewhat hooked. This form is extremely
characteristic and very easy to determine.

Pleurosigma japonicum, n. sp. (Plate XXIX. fig. 14.)

Minimum, sigmoideo-lanceolatum ; linea media apices cuneato-obtusos imequaliter
dividente ; grandiusculorum punctulorum lineis decussatis. Ad mare Japonicum.

1 W. Smith, op. tit, vol. i. p. 03, pi. xx. fig. 197.

EEPOET ON THE DIATOMACEiE. 39

This form is remarkable on account of its small size and the large granulations that
occur on its surface. In profile it is slightly sigmoid, the raphe beiug much more so, so
that the extremities are unequally divided by the raphe. The longitudinal axis is almost
four times as long as the transverse axis.

This type is not uncommon in the Sea of Japan, and its specific name has reference to
this circumstance.

Toxonidea, Donkin.

Under this head are comprised a few navicular forms -which, whilst they present a
granulation identical with that of Pleurosigma, have the extremities of their valves more
or less curved in the same direction, the middle line being, at the same time, decidedly
arcuate.

Toxonidea challengeriensis, n. sp. (Plate XXVI. fig. 15.)

Forma gracillima lanceolata ; apicibus elliptice inflatis et symmetrice curvatis ; raphe
medio subcentrali ad apices excentrico ; striis densissime decussatis. Ad portum Thaiti.

Amongst the many interesting types which were obtained from the washings of sand
brought from the port of Tahiti, and from no great depth, the singular navicular form
now being considered was observed. It possesses a very long and slender lanceolate form,
having the two extremities strongly bent in the same direction. The raphe remains almost
central throughout the greater part of the valve, but becomes eccentric near the more
attenuated extremities.

Toxonidea challengeriensis, n. sp., var. nov. (Plate XXVI. fig. 14.)

We have here represented a small but interesting form first observed by Cavaliere
Alessandro Garbi of Florence, among some algaa from the Mediterranean. Although
differing sensibly in size, no distinction between the striation of the Tahitian and Mediter-
ranean specimens was observed, so that the latter has been regarded as a variety of
the former. It was first recorded by Garbi in the Tirreno Sea.

Amphiprora, Ehrenb., Kg.

Of all the genera of Diatoms that of Amphiprora is one of the most difficult to under-
stand. Pritchard defines it as follows : — " Frustules free, simple in front view, constricted
at the middle ; valves convex, with a longitudinal wing, and central and terminal nodules ;
strise, when present, transverse." I am of opinion that the frustule is naviculoid, and
that on the axis, and as an emanation from the axis or from the raphe, bilobate wings
appear.

40 THE VOYAGE OF H.M.S. CHALLENGER.

Those species which, like Amphiprora alata, Ehrenb., Kg., are bent spirally, are the
most difficult to understand, being very frequently seen with a cone-shaped profile. Those
that are not spirally bent have the zonal side more or less panduriform owing to a central
constriction resulting from the bilobate form of the wings. The genus may, accordingly,
be conveniently divided into two sections, one comprising the non-spiral and the other the
spirally bent forms.

Amphiprora plicata, Greg., var. japonica, nov. (Plate XXX. fig. 8.)

A small specimen from the Sea of Japan is here represented. It greatly resembles
Amphiprora plicata,1 Greg., having a panduriform profile owing to the existence of a deep
central constriction. In both, too, the extremities are rounded and the connecting zone
is marked by longitudinal lines or folds. In the species determined by Gregory, however,
no noteworthy marks occur at the lines of junction between the wings and the valves,
whereas in the frustule from the Sea of Japan more salient points occur at intervals of
about four strise. These puncta are probably formed by a thickening of the stria?, so that
a kind of longitudinal keel somewhat similar to that which is found in the genus Nitzschia
is the result. Although this characteristic is well marked it cannot be viewed as of specific
importance, so that the Japanese form must be looked upon as a variety of Gregory's species
belonging to the non-spiral section of the genus.

Amphiprora fimbriata, n. sp. (Plate XVII. fig. 15.)

Spiralis, oblongo-elliptica ; apicibus late rotundatis ; ala vel carina undulata ; striis
imperspicuis. In mari Pacifico.

This very elegant oblongo-elliptical form belongs to the spiral section of the genus. The
wings, which have their origin in the subcordate inconspicuously striated valve, are also
bent spirally, and, as if to adapt themselves to the curvature of the frustule, they are
transversely undulated. This remarkable disposition of the alee has not hitherto been
recognised in any other species, and cannot but be regarded as of specific importance. The
apices are broadly rotundate.

Achnanthes, Bory.

Among the pedunculate Diatoms, Bory of St. Vincent established this genus in 1822.
It has been defined by Professor W. Smith2 as follows: — "Frustules geniculate, united
into a filament which is stipitate or attached, valves striated, unsymmetrical, the lower with
a longitudinal and transverse line, and central and terminal nodules, the upper with a
longitudinal line only."

1 Gregory, op. cit., p._33, pi. iv. fig. 57. 2 Op, cit., vol. ii. p. 25.

REPORT ON THE DTATOMACE.E. 41

Achnanthes kerguelenensis, n. sp. (Plate XX. fig. 15.)

Valvis lanceolato-rhombicis ; apicibas obtusis rotundatis; striis moniliforrnibus sub-
radiantibus. Ail insulam KergueleD.

This frustule has a lanceolate-rhomboidal outline, thus agreeing with Achnanthes
rhoniboides * of Ehrenberg, but instead of possessing acute extremities like the latter
it has absolutely rounded and obtuse apices.

It was found in a sounding taken near Kerguelen Island.

Achnanthes parallela, n. sp. (Plate XIX. fig. 11.)

Valvis linearibus ; apicibus rotundatis ; punctulorum liueis subr adiantibus. In portu
Thaiti.

This frustule was found in sand which proved to be rich in various forms of Diatoms
and which was obtained from a great depth at the port of Tahiti. The valves are bounded
peripherally by straight parallel lines, the extremities are rounded, and the lines of puncta
are subradiatinsr.

No hitherto recorded species agrees with this form, whose specific name has reference
to its parallel margins.

TRIBE II.— PSEUDORAPHIDIE^:.
Gephyria, Arnott.

This genus, as well as that of Eupleuria, was established by Professor "Walker Arnott,
but as it not unfrequently happens that the definitions of new genera when first con-
stituted are very imperfect, and as further examination often demonstrates not their
generic but only their specific value, the Eupleurice and Gephyrice were subsequently
united by Professor H. L. Smith, the peculiarities of their structure not being sufficiently
well marked to indicate distinct generic characters.

In both types the frustules, which are united together, consist of dissimilar valves, and
the transversely disposed lines on these are cut by longitudinal lines. The transverse
ribs, moreover, are so arranged that the areas included between them on the two sides of
the central band alternate (Plate XV. fig. 1 0). Both have arcuate valves, but in the case
of Gephyria'2 the lower valve differs from the upper " in having a smooth circular space at
each end," while in Eupleuria3 the inferior valve has "the costse and strise disappearing
below the extremities of the valve," although these structures reach the extremity in the
superior.

1 This form is identical with Achnanthes ventricosa, Kiitz., Baa, p. 76, pi. xx. fig. 7, and with Mono/jramma
ventricosa, Ehrenh., Monatsber. d. k. Akad. d. Wiss. Berlin, 1843.

2 Pritchard, op. cit., p. 809. 3 Quart. Juum. Micr. Sci., vol. vi. p. 89.
(bot. chall. exp. — part iv. — 1886.) D 6

42 THE VOYAGE OF ELMS. CHALLENGER.

According to Arnott, the difference between Gephyria and Eupleuria consists in the
fact that in the latter the zonal side is formed of rings of short septa while the margins
are beaded, whereas in the former the connecting zone is sublamellate and finely striated on
the surface.1 Since, however, it is true that the shape and structure of the two types are
essentially identical, they must be regarded as being generically equivalent, although the
EupleuricB may be considered as a distinct section of the Gephyrice.

Gephyria gigantea, Grev. (Plate XV. fig. 10.)

This figure represents a magnificent Gephyrian valve from the neighbourhood of
Kerguelen Island, where numerous interesting Diatoms have been found. It reminds
the observer of Gephyria gigantea, Grev., which is shown on plate xi. figs. 7 and 8 of the
Transactions of the Microscopical Society of London, vol. xiv., new series, 1866. Eugenie
O'Meara also describes 2 and figures under the name of Gephyria dyerana, in his work on
the diatomaceous gatherings made at Kerguelen Island by the Expedition, a form
of valve which seems to be identical with that recognised by me from the same spot and
very probably from the same collection.

On comparing the three figures above referred to as well as the descriptions given by
O'Meara and Greville, no difference is apparent except the circumstance recorded by the
former, that the valves are flat in the centre and arched at the extremities. This, however,
cannot by itself be regarded as a good foundation on which to establish a new species, so
that the form shown on our Plate XV. fiw. 10 must be regarded as identical with that
described and figured by O'Meara. Moreover, as no distinction is apparent between the
latter and the Grevillean type, the designation adopted by the last-named author should
be selected on account of its priority.

Gephyria, n. sp. (?) (Plate XXV. fig. 19.)

We have here a representation of the superior valve of a very small Gephyrian
frustule, marked by transversely disposed costules which alternate on the two sides. The
two halves of the valve are separated by an undulating central line.

It is indeed true that similar characteristics are seen on the valve of Eupleuria, and
particularly on the superior valve — a circumstance which goes to justify the reunion of
the two genera Gephyria and Eupleuria into one — but since the two valves of the genus
are constantly dissimilar, and since, in the present case, the characteristics of the con-
necting zone are unknown, it may in the meantime be indicated as probably a new species
of Gephyria, although its true specific or non-specific value can be determined only when
its characters are more fully known. This frustule was gathered at the port of Tahiti.

1 Quart. Journ. Micr. Sci, voL viii. p. 20.

2 Journ. Linn. Soc. Lond. (Botany), vol. xv. p. 59, pi. i. fig. 10.

REPORT ON THE DIATOM ACE. K. 43

Glyphodesrnis (Grev.), Cstr.

It is by no means easy to define the two organisms which are represented on Plate
XVIII. figs. 12 and 13, in their valval and zonal aspects. Both are navicular, but the
one, which is elliptical iu form, is very much elongated, while the other has two lateral
protuberances. In the centre they possess a very pronounced nodule, as may be well
observed when they are viewed from their zonal sides. At each extremity the valve is
provided with a small but somewhat elevated cushion, which is extended and flat, and
acts as a support between two adjacent frustules of the series. The valve, which is
clathrate, is ornamented with quadrate granules disposed in parallel rows.

Naviculoid clathrate valves provided with central nodules are to be found in the
Grevillean genus Gl yphodesmis, but in the latter there are no small terminal cushions
such as those present in the frustules represented in our present figures. In Plate x.
fig. 7, of the Quarterly Journal of Microscopical Science, vol. ii., new series, 1862, a series
of frustules of Gl yphodesmis eximia, Grev., is represented. These present in their zonal
aspects three projecting points which bring about a contact between adjoining frustules,
namely, a central one corresponding to the nodule, and two terminal ones. The last are
small smooth terminal cushions, and, although not distinctly marked, the presence of
terminal protuberances in all Glyphodesmides may be regarded as verified, and on
this account should be referred to in the definition of the s;euus. Hence the following
definition may be taken instead of that given by Greville : 1 —

Frustulis seriatim conjunctis, valvis navicularibus, nodulo centrali et pulvinulis
terminalibus, structura clathrata, granulis transverse quadra to ordine dispositis, linea
centrali subconspicua.

G-lyphodesmis murrayana, n. sp. (Plate XVIII. fig. 12.)

Maxima, longe elliptica (novies lougior quani latior) ; structura clathrata, granulis
subquadratis quadrato ordine dispositis; pulvinulis terminalibus kevibus ellipticis ; area
centrali transverse elliptica, cui nodulum rotundum inscribitur. Ad insulas Philippinas.

This frustule is remarkable on account of its size, elegance, and clearness of detail.
It is ornamented with large subquadrate granules which are arranged in parallel longi-
tudinal and transverse lines and give relief to the smooth oval extremities as well as to
the transversely elliptical central area where the large round nodule is conspicuous. When
viewed in the zonal aspect the central nodule is seen to be large, elevated, and somewhat
flattened at the top. It may also be noted that the nodule does not present an opening
or osteole so as to produce a communication between the surrounding medium and the

1 Greville's definition ran as follows : " Frustules united into a filament ; lateral view naviculoid, with a
central nodule, median line, and transverse rows of granules ; structure clathrate, the granules being developed
within square cellules, arranged in parallel series." — Quart. Joum. Micr. Sci., vol. ii., new series, p. 234.

44 THE VOYAGE OF H.M.S. CHALLENGER.

interior of the frustule, as maintained by Kiitzing, but that, on the contrary, the walls
are thick at this place, so that the general opinion that in Diatoms the siliceous nodules
are more or less solid protuberances is here confirmed.

The size of this new species, which has been named in honour of Mr John Murray, the
Editor of the Challenger Eeports, is not less than 200 m.

Glyphodesmis challengerensis, n. sp. (Plate XVIII. fig. 13.)

Magna, medio constricta ; apicibus productis cuneato-rotundatis ; pulvinulis ter-
minalibus lsevibus ellipticis ; nodulo centrali erectiore in area lsevi elliptica ; punctulis
quadrato ordine dispositis ; linea axialis in valva latior. Ad mare Philippinarum.

This frustule, which is represented in its valval and zonal aspects, possesses small
terminal cushions and a central nodule which are decidedly elevated. The central nodule
is not a mere' granule like that of the Naviculacese, but exists in the form of a large
elevated process with a flattened top. This nodule and the terminal cushions are the points
by which the adjoining frustules of a series are united to each other.

The valve presents a median constriction and two lateral protuberances. The
extremities are prolonged but rounded. The granules, which are punctiform, are disposed
in a quadrate manner in parallel rows, and the longitudinal axis is represented by two
rows of dots somewhat further apart and more prominent than those in the adjoining rows.

Glyphodesmis margaritacea, n. sp. (Plate XVIII. fig. 10.)

Valvis ellipticis ; granulis distinctioribus quadrato ordine dispositis et transverse
per plicas distinctis, ad apices et ad centrum areae lseves, in area centrali noduluin sub-
obscurum. Ad Zebu, in mari Philippinarum.

This frustule, which is only represented in its valval aspect, possesses a very elegant
elliptical form. The valve is ornamented with beautiful round granules disposed in
a quadrate manner. It is crossed by transverse folds, and in its centre as well as at
its extremities there are smooth areas which, however, are less clearly defined than in the
two preceding species. In the central area the nodule is not prominent, and greatly
resembles that of Glyphodesmis cximia, Grev., but the nature of the valvular sculpturing
is'quite distinct.

The frustule figured is of large dimensions.

Glyphodesmis_(?) an Dimeregramma (?) sp. (?) (Plate XIX. fig. 10.)

This figure represents a series of small frustules seen in the zonal aspect, but, since the
genera Glyphodesmis and Dimeregramma can be distinguished only when viewed in their
valval aspects — the essential distinction being found in the presence of a central nodule

REPORT ON TlIK DIATOMACE.E. 45

on the valve of Glyphodesmis — its generic determination must remain uncertain. This
being so, the specific characters, which can Le better divined than decided with certainty,
may at present be left out of consideration.

Plagiogramma, Grev.

Greville1 instituted this genus in 1859, and defined it as follows: — Frustules quad-
rangular, direct, two or more united into a filament ; valves linear or elliptical ; striae
moniliform ; vittae two or more, pervious, parallel with the striae.

Plagiogramma thaitiense, n. sp. (Plate XIX. fig. 4.)

Frustulis rectangularibus subinflatis ; vittis duabus centralibus perviis ; striis trans-
versis tenuioribus ad extremum usque. Ad portum Thaiti.

This Diatom was obtained at the port of Tahiti. It recalls the Plagiogramma
jamaicense" of Greville in the fact that the striae, which appear on the margins of the
connecting zone, are continued to the angles. In the type now under examination,
however, whde the striae extend absolutely as far as the extremities, they are much more
delicate than in the Grevillean species, while the margins of the frustule, when viewed in
the zonal aspect, are somewhat protuberant in the centre, whereas in Plagiogramma
jamaicense they are straight if not even slightly concave.

Although each of these points of distinction when taken by itself may be of small
value, when they are considered collectively they must be regarded as sufficient for the
establishing of a new species.

Plagiogramma margaritaceum, n. sp. (Plate XIX. fig. 13.)

Valvis late ellipticis, convexiusculis ; granulis distinctioribus quadrato ordine dispositis ;
granuli inter medias costas vacant, ad apices evanescunt. In portu Thaiti.

This frustule, which is also from the port of Tahiti, possesses an elliptical outline, and,
as indicated by the curves of the transverse lines which separate the rows of granules, is
markedly convex. The striation is continued to the extreme apices but the striae are not
moniliform — a fact which, however, cannot be regarded as of essential importance. On
this ground Plagiogramma tessellatum, to which Gregory affixed a note of interrogation,
must also be admitted to be a true Plagiogramma.

1 Micr. Journ., vol. vii. p. 20S. Greville here divided the genus into three sections, according to the
character of the vittse, viz., Sect. I. Vittse two, centrical; Sect. II. Vittae two, centrical, and one at each end
of the valve ; Sect. III. Number of vittae between the two centrical ones and the ends of the valve indefinite.

2 Micr. Juurn., vol. vii. p. 208, pi. x. fig. 3.

46 THE VOYAGE OF H.M.S. CHALLENGER.

Dimeregramma, Ralfs.

Among filamentous Diatoms the genus Dimeregramma is closely allied to that of
Denticula, the only distinction consisting in the valves of the former always possessing a
smooth line which marks the longitudinal axis, while those of the latter do not exhibit
such a structure.1 The majority of the known species which are now ascribed to
Dimeregramma, and which were discovered by Gregory before this genus was instituted,
were accordingly classed by him as belonging to the group Denticula, the forms recorded
by that observer, however, being marine, while the others were all freshwater species.

Dimeregramma nanum, Greg., var. thaiticnsis, nov. (Plate XIX. fig. 5.)

Among the Dimcregrammata collected by H.M.S. Challenger, the only noteworthy
specimens are those here figured. These greatly resemble the Dimeregramma nanum of
Gregory,2 differing from it only in the form of the extremities, which in the present case
are more dilated and round. Such a distinction, however, cannot be viewed as possessing
more than a varietal importance.

Terebraria, Greville.

This genus was established by Greville in 1864, when he made us acquainted with the
Terebraria barbadensis, from the famous Barbados deposit in Cambridge. The definition
of the genus as given by him3 ran as follows : — "Frustules in front view quadrangular,
binately conjoined, with transverse rows of conspicuous pseudopores and a longitudinal
serrated structure. Valve elliptical, with transverse rows of similar pseudopores."

To the species made known by Greville, O'Meara added his Terebraria kerguelensis,
which was obtained from a gathering made by the Expedition in the vicinity of
Kerguelen, and which has been figured in the Journal of the Linnean Society (Botany)
vol. xv. pi. 1, fig. 4. I am, however, in doubt with regard to the correct determination of
this frustule. By referring to the description of the genus, or to the figure given by
Greville, the indented line of suture is found to be so evident as to constitute one of the
principal characteristics, whereas in the figure given by O'Meara no such lines of suture are
to be perceived ; but, on the contrary, on the zonal side the granules alternate, so that in
the line of division between the two adjacent valves there is an appearance resembling
that presented by an indented suturation. It would seem, therefore, from the fact that
O'Meara, in his definition of Terebraria kerguelensis, omitted to notice this circumstance,
that he took the central line of the zone in Terebraria barbadensis, Grev., not for what it

1 Pritchard, op. cit., p. 790.]

2 = Denticula nana, Greg., Diatoms of the Clyde, p. 23, pi. ii. fig. 34.

3 Trans. Micros. Soc. Lond., new series, vol. xii. p. 8., figs. 12 and 13.

REPORT ON THE DIATOMACEJG. 47

really is, but for a profile of alternating granules belonging to two different valves. More-
over, judging from the figure of the zonal side of Terebraria herguelensis, O'Me., we cannot
recognise in this type the condition of a frustule "binatcly conjoined." Nor can it be
admitted that the so-called Terebraria herguelensis, O'Me., should be classed with Tere-
braria barbatle?isis, Grev., and it is not easy to decide to what other genus the form in
question should be ascribed, so that, though with reservation, the name given by O'Meara
may for the present be retained.

Terebraria (?) sp. (?) (Plate XXV. fig. 2.)

This figure represents an organism obtained to the south of Heard Island, which
shows seven lines having the appearance of indented lines of suture similar to those which
have been observed on the zonal side of Terebraria herguelensis, O'Me. It may be
observed that the undulating lines are disposed around corresponding straight lines, and
that at each of these regions a solution of continuity takes place as a result of the fitting
together of the two valves which seem to alternate with one another as in O'Meara's
species. In this figure, too, it is to be noted that the granules or costules which are
seen in profile differ greatly in number in the different valves. Yet it cannot be doubted,
as some have done, that the striation or the granulation of a frustule, presents within
certain limits good characteristic features for diagnostic purposes, although the necessity
for exercising caution in generalising on such appearances is at once obvious. This
has already been referred to by me in the Transactions of the Pontifical Academy,1
where the fact that other Diatoms, such as Eunotia formica,2 Ehrenb., have the striae
irregularly distributed not only on different frustules, but also on the same valve, has
been recorded. Apart, however, from the circumstance of the magnitude of the striaa
based upon the comparison of the undulated lines of division with those that are found in
Terebraria herguelensis, O'Me., this new type may for the present be named Terebraria,
although no specific designation can be given, as the characteristics of the valves have
not yet been adequately determined.

Plate XXV. fig. 1 also represents a series of frustules which, although probably not
members of the genus Terebraria, but of the genus Fragilaria, yet exhibit, when seen
in the zonal aspect, some relation to the Terebraria herguelensis of O'Meara, although they
cannot be regarded as identical with that species. The frustules have evidently been ob-
tained at different stages of development. The uppermost in the figure has recently under-
gone fissiparous division, but the resulting cellules have not yet become detached, although
the margins of the adjoining valves are now clearly indicated by the two transverse
central rows of granules. This phase would seem to be a repetition of the condition re-

1 Atti Aceacl. Pontif. d. nuov. Lincei, vol. xxxi. sect. 6.

2 Ehrenberg, MikrogeoL, pL iii. iv. fig. 18.

48 THE VOYAGE OF II. M.S. CHALLENGES.

cognisable in the frustule figured by O'Meara, but the alternation of the granules of the
two new valves which cause the latter to possess an apparent indented line of suturation is
not here visible.

The next three frustules of the series have gone on developing after fission, and now
present two symmetrically convex valves, which are again ready to divide. The fifth,
which is probably younger than the second, third, or fourth, as its cingulum is less
developed, is obviously advancing towards the same mature state. The sixth, on the
other hand, is evidently imperfect, having originated from that immediately preceding it.
Its abortive condition presents a phenomenon similar to what not unfrequently occurs in
the almond and other fruits or seeds, where one part may develop at the expense of the
others. The remaining frustules, again, show somewhat advanced phases, recalling the
appearances presented by the second, third, and fourth.

From the developmental stages here seen, great interest attaches to this figure, but it
is to be carefully observed that the granules, instead of alternating, are arranged in con-
tinuous and parallel series along the lines of division.

At present even the generic determination of these Diatoms must remain uncertain
until the form of the valve from its valval aspect has been fully studied.

Rhaplioneis, Ehrenb.

Authors are not agreed as to the definition of this genus. Pritchard,1 among others,
limits the name to those navicular forms in which the centre is unprovided with noduli
and transversely directed lines, whilst Eabenhorst2 admits those forms that possess costules.
Although it is desirable to limit the number of species in the various genera as far as
possible in order to facilitate study, it is better in the present case to follow the opinion
of Rabenhorst, especially as the species are not very numerous. The following is the
definition of the genus as given by that author : — " Frustula libera vel stipitata, quadrangula
simplicia navicularia, non conjuncta, fronte late lanceolata vel ovata, nodulis nullis, striis
vel costis transversis, ubi adsunt, vitta vel area longitudinali interruptis."

Rhaplioneis mammalis, n. sp. (Plate XXVI. fig. 3.)

Elliptico lanceolata ; apicibus productis mammiformibus ; costis raris radiantibus ;
vitta sublineari medio interrupta. Ad portum Thaiti.

This beautiful frustule was found in an interesting collection from the port of Tahiti.
Its form is elliptico-lanceolate, and its extremities are mammiform, while its transverse
and longitudinal axes are in the proportion of two to three. Its widely disposed costules
are radiating and interrupted in the middle by a narrow lineal smooth area.

This species has been named from the appearance presented by its extremities.

1 Op. at., p. 701. 2 Flora Europaea Algarum Aquiv dulcis et subruarinse, sect. 1, p. 125.

KErORT OX THE IUAToMACE.E. 49

Rhaphoneis elliptica, n. sp. (Plate XXVI. fig. 13.)

Parva, elliptica, duplo longior quam latior ; costis crebriusculis radiantibus, area
lineari lauceolata medio interruptis. Ad mare Japonicum.

This small elliptical Diatom is twice as long as broad. Its costules, which radiate like
those of Rhaphoneis mammalis, n. sp., are more closely arranged, and more prominent than
in the latter frustule.

The smallness of its size, and the absence of granulated striae, distinguish it from the
Rhaphoneis fasciolata of Ehrenberg,1 but it bears some resemblance to the lower valve
of a Cocconeis. Since, however, no upper valve belonging to this genus was observed
in the rich collection, and since the smooth central space characteristic of the genus
Rhaphoneis is present, it has from its shape been named Rhaphoneis elliptica. It was
gathered in the Sea of Japan.

Rhaphoneis japonica, n. sp. (Plate XIX. fig. 12.)

Mediocris, elliptica ; apicibus cuneato rotundatis ; distinctiorum granulorum lineis
subradiantibus ; granulis circum regulariter stipatis, medio irregulariter rarioribus. In
mari Japonico.

The determination of this species has been a work of no small difficulty. In it no
trace of a smooth area or of a median longitudinal line is to be found, so that, although
apparently not a Rhaphoneis, no other genus exists to which it could be more readily
ascribed. On consulting the various writers on Diatoms, and comparing their figures
with the present type, an oblong, elliptical form — different, indeed, from the form now in
question, but evidently belonging to the same genus, — which resembles the Rhaphoneis
australis of H. L. Smith, is represented at fig. 54 given in a paper on the "Diatoms
collected during the Expedition of the Vega," by P. T. Cleve, under the name of
Trachisphenia australis.

In order to understand how a form that presents no trace of central division can be
called a Rhaphoneis, reference may be made to the Synopsis of the families and genera of
Diatoms by H. L. Smith, as given at the end of Dr Van Heurck's work on the Microscope 2
(third edition), where, in the definition of the genus, after having said, " ayant une ligne
mediane ou un espace blanc," he parenthetically adds, "souvent obscur ou manquant," so
that amongst the characteristics of a genus the essential should always be distinguished
from the less essential features.

Since, however, this frustule from the Sea of Japan cannot be confounded with
Rhaphoneis australis, H. L. Sm., as it differs both in form and in the character of its
granulation, it must be regarded as a new species, and has been named from the locality
in which it was found.

1 Mikrogeol., T. xxxv. A. 22, fig. 16.

2 Le Microscope, sa Construction, son Mainement, et son Application a l'anatomie Vegetale et aux Diatonie'es
(troisieme Edition), Bruxelles, 1878.

(bot, chall. ezp. — part iv. — 1886.) D 7

50 THE VOYAGE OF H.M.S. CHALLENGER.

Asterionella, Hassall.

The genus Asterionella, instituted by Hassall in 1855, is characterised by the following
features : (l) Its frustules are grouped in a stellate manner, and (2) when viewed in profile
they are constantly inflated at the base, and sometimes also at their superior extremities.

Up to the present time all the species included in this genus have been freshwater
forms, with the exception of the Asterionella bleakeleyii of W. Smith and the Asterionella
frauenfeldii of Grunow, the latter of which was afterwards ascribed to the genus
Thalassiothrix. Among the collections procured by the Challenger, however, two different
types, which were gathered on the surface of the Antarctic Ocean near the ice-barrier of the
South Pole, must be ascribed to the Aster ionelloe, but it is to be borne in mind that, as
numerous frustules of Eunotia arcus, "W. Sm.,1 were found in the same collection, it is
unquestionable that in many cases terrestrial Diatoms which vegetate on the glaciers of
the Antarctic, and which represent the last traces of the icebergs that have been melted
by coming into contact with warmer water, float upon the surface of the waters in these
localities. Moreover, since it has been clearly proved that the other species of the genus
Eunotia do not vegetate except at an elevation of several hundred feet above the level
of the sea, it is manifest that it is impossible here to decide whether the Asterionella,
which have now to be recorded, should be regarded as terrestrial forms, or whether they
must be viewed as being strictly marine.

Although these new forms cannot be identified with Asterionella bleakeleyii of W.
Smith, which was till recently the only marine species that had been observed, it may here
be remarked that I have observed other marine Asterionella; at the island of Lesina in
Dalmatia — a region which is exceedingly favourable for research in this department of
marine biology. In this locality, by means of small nets, there was found on the surface of
the sea, and floating among many other organic forms, an Asterionella which could be
readily distinguished by the linear form of the frustules ; these were at the same time
provided with a swelling at the inferior extremity, and were radiately disposed so as to
present a stellate appearance. This form, too, was remarkable from the fact that the series
did not form a solitary star with only a few radii, but presented, on the contrary, a long
chain of very numerous frustules, which were disposed in such a manner as to form several
spiral turns, so that the name Asterionella spiralis, which has been applied to the species,
is exceedingly characteristic.

Asterionella glacialis, n. sp. (Plate XIV. fig. 1.)

Frustula a basi inflata exiliter linearia ; valvis inferius rotundato-inflatis, et binis
lineolis notatis. In mari Antarctico.

1 Synopsis of the British Diatomacese, vol. i. p. 15, pL ii. fig. 15. This Diatom is equivalent to Navicula
arcus, Ehrenb., Inf., pi. xxi. fig. 10, Cymbella (?) arcus, Hass., Alg., C. 6, and to Ceratoneis arcus, Kg., Baa, pi. vi.
fig. 10.

REPORT OX THE DIATOMACE.E 51

The frustules of this type, which was collected in the Antarctic Ocean, show a charac-
teristic peculiarity at their inferior extremities, and one which is not found in any frustule
of Asterionella bleakeleyii, W. Sm., namely, the presence, on the zonal side, of two small
symmetrical lines, which run in the direction of their long axes. These probably represent
the incapsuling of the two valves of the frustule, and if so, ought to have been visible
in Asterionella bleakeleyii, though contracted as a result of incineration.

Nor can Asterionella glacialis be confounded with the freshwater Asterionella
Jbrmosa of Hassall,1 as the latter has its superior extremity somewhat swollen, although
to a less extent than its inferior extremity, while the former has a swelling only at the
inferior extremity of the valve, and on the zonal side of the frustule, so that it must be
regarded as a new species.

Asterionella gracillima (Hantzsch.), Heib. (Plate XXV. fig. G.)

We have here represented a form which cannot be confounded with ^the Asterionella
bleakeleyii of W. Smith, as, instead of being linear in the parts succeeding the dilated base, it
is sensibly swollen at its superior extremity. The specimen, however, which unfortunately
has suffered somewhat from the bleaching process, agrees better with Asterionella gracillima"
(Hantzsch.), Heib., in its morphological characteristics ; but, instead of being found in fresh
water, like the latter, it occurred on the surface of the sea. As it may have been borne
to that locality by glaciers and icebergs, it may, for the present, be viewed as identical
with that species.

Synedra, Ehrenb.

This genus is chiefly characterised by having its small frustules joined by means of a
cushion, or sometimes of a more or less developed peduncle to seaweeds or other objects.
The attachment thus effected is of a feeble kind, so that it is a very common occurrence
to find living specimens in a free state, and in the examination of geological diatomaceous
deposits, or of frustules that have been treated with acids, it is impossible to determine
whether any given form existed in the free or adherent condition. But apart from its
attached or non-attached habit, the genus cannot, on account of its structural peculiarities,
be confounded with any other. Its closest affinity is to be found in the Bacillarice, from
which, however, it is readily distinguishable by the presence in the latter of a characteristic
punctated keel. Nor can it be confounded with the Asterionella, Hass. , or with members
of the genus Thalassiothrix (Grun.), Cstr., inasmuch as it does not possess large broad or
dilated extremities, but is provided either with a small cushion, which unites the frustules
in a star-like manner, or with small triangular pads placed laterally at the two extremities,

1 Miar. Journ., vol. viii. pi. vii. fig. 8; Heiberg, Conspec, pi. vi. fig. 20.

2 Heiberg, Conspec, p. 68, pi. vi. fig. 19 ; -Diatoma gracillimwn, Hantzsch. in Rabenh. Alg., Xo. 1104, c.
icone, Krypt. Flor. von Sachsen, p. 32.

52 THE VOYAGE OF H.M.S. CHALLENGER

and serving as means of union into zig-zag chains. In acid preparations the differentia-
tion of the Synedrce from species belonging to the genus Thalassiothrix is a matter of
greater difficulty, but in general this may be done by remembering that the latter are
distinguished by lateral rows of much raised thorn-like points so that each of the outer
extremities terminates in two projecting apiculi.

Synedra capitulata, n. sp. (Plate XXV. fig. 13.)

Lineari-lanceolata, apicibus dilatatis, rotundatis, valvis medio laevibus, ad marginem
ordine brevissimarum linearum circumdatis. In mare Philippinarum.

This very elegant Diatom, which was obtained in a sounding made in the neigh-
bourhood of the Philippine Islands, possesses a linear-lanceolate outline. Its extremities
are dilated and rounded, and the valve is surrounded by thick short lines, whilst the
remainder of the surface is smooth.

The specific name has reference to the capitate appearance of the extremities.

Synedra philippinarum, n. sp. (Plate XXV. fig. 15.)

Gracillima ; valvis angustissime lanceolatis ; apicibus linearibus obtusis ; striis trans-
versis continuis. In mari Philippinarum.

This very slender frustule, also from the sea in the vicinity of the Philippine Islands,
presents a slightly arched appearance. The extremities are linear and obtuse, and the
strise are transverse and continuous. The fact that this species was represented by many
specimens is opposed to the idea that it was originally a freshwater form which had
been floated into the ocean, although it may be noted that most slender and crooked forms
of the genus are from fresh water.

Synedra fimbriata, n. sp. (Plate XXV. fig. 14.)

Ehomboideo-lanceolata ; apicibus rotundatis ; striis transversis, ad marginem evidenti-
oribus et vittam linearem constituentibus. Ad oras Philippinarum.

The valve in this type is sub-rhomboideo-lanceolate. The centre is convex, the
convexity lessening towards the rounded extremities. The surface of the valve is orna-
mented with continuous transverse strise, which become salient near the margins, and
thus form a more prominent outline — a circumstance which has suggested its specific
name.

Synedra lanceolata, n. sp. (Plate XXV. fig. 20.)

Parva, lanceolata ; apicibus cuneato-acutis ; striis perspicuis transversis, linea lsevi
medio interrupts ; striae 1400 in millirnetro numerantur. Ad insulam Heard.

This frustule was collected near Heard Island in the Antarctic Ocean. It may be

REPORT ON THE DIATOMACEiE. 53

distinguished from a Navicula by the absence of the characteristic nodule of the latter,
but its stria3 are interrupted by a simple longitudinal line.1 This feature, taken in con-
junction with the cuneato-acute appearance of its extremities, constitutes a very distinct
specific type.

Synedra lanceolata, n. sp., var. thaitiensis, nov. (Plate XXV. fig. 18.)

This frustule, which is similar to the preceding, was found in a gathering from the port
of Tahiti. It may, however, be distinguished from Synedra lanceolata by its somewhat
finer stria?, which are interrupted by a fine line, but it cannot be regarded as more than
a variety of that type.

Synedra atlantica, n. sp. (Plate XXV. fig. 16.)

Lanceolata, elongata ; apicibus rotundatis ; striis evidentioribus transversis continuis,
binis lineis submaroinalibus unilateraliter sectis. In Oceano Atlantico meridionali.

This Diatom was obtained from a soundino; made in the middle of the South Atlantic.
Its valves are lanceolate, and are provided with rounded apices. The striation is well
marked transversely directed and continuous, but is cut on one side by two submarginal
lines, the signification of which is not manifest, although they must be regarded as
characteristic of the species.

Thalassiothrix (Grun.), Cstr.

Among the surface collections made in the Bay of Yedo and in the waters around
Hons-Kongr there has been found in great abundance a Diatom of a bacillar form, which
is either long and straight, or short and slightly curved in appearance, and which
generally occurs in small groups, the frustules being arranged either in a radiating
manner or forming a zig-zag series. The individual Diatoms are bordered by lines of very
elevated granules, so that the outer extremities are crowned by two apiculi. Similar frus-
tules have also been observed by me in surface collections made in the Adriatic off the
coast of Dalmatia, and on the Italian coast at Rimini and Fano. In specimens procured at
the last-named localities, and which had not been subjected to any incinerating process, the
frustules were found in groups of not more than eight, and were simple or double and
united in a radiating manner, the inferior extremities being sunk in a small transparent
cushion of the form of an armilla, which is destroyed by the action of heat. Not un-
frequently a few similar frustules were seen in zig-zag position, and it was of interest to
observe that when a normal radiating group of simple frustules was undergoing fission,
and so becoming double, the cushion broke by the separation of two contiguous frustules.

1 The double central line has been figured by mistake. '

54 THE VOYAGE OF H.M.S. CHALLENGER.

These remained united, however, by a small triangular cushion at the superior extremity
of the two recently formed valves.

This same Diatom was also found among the collections made at Tilanshang, one of
the Nicobar Islands, by the Austrian scientific frigate " No vara," and was named by A.
Grunow Asterionella frauenfeldii ; ' but, with this exception, no other naturalist has
recorded its occurrence, although it is found in almost all seas.

Although at first sight this curious Diatom recalls the genus Asterionella, it is to be
remarked that in the latter the inferior extremities of the frustules are dilated and are in
contact with one another instead of adhering by means of small cushions.

In the Synopsis des Diatomees de Belgique, by van Heurck and Grunow, which
has recently appeared, and in the section of Pseudoraphidiese (plate xxxvii.) the forms
which Grunow named Asterionella frauenfeldii are referred to the new genus
Thalassiothrix, to which there is also ascribed the very singular Synedra thalassiothrix
of Cleve. But, after a careful comparison of these two forms, they cannot be regarded as
congeners, the one — Asterionella frauenfeldii — having its frustules united by small
membranous cushions, while the other — Synedra thalassiothrix — is solitary or disposed in
parallel bands.

With regard to the genus Thalassiothrix, Grunow says that it embraces some species
resembling Asterionella, but having on the margins spines or elevated points between
which one sees a short marginal striation. The structure here described is indeed pre-
cisely that of the Synedra thalassiothrix of Cleve ; but it cannot in any way be compared
to that of Asterionella frauenfeldii, the bacillar Diatom in which there is only to be
discovered a line of very salient marginal granules, which render the two small apices
indistinct. Hence it would be better, at least for the present, to exclude Cleve's Synedra
thalassiothrix from the genus Thalassiothrix altogether.

But the Thalassiothrix (Asterionella) frauenfeldii of Grunow may, on the other hand,
be retained in that genus, although the generic definition must be somewhat modified. It
may be amended as follows : — Frustulis linearibus radiatim per pulvinulum gelineum
armilliforme unitis, bino erectiorum punctulorum ordine instructis ; post frustulorum
deduplicatione armilla disrumpitur, et frustula in seriem alternam per isthmum triangu-
larem cohalescunt.

Thalassiothrix frauenfeldii (Grim.), Cstr. (Plate XIV. figs. 7 and 8.)

Frustulis radiatim conjunctis, hinc in catenas alternas dispositis ; a latere zonali

anguste linearis, valvis arctissime linearibus, utroque margine serie unica punctulorum

(spinularum ?) evidentium ornatis. Ad insulas Nicobar, ad Java, in mari Sinensi et

Japonico, in Adriatico, in Tyreno et alibi.

1 Verhandl. d. L Jc. zool.-hot. Gesellsch. in Wien, 1S63, p. 140, pi. xiv, — Diatoinaceen der JNovara-Expedition
—fig. 18.

REPORT ON THE DIATOMACE.E. 55

In fig. S the frustules are simple and singularly slender, and the borders are provided with
fine but very salient points — spines or thorns — which at the superior extremity assume
the appearance of two apiculi. The frustules are never in groups of more than eight, and
the angles included between any adjoining pair are approximately equal, with the excep-
tion of that which corresponds to the opening of the armilla, which is always somewhat
larger.

This also applies to the frustules shown in fig. 7, which differ from the last only in
bein^ double, an appearance which has been caused by the temnogenetic changes which
have tak^n place. Sometimes promiscuous groups of double or simple frustules are to be
met with, but much more frequently, after division has occurred, the armilla, in which
the inferior extremities of the frustules are planted, breaks up by separation of the double
frustules, a result which is probably brought about by the development and swelling of
a triangular isthmus, which unites the superior extremities. In the course of this curious
change in the position of the locus of union the frustules are frequently found to be
simply united in short zig-zag series.

Thalassiothrix curvata, n. sp. (Plate XXIV. fig. G. )

Frustulis linearibus, subcurvatis, crasiusculis, radiatim vel alterne pulvinulo conjunctis;
punctulis frustulorum marginalibus duplo quam in Thalassiothrice frauenfeldii rarioribus.
In mari Japonico.

In this figure no traces of the small connecting cushions are to be found, as these have
been removed during the process of incineration. The frustules are bacillar, and flanked
on each side by a line of small points resembling those which are found in the above
described Grunowian species, but they are slightly curved, much shorter and wider, and
the punctations are only half as abundant as in Thalassiothrix frauenfeldii.

From such considerations the specific value of this Diatom cannot be doubted, especially
as it occurs in sreat abundance in some collections.

Fragilaria (Lyngb.), Agardh.

It has already been indicated that the enormous glaciers which cover the mountains
of polar lands, and which ultimately reach the ocean and form icebergs, are the means of
transporting freshwater Diatoms into the sea, where these organisms float freely when the
ice has been melted. Frustules of Asterionella formosa, Hass., Ceratoneis arcus, Kg., and
various species of the genus Eunotia, wdiich is not only peculiar to fresh water, but which
only vegetates at an elevation of several hundred feet above the level of the sea, have
been thus transported,1 and the same phenomenon has taken place in the case of the genus
Fragilaria, of which some new species must now be recorded.

1 Confer., p. 50.

56 THE VOYAGE OF H.M.S. CHALLENGER.

Fragilaria linearis, n. sp. (Plate XIX. fig. 9.)

E maximis ; valvis linearibus, transverse striatis, apicibus rotundatis. la mari
Antarctico.

This Diatom was collected in the Antarctic Ocean. The form of the valve is not
"linear-lanceolate or fusiform," but is absolutely linear, and provided with rounded
extremities, so that in this respect it differs from that of any previously recorded species
of the genus. Hence Pritchard's definition,1 as given in his History of Infusoria (p. 776),
must be extended, there being no doubt as to the fact that this frustule belongs to the
present genus. The generic description may be amended as follows : — " Frustules linear,
united into a filament, lateral valves smooth or faintly striated, linear, liuear-lanceolate
or fusiform."

Fragilaria antarctica, n. sp. (Plate XXV. fig. 12.)

Frustulis seriatim dispositis ; valvis plus minus subcurvatis, transverse raro plicatis ;
inter plicas duplex punctulorum ordo ; punctula invicem alternantur. In mari Antarctico
passim.

The two valves shown in the present figure are of very different sizes, yet the identity
of their structure indicates that they belong to the same genus, although in the larger
specimen the want of symmetry along the longitudinal axis is more pronounced than in
the other. This want of symmetry is not found in other forms of Fragilaria, except in
the case of Fragilaria (?) pacifica,2 Grun., which is figured in the Synopsis des Diatomees
de Belgique, Plate xliv. figs. 20-22, and of Fragilaria (?) schwarzii, Grun., represented
at fig. 24 of the same plate. With these two examples before us, however, the present
frustules may be classed as FragilaricB, especially as the valves are striated by un-
interrupted rows of points — a character which is generally found in the Fragilaria'.
Strictly speaking, however, this t)-pe should not be placed among the Pseudoraphidieae,
but among the Cryptoraphiiedse.

The structure of the valves is very characteristic. ' The surface is traversed trans-
versely by thinly-set but deep folds, so that the intervals resemble large granules when
viewed from the zonal side. Between the folds or plicae there are two lines of sparsely
disposed punctiform granules, which alternate with one another. The Diatom is very
common, and is quite characteristic of all the collections from the Antarctic, being found
sometimes isolated and sometimes in series. When in series the frustules are not linear,
but somewhat inflated, owing to the convexity of their outline.

1 This definition is as follows : " Frustules linear, united into a filament ; lateral valves smooth or faintly-
striated, linear-lanceolate or fusiform."

2 The localities given for Fragilaria (?) pacijica, Grun., are " Cap. de Bonne-Esperance et lies Samoa," and
for Fragilaria (?) schwarzii, Grun., " lies Seychelles."

KEPORT ON TIIE DIATOMACKEL 57

Plate XXV. fig. 11, represents two Fragilarian valves, the first of which belongs to the
new Fragilaria linearis1 above described, but the determination of the second, which
is of lanceolate form and differs markedly in its striation, is uncertain, as nothing except
the valve is known.

The frustules seen at fig. 17 of the same plate must also remain undetermined. They
possess a very elegant oblongo-lanceolate form, are provided with acute apices, and are
transversely marked with alternating light and dark zones. These rare specimens were
found in mud obtained from a sounding in the neighbourhood of Japan, and their charac-
teristic features are not yet fully known.

Diatoma, De Candolle.
Diatoma rhombicum, O'Me. (Plate XXV. fig. 22.)

We have here represented a small elliptical valve, which was found in a gathering
made to the south of Heard Island. In a paper on the Diatomaceous Gatherings made at
Kerguelen Island, by O'Meara, and published in the Linnean Society's Journal (Botany), vol.
xv., this frustule is seen on Plate i. fig. 2. Of the accuracy of the generic determination
there can be no doubt, especially as O'Meara has also figured a series of many frustules
disposed in a zig-zag manner, nor can there be any question as to the identity of the
Diatom now represented with O'Meara's specimens. The definition given of the typical
Diatom is as follows : " Frustules small, about -000G in length. On front view quad-
rangular, the costse appearing as a narrow band of puncta. On side view rhombic ; costae
very fine, pervious."

Grammatophora, Ehrenb.

Grammatophora stricta, Ehrenb., var. nov. (Plate XXIX. fig. 12.)

"We have here represented the only frustule belonging to this genus which calls for
particular attention, and which was collected in the Sea of Japan. It is closely related to
the Grammatophora stricta of Ehrenberg,2 as figured by Kiitzing in his work entitled
Die kieselschaligen Bacillarien oder Diatomeen, Plate xxix. fig. 76 ; but the latter is
somewhat larger, the septa are not so straight, and at the polar extremities these are not
flanked by a small appendage. Such points of difference, however, cannot be regarded as
of more than varietal importance.

Cyclophora, n. gen.

In February 1878 I read before the Pontifical Academy3 a communication, in which
I made known a new Tahellaria, which I had met with some years before adhering to

1 Compare Plate XIX. fig. 9.

2 Ehrenb., Verb. T. L i. fig. 22, and T. III. vii. fig. 31 ; Eabenhorst, Flora Europaea Algarum, p. 305.

3 Atti. Accad. Pontif. d. nuov. Lincei, 1878.

(bot. chall. exp. — part iv. — 1886.) D 8

58 THE VOYAGE OF H.M.S. CHALLENGER.

seaweeds on a rock called St Clement's near the port of Ancona, and which was also
abundantly represented in a collection of Diatoms from the aquarium of the Zoological
Station at Naples. This organism might easily be confounded at first sight with
Diatoma hyalinum,1 Kg. ; but a careful examination of the former reveals the fact
that the markings are more frequent on the zonal side, and that in the middle of one of
the longitudinal lines an appearance resembling the letter C is presented. The frustules, as
iu all Tabellarice, are united in a zig-zag series, and it is exceptional to see one frustule
by the side of another or lying isolated. They are connected with each other by means of
small membranous cushions generally placed at the angles. In the longitudinal series
the position of these points of union alternates from right to left, but in the lateral series
they are on the same side in all. Hence in this genus, as in Cocconeis and Achnanthes,
the valves are dissimilar to one another, the loculus of one being round.

The genus Cyclophora may accordingly be defined as follows : — Frustula tabulata,
rectangula, in fascias conjuncta, rarius soluta ; isthmo gelineo alterne concatenata ; a
fronte linearia vel parum inflata ; valvis inasqualibus, quarum una loculo centrali
instructa.

This genus, of which Cyclophora tenuis, Cstr., is the typical species, has been accepted
by all micrographers. Prof. H. L. Smith has recognised it in a recent collection made in
America, and Grunow, in the Synopsis des Diatomees de Belgique, has given on Plate
xxxvi. figures of Cyclophora tenuis, Cstr., and of a variety — Cyclophora tenuis, Cstr., var.
tropica — which was collected in the vicinity of Honduras, the Barbados, and the Isle of
France.

In a note annexed to his table Grunow2 remarks that in addition to the lonp-itudinal
lines, the valves show very fine transverse liues and two terminal nodules ; but it is to
be noted that the longitudinal lines, at least, are only on the zonal side. If these two
characteristics really exist the definition should be modified, and the existence of the two
terminal nodules would render the systematic position of the genus very problematical.
On the valves I have hitherto been able to distinguish nothing except the loculus and a
slight central line on one of the valves.

Cyclophora tenuis, Cstr., var. nov. (Plate XXV. fig. 3.)

We have here delineated two frustules from the Philippine Islands, which were
united together and evidently belong to the same genus. These frustules resemble
in their profile the figure of Cyclophora tenuis given by Grunow, but do not agree

1 Kiitz., Lac, p. 47, pi. xvii. fig. 20; Smith, Synopsis of the British Diatomacese, vol. ii. p. 41, pi. xli. fig.
312; Pritchard, op. cit., p. 778, pi. iv. fig. 16.

2 The original note is as follows : " Les valves ont des lignes ru^dianes et des nodules terminaux bien
marques. Ces derniers sont un peu eloignds des extremity qui sont obtuses. Les striae transversales depassent
le nombre de 30 en 0-01 mill. Les lignes longitudinales sont dedicates et un peu ondulees."

IMPORT ON THE DIATOMAC1LE. 59

with the typical and most common form already figured in its zonal and valval aspects
by me.1 It is rare to find a zonal presentation of such a frustule, but the appearance
presented by the extremities discovers to the observer that the valve is not linear but
inflated, and so probably represents a distinct species. In the absence, however, of more
conclusive arguments I have chosen to regard it as possessing nothing more than varietal
characteristics.

Surirella, Turp., Ehrenb., W. Sm.

Although the recognition of this genus is an easy matter, the determination of its
species is a work of great difficulty. The structural details of the Diatoms are constantly
found in a more or less developed condition, being sometimes scarcely even outlined, so
that it is common to find frustules having the same structural peculiarities, though
possessing widely different external contours.

The genus was first established by Turpin2 in 1827, and afterwards adopted with
modifications by Ehrenberg. Its conception was finally limited by W. Smith, who defined
it thus : — " Frustules free, ovate, or elliptical, valves with a longitudinal central line, and
margins produced into alas ; canaliculi distinct, usually parallel." But there are two consi-
derations which are obstacles in the way of our accepting this definition. Firstly, types
with a linear form, but which have all the other principal characteristics, such as Suri-
rella linearis,3 W. Sm., and still more Surirella arctissima, A. S., the valve of which is
figured in the Atlas of Schmidt, Plate lvi. figs. 13 and 14, are excluded. The linear form
may, however, be looked upon as derived from the oval or elliptical by a process of
elongation, while the panduriform outline results from the contraction of an oval form.
Secondly, the definition cannot be adapted to some specific types, which, although they
exhibit all the other characters of Surirellce, are not provided with a central line. Hence
not a few species at present classed as Surirellce should be eliminated, as, for example, the
Surirella fastuosa, Ehrenb., var. abludens* Grun., the Surirella intercedens,5 Grun., and
others. This difficulty may readily be overcome by excluding from the definition the
clause which demands the presence of a central line.

Surirella dives, n. sp. (Plate X. fig. 4.)

Forma elliptica, canaliculis radiantibus, margine denticulatis, a media superficie
nascentibus, area centrali arete lanceolata, et lineolarum corona cincta, et medio paucis
lineolis transverse signata ; margine striato. Ad Zebu, in mari Philippinarum.

This beautiful elliptical frustule is noteworthy on account of its well-developed

1 Atti. Accad. Pontif. d. nuov. Lined, vol. xxxv. sec. 6, May 21, 1882.

2 Mem. du Mus. d'hist. nat, vol xvi.

3 Synopsis'of the British DiatomaceaB, vol. i. p. 31, pL viii. fig. 58.

4 Schmidt's Atlas, pi. six. fig. 1. 5 Schmidt's Atlas, pi. xix. figs. 5 and G.

60 THE VOYAGE OF H.M.S. CHALLENGER.

canaliculi, which originate from the centre of the valve. In the middle there is a narrow
lanceolate area circumscribed by a corona of small lines, and marked transversely in the
centre by a few irregular lines. The border is delicately striated as well as the margins of
the canaliculi, and the centre of the interspace between two adjoining canaliculi is pro-
vided with a row of granules or denticules.

In this form, as in most of the Surirellce, increase takes place at the two poles of the
longitudinal axis.

The specific name has reference to the richness of the ornamentation which is found on
the valves.

Surirella japonica, n. sp. (Plate X. fig. 8.)

Elliptica ; canaliculis radiantibus, brevibus ; area centrali lanceolata lineolis terminata,
et nonnullislineis curvatis irregularibus quaudoque interrupts transverse signata ; margine
et areolis interstitialibus striatis. Ad mare Japonicum.

This species, which was collected in the Sea of Japan, is very similar to the preceding.
It is elliptical in form, but its canaliculi are somewhat shorter than those in Surirella
dives. The extremities of the interstitial area are almost pedunculate, and are notably
removed from the hoop of small lines which limits the central lanceolate areola. More-
over, the entire area between the canaliculi as well as the margin are delicately striated.
The central area is much less elongated than that of Surirella dives, and it bears inter-
rupted and somewhat curved transverse lines, which do not exist in that frustule.

Surirella argus, n. sp. (Plate X. fig. 9.)

Valvis ellipticis ; area centrali hyalina arete lanceolata, irregulari et interrupto
lineolarum ordine finita ; canaliculis constrictis, areolis interstitialibus granulo vel ocello
signatis ; margine striate In mari Japonico.

This frustule, which is also from the Sea of Japan, is provided with perfectly elliptical
valves. Its centre presents a narrow area, which does not exhibit any distinct markings,
and which is circumscribed by an irregular or interrupted series of small lines. Its
canaliculi exhibit a notable contraction in the middle, and the small interstitial areolae
that result are each marked by a granule or ocellus. The margin is delicately striated.

Surirella ocellata, n. sp. (Plate X. fig. 7.)

Panduriformis ; superficie centrali tereti, a qua canaliculi procedunt ; margine et alis
marginalibus striatis, spatiis interstitialibus ocello distinctis. Ad mare Japonicum.

This Diatom, which, like the preceding, was procured in the Sea of Japan, is panduri-
form in outline, and though somewhat large it is singularly elegant. It is not provided
with a central area circumscribed by small lines ; on the contrary, the centre is smooth,
and from it spring .the canaliculi that adorn the perimeter. The border of the valve as

KEPORT ON THE DTATOMACE^E. 61

well as its marginal alae are striated, while each of the area3 between the canaliculi is
ornamented by a grannie or ocellus.

Surirella multicostata, n. sp. (Plate X. fig. 6.)

Forma cuneata ; area centrali hyalina lanceolato-cuneata, et lineolarum corona
terminata ; numerosis costulis vel canaliculis (58) radiantibus ; margine et spatiis inter-
stitial ibus striatis. Ad Zebu.

This large cuneate valve, which was collected in the neighbourhood of Zebu, is
provided with a smooth central lanceolato-cuneate area, which is circumscribed by
irregular rows of small transversely disposed lines. It is distinguished by the possession
of a corona of very numerous canaliculi or radiating costules. The marginal alae, the
interspaces between the canaliculi and the border, are delicately striated.

Surirella thaitiana, n. sp. (Plate XIX. fig. 3.)

Valva obovata, area lanceolata centrali transverse undulata, et lineolis terminata,
canaliculis duplici ordine radiantibus ad verticem striatis ; margine evidenter granulate
In portu Thaiti.

This elegant frustule was found in a rich collection made at a small depth at the port
of Tahiti. It possesses an oval profile, and is provided with a central lanceolate area,
which is circumscribed by a row of small lines, and also presents a few slight undulations.
Its double row of medially constricted canaliculi are not separated by any interspaces,
and the canaliculi exhibit at their extreme ends an incipient serrated striation. The
margin is evidently granulated, and at its narrow extremity there are to be found a few
canaliculi, which are hardly outlined, and which indicate the locality in which the increase
of the Diatom occurs — a process which generally is bilateral and takes place at the poles
of the valve.

Surirella grandiuscula, n. sp. (Plate X. fig. 5.)

Forma subovata, e majoribus; area hyalina centrali lineolarum ordine interrupto cincta;
canaliculis inaequalibus per alas intercedentes et in alas marginales exeuntes distinctis ;
margine et alis striatis. In mari Pacifico.

This Diatom, which is of unusual size, was obtained in a sounding of great depth
made in the centre of the Pacific Ocean. Its long axis measures not less than 196 M-, and it
possesses a slightly ovate form with a large smooth central area,, which is circumscribed by
an interrupted irregular row of short lines. There is a circlet of large canaliculi, which
become a little shorter and narrower as they approach the inferior or narrow extremity, and
are separated by interspaces which terminate in marginal alae. These alas and the large
border are striated, while, at the same time, a small group of striae mark the centre of
each intercanalicular space.

62 THE VOYAGE OF H.M.S. CHALLENGES.

Campylodiscus, Ehrenb., Men.

As constituted by Ehrenberg and Meneghini, this genus embraced rotundo-elliptical
forms with subradiate costules and canaliculi. By a greater development at the periphery
than at the centre the frustule assumes a slightly tortuous aspect, and as this continues,
the general aspect becomes more or less deeply curved, so as finally to resemble a saddle.
The definition of the genus has been given by Pritchard 1 in the following words : — " Valves
equidistant, frustules solitary, disciform ; disc tortuous or saddle-shaped, rotundato-elliptic,
costate, costse mostly radiate."

The tortuous form of the frustules renders it difficult to obtain correct figures of them,
particularly from their zonal aspects.

In this genus, as in Surirella, the valves are surrounded by elegant submarginal
wings, which become elevated on parting from the zone or girdle, a circumstance which
has caused difficulty in the determination of species.

Campylodiscus japonicus, n. sp. (Plate XL fig. 1.)

Forma subrotunda, grandis, costularum vel potius canaliculorum brevium circulo
numeroso (fere 80) marginali cincta ; costulse vel canaliculi a lineis spinulosis disseptse,
quse dein evanescunt ; area hyalina centralis subrotunda. In mari Japonico.

This Diatom possesses the form of a large disc with a number of short radial costules
or canaliculi. These are separated by lines of short " thorns," which occur on the connect-
ing zone, and run centripetally from the extreme margin. There is also present a hyaline
central subrotund area, which approximately coincides in shape with that of the entire
frustule. The specimen was collected in the Sea of Japan.

Campylodiscus zebuanus, n. sp. (Plate XL fig. 10.)

Forma mediocris subrotunda, costis ab area lineari axiali hinc inde radiantibus, medio
angulariter curvatis. Ad Zebu, in mari Philippinarum.

This moderately large frustule was collected in the neighbourhood of Zebu in the
Philippine Sea. It possesses a rounded outline and a smooth narrow almost linear central
area, which runs along its axis. From the margin of this area radial costules run across
the surface of the valve, and are folded alon^ a line almost concentric with the ed^e.
The two extremities of the central area terminate at two points, which may be called the
poles of the frustule, and they seem to indicate two centres of development for the valve,
as structures which appear to be rudimentary costules are perceptible in these regions.
It is worthy of remark that when the developmental significance of these two points is
recognised it is easy to understand how the valve may become more and more winding
and ephippiform.

1 Op. cit, p. 798.

REPORT ON THE DIATOMACEiE. 63

The specific name lias reference to the locality in which the form was first
collected.

Campylodiscus bicinctus, n. sp. (Plate XI. fig. 2.)

Valvse subrotundae, epbippiformes, duplici brevium costularum corona aream ellip-
soideam includente. E mari Japonico.

This somewhat larger valve exhibits a large smooth rotundato-elliptical central area,
which is surrounded by a double row of short radial costules, the rows being separated by a
smooth narrow ring. As in the preceding case, the two poles present indications of the
development of new costules.

The presence of a double row of enveloping costules has been made use of in naming
this type.

Campylodiscus erosus, n. sp. (Plate XL fig. 3.)

Valva discoidalis, ephippiformis, costis subradiantibus numerosis prope marginem
circum divisis ad aream sublinearem terminantibus ; costularum pars centralis evanescens
et quasi usu attrita. Ad oras Japonicas.

The present frustule bears some analogy to that last described. It possesses a double
corona of costules, of which those round the periphery are short, while those situated more
centrally are three times larger than the former in the middle, but diminish somewhat
irregularly towards the poles. The smooth central area is long and narrow.

From the circumstance that the intermediate parts of the inner series of costules are so
slightly manifest as to appear as if they had been subjected to attrition, the specific name
of erosus has been given.

&*

Campylodiscus erosus, n. sp. var. nov. (Plate XI. fig. 5.)

We have here represented a frustule from the Sea of Japan, in which most of the costules
of the inner series are represented by two oblong cellules. There can be no doubt that
this is but a further stage of the attrition which is manifested in Campylodiscus erosus,
so that nothing remains of the costules except the two extremities. Hence, although
apparently very distinct characteristics are here found, these can only be regarded as
possessing a varietal significance.

Campylodiscus lepidus, n. sp. (Plate XL fig. 7.)

Forma subrotunda, mediocris ; costulis circumradiantibus a corona irregularium lineo-
larum, quse aream late ellipticam includit ; costulse lateraliter partim linea, partim spatio
quadratim punctato circumscribuntur ; margine granulorum linea signato, exterius spinuloso.
In mari Philippinarum.

64 THE VOYAGE OF H.M.S. CHALLENGER

This exceedingly elegant frustule possesses a subrotund form. It is slightly folded,
and has a series of long costules, which are, however, not quite half the length of the radius,
placed regularly round the periphery. These originate from a granulated circular line
which circumscribes a central smooth oval area. The costules are divided from one another
throughout half of their length by a simple line ; in their outer half they become constricted
and are rounded at their extremities, being at the same time separated by finely granu-
lated areas, ornamented by minute quadrately disposed points. The extreme edge
of the valve bears a circlet of minute granules, and the margin is provided with short
spines.

Campylodiscus humilis, n. sp. (Plate XL fig. 8.)

Forma parva, flexuosa, ephippiformis ; costulis vel canaliculis ab area sublineari
radiantibus ; costse utrinque a linea centrali profunde inflexa. Ad insulas Philippinas.

This small form bears a great affinity to Campylodiscus zebuanus, from which,
apart from a difference in size, it may be distinguished chiefly by the two following cir-
cumstances : (1.) It is far more elliptical than the latter, and (2.) the line of inflexion,
instead of being but slight, is well-marked. Although it is not possible to determine
the true specific value of these differences, the form now in question has been provisionally
regarded as a distinct species.

Campylodiscus philippinarum, n. sp. (Plate XL fig. 9.)

Mediocris, ephippiformis ; crebris costulis (centum et amplius) in tres ordines
inasquales irregulariter divisis, a quibus area subrotunda circumscribitur, medio punctulis
raris quadratim dispositis vix perspicuis signata. Ad mare Philippinarum.

This frustule, which was collected in the Philippine Sea, possesses a disc of moderate
size, which, by the flexion of two of its axes, appears somewhat quadrate in outline. It
has a corona of numerous radiating costules, which are separated by an undulating line into
two rows of unequal size, while the inner and larger is surmounted by a third row of short
cellules. The large central area is ornamented by a number of minute scattered granules,
which are arranged in a quadrate manner.

Campylodiscus orbicularis, n. sp. (Plate XVI. fig. 10.)

Valvis subrotundis vix flexuosis, area elliptica lsevi lineolarum serie cincta ; costulae
subradiantes (42) ad utrumque polum divisse. In mari Japonico.

At first this form seems to be a variety of Campylodiscus lepidus (Plate XL fig. 7), as
both possess approximately radiating costules or canaliculi, and in the centre of both there
is an oval hyaline area circumscribed by a series of small lines. The structure of the margin,
however, is completely different in the two types. In the former the border is coarsely

REPORT ON THE DIATOM ACEvE. 65

and irregularly striated, and the intercostal areas are punctated ; while in the latter, in
addition to the intercostal quadrately arranged puuctations, the margin is denticulated
and ornamented with a distinct corona of minute granules. As serving still more conclu-
sively to determine the independence of the two organisms, it may be noted that while
the symmetry of the radial costules of Campylodiscus lepidus indicate that augmentation
must go on all round the periphery, in the present frustule somewhat larger intercostal
areas are found at the diametrically opposite points, so that it is manifest that increase
takes place in these regions.

Campylodiscus oceanicus, n. sp. (Plate XI. fig. 4.)

Valvis flexuosis ; costulis longiusculis, quorum apicibus cellulce vel costulse inter-
ponuntur ; area centrali tereti subrotunda. In mari Pacifico.

This fiexuous Diatom was collected in the Pacific Ocean. In it the radiate costules
encircle a smooth and almost round area. At their summit these alternate with another
order of costules, which seem to be elongated cellules. These are continued to the
extremity of the submarginal wing, for which they form a support.

Campylodiscus nitens, n. sp. (Plate XL fig. 6.)

Forma parva, flexuosa ; costulis subradiantibus grandiusculis aream hyalinam
limitantibus ; costulse superius rotundo-complanatse brevi spatio veluti granulo terminato
distinctae. Ad Zebu, ex insulis Philippinis.

This small discoid form was collected in the neighbourhood of Zebu in the Philippine
Islands. It is distinguished by possessing a series of subradiate costules, whose outer
extremities are somewhat rounded, and separated from one another by small interspaces
that are bounded by small round granule-like bodies, which more centrally are replaced
by simple lines. In the centre there is a smooth area of moderate size.

Campylodiscus wallichianus, Grev., var. thaitiensis, nov. (Plate XVI. fig. 6.)

This magnificent frustule, of which several specimens were collected at the port of
Tahiti, surpasses all the others in the elegance of its ornamentation. It consists of a large
and slightly folded disc, the curve being reversed with respect to the two axes. Almost the
entire surface is occupied by subradiating canaliculi, which terminate marginally in rows
of distant erect spines. The canaliculi gradually disappear as they approach the centre,
where a biconcave smooth area exists. Around the two rounded extremities of this area
there are a number of more sharply defined punctations, the most external of which form
a semicircle, a circumstance which serves to give greater prominence to the area in these
regions.

On comparing this frustule with others previously known, it is found to be most closely
(bot. chall. exp. — part iv. — 1886.) D 9

66 THE VOYAGE OF H.M.S. CHALLENGER.

related to the Campylodiscus tcallichianus of Greville ; 1 but it differs from the latter in its
considerably larger size, its far greater number of canaliculi, and in the arrangement of its
spines, which, in the Grevillean species, appear in the middle part of the lines only, and
not throughout their whole lenjrth as in the form from Tahiti. Moreover, the central area
of Campylodiscus uxdlichianus is linear, and is sharply defined round its entire margin,
and not at the two extremities only as in the present case.

These points of distinction cannot, however, be regarded as essential, but depend
merely on the greater or less number of certain details of structure in the two types, so
that the Diatom now in question has been indicated merely as a variety of the Grevillean
frustule.

Campylodiscus anceps, n. sp. (Plate XVI. fig. 2.)

Valvis rotundis, vix curvatis ; e maximis ; triplici granulorum circulo bis opposite inter-
rupto, a quo plurimse (64) brevissimse costulse dimanant. Ad Zebu in mari Philippinarum.

This is the most singular and novel form of Campylodiscus which has to be recorded.
It possesses a large and almost perfectly spherical disc, which is bounded by a triple
granulated margin, from which there proceed centripetally numerous short costules. At
two diametrically opposite points the marginal bands are interrupted, and at these points
small embryonic costules occur. By means of oblique illumination it may be seen that
the internal border of the rim is very finely striated, while the centre is devoid of
ornamentation of any kind. It is also noteworthy that, contrary to what occurs in all
other known species of the genus, the valve is almost smooth. It differs too from the
other species by the absence of marginal wings ; but the non-existence of these is explained
by the all but entire absence of a valval curvature.

This curious frustule might be regarded as presenting a transition to the genus
Coscinodiscus, were it not for the diametrically opposite areolaa that occur upon its
rim, where augmentation takes place, and in this respect it approaches nearer to the genus
Surirdla.

The diameter of this novel form is 170 n-

Nitzschia, Hassall, W. Sm.

The genus Nitzschia, which was instituted by Hassall in 1845 in memory of one
of the first observers who called the attention of naturalists to the great family of the
Diatomacea?, embraces the free compressed more or less bacillar frustules, whose linear
keeled valves are provided with one or more longitudinal lines of puncta, the keel being
often excentric. This last character is sufficient to distinguish it from the genus Amphi-

1 This frustule is figured in' Schmidt's Atlas, pi. xiv. figs 15. and 16.

REPORT ON THE PIATOMACE.E. 67

prora, while the general characters of the shell prevent it from being confounded with the
Syned/rce.1

The NitzschicB are abundantly represented, both in freshwater and in marine col-
lections ; but in the latter case they especially belong to the littoral flora, a circum-
stance which sufficiently explains the fact that observers have in recent times failed
to meet with so many new species of this genus as in the case of many others which
do not flourish so abundantly in the neighbourhood of the sea-shore. Among the
numerous soundings made by the Challenger in different localities many species
of the genus were recognised, but of these only the following are worthy of special
attention : —

Nitzschia plana, W. Sm., var. zebuana, nov. (Plate XIII. fig. 10.)

This specimen greatly resembles the Nitzschia plana of Smith 2 when its outline is con-
sidered, although its extremities can hardly be called acute. Its valve is obscurely striated,
while the keel, instead of bearing a single band of larger granules, is marked at
intervals by small irregularly disposed but very prominent lines. It is also to be borne in
mind that the species of Smith is a brackish-water form, while the present frustule is
marine, having been collected in the canal of Zebu among the Philippine Islands. Not-
withstanding these differences the latter must be viewed as only a variety of the former,
especially when the great resemblance of the two forms and the identity of the interrupted
striation of the two valves is kept clearly in view.

Nitzschia obesa, n. sp. (Plate XIII. fig. 11.)

Duplo longior quam latior ; medio angulariter constricta ; apicibus obtusis ; striis
dense decussatis ; carina lata distinctioribus lineolis inordinate signata. In mari Philip-
pinarum prope Zebu.

This form, which is also from Zebu, is remarkable in being only twice as long as it is
broad. It is angularly constricted in the middle, its extremities are obtuse, and the striae
are densely decussate. The valve is flanked by a large keel, which is traversed by prominent
lines of irregular distribution and direction.

Nitzschia obesa, n. sp., var. nov. (Plate XIII. fig.. 13.)

We have here represented a frustule from the port of Tahiti, which is somewhat less
obese than that just described, and which is provided with somewhat cuneate extremities.
Its striation, however, is of equal delicacy, and the keel though narrower bears short
but equally prominent lines. From such characteristics it can only be looked upon as
possessing a varietal importance.

1 See Pritchard, op. cit., p. 779. 2 Synopsis of the British Diatomaceae, vol. i. p. 42, pi. xv. p. 114.

G8 THE VOYAGE OF H.M.S. CHALLENGER.

Nitzschia vermiculata, n. sp. (Plate XIII. fig. 12.)

Marma, triplo longior quam latior ; latere carinali per constrictionem angularem
bilobato ; carina granulis distinctioribus medio signata ; apicibus cuneato-obtusis ; valvis
vermiculatim striolatis. Ad Zebu.

This superb Diatom was collected in the channel of Zebu among the Philippine Islands.
Its frustule, which is three times as long as broad, is angularly constricted in the middle,
and presents a bilobed aspect, the lobes forming wide curves. It is also marked by a row
of large granules, and the extremities of the valve are cuneately obtuse. The general
valvular ornamentation occurs in the form of tortuous lines of granules which intersect
each other.

Nitzschia mammalis, n. sp. (Plate XXIX. fig. 5.)

Mediocris, oblonga, sinuato-constricta ; apicibis productis, mammiformibus, carina
excentrica ; striis transversis. In mari Arafura.

This frustule, which was obtained in the Arafura Sea, has an elegant oblong form and
a deep central contraction. Its transverse striation is uninterrupted longitudinally, and its
extremities are slightly prolonged and mammiform.

Bacillaria, Gmel.1

"When the movements of Diatoms are considered, the peculiarities presented by those of
the genus Bacillaria are the most remarkable and mysterious of any. The genus includes
forms which are generally united to one another laterally, and in great numbers, and in
the living condition the individual frustules are constantly gliding over one another without
ever becoming completely disunited. Although various hypotheses have from time to
time been advanced to explain such wonderful phenomena, no observer has been able by
the most ingenious methods, by the use of the most perfect object-glasses or the most
efficient methods of illumination, to discover any special organs that might bring about
such curious results.

When, by treatment with acids or by mechanical action, the frustules of the Bacillarice
become isolated from one another, their form may at first sight cause them to be con-
founded with species of Syncdra, Diatoma, or Nitzschia. Yet in the case of surface
gatherings it is easy, by placing the organisms in a glass cell before isolating the frustules
by the action of heat, to distinguish them by the characteristic disposition of the frustules
already referred to.

By this means Bacillarian forms were frequently recognised in surface gatherings made
in the Sea of Arafura. The frustules were sublinear, lanceolate, and very slightly sigmoid.

1 This genus was established by Gmelin in 1788 when he founded the species Bacillaria paradoxa (Linnaeus,
Syst. Nat., ed. xiii. vol. vi., 1788; Hassall, Freshwater Algoe, pi. xciii. fig. 10; Kiitzing, BacilL pi. xxi. fig.
18; Smith, Synop. Brit. Diat. vol. ii. p. 10, pi. xxxii. fig. 279, suppl. pi. lx. fig. 279.

EEPOET ON THE DIATOMACE/E. 69

Their apices were very acute, and the valves finely striated, while they were at the same
time surrounded by a row of small granules or lines similar to those which occur in the
NitzschioB.

Bacillaria socialis, Greg., var. indica, nov. (Plate XXV. figs. 9 and 10.)

The known species of Bacillaria are few, and amongst these Pritchard (op. cit., p. 784)
records as Bacillaria socialis a form determined by Gregory1 as Nitzschia socialis. On
comparing this frustule with those from the Arafura Sea they are found to agree (1.) in
having their valves of a linear-lanceolate form ; (2.) in being ornamented with a very fine
transverse striation ; and (3.) in presenting acute apices ; while they differ in the following
respects : (1.) The Arafuran frustules are slightly sigmoid, while the others do not exhibit
this peculiarity ; (2.) the presence of a central keel, mentioned by Gregory as occurring
in his Diatom, is not to be detected in the present case, in which there is (3.) a band of
granules which are somewhat more distinct and more sparsely distributed than the striae.

Although, in the absence of an authentic preparation of Gregory's species for purposes
of more accurate comparison, it is not easy to determine the true significance of the above-
mentioned points of difference between what are otherwise two closely allied forms, the
frustules now figured have in the meantime been regarded as a variety of his typical form.

TRIBE III — CRYPTORAPHIDIK&

Rhizosolenia, Ehrenb.

Judging from the various definitions that have been given of this genus, which was
first introduced by Ehrenberg,2 it does not appear to have been clearly understood hitherto
by any observer. The definition given by Pritchard in his History of the Infusoria (p. 865) is
as follows : — " Filamentous, frustules subcylindrical, greatly elongated, siliceous, annulate ;
annuli broadly cuneate ; surface striated, extremities calyptriform, pointed with a bristle."
In the Micrographic Dictionary, on the other hand, the following generic characteristics
are pointed out : — " Frustules elongate, subcylindrical, marked with transverse spiral lines,
ends oblique or conical, and with one or more terminal bristles." But these definitions
cannot be said to correspond more accurately to the reality than that proposed by
Ehrenberg himself, which, according to Pritchard, was as follows : — " Lorica tubular, with
one extremity round and closed, while the other is attenuate and multifid as if terminating
in little roots."

It is not easy to understand why it was not recognised from the beginning that the

1 Trans. Micr. Soc. Lond., vol. v. p. 80, pi. i. fig. 45.

2 Ehrenberg, Hikrogeol., plates xviii., xxxiii., and xxxv.

70 THE VOYAGE OF H.M.S. CHALLENGER.

cuneate hoops, which are seen in some species, and the spiral line, which is still more
frequent, result simply from the form of the distinct parts, which by their union constitute
the tube of the Diatom. That these frustules are composed of several parts that are more
or less liable to be detached from one another — a phenomenon which may also be witnessed
in the case of the numerous hoops of Rhabdonema, Striatella, and such like forms — can
hardly be doubted. The parts that form the tubular walls, instead of being like those of
the hoops, are generally somewhat rhomboidal or lozenge-shap3d, and in soma species, e.g.,
Rhizosolenia robusta, Norman, these, having two sides of the parallelogram extremely long
and united by means of two shorter sides, form a belt which is terminated by two
transverse lines. In other species every constituent part is exactly rhomboidal and
equilateral, while they are bent round and united together in such a manner that they form
a tube upon which the line? of suture appear to be arranged in a spiral manner. Very
frequently the sides of each rhombus are somewhat curvilinear, and the obtuse angles
truncated, yet they fit together in such a manner as to constitute a tube.

Among the Rhizosolenia, which were found frequently to abound in many surface
gatherings made by the Challenger, frustules were often observed in a broken condition on
account of their size and the relative tenuity of their walls, aud in such cases the
rhomboidal parts of which they were composed were frequently isolated from one another,
while at the same time the sutural lines could be disfciuo-uished. It must also be recorded
that in some species the rhomboidal plates were found to be very small, so that the
perimeter of the tube was formed of several plates, and in such cases (see Plate XX.X. figs.
11 and 14) the surface of the Diatom resembled the skin of a fish or of a serpent.

The genus Rhizosolenia was always regarded as marine until Professor Hamilton
Laurence Smith discovered his very singular Rhizosolenia eriensis1 and subsequently
Rhizosolenia gracilis, which more rarely accompanied the former, in a surface gathering
made on Lake Erie in North America.

In marine gatherings the genus, in addition to its occurrence as a surface form, has
also been recognised amongst the contents of the stomach and alimentary canal of
Mollusca and other marine animals. In soundings, and still more in deposits, the Rhizo-
solenice can only be recognised by means of their mucrones, which, like a solid or massive
substance, can resist the trituration and pressure that invariably reduce their thin walls
to very minute and unrecognisable debris.

Another feature of great importance, and one which has not hitherto been recognised,
is well seen in some of the species that have now to be recorded, namely, the existence of
a minute cavity towards the extremity of each frustule. This cavity — if it is not to be
regarded as a consequence of the union — no doubt serves for the more perfect adhesion of
the frustules, which are disposed in rows or series, as each corresponds in position to the
extremity of the terminal region of an adjoining frustule.

, 1 See figure in van Heurck's Synopsis des DiatonieWde Belgique, pi. lxxix. fig. 9.

REPORT ON THE DIATOMACE7E. 71

Ehizosolenia inermis, d. sp. (Plate XXIV. figs. 7, 8, 10, and 13.)

Annulis distinctis ; striis a?gre perspicuis ; processu calyptriformi terminali truncato
et lineola brevi loncitiulinalitcr sicnato. In mari Antarctico.

The four cylindrical frustules here shown are composed of rings of trapezoidal plates,
and terminate in oblique calyptrae. It is noteworthy that in all the terminal mucrones
are alsent, and that the extremities of the calyptrae are truncated, each being ornamented
by a short longitudinal line, the whole reminding the observer of the shape of a pen.
This line seems to represent the mucro which is peculiar to other Rhizosolenian species,
a belief which is substantiated by the fact that in fig. 8 a trace of the line may be
perceived in the cavity, which is so disposed as to receive the extremity of the succeeding
frustule. It is therefore of the utmost importance to remark that in the definition of
the genus the terminal mucro, although a very general, cannot be regarded as an
essential characteristic. Although varietal differences may exist in the four frustules
figured they have all been, for the sake of convenience, grouped under one specific name,
which has been suggested by the common character of their extremities just alluded to.

Rhizosolenia polydactyla, n. sp. (Plate XXIV. fig. 2.)

Annulis brevioribus" crebris ; striis imperspicuis ; processu calyptriformi brevi et acuto
mucrone armato. In mari Antarctico.

This frustule is peculiar on account of the shortness and multiplicity of its rings. The
terminal mucro is strong and very acute, even more so thau is indicated in the figure.

The specific name of this Antarctic form is intended to be indicative of the remarkable
character of its rings.

Bhizosolenia ingequalis, n. sp. (Plate XXIV. fig. 15.)

Forma subcylindrica, constans annulis transversis parallelis ; processu calyptriformi
composito, in acutissimum mucronem exeunte. In mari Sinensi ad Hong-Kong.

This Diatom was observed in a surface gathering which was made in the neighbour-
hood of Hong-Kong. It is remarkable on account of the inequality of its transverse
diameters at different parts, an anomaly which may be attributed to regional inequalities of
pressure on the rings, which, instead of being round, are oval in section. It is also
especially noteworthy that the calyptriform process does not consist merely of a single
piece, but is made up of two or three rings or distinct parts.

Rhizosolenia sima, n. sp. (Plate XXIV. fig. 11.)

Frustulis subcurvatis ; calyptra brevi in obtusum brevissimum mucronem desinente ;
imperspicue striata. In extremo mari Antarctico ad glacies impervios.

A singularly strange aspect is presented by the present frustule, which cannot be
regarded as an accidental or monstrous form, as it has been frequently met with in the

72 THE VOYAGE OF II.M.S. CHALLENGER.

same collection. It is true that much significance cannot be attached to the arcuate
curvature, which is here very manifest, but the small protuberant, massive, and almost
spherical mucro forms a well-marked characteristic, and must be viewed as of specific
importance.

Rhizosolenia sirna, n. sp., var. nov. (Plate XXIX. fig. 9).

This form differs from the type, not merely by its size — a circumstance of little import-
ance and attributable to pressure acting on the frustule — but rather in the construction of
the terminal calyptra, which is not simple but formed by the union of several pieces.

Rhizosolenia japonica, n. sp. (Plate XXIII. fig. 7.)

Forma subcylindrica, annulata ; annulorum divisione transversa ; striis imperviis ;
mucrone terminali longissimo tenuissimo. In mari Japonico.

Only one specimen of this interesting type, which was obtained in the Sea of Japan,
has been observed. It is characterised by its very long terminal mucro, a circumstance
in which it perhaps does not differ from Rhizosolenia stylif&rmis, Bright.1 — and by the
fact that its rings are divided transversely — a character which is not found in Brightwell's
species. Although its minute characteristics have still to remain incompletely determined,
it cannot be regarded as a teratological specimen, but must be looked upon as the type of
a true species.

The organism shown in Plate XXIII. fin;. 8, although bearing certain external
resemblances to the Rhizosolenice in its general form and in the nature of its extremity,
may not improbably ultimately be determined to be of an animal nature. The absence
of any trace of division and of any indication of the origin of the mucro, as well as the
presence of a very pronounced granulation, are opposed to its being regarded as a
Ehizosolenian Diatom, while the fact that the specimen has been found in a preparation
of fresh material made on board the Challenger has rendered it impossible to determine
whether its walls are provided with or are devoid of siliceous matter.

Rhizosolenia murrayana, n. sp. (Plate XXIV. fig. 12.)

Cylindrica, constans partibus rectangularibus reticulatim dispositis, et convexe
terminata, et mucrone terminali instructa. In mari Antarctico.

This very singular form presents an entirely novel aspect, yet possesses all the
characters necessary to its being regarded as a member of the present genus. It has the
form of a small cylinder, and is composed of a large number of distinct equal rectangular
parts, which are cemented together, the tube being closed at both extremities by a convex

1 Mier. Journ., vol. vi. p. 94, pi. v. fig. 5; Norman, Ann. Mag. Nat. Hist., vol. xx. 1857, p. 158; Schultze,
Mizr. Journ., vol. vii. p. 18, pi. ii. fig. 1 ; Pritchard, op. cit., p. 8G5, pi. vii. fig. 32.

REPORT ON THE DIATOMACE.E.j 73

portion surmounted by short sharp points. Although the extremities can hardly be called
calyptriform, the generic determiuation that has here been given cannot be questioned,
more especially as an abundant supply of siliceous matter has been proved to occur in its
walls by subjecting it to a process of incineration.

Rhizosolenia imbricata, Bright. (Plate XXIV. figs. 1 and 1 bis.)

The form which is rej)resented on Plate XXIV. fig. 1 under a magnifying power of
130 diameters, and at fig. 1 bis under a power of 640 diameters, belongs without doubt
to the Rhizosolenia imbricata of Brightwell.1 In connection with fig. 1 the unimportance
of attaching any significance to the proportion between the transverse axis and the length
of the frustule may be noted, and it may also be observed that in the middle of the
granulated series of rings there occurs a smooth ring. Here it is probable that the
process of division takes place, by the formation of a diagonally placed diaphragm, which
arises in such a manner as to bring about the formation of two new calyptrse.

The arrangement of the moniliform striae on the other rinses is radiating;, the radiation
taking place outwardly from a median axial line. Each ring is flanked by two small
triangular and similarly sculptured spaces, from which it may be inferred that the entire
annulus is composed of two trapezoidal parts, on each of which the striae run in different
directions.

Plate XXIV. fig. 1 4, represents a form bearing on its calyptriform extremity two long
and very acute mucrones. Whether this appearance is normal or teratological, or whether
it is merely accidental— one mucro alone belonging to the frustule, the other being
casually apposed — cannot be determined. Should further observations prove that both
belong to the same Diatom, it must be regarded as a new species, but at present there is
nothing to show that it is not a monstrous form.

o

Rhizosolenia robusta, Norman, var. nov. (Plate XXIV. fig. 5.)

Here there is shown the calyptriform extremity of a Rhizosolenia, which is remarkable
on account of its extraordinary size and the sculpturing of simple radial lines which it
presents. These characters coincide with those of Rhizosolenia robusta, Norman, which
is figured in Plate viii. fio;. 42 of Pritchard's Historv of the Infusoria. The form of the
latter, however, unlike that of the former, is compressed, and its mucro is described as
being " short, delicate, and nearly linear," instead of being short and cuneato-acute, as in
the present case. Yet these points of difference cannot be regarded as possessing more
than a varietal significance.

The parts entering into the composition of the rings of this variety are sub-
rectangular.

1 Micr. Joum., vol. vi. p. 94, pi. v. fig. 6.
(bot. chall. exp. — pabt iv. — 188G.) D 10

74 THE VOYAGE OF H.M.S. CHALLENGER.

Rhizosolenia (?) flaccida, u. sp. (Plate XXIX. fig. 4.)

Cylindrica, aniiulata, parietibus tenuissimis firrnitate destitutis ; annulis brevibus
parallelis. In mari Adriatico et Arafura.

Among the surface gatherings which have been made at intervals for some years past in
the Adriatic, I have observed the frustule here figured from a form found on the surface of
the Arafura Sea. It possesses a delicate structure, being bounded by two nearly parallel
longitudinal lines, and crossed by transverse parallel lines, which are disposed at approxi-
mately equal distances from one another. Towards each extremity there is an irregular
circle, whose diameter is to that of the frustule in the ratio of two to three. These circles
are the extremities of the tubular organisms, and, like the parietes, are so thin that
when dried on a glass they collapse. It may be noted, however, that the parietes, though
very delicate, contain siliceous matter, as they resist exposure to a red heat.

Although the tubuliform part of this interesting organism bears a resemblance to the
Rhizosolenice, no calyptriform extremities have ever been observed, hence its true generic
position cannot be determined with certainty. It can only be examined when mounted in
the dry state.

An organism which is pretty abundant in the surface of the Sea of Arafura is shown
in Plate XXIX. fig. 8. It is of ovoid form, and terminates at one of its slightly prolonged
extremities in a small flat circular area, from which a strong solid conical point arises. At
its opposite extremity an opening is seen, which doubtless serves for the reception of the
salient extremity of the succeeding frustule. The parietes are slightly and indistinctly
striated, but the small terminal circle is surrounded by short but somewhat more pro-
tuberant lines. Although the appearance presented by the pointed extremity recalls the
calyptra of a Rhizosolenia, it is at present impossible to determine whether the organism
really belongs to this genus or is a sporangial form. It is clear, however, that it
is a normal form, as many similar specimens have been observed.

Rhizosolenia arafurensis, n. sp. (Plate XXX. fig. 12.)

Frustula cylindrica, non annulata ; parietes partibus subtrapezoidalibus composita? ;
calyptra desinit in elongatum, obtusum, tubularem mucronem ; striae impervise. In mari
Arafura.

Although in all other recent and fossil Rhizosolenia? the terminal mucrones are solid, the
form shown at fig. 12 has this part tubular and obtuse. The frustular walls are composed
of subtrapezoidal pieces, the separating quasi-sutural lines being merely thinner and more
transparent strands. The striae ornamenting the various parts are exceedingly delicate,
but their presence is argued by the slight tint that may often be recognised under the
microscope.

On Plate XXX. figs. 11 and 14, two other forms of Rhizosolenia are represented, in

REPORT OX THE DIATOMACK.K. 75

Loth of which many scale-like parts go to form the circumference of the tubes. These
parts are in both somewhat irregularly quadrate in outline, but they are relatively much
broader in fig. 11 than in fig. 14. As in the case of Rhizosolenia arafurensis only frag-
ments have been observed.

Dactyliosolen, n. gen.

In a tow-net, which was wrought on the surface of the Antarctic Ocean on
March 3, 1874, lat. 53° 55' S., long. 108° 35' E., some interesting forms, akin to the
Rhizosolenia;, Ehrenb., were collected. In both the form of the frustules is cylindrical,
aud in both it is composed of plates of trapezoidal outline — the Dactyliosolen cylinder
(Plate IX. fig. 7) being manifestly the result of the union of a series of hoops, which,
if detached along the lines of suture and spread out on a plane, would present long
linear profiles terminated by oblique parallel lines. But, on the other hand, in Dacty-
liosolen there is no trace of the calyptriform extremities, which are essential to the
Rhizosolenia?, so that it cannot be ascribed to the latter genus. Apart, however,
from this circumstance, it is of importance to observe that at intervals in the course of
the filament hyaline belts occur. Each of these seems to be a terminal belt of union
between the various frustules, any adjoining pair being separated by a space equal to
about one-tenth of the diameter of the frustule. It is also to be remarked that at the
lines of junction a few incipient granules or denticules may be perceived, and it is very
important that the cellules, which ornament the component rings of two adjoining
frustules, stand on the opposite and not on the corresponding margins of the rings.

That the combination of such characters is sufficient to warrant the estabbshment
of a new genus must be admitted, and it has been named Dactyliosolen1 from the
circumstance that the perfect tube is composed of a series of rings or hoops.

Dactyliosolen antarcticus, n. sp. (Plate IX. fig. 7.)

Forma cylindrica ; frustulum compositum ex pluribus annulis cellulatis ; cellulis
linearibus oblongis. In mari Antarctico.'

This is the only species of the present genus that has yet been observed, and its
characters are consequently those of the genus.

Chaetoceros, Ehrenb.

The following is the definition of this genus given by Ehrenberg (vide Pritchard's
History of the Infusoria, p. 861): — "Frustules without striae, united with the adjacent
ones by the interlacing on each side of awns proceeding from the frustule, or from a
cingulum between the frustules, and so forming a filament."

A more exact and at the same time more concise definition is found in Lauder's

1 day.TuXiog, a ring; tfwX^u, a channel or pipe.

76 THE VOYAGE OF H.M.S. [CHALLENGER.

Memoir on the Marine Diatoms of Hong-Kong, annotated by Ealfs, and printed in the
Transactions of the Microscopical Society, new series, vol. iv. It runs thus : — " Frustules
smooth or minutely punctated, united with the adjacent ones by the interlacing of awns
proceeding from the frustule." I

All the members of the genus CJi&toceros are extremely fragile, so that it is very diffi-
cult to distinguish them in deposits, especially in marine mud, in which only the more
robust forms are to be recognised, such as sporangial Goniothecia, Dicladise, or Syndendria.
On the other hand, filaments of Chsetoceros are much more abundant in surface collections,
being perhaps even the predominant forms, and among those collected in the Challenger
tow-nets, specimens belonging to the greater number of known species have been
observed.

Chsetoceros" protuberans, Lauder, var. nov. (Plate VIII. fig. 2.)

That a form of the Cli&toccros protuberans of Lauder;; is here represented is indicated
by the mammillar swelling on each valve, the outline of this protuberance being clearly
seen in the oval space intervening between two adjacent frustules. This species is described
by its founder as possessing minute bristly awns, but these are not to be observed in
the present frustule, and on consulting the plate of the typical form given by Lauder
minute bristles are only seen in connection with the terminal valves. If this is the
normal condition the specific definition given must be regarded as imperfect. Since, how-
ever, the series of frustules represented in our plate may be incomplete, the precise differ-
ences between the two organisms cannot be definitely fixed, although the varietal
character of the series at present in question may be provisionally accepted.

Chaetoceros dispar, n. sp. (Plate VIII. fig. 6.)

Frustula compressa, in seriern per longas setas teretes connexa a valvis procedentes, et
ad orio-inem constrictas : valvis alterne concavis et subconcavis. In mari Antarctico.

This form shows several analogies to the Chsetoceros decipiens of Cleve,3 but no traces
of striation or of punctation are to be found in the filaments or awns as in the allied type.
Moreover, the awns arise, not from the angle of the frustule, but exclusively from the
plane of the valve. At a short distance from their origin each filament presents a slight
swelling, and one of the valves is always more concave than the other, a circumstance
which has suggested the specific name.

This species was found in the Antarctic Ocean.

1 This latter definition, which does not exclude punctated forms, again illustrates the necessity that frequently
arises and that has already been referred to, for extending original definitions.

2 Remarks on the Marine Diatomaceoe found at Hong-Kong, with descriptions of new species, Trans. Micr.
Soc. Lond., new series, vol. xii. p. 79, pi. viii. fig. 11, 1864.

3 Diatoms from the Arctic Sea, Bihang k. SvensJca] Vet. Alcad, Handl., Band 1, No. 13, p. 11, pi. i.
figs. 5 a and 5 b.

RETORT ON TIIE DIATOMACEiE.

77

Chsetoceros janischianum, n. sp.

Frustula cylindrica per cingulum binatim in laxam seriem conjuncta ; cornubus binis
prolixioribus mediis valvis assurgentibus et eleganter curvatis. In mari glaciali Antarctico.

Tbis is without doubt the- most
elegaut typo of the genus Chaioceros.
It -was gathered from the surface of
the Antarctic Ocean in the form of
long seriate frustules, which were
united together in pairs by a common
belt. The valves are oval and notably
convex, and their awns, which are very
long and smooth, originate from their
centre, proceeding first almost at right
angles to the surface of the valve, and
afterwards interlacing by an elegant
curve with the corresponding awns
of an adjoining frustule. "Where this
interlacing occurs the awns are curved
horizontally and bent in such a manner
that each is alternately prolonged in a
direction at right angles to the plane
of the other. At the inferior valve of
the terminal frustule the two setae are
somewhat shorter, and are symmetri-
cally curved downwards.

The belt already alluded to is
siliceous, and its presence is difficult
to reconcile with the independence of
the two frustules, nor is it easy to
imagine how it can disappear, and so
leave the latter free.

The species has been named in
honour of Mr Janisch the well-known

UiatOmiSlj. Chwtoceros ianischianum, n. sp.

Cheetoceros curvatum, n. sp.

Frustula solitaria ; valvis ovalibus, curvatis, superiori convexa, inferiori concava ; satis
a medio valvarum utrinque inferius curvatis, brevioribus ; zona quadruplo transverse lon-
gior quam latior. Ad meridiem insulae Heard.

78

THE VOYAGE OF HALS. CHALLENGER.

This isolated frustule possesses four awns or cormia, originating from the centre of
the concavo-convex valves. Those from the externally concave or inferior valve are slightly
tortuous, while all are short, and bent downwards. The zonal side
is about four times as long as broad.

The specific name has reference to the curvature of the valve.

Chsetoceros criophilum, n. sp.

Frustula oblonga, valvis convexis (superior inferiori convexior) ;
setis spinulosis longissimis a rnediis valvis orientibus, et inferius
curvatis. In mari Antarctico.

In this Diatom it is to be noted that the two valves differ from

each other in two important respects : (1.) The height of the

c hwtoceros curvatum, n. sp. upper seems to be double that of the lower; and (2.) the upper

valve is decidedly more convex than the lower. The setse are very long, spinulose,

.spring from the middle of the valves, and are slightly curved near their origin. In the

Chatoceros criophilum, n. sp.

Chcetoceros convolutum, d. sp.

course of their length they are ornamented with sparsely and somewhat irregularly
disposed apiculi, all of which are directed towards their distal extremities.
This species was gathered near the ice-barrier of the Antarctic Ocean.

Chaetoceros convolution, n. sp.

Valvis ovalibus, una convexiuscula, altera plana ; setis longissimis a media valva
convolute orientibus. In mari Antarctico inter insulas Kerrmelen et Heard.

KETOKT ON THE DIATOMACE.E.

79

This Diatom was obtained in a surface collection made in the Antarctic Ocean between
Kerguelen and Heard Islands. The frustules are subquadrate, and arranged in a series.
The valves are oval, the one being convex and the other flat. The setae originate in
the centre of the valve, and, after curving round each other, diverge horizontally. In
the figure only two awns are seen between two adjoining frustules, the remaining two
beinc; either broken off or shortened. On the terminal valve the setae are seen in an
embryonic form as a pair of short protuberances. At their origin they are always
convolute, a characteristic which is referred to in the name of the species.

Chsetoceros radiculum, n. sp.

Frustulum solitarium ; valvis ovatis, bino processu submarginali, inflato, costulato,
brevi. Ad meridiem insulae Heard.

This very puzzling form was collected to the south of Heard Island in the Antarctic
Ocean. It is always solitary, and its setae, instead of extending in a long line, form
short swollen appendices, like bulbous roots, which are to a greater or less extent marked
by longitudinally directed costae that sometimes have their origin on the plane of the
valve. To increase the difficulty of fully understanding this type it is to be noted that
in addition to two setae or cornua of the form just described there are in some cases

Ch&toceros radicu-
lum, d. sp.

Chcctoccros radiculum, n. sp., rar. a.

Chatoceros radiculum, n. sp..
var. /3.

protuberances, which are twice as long, but are provided with less pronounced swellings,
and are devoid of costae. Such forms may be provisionally characterised as constituting
variety a of the typical species. In a second group — variety /8 — there is found a trans-
versely striated median zone and a single median projecting knob between each pair of large
bulb-like marginal protuberances. Whether these differences from the typical species are
to be regarded as varietal or as indicative of so many distinct species it is difficult
to decide, although probably the former view is the more correct, and has been here
adopted.

80 THE VOYAGE OF H.M.S. CHALLENGER.

The small but distinct projecting processes which are sometimes present in the centre
of the valves have also been regarded as having but a varietal significance.

The specific name which has been adopted, will serve to remind the observer of the
curious form of the setse or processes.

Chaetoceros, sp. (?) nov.

The annexed woodcut represents a Chsetoceran valve with long denticulate awns,

but hitherto insufficient material has been observed
to enable a complete account to be given of its
specific characters. It must, therefore, be indicated
as a doubtful species. It was obtained in a surface
collection made to the south of Heard Island.

Chaetoceros incurvum, Bail., var. umbonatum, nov.
(Plate XXIX. figs. 10 and 16.)

Chaetoceros, sp. (?) nov, _

An examination of the contents of the alimentary
tube of two Echini dredged in the North Atlantic at Station 47, from a depth of 1340
fathoms, revealed the presence of numerous Diatoms, which were no doubt serving as
food to the animals in question. One of the most frequent of these was the very small
Chaetoceros shown in the present figures. Its oval valves are provided with short re-
curved filaments, and the frustules are never found in series. Hence the entire genus
Clisetoceros may be divided into two sections, namely — (a.) Clisetocerotidse gregarise, com-
prising all those forms that occur in chains, and (b.) Chsetocerotidse solitaries, embracing
all free species — as, e.g., Clisetoceros radiculum, and the form now before us.

The chief point of distinction between this form and Chaetoceros incurvum, Bail.,1
consists in the notable swelling at the centre of the valve which is to be found in the
former. This, however, can hardly be regarded as a difference of specific importance.

In the genus Chaetoceros, I include three genera established by Ehrenberg,2 namely —
Dicladia, Goniothecium, and Syndendrium. These have from the first presented great
difficulties, and, in the words of Brisrhtwell,3 " much must vet be brought to light before
a satisfactory classification of this group can be effected." The uncertainty in connection
with them is owing to the fact that they have hitherto only been observed in a fossil
condition in deposits, so that it has been found impossible to understand their form
when in a state of actual vegetation sufficiently well to enable the observer to ascribe to
them their proper place in a system of classification.

1 Micr. Journ., voL iv. pL vii. figs. 9-11. • Ehrenberg, Mikrogeologie.

3 Quart. Jirum. Micr. Set., vol. iv. p. 105-109, pi. vii.

KEPORT ON THE DIATOMACEiE. 81

It has already been observed that in some living species of C/wetoceros, Goniothecia
in embryonic stages occurred, and for this reason Brightwell,1 by analogy, has main-
tained that the three above-named genera are not organisms per se, but dependent and
subordinate parts of species of Chwtoceros.

In confirmation of this conclusion, I have repeatedly observed conditions such as are
represented on Plate XIX. figs. 7 and 8, in which there may be seen frustules of Dicladia
capreolus, Ehrenb.,5 enclosed in frustules of Chcetoceros, and as no substantial difference
is found between Dicladia and Syndendrium, the non-independent character of these
genera is manifest. But what can be the significance of this singular organic siliceous
body enclosed in a Diatom cell, and itself constituting another cell ? It would seem as
if its sporangial nature was beyond doubt, and it must accordingly be regarded as designed
for the reproduction of the species.

This interpretation of the so-called Dicladia, Gonioihecium, and Syndendrium, is
a confirmation of the opinion of Dr Wallich, who, instead of, like some others, viewing
the sporangial frustule as a means of restoring to its original dimensions a Diatom which
had been diminished in size by successive divisional processes, viewed it not as a normal
and independent organism, but as a transient monstrous form designed for the elaboration
of sporules or embryonal forms, and therefore capable of reproducing the species.

" In fact it had been observed that every species, by the process of conjugation, must
be represented under two forms, one large and the other small, between which a gap
exists, over which we have at present no means of bridging except by supposing that the
two halves formed in cell division need not always be equal, and that by dwindling away
through a succession of steps of this kind, the progeny of the sporangial frustules may be
reduced to the original size." But that this does not happen may be shown by recalling
what takes place in Cocconema, Ehrenb., in a collection of which it is easy to observe the
sporangial, accompanied by the normal and much smaller form, without being able at the
same time to recognise the gradual diminution of the frustules by which the extreme
dimensions are reunited. (See vol. i. pi. xxiii., and vol. ii. pi. C, Synopsis of the British
Diatomacese.) Moreover, the same fact is shown in a still more convincing manner by an
observation made by Prof. Hamilton L. Smith on the Stauroneis gracilis of Ehrenberg.3
This well-known naturalist recognised that the sporangial form is the same as that which
Ehrenberg called Stauroneis phoenicentcron * — a type which greatly differs from the former
in the delicacy of its striation, so that, instead of being regarded as a specific and
independent form, the latter must be looked upon as but a transitory reproductive form
of Stauroneis gracilis, Ehrenb.

1 Brightwell, Quart. Journ. Micr.Sci., vol. iv. pp. 105-109, pi. vii. figs. 53-60.

2 Ehrenberg, Mikrogeologie, pi. xxxv. A. 17, fig. 8.

3 Ehrenberg, in Kiitzing's Bacill., pL xxix. fig. 3 ; Smith, Synopsis of the British Diatomaceae, vol. i. p. 59,
pi. xix. fig. 186.

4 Kiitzing's Bacill., pi. iii. fig. 53 ; = Navicula phoznicenteron, Ehrenberg Infus. pi. xiii. fig. 1 ; Smith, op.
tit., vol. i. p. 59, pi. xix. fig. 185.

(bot. chall. exp. — part iv. — 1886.) 1) 11

82 THE VOYAGE OF H.M.S. CHALLENGER.

Chsetoceros dicladia, n. sp. (Plate VIII. fig. 1, and Plate XIX. figs. 7 and 8.)

Frustulis in latere zonali transverse triplo longioribus, undulatis, in series per spatium
medio constrictum divisis ; quae binis utrinque cornubus breviusculis subclavatis quater
costatis connectuntur, setarum costis spinulosis. Inter Kerguelen et Heard.

The species of Chsetoceros in which specimens of Dicladia capreolus, Ehrenb., have
been repeatedly observed, was often found in gatherings made between Kerguelen and
Heard Islands, near the polar ice-barrier in the Antarctic Ocean.

The frustules have their zonal sides three times as long as broad, while the transversely
directed bounding line is undulating, so that an oblong space with a central contraction is
left between the adjoining frustules. The setae, which spring from the surface of the
valve, are somewhat short, and are provided with somewhat club-shaped outer extremities
which bear four thorny ribs.

Bacteriastrum, Shadb.

This very remarkable genus was instituted by Shadbolt in 1860 to embrace several
curious small sun-like organisms, which are surrounded by a few radiating protuberances
of considerable length, and which are very often met with in the stomach of Mollusca.
Shadbolt defines his genus in the following words:1 — "Frustules awned, united into a
jointed conferva-like cylindrical filament, valves discoidal, with marginal radiating awns."

Although there is a great analogy between the present genus and CJisetoceros, I am
of opinion, notwithstanding the opposite view of Professor H. L. Smith, that the differ-
ences are sufficient to warrant the preservation of both, apart altogether from the less
important question already referred to of the desirability of avoiding too great multiplica-
tion of the species belonging to any given genus. Among these differences may be noted the
following : — (1.) In Bacteriastrum the valves are always perfectly round, while in Chseto-
ceros they are generally oval ; (2.) in the former there is a greater number of awns, which
always radiate round the border, than in the latter ; and (3.) in seriate CJisetocerotidse
the frustules constantly interlace by the alternate crossing of the awns — a character which
is not found in Bacteriastrum.

Only a few species belonging to this genus are yet known, and these were diminished
when Lauder,2 after observing that Bacteriastrum curvatum, Shadb., constantly ter-
minated the series of Bacteriastrum furcatum, Shadb.,3 united the two forms under the
name of Bacteriastrum varians. This important observation, however, increases the
difficulty of determining the exact limits of specific forms, and, although among those

1 Pritchard, op. cit., p. 803.

2 Trans. Micr. Soc. Loud., new series, vol. sii. p. 7, pi. iii- figs. 1-6, 1864.

3 = Adiniscus sexfurcatus, Ehrenb., Mikrogeol., pi. xxxv. B. 4, fig. 15; Adiniscus biseptenarius, Ehrenb. ;
Adiniscus bisodonarius, Ehrenb., loc. cit. See also Pritchard, op. cit., p. 863, pi. vi. fig. 26 ; Trans. Micr. S<><:
Lond., vol. ii. pi. i. figs. 1 and 2.

REPORT ON THE DIATOM ACE.E. 83

new forms which have now to be recorded from the Challenger collection, some may
ultimately be found not to be specifically distinct, but to be different conditions of a species
already known, distinctive names have been provisionally applied in all cases where the
forms differ notably from one another, or from any hitherto recorded member of the genus.

Bacteriastrum brevispinum, n. sp. (Plate XV. fig. 8.)

Valvse setis submarginalibus, numerosis (29) brevibus, laevibus, cuneatis, rectis. In
mari Sinensi ad Hong-Kong.

This species occurs abundantly in surface gatherings made in the neighbourhood of
Hong-Kong. It has the form of a large round hyaline disc, which is surrounded by
numerous (twenty-nine) short, straight, radiately disposed, and gently cuneated thorns
or setse. These setae cannot be regarded as the representatives of longer filaments that
have become accidentally abbreviated, as their outermost extremities do not present the
appearance of any truncation.

The specific name that has been applied has reference to the shortness of the spines.

Bacteriastrum brevispinum, n. sp., var. nov. (Plate XV. fig. 6.)

From the same locality the form represented in the present figure was obtained. It
may be distinguished from the preceding by its smaller size, by being furnished with only
ten radiating seta?, and by having the centre of the valve ornamented by a single granule,
which may probably indicate that it has been terminal. Notwithstanding these differ-
ences, however, it can only be regarded as possessing a varietal significance.

Bacteriastrum wallichii, Ealfs, var. hispida, nov. (Plate XXIII. fig. 3.)

This frustule cannot be viewed otherwise than as a variety of Bacteriastrum wallichii,1
Ealfs, which is provided with "smooth, simple, divergent awns." In the Challenger form,
on the other hand, the awns are not smooth, but are provided with manifest apiculi. This
circumstance may be recalled by the varietal name hispida, which has been proposed for
this interesting form from the Arafura Sea.

On Plate XXIX. fig. 6, the same variety is represented from the neighbourhood of
Hong-Kong.

Bacteriastrum spirillum, n. sp. (Plate XIX. fig. 2, and Plate XXIX. fig. 1.)

Setis submarginalibus tubulosis et spinularum spira decoratis. In mari Arafura.
The present figures show what is probably the terminal valve of the same species, of
which perhaps they are two varieties. The curves of the awns in the one are slight and in

1 = Chxioeeros bacteriastrum, Wallich, Micr. Journ., vol. viii. pi. ii. figs. 16 and 17 ; Pritchard, op. cit,
p. 863.

84 THE VOYAGE OF H.M.S. CHALLENGER.

the same plane as the valve, while in the other the awns are flexuous, being first bent
downwards and then curved round. The circumstance that one valve possesses a single
central granule, while the other has several small irregularly distributed granules, is unim-
portant, but it is a matter of greater significance that in both the awns are not filiform,
and that in both they are adorned by elegant spirals of very small apiculi, which give a
singular appearance to the valves. In contradistinction to the last-named characteristic,
it may be noted that in Bacteriastrum varians, Lauder, the awns are filiform and smooth.

Bacteriastrum varians, Lauder, var. princeps, nov. (Plate XIV. fig. 2, and Plate XXIX.
fig. 3.)

This interesting form occurred in the rich Qjatherin<r made on the surface of the Sea
of Arafura, an intermediate frustule being shown in Plate XIV. fig. 2, and a terminal
frustule in Plate XXIX. fig. 3. The former possesses the singularity of having the two
filaments into which each ray bifurcates different from one another, the one being slightly
arcuate and the other spirally bent, while the disposition of these filaments is such that
the curved always alternate with the spiral. The latter has the rays undulating and
slightly curved in the same direction. The rays of the terminal frustule do not arise from
the perimeter of the valve, but somewhat more centrally, a circumstance which indicates
that the valve is somewhat convex ; moreover, a single central granule is very distinctly
seen in this case.

Although the association of such characters gives the entire series a somewhat peculiar
appearance, the present singularly distinct and elegant forms cannot be regarded as
specifically distinct from Bacteriastrum varians, Lauder, from which indeed they differ
only with regard to the condition of the rays.

Bacteriastrum varians, Lauder, var. nov. (Plate XXIII. fig. 1.)

Another new variety of Lauder's typical form is here shown under a magnifying power
of 460 diameters. The round central area is entirely devoid of ornamentation, but near
the periphery eleven stout rays project round the central disc. These rays are approxi-
mately at a uniform distance apart, and at their origin are somewhat swollen, a slight
involution of the proximal bounding line of the ray occurring here just in the line of the
longitudinal axis of the rays. At their distal extremities the rays bifurcate, the main shaft
of each having, however, a length considerably greater than the diameter of the central
disc. The branches of each bifurcation are bent in a graceful curve towards one another ;
their diameter is somewhat less than that of the main shaft, and their length is somewhat
less than the radius of the disc. Where the bifurcation takes place a short, straight, but
well-defined line runs down the shaft of each ray for a short distance. The. almost perfect
uniformity of the diameter of the rays from their origin to the point of bifurcation is
especially noteworthy.

REPORT ON THE DIATOMACE/E. 85

Corethron,1 n. gen.

The new forms upon which this genus is established are exclusively Antarctic. They
have a more or less cylindrical form, and are terminated by hemispherical surfaces, the
base being surrounded by a corona of awns, which are more or less thick and long, and are
smooth or thorny. Although from the rarity of these interesting organisms it was not pos-
sible to determine their siliceous or non-siliceous character by means of acids, this determina-
tion was made by the application of strong heat for a protracted period. This produced
no visible change on their form, so that their diatomaceous nature could not be doubted.

Although the presence of long awns radiating round the valves might affiliate the
present organisms to the genus Bacteriastrum, the two groups of forms differ very
notably in other respects. Thus in the former there is as yet no proof that the frustules
are arranged in series, while their general form is cylindrical or cylindroidal — one axis
being much longer than the other — and the awns are never tortuous or dichotomous.
But these characteristics are not wanting in the genus Bacteriastrum, hence the new
series of forms, which unquestionably have certain affinities to Chsetoceros, must be
regarded as forming a transition to the tubulate Rhizosolenise.

The name that has been proposed for the genus has reference to the " broom "-like
appearance of the frustules, and the following generic definition may be given : — Frustula
cylindrica, libera (?) ; valvis convexis, setarum radiantium corona cinctis.

According to the character of the connecting zone and awns this genus may be divided
into two sections, embracing —

A. Forms with simple smooth connecting zones and smooth awns.

B. Forms with complex annulate connecting zones and echinated awns.

Section A. Connecting zones simple and smooth ; awns smooth.

Corethron criophilum, n. sp. (Plate XXI. fig. 14.)

Forma longe cylindrica, valvis producto-convexis, setis tenuissimis. In Antarctico ad
glacies impervios.

This long and perfect little cylinder has a longitudinal axis, which bears to its diameter
the ratio of 14 to 1. The awns are long and very delicate, smooth and radiating in the
same direction at the two extremities. The two valves are extremely convex. This
type occurs not unfrequently at the Antarctic ice barrier, and on floating fields of ice.

Corethron criophilum, n. sp., var. nov. (Plate XXI. figs. 12 and 15.)

Probably the frustules here shown only represent two varieties of the preceding species.
Both differ, however, from the latter (1.) in having the longitudinal axis shorter and

1 xigriQw, broom.

86 THE VOYAGE OF H.M.S. CHALLENGER

the diameter greater, and (2.) in possessing a cylindroidal rather than a cylindrical outline,
while (3.) one has the valves exactly hemispherical, whereas in the other more pro-
tuberant form they are somewhat globular. Whether such distinctions have more than
a varietal significance must at present remain somewhat doubtful, although the evidence
would seem to indicate that a specific value may be ultimately assigned to them.

It may also be noted that the line bounding the upper valve in fig. 12 is undulating,
while in fig. 15 it is straight, but provided with three well-marked involutions which are
directed towards, the centre of the frustule. The radiating awns in fls;. 12 are also
relatively larger and somewhat more elegant than in fig. 14.'

Corethron, sp. (Plate XV. fig. 7.)

The organism here represented, which has been figured from a preparation made on
board the Challenger, is also manifestly a member of the present genus. In it the
awns, which are not represented in full length in the figure, are very strong, long,
and ribbed, and — a circumstance which is characteristic of the genus — they are never
numerous, while their diameter is almost uniform throughout.

As this form has not been observed from its zonal aspect its real specific value must
remain doubtful.

Section B. Connecting zone complex, annulate ; awns echinated.
Corethron hispidum, n. sp. (Plate XXI. figs. 3 and 5.)

Forma cylindrica, annulata ; valvis hemisphsericis spinulosis ; setis costatis, echinatis.
In mari Antarctico.

Although the frustules shown in the present figures undoubtedly belong to the same
species, in only one are the details of structure represented. The connecting zone is
composed of many hoops which are united together, and in this respect, through the
genus Lauderia of Cleve, the genus Corethron approaches that of Rhizosolenia. The awns
are strong, thorny, and costate, become gradually attenuated from their origin to their
distal ends, where they are reduced to mere lines, and spring from the margin of the hemi-
spherical valve, the surface of which is roughened by very minute thorns. The presence
of these thorns has been employed in naming this interesting species.

Corethron murrayanum, n. sp. (Plate XXI. fig. 4.)

Forma cylindrica, annulata ; valvis convexis, la3vibus ; setis costatis, echinatis. In mari
Antarctico.

This somewhat large beautiful cylindrical organism — the longitudinal axis of which is
equal to the diameter — has two convex valves, which have the form of the segment of a
sphere. The cylindrical median zone is annulated, and a definite line of suture, as in

KEPOKT ON THE DIATOMACEyE. 87

Podosira, is seen in the centre. The awns are ribbed, gradually attenuated, and bear
sparsely disposed but -well-defined thorns along their margins. The valves are perfectly
smooth.

This exceedingly elegant Diatom was found in a preparation made on board the
Challenger, and has been named in honour of Mr John Murray.

Corethron (?) sp. (?). (Plate XXL fig. 6.)

The singular spherical organism here represented is provided with a raised zone,
showing a double row of distinct round granules, and a corona of very fine echinated
and slightly arcuate awns. Although it is by no means an easy matter to determine
to what this straDge form belongs, I am of opinion, from the fact that it has been
found associated with other Corethral frustules, that it is a young form of a species of
Corethron, the adult form being reached by the development of a connecting zone. The
entire frustule would thus become cylindrical, and be terminated by two convex valves,
which would be surrounded at their margin by a corona of granules and by circlets of
awns.

Stephanopyxis, Ehrenb.

Greville1 and Professor Walker Arnott, departing from the recognised laws of
nomenclature, have desired to substitute the name Cresswellia for that of Stephanopyxis
established by Ehrenberg. This substitution, however, has not been generally adopted,
notwithstanding the claims that have been advanced in its favour, so that the old
designation remains, the genus being defined as follows : 2 — " Frustules simple or united
into short filaments, in front view orbicular or oblong, composed of two cellulose valves,
each having a crown of teeth, spines or membrane ; central portion obsolete ; lateral view
circular."

Stephanopyxis kittoniana, n. sp. (Plate IX. fig. 5.)

Frustulis globoso-cylindricis, lineariter punctulatis, et per apicum coronam inter se
connexis. Ad insulas Philippinas.

This organism presents a series of four suborbicular frustules, which are united by
means of a corona of capitulate processes, and are granulated or punctated in parallel lines
running in the direction of the longitudinal axis of the series. Although some affinities
are here presented to the Cresswellia turgida of Greville, for example, in general outline

^ 1 See Gregory, Diatoms of the Clyde, p. 64, pi. vi. fig. 109; Greville, Micr. Journ., vol. vii. pp. 165, 166,
pi. viii. figs. 14-16; Trans. Roy. Soc. Edin.,vo\. xxi. p. 536; Ehrenberg, Die Iufusionsthierehen, p. 165;
Monatsber. d. k. Altad. d. Wiss. Berlin, 1844, p. 264; Mikrogeologie, pi. xviii. fig. 4, pi. xix. 13, fig. 6. i
2 Pritchard, op. cit, p.- 826. .

88 THE VOYAGE OF H.M.S. CHALLENGER.

and in the form of the processes, it is to be noted that the latter is areolated instead of being
delicately punctated, and instead of being globoso-cylindrical it is oblongo-cylindrical.
The specific name of this Diatom, which was collected in the neighbourhood of the
Philippines, has been given in honour of Mr Kitton, the well-known English micrographer.

Stephanopyxis rapax, n. sp. (Plate IX. fig. 9.)

Valvis convexis, laevi margine hinc grandiuscularum cellularum ordine cinctis, medio
granulis rarioribus radiantibus ornatis ; a cellularum corona decern aduncse spinae adsur-
gunt. In mari Antarctico.

This small discoid organism was obtained at Station 153,1 lat. 65° 42' S., long. 79° 49'
E., from a depth of 1675 fathoms, in a bottom of blue mud. It is terminated by a wide
smooth border, within which a circle of very large cellules occurs. The disc is orna-
mented with delicate subradiating granules, which diminish slightly in size from the centre
to the periphery. Between the central disc and the cellular corona ten strong claw-like
structures arise — a characteristic which must be held as of specific importance.

Stephanopyxis turris ( = Cresswellia turris, Grev.)

A specimen of this Diatom, which is unquestionably identical with the Cresswellia
turris of Greville,2 was obtained in the Arafura Sea. Although the species may be
readily recognised when viewed in its zonal aspect, it is much more difficult to do so
when seen from its valval side.

Stephanopyxis campana, n. sp. (Plate XIX. fig. 14.)

Oblonga, subcylindrica, valvis campanulatis, cellulosis, hemispherice terminatis et per
paucos obtusos processus coronatis ; cellulis confertis parvis, et marginem versus minu-
entibus. In mari Japonico.

This organism, which is figured in its zonal and most characteristic aspect, was observed
among the Diatoms collected from the Sea of Japan. The valve is subcylindrical and
oblong, and is twice as long as it is broad. The surface of junction is hemispherical and sur-
mounted by a crown of truncated processes. The cavity of the valve is bell-shaped, and
its external surface is densely cellular, the cellules diminishing as they approach the edge.

In many of the above characters this organism agrees with Stephanopyxis apiculata,
Ehrenb.,8 but the cellulation of the latter is described as "not crowded," and as being
"arranged in longitudinal rows." But the most important character of Stephanopyxis
campana, and one which, while it has not hitherto been recognised in any species, must

1 This station was the nearest to the South Pole.

2 Trans. Roy. Soc. Edin., vol. xxi. p. 538, pL xiv. fig. 109.

3 Ehrenberg, Mikrogeologie, pL xix. 13, fig. 6 ; Pritchard, op. cit., p. 826.

REPORT ON TEE DIATOMACE^E. 89

nevertheless be viewed as of specific value, is the gradual diminution in size which the
cellules undergo as they approach the margin.

Lauderia, Cleve.

Professor P. T. Cleve, in his Examination of Diatoms found on the Surface of the
Sea of Java,1 established this genus in honour of Mr Lauder, who has contributed so much
to our knowledge of the genus Cliwtoceros, giving the following definition : — " Frustules
cylindrical, side view orbicular ; covered, at least near the margin, with numerous short
hair-like processes or spines ; front view annulated. Sculpture consists of very fine
puncta."

The Lauderise are very well marked, and cannot readily be confounded with any other
genus ; it is, therefore, not easy to understand the grounds on which Professor H. L.
Smith " failed to mention the genus in his Synopsis of the Families and Genera of Diatoms.

Lauderia annulata, Cleve. (Plate VIII. fig. 7.)

We have here represented an organism, found in a surface gathering made in the
Antarctic Ocean, which is obviously identical with the Lauderia annulata of Cleve.3
It need not therefore be further alluded to.

Lauderia elongata, n. sp. (Plate IX. fig. 4.)

Forma cylindrica, annulata, quatuor vicibus longior quam latior, per superficiem valde
convexam terminata ; frustula in seriem per apicum coronam coherent. In mari Philip-
pin arum.

This organism was found in a surface gathering made in the neighbourhood of the
Philippine Islands. It differs from Lauderia annulata, Cleve, in the following note-
worthy points : (1.) It is four times as long as broad, while in the latter the length
is equal to the breadth ; (2.) its surfaces of junction are almost hemispherical, while in
the latter they are hardly convex ; and (3.) it presents a line of junction around the middle
of the annulated connecting zone — a feature which is not indicated in the figure of
Lauderia annulata given by Cleve.

That these distinctions must be regarded as of specific importance is at once obvious.

Lauderia pumila, n. sp. (Plate IX. fig. 8.)

Frustula annulata cylindrica triplo longiora, punctulorum corona marginak' seriatim
disposita. Ad insulas Philippinas.

The four small cylindrical frustules here represented are annulate, and are in this
respect similar to the species last described. In the present case, however, the contiguous

1 Bihang k: Svensk. Vet. AJcad. HandL, Band I. No. 11, Stockholm, 1873.
" The Lens., vol. i., 1872. » 0p. cit.f p. 8, pi. i. fig. 7.

(bot. chall. exp. — part iv. — 1886.) D 12

90 THE VOYAGE OF H.M.S. CHALLENGER.

surfaces do not show the well-defined points which are so manifest in Lauderia
anmdata, Cleve, and Lauderia elongata, described above ; on the contrary, a crown of
marginal granules is alone represented. If then this organism be accepted as a Lauderia
the definition of the genus must be extended to embrace both these conditions.

Lauderia (?) moseleyana, n. sp. (Plate XXIV. fig. 9.)

Forma cylindrica, annulata, parum longior quam latior ; liuea suturali excentrica ;
margine granulato. In mari Arafura.

This cylindrical Diatom from the Sea of Arafura is annulated and terminated by fiat
surfaces. The rings, however, though still of considerable diameter, are markedly narrower
than in the preceding types. The individual frustules of a series are terminated at each
extremity by a distinctly granulated line, and each at the same time presents a well-marked
transverse line of suture which crosses the frustule nearer one end than the other. These
lines are approximately at the same distance from the granulated terminal bands on both
sides of each granulated area, so that two adjoining wider and two adjoining narrower
zones succeed each other regularly.

The specific name has been given in honour of Professor Moseley of Oxford, who
accompanied the Challenger Expedition as one of the Naturalists.

Rutilaria, Grev.

In 1863 Greville1 established this genus for some rare and singular fossil organisms
discovered in the deposits of Monterey and Barbados. Having only seen the valve of
the frustule, the original definition was not, as might be expected, exact ; but, after the
discovery of other frustules adhering to one another, he amended this in 1866 in the
following manner :2 — "Frustules very compressed, cohering into a short filament; valves
slightly elevated at the angles, with a central glistening nodule prolonged into two short
linear, obtuse processes ; the margin pectinate-ciliate."

The affinity of this genus with that of Biddulphia, which has also been recognised by
Greville, presents no difficulty after a simple examination of its valval side, while the
presence of a large central convolute nodule prevents it from being confounded with the
genus Nitzschia. On the valval side of Rutilaria there may also be seen a well-defined
elevation at each of the extremities. In Biddulphia, on the other hand, terminal processes
or cornua occur, while generally the centre is convex and salient, and the adjoining frustules
become united to one another by means of two long spines or setae.

1 Quart. Journ. Micr. Sci., n. s., vol. iii. p. 227. The original definition ran as follows: "Frustules free,
elongated, compressed, with a convolute or nodulose central nodule (no median line or terminal nodule), and
minute radiate or decussato-punctate structure. Valve (linear, keeled?) with a longitudinal row of puncta."

2 Trans. Micr. Soc. Lond., n. s., vol. xiv. p. 124.

REPORT ON THE DIATOMACE.-E. 91

Owing to the great rarity of the Rutilarise, the present forms have not been observed
from the zonal aspects; and, with the exception of Rutilaria epsilon, Grev.1 — whose
convolute nodule recalls the shape of the Greek letter from which its specific name has
been derived — it has been found impossible to determine the form of the nodule.

Hitherto no living specimens belonging to this very interesting genus have been
recorded, and it is therefore of the highest importance that such specimens should have
been procured during the Challenger Expedition. It is true, indeed, that Professor Cleve,
of Upsala, records in his Memoir f on some New and Little Known Diatoms, Plate iv.
fig. 57, a, b, under the name of Rutilaria recens, Cleve, two curious lanceolate forms
which were brought home by the "Eugenie" Expedition3 from the Galapagos Islands.
These, however, have not the slightest trace of the convolute central nodule — a circum-
stance which should serve to exclude the organisms in question from the genus to which
they have been ascribed by him — unless, perchance, it can be proved that one of the
valves is normally devoid of such a nodule, as happens in Cocconeis. But this peculiarity
has not hitherto been substantiated, and it seems an improbable one in the case of a genus
whose frustules are disposed in series.

Of the specimens collected by the Challenger, one was obtained in a surface gathering,
and so must have been in a condition of active vitality.

Rutilaria tulkii, n. sp. (Plate XVIII. fig. 11.)

Valvis elliptico-lanceolatis, apicibus productis, rotundatis et in 'partem elevatiorem
exeuntibus ; denticulis rariusculis marginalibus circumductis, et aliis medio irregulariter
distributis ; striis tenuissimis medio radiantibus, ceterum decussatis, nodulo centrali con-
volute Ad Samboangan ex insulis Philippinis.

This organism was found in a surface collection, in which hardly anything but small
animals occurred. After washing these with distilled water, a preparation composed
chiefly of spicula wa3 obtained, but a more minute examination revealed the presence
of the two small elegant frustules here figured, which were similar to one another, and
arranged in a cruciform manner. I at first believed that this peculiar arrangement corre-
sponded to the intersection of the axes of the figures, as occurs in some other genera, such
as Camjiylodiscus, but after submitting this view to my friend, Mr Tulk, I became con-
vinced that the frustules are enabled to have this disposition because of the peculiar form
of the central nodules which form the real points of union between them.

A similar arrangement was also observed by Dr Gray, who had artificially caused two
Kutilarian frustules to move from their normal parallel position.

Rutilaria tulkii possesses a small but elegant elliptico-lanceolate form, having its two

1 Quart. Journ. Micr. Sci., n. s., vol. iii. p. 228, pi. ix. fig. 1.

2 Kongl. Svenska Vetensh.-Akad. Handl, Band xviii. No. 5, Stockholm, 1881.

3 The Expedition of the Royal Swedish Frigate "Eugenie" took place during the years 1851-53.

92 THE VOYAGE OF H.M.S. CHALLENGER.

extremities notably prolonged, and terminated by a well-defined elevation in the form of
a process. Around the margin there is a line of sparsely disposed granules or denticules,
and the surface of the valve is delicately granulated in a radiating and decussate manner,
a few more prominent granules being irregularly disseminated over the same area. The
larse central convolute nodule is so arranged as to render the exact determination of its
form impracticable.

The specific name which I have adopted is in honour of my friend, Mr Tulk, who has
aided me in many ways with my work.

Rutilaria edentula, n. sp. (Plate XVIII. fig. 14.)

Forma grandiuscula. lanceolata ; apicibus longe porrectis, rotundatis in elevatiorem
partem terminantibus ; denticulis partem centralem cingentibus, hinc rariusculis, dein
deficientibus ; striis egerrime perspicuis, medio radiantibus ; nodulo convoluto distiuctiore.
Ad insulas Philippinas.

This form was found among the Diatoms obtained in a sounding made at Station 211,
in a depth of 2225 fathoms. It is much larger than the preceding, and is very different in
outline and structural detail. It is lanceolate, but both extremities are much prolonged,
and bear at their tips elevations which rise above the plane of the valve. The striation
of the latter is extremely delicate in the central part, and can only be determined by the
use of oblique light, and the exquisite homogeneous immersion objective of Zeiss. The
marginal denticules are thicker than in Rutilaria tulkii, but they become more rare towards
the extremities, and finally disappear. The central nodule is less prominent, and appears
to be doubled up.

The specific name which has been applied is intended to recall the absence of marginal
denticules at the extremities.

Melosira, Ag. ( = Gallionella, Ehrenb.)

This genus was established by Agardh in 1824 for a series of frustules, having cylin-
drical, discoid, or globose outlines, and connected into cylindrical Conferva-like filaments,
one or two lines passing round each frustule near the centre. Since the frustules are
arranged in series, and assume somewhat the appearance of Confervas, it is not surprising
that, towards the end of the last century, 0. F. Miiller described, under the name of
Conferva armillaris,1 the Melosira varians of Agardh,2 the siliceous nature of the walls
not being then suspected.

The union of one frustule with another is either effected by means of convex surfaces or

1 Nov. Act. Holm., 1873, pi. iii. fig. 67. Rabenhorst places this frustule as a synonym under Aehnanthes
longipes, Ag. (Flora Europaea Algarum, p. Ill, 18G4).

2 Ralfs, Ann. and Mag., pi. ix. fig. 5; Rabenh., Siissw. Diat., pi. ii. fig. 4; Smith, Synop. Brit. Diat.,
vol. ii. p. 57, pi. Ii. fig. 332.

[KEPORT.ON THE DIATOMACE.K. 93

of plane spinous or denticulate surfaces, and on this account W. Smith, acting on the sug-
gestion of Thwaites,1 separated from the genus Melosira those forms which are connected
in the manner last named, grouping them into the genus Orthosira. This scheme has,
however, not been generally adopted, the character in question not being regarded as of
Generic significance, nor can it ever be defended unless the jrenus Melosira should
ultimately become so rich in species that its separation into two parts should become
necessary for the facilitating of systematic work.

Melosira sol, Ehrenb., var. nov. (Plate X. fig. 3, Plate XVII. fig. 13, and Plate XXI.
fig. 7.)

Among the Melosira? contained in the Challenger collection, the present forms, which
were collected in the neighbourhood of Kerguelen Island, are of great elegance. The two
complete frustules, which are united to one another by their valval sides (Plate XXI.
fig. 7), although having very close affinities to Melosira sol, Ehrenb.,'2 cannot be regarded
as identical with the latter. The description given by Pritchard3 of Melosira sol, as
determined by Ehrenberg, corresponds with the characters of the variety now in question
with the exception of the fact that, in the latter, one additional feature has to be added
— namely, the circumstance that between the radiating lines or plicae, which run centripe-
tally from the periphery and quickly disappear, more minute radial lines occur and also
soon vanish. Although this is not clearly indicated on Plate XXI. fig. 7, owing to the
insufficiency of the magnifying power employed, it is shown in Plate X. fig. 3 under a
power of 800 diameters.

With the comparatively imperfect instruments used by Ehrenberg, it is not surprising
that these fine fines escaped his observation.

Melosira costata, Grev., var. nov. (Plate XXIII. fig. 5.)

Among the pelagic organisms collected in the Bay of Yedo, the very elegant series
of frustules represented in the present figure were not unfrequently met with. That
these frustules belong to the genus Melosira cannot be questioned, and the presence of
the longitudinally disposed costules, which are contiguous to the surface of junction of two
adjoining valves, seems to point to their identity with Melosira costata, Grev. On
comparing, however, Greville's figures4 with the present one, the following differences
may be noted : — (1.) The smooth areas around the different frustules of Melosira costata
are almost equal, and in many more restricted than in the present form, where they are
much larger and unequal ; and (2.) in the typical Grevillean species these smooth areas

1 Thwaites, Ann. Nat. Hist, March 1848.

2 = Gallionella sol, Ehrenberg, Mikrogeol., pi. xxxv. A. 22, fig. 12. This species rivals Melosira arenaria,
Moore, in size, Ealfs, Ann. Nat. Hist., vol. xii. pi. ix. fig. 4.

3 Pritchard, op. cit, p. 819. * Trans. Micr. Soc. Lond., n. s., vol. xiv. p. 77, pi. viii. figs. 3-6, 18G6.

94 THE VOYAGE OF H.M.S. CHALLENGER.

are oval, thereby indicating that the costules are of different lengths. Such distinctions,
however, cannot be viewed as having more than a mere varietal significance.

Melosira westii, W. Sm. (Plate XXI. fig. 16.)

This frustule was observed on the coast of Australia, near Sydney, in a sounding
made at a depth of 950 fathoms; but it is to be noted that the valves, instead of
having "truncated apices," as described by Smith,1 are depressed, a slight hemispherical
projection occurring in the centre of the depressed area. These minor differences, however,
are not essential, but they may be well seen on comparing the figures on the present plate
representing the valval and zonal aspects.

We now pass to those species that are provided with a double furrow, and have
their zonal sides divided into three parts.

Melosira hyalina, n. sp. (Plate XXI. fig. 1.)

Frustulis cylindraceis quadruplo longioribus, medio binis sulcis distinctis, apicibus
hemisphericis ; granulis imperspicuis ; zona connectiva binatim junctis. In mari Pacifico
meridionali.

Here each frustule is of cylindrical form, and is four times as long as it is broad.
Two furrows occur towards the middle, and are placed at a distance from one another-
equal to the length of the transverse axis of the frustule. In the figure two frustules
are seen to be united to one another by a common belt, and both are terminated by
hemispherical surfaces. It is noteworthy, however, that the surface of junction between
the adjoining frustules is flat. This curious circumstance is of little importance, and may
probably be explained in the following manner. In the final stages of increase in size, the
extremities are prevented from dilating by the resistance which is offered by the common
belt which has already ceased to lengthen, so that lateral expansion occurs, and a flat
septum is the result.

The prolonged cylindrical form of the frustule and the prominent hemispherical
extremities are, however, of specific importance.

This interesting form was collected near the Pacific coast of South America in a depth
of 2225 fathoms.

Melosira thaitiensis, n. sp. (Plate XXI. fig. 2.)

Cylindracea, triplo longior quam latior ; binis sulcis medio signata ; apicibus convexo-
complanatis, exterius granulatis ; ordine quadrato punctulatis. In portu Thaiti.

This Diatom, from the port of Tahiti, is composed of cylindrical frustules, which are
three times as long as wide, and terminate in convex depressed extremities or surfaces of

1 Synopsis of British Diatomacese, vol. ii. p. 59, pi. lii. fig. 333.

RETORT ON THE DIATOMACE.E. 95

junction, -which are crowned by denticules. Each is provided with a double furrow near
the centre, and the partitions are densely granulated in a quadrate manner.

This new species has been named from the locality in which it was first observed.

Melosira glomus, n. sp. (Plate XXI. fig. 10.)

Frustulis cylindraceo-suborbicularibus, dense quadrato ordine punctulatis, medio zona
laevi cinctis. Ad portum Thaiti.

The spheroidal frustules which compose this Tahitian Diatom present very many
granulations save in the smooth linear zone which surrounds each in the middle region.
The entire series is irregular on account of the different conditions of development of the
component parts, and the valves also vary from the subhemispherical to the depressed
convex form. But for the condition of the median band this form would ao-ree with the
Melosira labuensis of Cleve ; but its true specific value cannot, in view of the characters
quoted, be called in question.

The specific name has been given from its resemblance to balls of coarse thread.

Plate XXIII. fig. 4, represents a variety of Melosira glomus from the port of Tahiti ;
it may be named Melosira glomus, n. sp., var. major nov. ; while Plate XXI. fig. 11, is a
doubtful species of the same genus collected in the Sea of Japan.

Thalassiosira, Cleve.

This genus was established by Professor Cleve of Upsala, in his paper On Diatoms
from the Arctic Sea,1 and was characterised by him in the following manner : — " Side
view circular, with a row of submarginal spines ; sculpture very minutely cellular,
cellules arranged in radiating and curved lines, crossing each other. Front view quad-
rangular, with truncate angles, connecting membrane broadly linear, without any distinct
sculpture. Frustules, in the living state, connected by means of a central fine thread of
mucus into long filaments."

From specimens of remarkable purity obtained from the Arctic Sea, and more
especially from Davis Straits, Cleve established the species Thalassiosira nordenskioldii.'2
This Diatom occurs in these regions in such enormous quantities as to colour the sea for
many miles, but I have not observed it in the collections made by the Challenger in the
Antarctic Ocean among the South Polar icebergs. It is, however, of the greatest importance
to note that I have recently found numerous specimens of Thalassiosira along with many
other Diatoms in the alimentary canals of two Echini which were procured at a depth of
1340 fathoms in lat. 41° 15' N., long. 65° 45' W. These frustules constituted the food of

1 Bihang b. Svensb. Vet. Abaci. Hanoi. , Band i., No. 13, p. 6, pi. i. figs, la, 16, \c, Id, Stockholm,
1873. 2 Op. at., p. 7.

96 THE VOYAGE OF a M.S. CHALLENGER.

the animals in question, and so must have been growing in their immediate vicinity at a
depth at which, notwithstanding the great distance from the poles, the temperature of the
water approaches zero.

Thalassiosira nordenskioldii, CI eve, var. nov. (Plate XXX. fig. 4.)

This figure represents a Diatom with a characteristic corona of submarginal spines
which was found in the vicinity of the Bermudas ; while fig. 4 bis of the same Plate is a
reproduction from the Synopsis des Diatomees de Belgique of Dr van Heurck1 of the
Thalassiosira nordenskioldii of Cleve, the former being magnified 850 diameters, and the
latter 600 diameters. In the frustule, now for the first time recorded, the radial striation
is so minute that it is difficult to observe even in dried specimens, while the number of
prickles or submarginal spines is much greater than in the Arctic form. Hence, although
the two might be regarded as distinct species, I prefer to look upon the deep-sea form as
a variety of the surface form of Cleve until such time as the former is procured in a living
condition, and its history more accurately known.

Isthmia, Ag.
Isthmia enervis, Ehrenb., var. japonica, nov. (Plate XXV. fig. 5.)

The only form belonging to this genus which has to be recorded is one from the Sea of
Japan, which presented an evident analogy to Isthmia enervis, Ehrenb.,2 yet differed from
it in several respects. Thus (1.) the profile of the frustule on the zonal side is simply
trapezoidal in Ehrenberg's species, while in the Japanese form the two oblique terminal
lines are more or less undulating — a distinction upon which much weight cannot be placed.
(2.) At the margins of the median band, and adjoining the two sutural lines, a row of
larger granules occurs in both cases, as in Biddulphia ; but the band itself is ornamented,
in Isthmia enervis, by small round quincuncially-arranged granules, while in the present
case the granules are somewhat oval, and are not arranged in a quincunx manner. (3.) It
may also be observed that in the Japanese Diatom the strong granules at the two
extremities are somewhat different from one another, being on the more obtuse side sub-
quadrate and of larger size than on the other side, where they are subrotund.

Notwithstanding these differences, the frustule from the Sea of Japan must be re-
garded as possessing merely a varietal importance, and the distinctive name that has
been applied has reference to the locality in which it has been first collected.

L1 PL Ixxxiii. fig. 9.
2 Ehrenb., Inf., p. 209, pi. xvi. fig, 6 ; Kutz., Bacill., pi. xix. fig. 4 ; Smith, Synop. Brit. Diat, vol. ii. p. 52,
pi. xlviii. ; Jan. et Eabenh., Hondur., p. 9, pi. iv. fig. 13 ; Hohenack., Alg. Mar., N. 454 ; Ealfs, Ann., vol. xii.
pi. viii. fig. 1. Compare also Isthmia obliquata var. tciiuior, Ag. Conspec, p. 55, and Conferva obliquata, Eng.
Bot., tab., 1869.1

REPORT ON THE DIATOM ACE/E. 97

Eucampia, Ehrenb.

This genus is defined by Fritchard in his History of the Infusoria (p. 937) as follows : —
" Frustules hyaline, imperfectly siliceous, cuneate, without terminal puncta, united in a
jointed spiral filament." With regard to this definition, it may be noted that although,
with the relatively imperfect microscopes used by Ehrenberg and W. Smith, by means of
which the minute sculpturing could not be detected, it was admissible to speak of the
frustules as hyaline, it is now possible to detect minute rows of points, so that that term
can no longer be made use of.

Akin to the genus Eucampia, Professor Cleve of Upsala, in his paper entitled Exami-
nation of Diatoms found on the Surface of the Sea of Java,1 instituted another genus
which he named Molleria, in honour of the well-known mounter of Diatoms, and crave a
definition which agrees with that of Eucampia, except that, in Molleria, the extremities
of the valves are produced into long processes, and the connecting membrane is orna-
mented with numerous costse (rudiments of diaphragms ?). That the presence of long
processes is an insufficient character upon which to establish a genus is manifest, and in
his Synopsis of the British Diatomaceas, W. Smith,2 in representing a superb specimen of
Eucampia zodiacus, Ehrenb., indicates a slight puncturing on the valve, and traces of
several costae on the connecting zone. Hence the genus Molleria cannot be accepted on
the grounds quoted by Cleve. But in Molleria cornuta, Cleve, there exists a central nodule
which — as I observed in the case of a valve found in the original collection from the Sea
of Java, which was kindly placed at my disposal by Professor Cleve — is ornamented by
minute radiating puncta. Hence it only remained to be shown that, in Eucampia, no
central nodule occurred, in which case the genus Molleria could be reserved for those forms
of Eucampia possessing such a structure, and defined as indicated on page 98.

Among the Challenger collections, series of cuneate frustules, without the terminal
point or appendage, and having all the peculiarities required by the definition of Eucampia,
Ehrenb. — except that, instead of being imperfectly siliceous or hyaline, they were fur-
nished with large granules — have been seen on several occasions. Hence the characters
of Eucampia, as established by Ehrenberg, may be emended thus : — Frustula cuneata in
duas insequales extremitates abrupte desinentia et in seriem spiralem conjuncta.

Eucampia balaustium, n. sp. (Plate XVIII. fig. 5.)

Valvis ellipticis convexis superne in duos insequales processus truncatos desinentibus et
medio inflatis ; granulis grandiusculis nullo certo ordine stipatis. In mari Antarctico.

Each frustule here consists of two elliptical valves, which terminate above in two
unequal imperfect processes, between which a slight protuberance exists. The valves,

1 Bihang k. Svensk. Vet. Akad. Handl, Bel. I. No. 11, p. 7.

2 Smith, Synop. Brit. Diat,, vol. ii. p. 25, pi. xxxv. fig. 299., and Suppl., pi. lx. fig. 299.

(bot. CHALL. EXP. PART IV. 1SSG.) J) 13

98 THE VOYAGE OF H.M.S. CHALLENGER.

when seen in the zonal aspect, are symmetrical in themselves, but each is unsymmetrical
with reference to the transverse axis. Their surface is ornamented with large irregularly
disposed granules. No distinct connecting zone is seen, the two adjoining valves being
immediately united. In some frustules a smooth belt may be observed between the
two valves.

The specific name has reference to the form of the frustule.

Eucampia balaustium, n. sp., var. minor, nov. (Plate XVIII. fig. 6.)

We have here represented a form which must be regarded as a mere variety of the
preceding type, differing from it in its smaller size, and in having the extremities some-
what blunted.

It is important to note that, as may be observed in this figure, the last valve in the
series differs from the others, having the two extremities less prominent and of equal length.
In this circumstance Eucampia resembles Bacteriastrum, Clisetoceros, and probably also
Hemiaulus.

Molleria (Cleve), Cstr.

As above referred to (see Eucampia, p. 97), this genus, instituted by Cleve in honour
of M. Moller, of Wedel, Holstein, does not differ essentially from Eucampia, except by
the presence of a nodule in the centre of the valve, and by possessing two unequal ter-
minal processes. Hence the genus may be defined as follows : — Frustula cuneata in
spiralem seriem conjuncta ; valvis ovalibus centrali nodulo instructis, et in duos inajquales
processus desinentibus ; zona connectens plerumque crebre costata vel plicata.

Molleria cornuta, Cleve. (Plate XXV. fig. 8.)

We have here shown a small specimen of Molleria cornuta, Cleve,1 which has hitherto
been unique in its kind. Two frustules are seen in series, and the shorter of the two
has been checked in the act of fission, so that two new valves, in process of formation,
may be observed through the belt. Similar observations have already been made in the
case of Biddulphia, and in other types possessing two symmetrical valves.

Molleria antarctica, n. sp. (Plate XVIII. fig. 8.)

Forma cuneata grandiuscula, valvis ellipticis vix perspicue punctulatis et nodulo cen-
trali instructis ; cingulo costulato vel plicato, bipartito ; processibus terminalibus parum
productis. In mari Autarctico.

1 Cleve, Examination of Diatoms found on the surface of the Sea of Java, Bihang k. Svensk, Vet. Akarf.
IlandL, Bd. I. No. 11, p. 7, Stockholm, 1873.

REPORT ON THE DIATOMACE/E. 99

This Antarctic Diatom differs from the Mollcria cornuta of Cleve in the following
respects : — (1.) The greater size of the frnstules ; (2.) the greater shortness of the terminal
processes; (3.) the inferior degree of development of the connecting zone; and (4.)
the bipartite condition of the connecting zone ; the costa?, moreover, of the two hoops of
adjoining valves often diverge more or less — a peculiarity which has not been observed
in Cleve's species. Although taken singly, these differences probably are not of very
great importance ; it cannot be doubted that the combination of all of them is sufficient
to warrant the establishment of this new species.

Hemiaulus, Ehrenb.

Seeing that the frustules of this genus are provided with two processes terminated by
a thorn or claw, we are led to conjecture, in opposition to the assertion made by Pritchard,1
that the frustules exist in a concatenate or seriate form, since the spine can only be inter-
preted as a connecting organ. This conjecture I have been able to verify more than once
by observing two or more frustules linked together, and in such a manner that their ter-
minal spines constantly alternate with one another. Thus of the two spines of one valve,
one is superposed to the corresponding spine of the adjoining frustule, while the other is
covered by its corresponding spine. This peculiar method of union may also be studied
in the genus Ch&toceros, but its signification is by no means easy to understand. The
concatenation of Hemiaulus was observed, and accurately described, by Heiberg, in his
work entitled Conspectus criticus Diatomacearum Danicarum, plate i. a.

In Hemiaulus, as in Cliietoceros and Cocconeis, the terminal valve differs somewhat
from the others — a circumstance which must be carefully borne in mind in order not
to multiply species unduly. The overlooking of this fact explains why, in Pritchard's
History of the Infusoria, the different condition of the processes, as elongated in one valve
and truncated in the other, is given as a specific character of Hemiaulus antarcticus,2
Ehrenb. But the processes of the terminal valve are truncated in all the species, since
the spine becomes superfluous where union with another frustule does not take place.

Hemiaulus, sp. ? (Plate XXI. figs. 9 and 13.)

The two valves of Hemiaulus here figured must be regarded as terminal, but their
specific determination must for the present remain doubtful. The only circumstance — ■
apart from the more crowded condition of the granules — which would seem to indicate
that the valve represented in fig. 13 belongs to a distinct species from that shown in
fig. 9, is the existence of a well-defined granule, surmounting its two obtuse terminal
processes, instead of the simple spine or process which is peculiar to the latter.

1 Pritchard, op. cit., p. 851.

2 Ehrenberg, Mikrogeologie, pi. xxxv. A 22, fig. 15 ; Pritchard, op. cit., pi xi. fig. 54.

100 THE VOYAGE OF H.M.S. CHALLENGER.

Hemiaulus glacialis, n. sp. (Plate XXV. fig. 4.)

Processus laterales elongati truncati, mucronati ; intermedia frustuli inflatione utrinque
septata ; granulis perspicuis stipatis. In mari Antarctico.

Tins valve, which is perfect, is furnished with two long parallel processes, which are
truncated at the ends, and upon each of which a well-defined claw-shaped terminal spine
projects from the internal part. Between the long processes, and in the centre of the
valve, a slight protuberance or swelling occurs, indicating, when seen in profile, that the
plane of the valve must be crossed by two septa. This frustule, which has not been
hitherto figured, has been named glacialis, on account of its Antarctic habitat.

Cerataulus, Ehrenb.

Professor H. L. Smith1 has united the genera Cerataulus, Odontella, Zygoceros, and
others, under the single genus Biddulphia — a circumstance which has rendered the classi-
fication much more intricate, and which therefore should not be adopted, notwithstanding
the intimate relation that exists between all these organisms. The oval shape of the
valves, the alternation of the two sublateral processes, and the strong horn-like spines,
in the case of Cerataulus, compared with the disposition of the frustules in zigzag chains,
or their adhesion to each other by alternate angles, the elongated and often septate
character of the valves, and the presence of subterminal processes in Biddulphia, must be
regarded as essential distinctions between the two genera. This view has also been adopted
by other naturalists, among whom Dr Eulenstein may be mentioned. This observer, in
issuing the first hundred of his typical preparations, gave, under Nos. 9 and 10, Cerataulus
Isevis, Ehrenb.,2 and Cerataulus turgidus, Ehrenb.,3 and under No. 11 Biddulphia pidchella,
Gray,4 recognising, as an essential characteristic of Cerataulus, the possession of strong

1 Pritchard, op. tit., p. 847. The Lens, vol. i. 1872, p. 89.

2 = GalUonella, sp..1 Bailey, 1842; Sil. Journ., vol. xlii. pi. ii. fig. 8.

Biddulphia lasvis; Pritchard, op. cit., p. 847, pi. vi. fig. 7; Roper, Micr. Journ., vol. vii. p. 18, pi. ii.

figs. 24-26.

Odontella pohjmorpha, Kiitz., Baa, 1844, pi. xxix. fig. 90; Kiitz., Spec. Alg., 1849, p. 136.

Jsthmia polymorpha, Montagne ) n , ,- -„-... 0 ., ,„,n ,OP

,,, . , ,. ,.- , 7 . } Quoted in Kutz., Spec. Alg., 1849, p. 136.

Melosira thermahs, Meneghuu J

a Ehrenberg., Monatsber d. k. Akad. d. Wiss. Berlin, 1843, p. 270; Bailey, 1850, Micr. Obs., pi. ii. figs.

26 and 27.

= Biddulphia turgida, Smith, Synop. Brit. Diat., vol. ii. p. 50, pi. lxii. fig. 384; Roper, Micr. Journ.,

vol. vii., p. 17, pi. ii. fig. 23.

4 Gray, Nat. Arr. of Brit. Plants, vol. i. p. 294; Ralfs, Ann. Nat. Hist, 1843, pi. viii. fig. 3.

= Conferva biddulphiana, Eng. Bot., 1807, vol. xxv. tab., 1762; Dillwyn, Brit. Conf., 1809, p. 52.

Diatoma biddulphianum, Ag. Syst. Alg., 1824, p. 5, and Conspec. Crit. Diat., 1830, p. 54.

Hooker's Brit. Flor., 1833, p. 404; Harvey, Man., 1841, p. 201.

Biddulphia trilocularis, Kiitz., Baa, 1844, pi. xix. fig. 89, and Spec. Alg., 1849, p. 137.

Biddulphia quinquelocularis, Kiitz., Baa, 1844, pi. xix. fig. 1, and Spec. Alg., 1849, p. 137.

Biddulphia septemlocularis, Kiitz., Baa, 1844, pi. xix. fig. 2, and Spec. Alg., 1849, p. 138.

Biddulphia australis, Mont., PI. CeL de Cuba, 1845, p. 5.

REPORT ON THE DIATOM ACE;E. 101

born-like processes, which are not situated on the protuberances between the two processes,
but alternate with them.

Cerataulus turgidus, Ebrenb., var. polyceros, no v. (Plate XXVI. figs. 6 and 8.)

The organisms here figured belong to the genus now before us, although the valve is
of considerable size, and is distinguished by the presence of some well-defined irregularly
scattered granules. That both forms, which were collected in the same locality — the Sea
of Japan — represent the same species is unquestionable.

The general aspect and sigmoid curve of the frustule and connecting zone point to
affinities with the Cerataulus turgidus of Ehrenberg. On comparing, however, our two
figures with the Cerataulus turgidus of the Typenplatten of ■ Moller, and with the
different specimens which we find in the typical preparations of Eulenstein, they agree in
all respects, except that Moller's specimen is provided with a remarkable crown of thorns,
along with very small irregularly scattered valval thorns, while in the frustules of Eulen-
stein's preparation only the very minute thorns occur — differences that are of little
moment. In the case of the Challenger forms, on the other hand, the minute thorns are
entirely absent ; and, instead of a strong long point or cornu alternating with the two
short processes, there is present a group of strong, short obtuse cornua. These marks
of distinction have been looked upon as sufficient to justify the establishment of a new
variety.

Biddulphia, Gray.1

Although, as above stated, I cannot adopt the view of Professor Smith, according to
which he would unite the genus Cerataulus, Ehrenb., to that of Biddulphia, I am of
opinion that the union of the genera Zygoceros, Ehrenb.,2 Odontella, Ag.,3 Porpeia, Bailey,4
and Amphitetras, Ehrenb.,5 to that genus as advocated by him is well founded. Thus the
only difference of any importance between Zygoceros and Biddulphia is the circumstance
that the frustules of the former are isolated. A Biddulphia, however, may also be found
solitary, not only as a result of having been casually detached from a series, but because
its mode of reproduction supposes that, at some time, the frustule should occur free.
Again, only the absence of spines on the margins of the valve distinguishes Odontella

1 Gray, Nat. Arr. of Brit. Plants.

2 Compare Zygocerus mobilensis, Bail., Smithson. Contrib., 1859 ; Zygoceros rhombus, Ehrenb., Kreideth.,
p. 80, N. 61, pi. iv. fig. 11; Kiitz., Baa, pi. xviii. fig. 9; Zygoceros surirella, Ehrenb., Kreideth., loc. cit.,
fig. 12, Kiitz., loc. cit., fig. 12; Roper, Micr. Journ., voL vii. pi. vi. figs. 11 and 12.

3 Ag. Consp., 56, Kiitz., Bacill, pi. xviii. fig. 89, pL xxix. figs. 88 and 90 ; Wigand in Hedwigia, vol. ii.
p. 45, pi. vii. fig. 21, Montague, SylL, p. 473.

4 Bailey, MS. Pritchard, op. cit, p. 850, pi. vi. fig. 6.

5 Ehrenb., Kreideth., p. 62, N. 22 ; Ralfs, Ann. and Mag., vol. xii. pi. viii. fig. 5 ; Smith, Synop. Brit.
Diat., vol. ii. p. 47, pL xliv.=fig. 318 ; Jan. et Rabenh., Hondur., p. 4, pi. i. fig. 3 ; Heiberg, Conspec, p. 42, &c

102 THE VOYAGE OF H.M.S. CHALLENGER.

from Biddulphia; while, like Zygoceros, Porpeia differs from it merely by livincr in an
isolated condition. In Amphitetras the valves are subquadrate, and provided with four
pseudo-openings at the angles, while in Biddulphia there are only two. Yet the struc-
ture of both of the latter types is the same, while the number of pseudo-openings depends
on the form of the valve — there being one at each angle ; thus some triangular forms, or
forms with four or more angles, each provided with a pseudo-opening, occur, and have
hitherto been named Triceratium, but these must in future be enrolled among the Bid-
dulphia. Amphitetras agrees with Biddulphia in its zigzag appearance, the union
between its frustules being effected by means of an isthmus or small angular cushion ;
and, in its structure, it agrees with the latter so far as to possess a pseudo-opening at
each extremity. Moreover, Amphitetras antediluviana, Ehrenb.,1 i3 frequently found
associated with the Biddulphia pulchella of Gray, which is but another proof of the
affinity of the two types.

Biddulphia pulchella, Gray, var. major, nov. (Plate XXIII. fig. 6.)

We have here represented a very large valve divided by six transverse lines or septa,
the characters of which exactly agree with the definition of the Biddulphia pxdchella of
Gray, except that the strong thorns, which are constantly found in groups of two or three
in the latter, are absent in the former. For this reason, the present form has been referred
to as a variety, a view which is, in part, substantiated by its unusually large size. It was
collected in the Caroline Archipelago.

Biddulphia reticulata, Roper, var. inermis, nov. (Plate XXVI. fig. 9.)

Another beautiful form of this geuus is here shown. It possesses reticulated parietes,
and two opposite capital processes, towards the extremities of which the network dis-
appears. Between the two latero-terminal processes the end of the frustule is slightly
inflated or convex. In these respects it resembles Biddulphia reticulata, Roper,2 but it
may be distinguished from the latter by the absence of superficial puncta.

It may be observed that in the present figure the two frustules are united together by
means of two hoops or belts, which embrace each other, and each of which shows a dif-
ferent internal structure from that of the valve. This appearance enables us to understand
the transitory nature of the belt in some forms, as has been already pointed out by Dr
Wallich, who observed that it did not form an integral part of the frustule, but was joined
to it by means of several distinct oblong cellules.

This elegant organism was collected at the Philippine Islands.

1 Pritchard/o/i. cit., pi. xi. figs. 21 and 22.

2 Trans. Micr. Soc. Lond., vol. vii. p. 14, pi. ii. figs. 13-15.

REPORT ON THE DIATOMACEuSl 103

Biddulphia parallela, n. sp. (?) (Plate XXIII. fig. 10 ; Plate XXVI. fig. 7.)

To the genus Biddulphia, aud probably to the section Amphitctras, belong the
forms here represented. In both cases one valve has a shorter diameter than the
other — an appearance that may be explained by the fact that the two valves are not
exactly in the same plane. The system of granulation, however, and the form of the
processes, each of which terminates in a pseudo-opening, shows that the two specimens
belong to the same species.

These interesting types were collected in the Sea of Japan.

Biddulphia pellucida, n. sp. (Plate XXVI. fig. 5.)

Valvis late ellipticis, inconspicue punctulatis ; apicibus rotundatis parum productis ;
valvis ad centrum inflatis. Ad insulas Philippinas.

This organism, which is from the neighbourhood of the Philippine Islands, belongs to
the Odontelloid section. Its sculpturing is so delicate that punctations can only be observed
with great difficulty under oblique illumination, and with the best microscopes, so that it
could only have been shown in the figure by the use of a much greater magnifying power
than has been there employed. The encapsuling of the two valves by means of two hoops
or belts, which extend over one another, is well seen, but no third protective hoop
accessory to the process of deduplication, and possessing a distinct structure, is here found.
The angles of this form are sharply prominent and rounded, while the centre of the valve
is very protuberant.

The specific name, which has been applied to this new species, has reference to the
great delicacy of its ornamentation.

Biddulphia purnila, n. sp. (Plate XXIII. fig. 12.)

E minimis ; valvis hirsutis-convexis immediate connexis, quadrato ordine punctulatis ;
angulis productis conicis. E mari Japonico.

The series of frustules here represented are akin to Biddulphia aurita (Lyngb.) l
Breb., but their extraordinary minuteness and the difference in the outline of the valval
side, as well as the fact that the two valves join each other directly without a cylindrical
intermediate zone, prevents me from uniting the two types. Moreover, in the present

r, 1 Brebessin, Cons, sur les Diat., 1838, p. 12; Ralfs, Ann. Nat. Hid., vol. xii., 1843, pi. viii. fig. 4; Smith,
Synop. Brit. Diat., vol. ii pi. xlv. fig. 319; Microg. Diet., 1856, pi. xiv. fig. 10; Roper, Micr. Journ., vol. vii.
p. 10; Heiberg, Conspec, p. 40; Jan. et Rabenh., Hondur., p. 5, pi. iii. fig. 14; =(1.) Denticella aurita,
Ehrenb., Mikiogeol., 1854, pi. xxxv. A 23, fig. 7; (2.) Denticella gracilis, Ehrenb., Monatsber. d. /:. A/rad. d.
Wiss. Berlin, 1840, p. 207 ; (3.) Diatoma auritum, Lyngb., Tent. Hydro. Dan., 1819 ; Hooker, Brit. Flor., 1833,
p. 404; (4.) Odontella aurita, Ag., Consp. Crit. Diat., 1830, p. 56; Smith, Eng. Bot, t. 2842, fig. 2; Harv.,
Man., 1841, p. 201; Kutz., Baa, 1844, pi. xxix. fig. 88; Kiitz., Spec. Alg., 1849, p. 136.

104 THE VOYAGE OF H.M.S. CHALLENGER.

case, none of the long thorns characteristic of Biddulphia aurita occur. The valves are
hirsute, convex, and ornamented with quadrately disposed punctations, while the angles
are of conical form.

The series, which is formed by the direct union of the frustules, was collected in the
Sea of Japan.

Biddulphia japonica, n. sp. (Plate XXIII. fig. 14.)

Forma parva ; valvis convexis, areolato-punctulatis ; apicibus productis acutiusculis,
ad basim inflatis ; cingulo cylindrico punctulato. In mari Japonico.

This frustule, from the Sea of Japan, is not mucb larger than those of Biddulphia
pumila, n. sp., from which, however, it is at once distinguishable by its areolate ornamen-
tation, and by the fact that the angles which are equally prominent are not simply
conical, but present a protuberance near the base. The cylindrical cingulum is punctated.

Biddulphia, sp. (?) (Plate XXVI. fig. 1.)

The valve here shown is from the Arafura Sea. It possesses an elliptico-lanceolate
form, with rounded, somewhat mammiform vertices, while the surface has distinct granules
which radiate from a very excentric point. Near the extremities there are two small
pseudo-openings. Although this valve cannot be identified with any species hitherto
recorded, it does not afford the means of fully determining the characters of the new
species which it represents.

Biddulphia, sp. (?) (Plate XXIII. fig. 13.)

This organism, like the one last mentioned, cannot be identified with any known
species, and is insufficient to enable a complete conception of the new species, which it
typifies, to be obtained. It is of elliptical form, and its punctations radiate from the
centre, at which a small group of more salient granules or denticules occur.

Biddulphia weissflogii, Grim. (Plate XXVI. fig. 2.)

Forma subquadrata, decussatim punctulata ; apicibus parum productis linea axiali vix
perspicua unitis ; valvis ellipticis, ad centrum subinflatis, duplici brevi cornu excentrico
ad apices instructis.

This perfect frustule is adorned with very minute granules, disposed in a quincuncial
manner. Its extremities are slightly prolonged and rounded, and, when seen from the
zonal aspect, each presents two short strong conical and slightly curved points, which are
quite characteristic of the species.

In Plate c. figs. 1 and 2, of the Synopsis des Diatomees de Belgique, by Dr H. van
Heurck, there is represented a frustule which is identical in its principal characteristics
with that brought home by the Challenger which is now being considered. It has been

REPOKT ON THE DIATOMACE^E. 105

named Biddulphia weissflogii by Grunow, who gives South Africa as its locality. The
Challenger Diatom must accordingly be designated by the same specific name.

Biddulphia (Amphitetras) ornata, Shadb., var. hirsuta, nov. (Plate XXIII. fig. 9.)

The frustule here shown recalls the Amphitetras ornata of Shadbolt,1 as it agrees with
the latter in its quadrate form, its concave sides, the evident granulation, and the reticu-
late linear veining. On comparing the present form, however, with the specimen of
Amphitetras ornata of the Typenplatten of Moller, it may be observed to differ in the
somewhat greater concavity of the walls and the consequent prolongation of the extremi-
ties, and also in the presence of a corona of spines or puncta. Moreover, the precision
with which the details of sculpture are represented in the Typenplatten show that the
valve is not flat, a feature which could not be seen in the present case. Notwithstanding
these differences, the Challenger frustule, which was collected in the Sea of Japan, can
only be viewed as a variety of Shadboldt's species, which, however, should receive the
generic designation Biddulphia, instead of Amphitetras.

Biddulphia (?) sp. (?) (Plate XIX. fig. 1 ; Plate XXX. fig. 9.)

A very singular Diatomaceous frustule is shown on Plate XIX. fig. 1, and again under
a somewhat higher magnifying power on Plate XXX. fig. 9. The two figures have been
given because of the omission of a character of great importance in the former, namely,
the occurrence of a small area, ornamented with small crowded granules, at each of the
rounded angles of the slightly concave side, a peculiarity which occurs in Triceratium,
arcticum, Bright. As only one specimen of this form has been observed, it has been
found impossible accurately to determine to which genus it belongs. The slight pro-
minence of the two adjacent extremities of the concave side seems to point to its being a
Biddulphia, but the marked want of symmetry in the two valves opposes this view,
although this last circumstance may point to the fact that it may have formed part of a
seriate Diatom, of which it may have represented the terminal frustule, thus recalling the
irregular appearances that are found in many forms of Chastoceros and Bacteriastrum.
The characters of the frustule are the following : — Zonal profile of frustule subquadrate,
with one side slightly concave ; angles rounded ; two transverse diaphragms or septa
dividing the frustule into three approximately equal parts ; surface of parietes covered
with thinly disposed subparallel rows of oval granules ; granulation becoming abruptly
crowded and more delicate near the angles on the concave side. Locality : Sea of
Japan.

The frustule was found in a preparation generously placed at my disposal by Dr James

Rae, R.N.

1 Trans. Micr. Soc. Lond., vol. ii. p. 16, pi. i. fig. 10.
(bot. chall. kxp. — part iv. — 1886.) D 14

106 THE VOYAGE OF H.M.S. CHALLENGER.

On Plate XXVI. fig. 10, there may be seen Biddulphia tuomeyi, Bail.,1 and on Plate
XXX. fig. 6, a variety of this — Biddulphia tuomeyi, Bail., var. pacifica, nov., from the
Pacific Ocean.

Biddulphia roperiana, Grev.,2 is represented on Plate XXVI. fig. 4.

Triceratium, Ehrenb.

Soon after Ehrenberg established the genus Triceratium among the anguliferous Diatoms,
frustules- were found which exactly corresponded to its essential characters, but which had
valves with four or five angles, and these frustules were admitted into the genus as excep-
tional forms — Triceratium being defined as possessing " valves rarely with four or five
angles." s Later on, frustules with six, seven, and even twelve angles had to be enrolled, as, for
example, the Triceratium eulensteinii of Grunow (see A. Schmidt, Atlas, Plate lxxv. fig. 6).
Although this last species is probably a real Stictodiscus, it can hardly be doubted that,
if Ehrenberg had foreseen the extension of his genus to embrace polyangulated frustules,
he would have chosen the name Polyceratium. The following is the definition of the
genus Triceratium as given by Pritchard : — " Frustules cellular, free, simple, in lateral
view triangular (rarely with four or five angles)."

More than 200 species have been ascribed to this genus, about 150 being figured in
the Microscopical Journal alone. It is to be remarked, however, that many of these so-
called " species " have no real claim to be regarded as true specific types, and several forms
which have been observed amongst the Challenger material have been relegated to poly-
gonal types of Stictodiscus, although they might at one time have been placed in the
present genus.

Triceratium pulvillus, n. sp. (Plate VI. fig. 8.)

Valvis quadratis, areolatis vel cellulatis ; laterales linese late concavse ; apices acuto-
rotundati ; areolae aequales, hexagonales ; processu terminali nullo. Ad mare Philippi-
narum.

This quadrate frustule from the Philippine Sea possesses greatly hollowed-out sides and
acute rounded extremities. The whole surface is occupied by equal areolae or hexagonal
cellules, there being no trace of any process on the extremities.

Triceratium thaitiense, n. sp. (Plate XIII. fig. 14.)

Hexagonum, lateribus concavis, apicibus rotundatis ; valva punctulis rariusculis signata.
In portu Thaiti.

This hexagonal form was observed in a collection made with a tow-net, at a small
depth in the port of Tahiti. It is provided with concave sides and rounded extremities, but

1 = Zygoceros tuomeyi, Bail., Amer. Journ., vol. xvi. pi. iii. figs. 3-9; Pritchard, op. cit, p. 848, pi. vi. fig. 10.

2 Micr. Journ., vol. vii. pi. viii. figs. 11-13. 3 Pritchard, op. cit., p. 853.

REPORT ON THE DIATOMACEyE. 107

has no processes of any kind from the valve, which is ornamented with small, ill-defined,
thinly placed, almost punctiform granules, which decrease towards the extremities.

Triceratium ferox, n. sp. (Plate VI. fig. 4.)

Triangularis ; apicibus cuneato-rotundatis, prominulis ; lateribus rectis ; dentibus non-
nullis marginalibus erectis ; valva cellulis vel areolis grandiusculis hexagonis. In mari
Japonico.

This frustule is figured both from its zonal and valval aspects. It is small, triangular,
and provided with large hexagonal areolae. Its sides are rectilinear, and its extremities
are smooth and slightly prominent. The specific name has reference to the few long
teeth or points that are disposed on the sides, and which are probably intended for the
union of several frustules into a series.

Triceratium arcticum, Bright., var. kerguelenensis, nov. (Plate XIII. fig. 7.)

The specimen here shown must be regarded as a variety of Triceratium arcticum,
Bright.,1 as it only differs from that Diatom in not possessing so markedly concave sides.
The cellules or areolse are also larger than in the typical specimen.

Triceratium arcticum, Bright., var. kerguelenensis /3, nov. (Plate XXII. fig. 5.)

This frustule is also a variety of Triceratium arcticum, Bright. Its radiating rows of
granules which spread outwards from the centre, the granules becoming larger and more
distinct as they approach the periphery, distinguish it from the Brightwellian species.

Triceratium arcticum, Bright., var. Icerguelenensis y, nov. (Plate XIII. fig. 5.)

This triangular form, with concave sides and rounded extremities, was obtained in the
neighbourhood of Kerguelen Island. The valve is covered with radiating lines of cellules,
which decrease towards the extremities, becoming small granules, arranged in a quincuncial
manner. Although thus aQreeinp; in the nature of its cellulation with Triceratium
arcticum, Bright., it differs from the latter (1.) in the concave character of its sides,
and (2.) in the occurrence of a small central non-granulated or smooth area. These
distinctions, however, cannot be looked upon as possessing more than a varietal importance.

Triceratium calvescens, n. sp. (Plate IX. fig. 1.)

Grandiusculum, triangulare ; lineis lateralibus concavis ; apicibus rotundatis ; cellulis
grandiusculis in lineas radiantes distributis, et ad apices subito in lineolas punctulatas
decrescentibus, media valva raris et minuentibus. In mari Japonico.

This form, like Triceratium arcticum, Bright., is also covered with radiating granules or

somewhat irregular cellules, which quickly diminish as they approach the extremities, and

1 = Triceratium wilkesii, var. /?, -with four angles ; Amphitetras wilkesii, Bright-well, Micr. Journ., vol. i.
p. 250, pi. iv. fig. 11 ; Roper, Trans. Micr. Soc. Lond., vol. viii. p. 58.

108 THE VOYAGE OF H.M.S. CHALLENGER.

assume the appearance of minute points, forming very delicate striae. The walls, however,
are slightly concave, the general valval granulation larger, while at the centre the granules
are few and small, so that the genus established by Brightwell cannot be so far extended
as to embrace this form, which must therefore become the type of a new species.]

Triceratium pavimentosum, n. sp. (PI. XIII. fig. 8.)

Quadratim areolatum ; lateribus concavis ; apicibus cuneato-rotundatis ; areolis sub-
hexagonalibus ; apicum processu nullo. Ad mare Japonicum.

This frustule possesses concave walls and valves ornamented with subhexagonal
cellules or areolae. It may be distinguished from the allied Triceratium favus of Ehren-
berg1 principally by the absence at each extremity of the horn-like processes, which
are characteristic of the latter. Triceratium favus also possesses rectilinear or slightly
concave sides.

Triceratium cariosum, n. sp. (Plate VI. fig. 6.)

Valvis trigonis, cellulosis ; lateribus subconvexis ; apicibus rotund atis ; cellulis sequali-
bus nullo certo ordine dispositis et vacuis lineolis irregulariter divisis. In mari Pacifico.

This singular Diatom possesses a triangular form, with slightly convex walls and
rounded extremities. The cellules that ornament the valve are equal, but have no definite
arrangement, although a tendency to form excentric curves may be recognised at some
places. Lacunae occur at irregular intervals amongst the granules, and give the valve an
eroded appearance.

Triceratium punctigerum, n. sp. (PI. XIII. fig. 4.)

Parvum, triangulare ; apicibus rotundatis ; lateribus subconcavis ; denticulo vel
punctulo erectiore marginali ad apices ; valvis inordinate punctulatis. In mari Pacifico.

This small but elegant triangular Diatom was discovered in a sounding made
in the Pacific Ocean. It possesses rounded extremities and slightly concave sides. The
valve is very delicately punctated, the puuctations assuming no definite arrangement, and
at the margin of each extremity a more prominent granule or point may be recognised.

Triceratium coronatum, n. sp. (Plate VI. fig. 7.)

E maximis ; triangulare, areolatum ; lateribus convexis ; apicibus elongatis, subulatis ;

areolis sequalibus, hexagonis granulo erecto ad quemque angulum; corona marginali

erecta. Ad Zebu in mari Philippinarum.

1 Ehrenberg, Kreideth., p. 79, N. 58, pL iv. fig. 10; Mikrogeol., pi. xix. p. 17; Smith, Synop. Brit. Diat.,
voL i. p. 26, pi. v. fig. 44, and Suppl., pi. xxx. ; Heiberg, Conspec, p. 41 ; Jan. et Rabenh., Hondur., p. 14,
pi. iii. fig. 10; =(1.) Triceratium megastomum, Brightwell, Micr. Joum., voL i. ; (2.) Triceratum fimbriatum,
Wallich, Micr. Joum., voL vi. p. 247, pi. xii. figs. 4-9.

RETORT ON THE DIATOMACE^E. 109

In this very striking frustule a marginal corona exists on one edge. This usually
escapes notice, because the flat form of the Diatom prevents it from being recognised,
save on the valval side. The form of the entire frustule is triangular, and the margins are
convex, while each extremity terminates in an elongated subulate process. The areolae are
hexagonal, and remarkable in having small rounded granules at their angles. This circum-
stance cannot be considered fortuitous, and, in conjunction with the other characters of the
frustule, must be regarded as adequate for the establishment of a new species.

Triceratium sarcophagus, n. sp. (Plate VI. fig. 3.)

Trigonum, lateribus convexiusculis, apicibus cornutis, asperis; valvis convexis, late
areolatis vel cellulatis ; areolis vel cellulis hexagonis. In mari Arafura.

The specific name of this form is applied from the peculiar appearance presented when
seen from its zonal aspect. It is triangular, and possesses slightly convex sides and
terminal horn-like processes. In its general aspect it shows some affinity to Triceratium
Javus, Ehrenb., but the areolae are smaller in the latter, and the terminal processes are not
rough, as they are in this frustule.

Triceratium favus, Ehrenb., var. late-areolata, nov. (Plate IX. fig. 3.)

This form only differs from Triceratium favus, Ehrenb., in having the areolae one-third
larger. It was obtained in the Sea of Japan.

Triceratium favus, Ehrenb., var. pacijica, nov. (Plate VI. fig. 1.)

The areolae in this variety are still larger ; the ratio of their size to that of the type-
specimen being as 19 to 8. It possesses four concave sides, and a small horn-like
process at each extremity. It was obtained in the Pacific Ocean.

Triceratium tumescens, n. sp. (Plate VI. fig. 9.)

Triangularis, lateribus convexis; apicibus cuneato-obtusis ; processibus spheroidalibus ;
valvis areolatis ; areolis hexagonalibus, intequalibus. In mari Japonico.

This triangular form possesses very tumid sides, obtuse extremities and processes, and
unequal hexagonal areolae. It presents a superficial resemblance to Triceratium Jimbri-
atum, Wall.,1 but it may be readily distinguished from it by the character of its terminal
processes, which resemble a compressed spheroid, instead of being elevated and conical.
It was obtained in the Sea of Japan.

Triceratium armatum, Eoper, var. 8, nov. (Plate VI. fig. 2.)

This elegant triangular form from the immediate neighbourhood of Japan is provided
with straight sides, and its extremities are notably elongated, forming well-marked subu-

1 Micr. Journ., vol. vi. p. 247, pi. xii. figs. 4—9.

110 THE VOYAGE OF H.M.S. CHALLENGER.

late processes. The granules, which are of equal size, are arranged in rows at right angles
to the margins, but in a central triangular space they are disposed irregularly. The demar-
cation of this central area is clearly indicated by a few well-marked granules or denticules,
placed at intervals along its sides. In the area itself there are a few scattered minute
spines or apiculi. In these respects this frustule resembles the Triceratium armatum of
Eoper,1 but it may readily be distinguished from the latter by the denticules referred to
above. Brightwell has already (Quart. Journ. Micr. Sci., vol. iv. p. 274, Plate xvii. figs.
9, 9b, 10, 11, 116, 12) notified several varieties of the typical species, and the present
form may be similarly referred to as var. S.

Triceratium fimbriatum, "Wall., var. nov. (Plate IX. fig. 12.)

This form resembles in its general characters the Triceratium fimbriatum of Wallich,2
but its terminal processes are much more elevated, and its extremities somewhat more dilated.

Triceratium incrassatum, n. sp. (Plate IX. fig. 10.)

Trigonum, areolatum ; valva lateribus convexis tumida ; areolis hexagonalibus ;
apicibus obtuso processu terminantibus. In mari Japonico. •

This form corresponds in general outline to the Triceratium grande of Brightwell,3
and its hexagonal areolation is also similar. The extremities, however, are here remark-
ably depressed and obtuse, instead of being more acute and attenuated. The ocelli are
nearly in the plane of the valve, which is evidently tumid. The three sides are slightly
convex, and the hexagonal areolae vary somewhat both in size and shape even in the
central region of the valve.

Triceratium grunowianum, n. sp. (Plate XVI. fig. 5.)

Forma triangularis, latissime areolata ; apicibus aspero processu erecto instructis ;
areolis subhexagonalibus ; valvis finissime et radianter punctulatis ; margine convexo
lsevi. Ad insulas Philippinas.

This Diatom is remarkable from the enormous size of its unequal hexagonal areolae,
its strong terminal processes, covered with inequalities, and its very delicate radiating
punctiform striation. The three sides of the valve are slightly convex outwards, and in
immediate contact with these is a hyaline belt which bears irregularly elliptical or clavate
granules. At the bases of each of the three terminal somewhat infundibuliform processes
there are also several irregularly scattered angular granules which serve to give greater
prominence to this region. The narrower and more distal parts of each process bears a series
of fine striae which radiate into the proximal and wider end of the funnel. The specific
name has been given in honour of Mr Albert Grunow, the well-known observer of Diatoms.

1 Micr. Journ., voL ii. p. 283, fig. 1. 2 Micr. Journ., vol. vi. p. 247, pi. xii. figs. 4-9.

3 Micr. Journ., vol. i. p. 249, pi. iv. fig. 8; =(1.) Triceratium oriental?., Harv. et Bail.; (2.) Triceratium
faints (large), Pritcliard, op. cit., p. 856. The original specimens of Ticeratium grande, Bright., were found on
Tridachnidoe'and other shells from the Indian seas.

'KEPOKT ON THE JDIATOMACE^E. HI

Triceratium abyssale, n. sp. (Plato XIII. fig. 6.)

Forma orbiculato-triangularis, apicibus late rotundatis, in ocellum exeuntibus ;
lateribus convexis ; valva punctulis aequalibus radiantcr siguata. In mari Indico.

This form possesses large rounded extremities and convex sides, so that the frustule
becomes suborbicular. The valve has equal radiating lines of punctations slightly con-
ver°rin°r at the obtuse apices, each of which terminates in a large ocellus. The specimen
was obtained at a depth of 2050 fathoms.

Triceratium insutum, n. sp. (Plate XXV. fig. 7.)

E rnaximis ; valvis novem-angulatis, areolatis ; lateribus concavis ; arcolis hexagonali-
bus, grandiusculis, inasqualibus ; valva exterius areolata, inferius arctissime et radianter
striolata. In mari Philippinarum.

This species possesses nine angular points, and between each adjoining pair a concave
margin. The surface is covered by very large subhexagonal areolae, which gradually
diminish in size from the centre towards the periphery. The valve exhibits fine radiating
striae, and the areolae present at the bottom irregularly rounded spaces, which bear one or
more distinct and irregularly disposed granules or points. It appears certain that these
two kinds of ornamentation belong to two distinct strata of the diatomaceous cell wall,
which may sometimes become detached from one another. I became convinced of this
when Dr James Rae, E.N., drew my attention to a Triceratium in which only a part
of the inner striated layer still adhered to the valve. A similar doubly laminated wall
may be seen in Triceratium favus, Ehrenb., and in Triceratium grunoivianum above
referred to ; but the nine-angled frustule now in question cannot be united with these,
not only because of the little points in the centre of the areoles, but still more because
of the absence of the raised terminal processes which are presented by the others.

In consequence of this bilamination of the frustular wall, it becomes necessary to
exercise great caution in the establishment of new species of this genus.

Triceratium atlanticum, n. sp. (Plate XVII. fig. 3.)

Forma quadrato-globosa, cellulis grandiusculis a quinque majoribus mediis radiantibus,
et leviter ad quatuor ocellos terminales convergentibus. In Oceano Atlantico ad insulas
Azores.

This specimen possesses four very obtuse angles, between each of which the sides
are convex, so that the resulting form is suborbicular. The cellules are of somewhat large
size, and radiate from a small central rosette formed of five large cellules. Each of the
angles presents an elliptical ocellus, and towards each of the four extremities the cellules
become slightly convergent.

112 THE VOYAGE OF H.M.S. CHALLENGER.]

The specimen figured was obtained in a sounding made near the Azores in the
Atlantic Ocean.

Stictodiscus, Grev.

This genus was established by Greville in a communication read on 12th March 1861,1
and he classed in it Discoplea (?) rota2 and Discoplea rotula,3 Ehrenberg, as well as
Actinoptychus dives, i which, as already suggested by Ehrenberg, might be generically
associated with Cyclotella rota and Cyclotella rotula. The words of Ehrenberg are (under
his definition of Discoplea (?) rota) : " Proxime ad Actinoptychum divitem in Grsecia fossilem
accedens forma, et cum ea forsan, et cum sequente (Discoplea (?) rotula) in peculiari
genere reponenda."

To this genus Greville at this time ascribed six species, and later Kitton5 made us
acquainted with Stictodiscus crozieri. Grunow 6 added his Stictodiscus angulatus in the
Typenplatten of Moller, while Cleve 7 announced his Stictodiscus novarce in a paper com-
municated to the Eoyal Swedish Academy of Science on 15th September 1880.

The definition of the genus Stictodiscus, as given by Greville, ran as follows : —
"Frustules simple, discoid, divided by radiating lines into numerous plicate compart-
ments. Lines not reaching the centre. Compartments furnished with conspicuous trans-
parent, pore-like points. (In the four typical species, large scattered puncta also occupy
the blank central portion of the disc.)"

Among the materials collected by the Challenger Expedition a considerable number
of specimens of Stictodiscus has been observed, but these were only found in a few locali-
ties, the most noteworthy of which are the Japanese coast and Zebu in the Philippine
Islands. The examination of these specimens has convinced me that the structural
peculiarities, and particularly the folded compartments, are somewhat variable and do not
always afford sufficient ground for establishing new species. All the forms may be reduced
to two principal sections, namely (1.) radiate Stictodisci, in which the discs possess com-
partments and the folds are simply radiating ; and (2.) radiato-areolate Stictodisci, in
which the discs have radial folds which proceed from the circumference, but soon bifurcate
or are interrupted so as to bound numerous polygonal areas. It is to be noted, however,
that this classification, though of use in facilitating the description of forms so similar to

1 Micr. Journ., n. s., vol. i. p. 39.

2 Monatsber. d. Jc. Mead. d. Wiss. Berlin, 1844, p. 202; MikrogeoL, pi. xxxv. A 22, fig. 6.

3 Ehrenberg, MikrogeoL, pi. xxxv. A 22, fig. 7.

4 Ehrenberg, MikrogeoL, pi. xix. fig. 12 ; Ralfs in Pritchard's History of the Infusoria, 4th ed.,'p. 840.

5 Micr. Journ., 1873, pi. xxxviii. fig. 2.

6 In Schmidt's Atlas (explanation of plate Ixxiv. figs. 26-30), a note is appended to Grunow's Pseudosticto-
discus angulatus to the following effect : " Grunow bemerkt dazu; die beiden Schalen seien nicht ganz gleich,
die flachere habe keine Rippen, und immer in der Mitte einen rudimentaren Stachel, der in meinen Zeichnungen
fehle; die Ecken etwas vorspringend, darum den Biddulphien sich nahernd." <1

7 Kongl. SvensJc. Vetensk.-Akad. Handl., 1880, Bd. xviii. No. 5, p. 21, pi. v. fig. C6.

KErOET ON THE DIATOMACE/E. 113

one another as those belonging to this group are, cannot be rigorously applied, since, for
example, in Stictodiscus buryanus, Grev.,1 the radiating appearance is still retained by the
valve, although some of the radial folds are found to be interrupted.

Again, the radiate Stictodisci may be divided into two groups according as the
central area is or is not distinguished by a simple or double corona of fine points or
granules, and these groups may be indicated as coronate or non-coronate respectively.

It must further be remarked that among triangular frustules having the very greatest
affinities to Triceratium, some have been observed which, with the exception of their
non-discoidal form, exactly fulfil all the conditions of structure recpiired by the definition
of Stictodisciis as given by Greville. Thus among the triangular specimens brought home
by the ChalleDger, one had its surface folded in a radiating manner, and also showed the
central Stictodiscoid corona of points, so that notwithstanding its non-discoidal form it
must still be regarded as a true Stictodisciis. On this ground, therefore, the generic
definition as given by Greville should be modified, and may be stated in the following
words : — " Frustula simplicia, discoidalia, vel polygonalia, per lineas radiantes divisa in
areas plicatas, lineis centrum non attingentibus ; areis plerumque conspicuis punctis vel
granulis instructis."

We cannot, indeed, believe that the same species can assume sometimes one form
and sometimes another, or that from the same Stictodiscus sometimes discoid and some-
times triangular or polygonal forms arise — -before admitting such an anomaly it would
be necessary to have the experimental results of artificial cultivation of some given species
to go upon — yet from their structural characteristics I must conclude that a Stictodiscus
may exist a3 a discoid or triangular or even polygonal body. The two last groups of
forms may be indicated by adding to the specific name the words forma triangularis or
forma polygonalis respectively.

Having arrived at the conclusion stated above, it seems a fitting occasion to review
the long series of species of Triceratium to determine whether any of these forms might
not more properly be relegated to our present genus.

Greville in 18612 described under the name of Triceratium harrisonianum a magni-
ficent triangular form adorned with rows of pearl-like puncta, which formed a large
circumscribing belt. This frustule, moreover, was provided with sparsely disposed pearl-
like granules in the central space, in which there also occurred a conspicuous network of
large elongated radiating cellules from which lines passed outwards between the rows of
granules to the margin. But on examining the figure the cellules in question are areas
or compartments which are bounded by depressed lines that pass between the rows of
granules, and not by lines of relief as might readily be inferred from the bold manner in
which they are represented in Greville's figure. The fact that Greville had not found
examples belonging to any genus which had sometimes a discoidal, sometimes a triangular

1 Micr. Journ., n. s., voL i. p. 40, pL iv. fig. 1. - Micr. Journ., n. s., vol. i. p. 76, pi. ix. fig. 9.

(bot. chall. exp. — part iv. — 1886.) D 15

114 THE VOYAGE OE H.M.S. CHALLENGES.

or polygonal form, prevented him from recognising affinities between two groups of forms
so distinct from one another, and although from the disposition of the granules, in lines
radiating from the sides and sparsely disseminated in the centre, he thought of constituting
a special group for Triceratium harrisonianum and Triceratium margaritaceum, Ralfs, he
was unaware of the importance of the folded character of the former, which, taken in
conjunction with the special distribution of the granules, should have induced him to class
it with Stictodiscus.

Again, Greville1 determined as Triceratium irijlatum, a three-sided form furnished
with lines or veins radiating from the margin to the middle of the radii, and possessing
puncta or granules which were large and sparsely disposed in the central part, but smaller
and more numerous on the circumference. These characters, however, are precisely those
of the genus Stictodiscus, so that the organism should be named Stictodiscus infiatus.

The name Triceratium lineolatum, Grev.,2 cannot, however, be corrected in the same
manner, because though in this veins or folds occur on the plane of the valve at its
margin, the valve is elsewhere covered throughout with equal, minute, radiating granules
— a character which I have never met with in any of the different specific forms of
Stictodiscus.

That, on the other hand, Triccratium parallelum, Grev.,3 is a true Stictodiscus
cannot be doubted. It possesses four or six angles, its centre is areolated by furrows or
folds which are distributed in a retiform manner, and its border is furnished with granules
disposed in radiating lines. It should therefore be named Stictodiscus parallelus.

To the same genus should also be ascribed Triceratium quadratum, Grev. (loc. cit:,
fig. 19), because it shows short folds at the circumference, and the surface is covered with
radiating lines of granules, which however are few, and scattered in the central space. It
may also be observed that Greville reports having seen specimens with a sort of umbilicus,
or, at least, a somewhat irregular circle of smaller cellules, around which the ordinary
cellules are often more or less scattered before they pass into radiating lines — a circumstance
which would seem to indicate that the frustule in question is a coronated Stictodiscus.

In the case of Triccratium polygonium, Grev.,4 we have to deal with a true Stictodiscus.
This beautiful Diatom, with its six straight sides and obtuse angles, not only possesses
elegant lines of radiating granules except in the centre, which is slightly reticulate, but, as
shown in the figure, as many depressed lines or folds as exist in the recognised forms of
Stictodiscus occur at the margin. The name should accordingly be changed to Stictodiscus
polygonius.

The Triceratium quadricorne of Greville 5 is also probably a Stictodiscus, as its surface

1 Micr. Journ., n. s., voL iii. p. 232, pi. x. fig. 15.

2 Micr. Journ., n. s., vol. iii. p. 233, pi. x. fig. 16.

3 Micr. Journ., n. s., vol. v. p. 104, pi. ix. figs. 22 and 23.

4 Micr. Journ., n. s., vol. v. p. 105, pi. viii. fig. 14.

5 Micr. Journ., n. s., vol v. p. 103, pi. ix. fig. 16.

REPORT ON THE DIATOM ACE^K. 1 1 5

is marked by many areas or compartments which arc limited by folds or depressed lines,
while each compartment is ornamented by a pointor granule — the latter character being one
of the chief distinguishing features of the jjenus Stictodiscus. It is true that here no radiat-
ing folds occur at the circumference, and that the granules arc more dense at the edge than
at the centre ; hence the frustule should probably be designated Slictodiscus quadricornis,
although its precise determination must remain somewhat uncertain.

Professor Cleve, in his communication On Some Rare and Little-Known Diatoms,1 has
described two forms under the name of Triceratium, which, in my opinion, are also true
Stictodisci. The first — Triceratium gallapagense, is shown in his plate vi. fig. 72. It is a
triangular acute-angled form, with granules sparsely scattered in the middle, but more
densely disposed at the sides, where they form short rows, while fine branching lines of
more simple points are seen at the angles. In the figure the few radiating lines which
are indicated evidently denote the characteristic folds of Slictodiscus. It should accordingly
be designated Stictodiscus gallajoagensis. It is distinguished from my Slictodiscus
bicoronatus specifically by the absence of the central corona.

The second form referred to by Cleve as Triceratium margaritiferum (loc. cit., pi. vi.
fig. 76) is a coronated Stictodiscus. It has been described by Cleve as follows : — "Valve
quadrangular, with concave sides and rounded angles, without processes. Surface probably
plane. Structure tolerably coarse, granules arranged near the margins in short lines,
smaller in the angles, rare and scattered in the middle." In the centre of the figure there
is also seen a small corona of points which has been omitted in the above description.
By examining the various figures which I have given of new species of Stictodiscus, the
very close affinity of this species of Cleve is at once apparent, although the folded con-
dition of the valve in the latter is not evident. Hence this frustule should be named
Stictodiscus margaritiferus.

The above remarks on Stictodiscus and Triceratium were finished when the nineteenth
part of Schmidt's Atlas of the Diatomaceae was published. In that work very good
figures are given on plate lxxiv. of several new species of Stictodiscus as well as of some
already known, and on plates lxxv. and lxxvi. numerous types have been represented as
belonging to the genus Triceratium, among which I at once recognised many real
polygonal Stictodisci. Among the known discoidal forms Stictodiscus hardmanianus,
Grev., is shown on plate lxxiv. fig. 8, in a somewhat small form ; Stictodiscus kit-
tonianus, Grev., is seen in figs. 16, 17, and 18 ; while Stictodiscus califomicus, Grev.,
occurs as fig. 4, and several varieties as figs. 1, 2, 3, 6, 7, and 9.2 Among the new
species, on the other hand, Stictodiscus simplex, A. S., is shown in fig. 11 ; Stictodiscus

1 Kongl Svensk. Vetensk.-Akad. HandL, Ed. XVIII. No. 5, p. 25, pi. vi. fig. 72.'

2 These varieties are — Stictodiscus califomicus, Grev., var. areolata, Grun., pi. lxxiv. fig. 1.

Stictodiscus califomicus, Grev., var. nankoorensis, Grun., pi. lxxiv. figs. 2 and 3.
Stictodiscus califomicus, Grev., var. ecostata, Grim., pL lxxiv. tigs. 6 and 7.
Stictodiscus califomicus, Grev., pi. lxxiv. fig. 9.

116 THE VOYAGE OF H.M.S. CHALLENGER.

argus, A.S., in fig. 12 ; and finally, Stictodiscus morsianus, A. S., in figs. 19 and 20. The
forms marked figs. 21, 22, and 23 are given as doubtful, while those of figs. 13, 14, and 15
are interpreted as internal valves1 of Stictodiscus — a determination which seems to rest on
no other foundation than the presence of evanescent folds towards the centre, there being
no trace of any such reticulate arrangement as occurs in Stictodiscus. It would seem
rather that such discs belong to the genus Cyclotella, of which they would constitute
a new species.

On plate lxxv. fig. 1, Schmidt has represented a superb polygonal Stictodiscus under the
name of Triceratium multiplex, Janisch (?). It possesses a central corona, and the almost
regular short lines of granules form a belt at the circumference, while the middle of the
surface is covered with sparsely disposed granules of a similar kind. The very large valve,
which has arrived at its complete development, scarcely preserves traces of the folded
condition of its surface, but I do not hesitate to designate it Stictodiscus multiplex.

In fig. 2 of the same plate the organism represented as Triceratium jercmianum, A. S.,
is another subquadrate Stictodiscus with subradiating folds which alternate with granular
lines, that subsequently become rare and are irregularly scattered in the middle. Many
lines of small granules occur at the angles, and these point, as in other allied forms, to
its progressive development. This form should accordingly be designated Stictodiscus
jeremianus.

At figs. 6 and 7 Schmidt also shows two different forms of Triceratium, established
by Greville, which, as already said, must be named Stictodiscus eulensteinii ; while the
Triceratium karri sonianum, Grev., shown at Plate lxxv. figs. 14 to 16 should be
designated Stictodiscus harrisonianus.

On plate lxvi. of this Atlas several other forms which have been named Triceratium
appear to me to be true Stictodisci, although the essential characters of the latter genus
are in these less evident than in the above-mentioned cases.

In considering the new types belonging to the genus Stictodiscus which have been
brought home by the Challenger I shall first refer to the discoidal forms, and then pass
on to those which are polygonal.

I. Discoidal Forms of Stictodiscus.
Stictodiscus anceps, n. sp. (Plate I. fig. 5.)

Granulis raris in lineas subregulares radiantes distributis; centrum punctulorum corona
distinctum. Ad oras Japonicas.

1 "Nach meiner festen Ueberzeugung innere Schalen v. Stictodiscus. Solche innere Schalen, welche nament-
lich bei Asteromphalus und Asterolampra vielfach vorkommen und irrthumlich als besondere Arten benarmt sind,
sind leicbt daran zu erkennen, dass sie theils einen Stich ins Yiolette haben theils der feineren Sculptur ermangeln,
uberhaupt roher gestaltet sind. Grunow bestimmt die vorliegendon Figuren als Schalen von Metosira clavigera,
Grunow."

REPOET_ON^THE DIATOM ACEiE. 117

This elegant disc is ornamented at the margin with sparsely disposed lines of
granules. These become irregular towards the centre, and are replaced there by an
irregular corona of small points. The presence of this corona has induced me to include
this frustule among the Stictodisci, although the folded condition of the valve, which is
the principal generic character, could not be distinguished. It is to be noted, however,
that in the different species the folds are found more or less clearly defined, and when
they are very slight they become almost invisible, so that the impossibility of recognising
them here is not necessarily opposed to the determination that has been arrived at.

Sticto discus radiatus, n. sp. (Plate I. fig. 1.)

Granulis in lineas radiantes plicis divisas distributis ; plicis ad centrum evanes-
centibus ; centrum granulis caret. Ad Zebu ex insulis Philippinis.

This frustule, and its variety shown in Plate I. fig. 3, belong to the radiate section of
the genus Stictodiscus. They are closely related to the Stictodiscus californicus of
Greville,1 which in its turn is akin to Stictodiscus johnsonianus, Grev.,2 differing from it
in the indistinct and shorter radii, and in the condition of the peripheral zone from which
these radii spring. This zone is marked in Stictodiscus johnsonianus by a series of round
or somewhat elliptical granules which are situated opposite the peripheral ends of the radii,
while in Stictodiscus californicus it bears semi-elliptical areolae which usually alternate
with the radii. In the forms now under examination, on the contrary, the radii are
notably long and the centre always remains free from granules, while the great number of
the radii — although by itself a character of no great significance — cannot be overlooked.

Although the type-specimen shown in Plate I. fig. 1 bears a great resemblance to its
variety seen in fig. 3 there are some differences which must be noted. Thus in addition
to the greater number of granules which are radially disposed and ornament the valves in
fig. 3, there is also present here a somewhat prominent granule at the extremity of each
fold, and the radii begin with two small granules. In the course of the radii each granule
is elevated upon a subrectangular areola. Much importance cannot be attached to the
granules at the extremities of the folds, as their existence is doubtful in some cases —
similar appearances often occurring at the margin of the valve in changing the focal
distance. The subdivision of the radii into compartments serves to indicate the connec-
tion between the radiate and radiato-areolate forms of Stictodiscus. The following
definition may be given of the varietal frustule : — " Lineis crebriusculis granulorum in
areas rectangulares, granulum in margine ad ortum cujusque plicse. Ad oras Japonicas."

Stictodictus eulensteirdi, Cstr. (Plate I. fig. 7.)

Triceratium eulensteinii, Gran., A. Schmidt's Atlas, pL lxxv. figs. 6 and 7.]
In Plate I. fig. 7 there is represented an exceedingly elegant form of Stictodiscus.
which indicates the transition from the discoidal to the triangular and polygonal types.

1 Schmidt's Atlas, pi. lxiv. fig. 4. - Micr. Joum., n. s., vol. i. p. 41, pi. iv. tig. 3.

118 THE VOYAGE OF H.M.S. CHALLENGER.

It possesses the form of a disc with an undulating (almost dodecahedral) margin. A
triangle of smaller granules occurs in each of the protuberances, and hence I am of
opinion that the frustule originally was discoid, and that its subsequent development did
not take place uniformly all round the circumference, but only in a few circumscribed
areas. During this process small granules have been formed ; these have gradually
increased in size and become disposed in the large regular lines, which occupy the spaces
lying between the protuberances. A similar condition obtains more or less markedly at
the extremities of triangular frustules, as will be pointed out below.

Between the lines of the quadrately disposed granules radiating folds arise and extend
towards the centre. The compartments circumscribed by these folds are furnished with
sparsely arranged granules, and a beautiful corona of smaller points occupies the centre.

After this frustule had been figured and described I received the nineteenth part
of A. Schmidt's Atlas, in which the same form is figured at Plate lxxv. figs. 6 and 7,
under the name of Triceratium eulensteinii, Grun. Although therefore recognising the
right of precedence of the well-known Austrian microscopist in the case of the specific
determination, I cannot accept his generic name, but assign the organisms, for reasons
already quoted, to the genus Stictodiscus. The frustule was gathered both in the Sea of
Japan and in the Philippine Sea.

Stictodiscus radfordianus, n. sp. (Plate XVII. fig. 10.)

Forma octogona, lineis undulatis, et angulis subproductis rotundatis inclusa ; zona
regulari ordine quadrato granulata, plicis a margine radiantibus ; granulis medio raris ;
puuctulorum corona umbilicali. . In Archipelago Philippinarum.

This frustule, which is still more elegant than that last referred to, belongs to the
coronate section of the genus. It possesses an octagonal form with undulating sides,
while the angles are rounded and scarcely elongated. A belt formed of four regular rows
of granules runs round the periphery of the valve. These granules are disposed in a
quadrate manner, and folds extend between them in a centripetal direction. In the middle
region small granules are sparsely disseminated, and the centre is marked by a beautiful
corona of small points. Here too, as in the case of Stictodiscus eulensteinii, rows of small
granules occur at the angles, and as in the former case seem to indicate points of gradual
increase.

The specific name has been given in honour of Dr Radford, to whom, in part, I am
indebted for being invited to examine the Diatoms brought home by the Challenger
Expedition.

II. Radiato-Areolate Forms of Stictodiscus.

In all these types, folds of greater or less length radiate from the circumference towards
the centre, and the central space is subdivided into areoles in a reticulate manner.

REPORT ON TUE DI ATOMACE.-E.

119

Stictodiscus japonicus, n. sp. (Plate I. fig. 2.)

Granulorum lineis per plicas radiantes divisis prope marginem binatim oricntibus ;
areis mediis grandiusculis reticulatim dispositis singulo granulo plerumque distinctis. In
mari Japonico.

This organism may be distinguished from the others by the circumstance that in it the
granules do not occur on the perimeter, while there are always two granules standing side
by side at the outer extremity of every row. The central areas are somewhat large, of
irregular forms, and are in most cases ornamented with a siugle granule.

Stictodiscus affinis, n. sp. (Plate I. fig. 4.)

Lineis granulorum per plicas circumradiantes divisis ; media superficies in areolas
reticulatim distincta; area? nonnullse granulo instructae. Ad Zebu in mari Philippinarum.

This form is much akin to the preceding, but may be distinguished from it by the
fact that the lines of granules proceed from the margin — a peculiarity which may also
be seen in the variety described below. The two granules found at the outer ends of
every row in Stictodiscus japonicus do not occur here, where, moreover, the central
areas are of smaller sizes and somewhat more abundant. Although the majority of these
areas are ornamented by a single round granule, some of them are entirely devoid of this
structure, while others bear two.

Stictodiscus affinis, n. sp., var. now (Plate I. fig. 6.)

Differt a forma typica per lineas binis granulis ter vel quatuor vicibus repetitis
incipientes. Ad mare Philippinarum.

In this Diatom the granule-bearing areas sometimes exhibit four pairs of puncta
towards the circumference of the disc. The central areas are somewhat irregular in size
and form, and while several are perfectly smooth others bear a single round granule.
Hence although not identical with the Zebu type, this frustule cannot be regarded as more
than a variety of it.

Stictodiscus affinis, n. sp., var. latc-zonata, nov. (Plate XVII. fig. 11.)

Although well characterised by its distinct granulated zone, this Diatom must be
viewed as a distinct variety of Stictodiscus affinis. It is principally distinguished by
the marginal radiating folds, which become reticulated towards the centre, and by the
distinct and regular lines of granules interposed between the folds. Marginally the
granular lines are ornamented by pairs of puncta placed laterally. Two well-defined
rings of approximately equal size run round the periphery.

Stictodiscus affinis, n. sp., var. nov. (Plate XVII. fig. 8.)

Another variety of the same species is here shown. The smallness of its disc and
the smaller number of its radii can certainly have no specific significance, but it may be

120 THE VOYAGE OF H.M.S. CHALLENGER.

observed that the reticulated central space is almost destitute of granules. This variety,
like the previous one, was collected in the neighbourhood of the Philippine Islands.

Stictodiscus reticulatus, n. sp. (Plate I. fig. 8.)

Liueis granulorum plerumque per plicas a margine radiantes distinctis ; hinc plicae
reticulatim distribuuntur ; arese centrales grandiusculae, pluribus granulis plerumque
instructas. Ad Zebu.

This disc is ornamented with lines of granules which are for the most part double.
These lines are separated by folds which soon become reticulated after leaving the margin.
The meshes of the well-marked somewhat irresrular network thus formed are adorned with

O

more or fewer granules, disposed in small groups. The frustule was obtained at Zebu.

Stictodiscus margaritaceus, n. sp. (Plate XVII. fig. 12.)

Disciformis, granulis plerumque magnis, medio raris, ad marginem frequentioribus, et
subregulariter dispositis ; plicis a granulo marginali exeuntibus et granulorum lineas
discriminantibus. Ad insulas Philippinas.

In this specimen the granules are unusually large and pearl-like. They are sparsely
scattered in the centre and more closely near the margin, where they are divided by slight
folds, each of which originates in a well-marked marginal point. The disposition and
size of the granules are inconstant.

III. Triangular or Polygonal Forms of Stictodiscus.

Stictodiscus varians, n. sp. (Plate XVII. fig. 7.)

Forma subquadrata, lateribus convexis ; granulis a margine radiautibus, medio rari-
oribus et irregulariter distributis ; linese marginales plicis evanescentibus dividuntur. In
mari Siuensi.

This specimen is, in form, transitional between the round and the polygonal types.
The discoidal valve tends to become angulated at four points in such a manner that by a
further extension of the angulation the round form would become quadrate. The disposi-
tion of the granules is regular and somewhat dense at the circumference, but rare and
irregular at the centre. At the perimeter the lines of radiating granules are intersected
by folds, which soon disappear. The specific name which has been applied is intended to
recall the tendency to vary in form from round to quadrate which is exhibited by this
interesting Diatom.

Stictodiscus bicoronatus, n. sp. (Plate VI. fig. 5.)

Trigonus, lateribus subconvexis et apicibus rotundatis ; zona granulis raris et decus-
satim dispositis; granulis medio nullo certo ordine distributis; centrum duplici punctulorum
corona insiguitur ; ad apices multipla? punctulorum linese radiantes. Ad oras Japonicas.

REPORT.ON THE DIATOMACEiK 121

This beautiful triangular frustulc, from the neighbourhood of the coast of Japan, has
slightly convex sides and rounded angles. The granules are disposed at right angles to
each of the margins along regular lines, and so form a well-marked peripheral belt which
circumscribes a central triangular space ornamented with irregularly disposed granules,
very minute points, and short irregular wavy lines, but in the centre two coronse of smaller-
granules occur. At the three rounded angles of the triangle the regular lateral granulation
is replaced by many lines of small points, which are arranged in a fan-like manner.
Although no folds are to be found in this species its fundamental Stictodiscoid character
is indicated (1.) by the peculiar disposition of the granules along the sides; (2.) by the
general character of the granulation, which recalls that of Stictodiscus anceps, n. sp. ; (3.)
by the presence of central coronas ; and (4.) by the general similarity of the granulation
of the angles to that of the twelve salient points of Stictodiscus eulensteinii.

The specific name has reference to the existence of a double corona at the centre of the
valve.

Stictodiscus bicoronatus, n. sp., var. punctigera, no v. (Plate XIII. fig. 2.)

This is another triangular Diatom which recalls the general characteristics of the pre-
ceding, except that it possesses cuneate rounded extremities, punctiform granules, and a
more regular double corona of small points. The marginal granules, though small, are
arranged in well-defined lines, which are not sharply limited on the inner side. The
granules occupying the subcentral area are irregularly and very sparsely disseminated.
Inside of the double corona no granules occur. These characters, however, do not seem to
be of sufficient importance to justify the establishment of a new species.

The name of this variety is intended to recall the elegant character of the valvular
punctation.

Stictodiscus elegans, n. sp. (Plate XIII. fig. 3.)

Forma triangularis lateribus subrectis ; plicae a lateribus radiantur, medio reticulatim
dispositse ; granulis raris ordinate circum distributis, medio rarioribus et inordinatis ; ad
apices rotundatos punctulorum linese crebrae radiantur. In mari Japonico.

This beautiful triangular frustule has the marginal rows of granules separated by

folds, which become reticulate in the centre. In the central area the granules are sparsely

and irregularly disposed, while the three rounded extremities are occupied by radiating

lines of small points. Each of the margins is ornamented by a band of oval granules,

which are largest in the middle, and gradually become finer towards the extremities.

These may indicate a kind of elevated marginal corona, such as is found in some species of

Triceratium. The external limiting line of this frustule is very slightly convex on its

outer side.

(bot. chall. exp. — part iv. — 1886.) D 16

122 THE VOYAGE OF H.M.S. CHALLENGER

Stictodiscus trigonus, n. sp. (Plate XIII. fig. 1.)

Forma triangularis, apicibus obtuso-rotundatis, lateribus tumidufis ; granuli medio
irregulares et rariores, ad latera lineis per plicas separatis, ad apices punctuli densiores.
In mari Japonico.

The slightly tumid sides, and obtuse, rounded angles of this form are noteworthy.
The granules are few, and irregularly placed in the centre, but are more abundant and
arranged in lines divided by folds at the sides, while they are smaller and still more
numerous at the extremities. The gradually decreasing size of the granules from the
centre towards the periphery is a well-marked characteristic of this Diatom.

Stictodiscus hexagonus, n. sp. (Plate XVII. fig. 17.)

Valvis hexagonis, granulatis ; apicibus late rotundatis ; margine granulorum ordine
terminato ; granulis grandiusculis circum subregularibus, medio nonnullis ; superficies
medio plicata. In Atlantico septentrionali.

The present curious hexagonal frustule has very rounded angles, and a line of small
round granules at the extreme border. The convex surface is flattened at the margin, and
is ornamented medially by large granules. These occur in sparsely disposed subregular
lines at the periphery, but are almost entirely absent in the centre, where the surface is
reticulate, the meshes of the network being very variable in size and irregular in form. At
the angles a few closer lines of granules are to be observed. These are disposed in a
fan-shaped manner, as in all the polygonal forms of the genus. An intra-marginal thin
bine, almost parallel throughout to the edges of the frustule, forms a well-marked boundary
line between the peripheral and intermediate areas, on the latter of which a reticulation
such as is found at the centre is not manifest.

Stictodiscus hexagonus, n. sp., var. nov. (Plate XVII. fig. 14.)

A variety of the last-mentioned type is here represented. It differs from the typical
frustule in the smaller size of the disc and granules, in the relatively larger size of the
central network, and in the total absence of those lines of small granules that constitute
the margin of the valve of Stictodiscus hexagonus. Both the frustules were found in the
South Atlantic.

Cestodiscus, Grev.

This genus was instituted by Greville1 in 1865 tojnclude some very elegant discoidal
frustules, which differed from Eupodiscus only in possessing numerous marginal processes
instead of a few. It greatly resembles the genus Aulacodiscus, and might, indeed, be
identified with it if the intra-marginal shortly tubular processes of the latter were not
connected with the centre by means of a distinct furrow formed by a well-defined interval

1 Trans. Micr. Soc. Land., n. e., vol. xiii., p. 2.

REPORT ON THE DIATOM ACEjE. 123

occurring between two of the radiating lines of granules which proceed from the sub-
marginal processes and run towards the centre.

Greville defines the genus Cestodiscus] as follows : — "Frustules disciform (circular or
oval), disc with radiating granules or cellules, and a submarginal circle of obtuse processes
unconnected by means of special radiating lines of cellules with the centre."

To this description the following character, which may be constantly observed in all
specific forms, should be added, namely, the radiating granules or cellules of the central
part of the disc become more or less abruptly minute crowded granules or points, on
approaching the margin. This may be well seen in Cestodiscus trocos described below.

Cestodiscus trochus, n. sp. (Plate VII. fig. 1.)

Valvis rotundis ; distinctorum granulorum radiis a centro dimanantibus, pluribus brevi-
oribus insertis ; granulis rotundis, distinctis, sequalibus, ad marginem sensim diminuentibus
et stipatis ; ad extremum plurium (32) denticulorum extat corona. In oceano Pacifico.

In this beautiful disciform Diatom there is a group of small, somewhat sparsely dis-
posed granules in the centre, and from this region a series of distinct radiating lines of
granules pass out towards the periphery near to which other shorter lines are interposed
between these. The granules are usually large and equal, but become, smaller near the
margin, and finally in a narrow marginal belt become extremely fine and crowded so as to
present the appearance of very delicate striae. Around the striated border very small
processes occur. These are irregularly disposed, and accordingly divide the circumference
irregularly. They are thirty-two in number in the specimen figured. The specific name ~
has reference to the wheel-like appearance presented by the larger radiating granules
that form lines across the valves.

Cestodiscus trochus, n. sp., var. no v. (Plate VII. fig. 3.)

A variety of the preceding species is here shown. It differs from the typical form (1.)
in the size of the granules, which are larger ; (2.) in the condition of the centre, which is
marked by a small irregular group of somewhat smaller granules; (3.) in the smaller
number — twenty-four — of denticules in the submarginal corona; and (4.) in the more
abrupt change in the size of the granules at the border, where they assume the
appearance of minute strise. The general aspect of the two frustules, however, is the same,
and the above distinctions cannot be looked upon as of specific importance.

Cestodiscus a con vexus, n. sp. (Plate VII. fig. 6.)

Forma rotunda, convexa ; grandiusculorum granulorum ordinibus ab umbilicali area
dimanantibus, et continuo crebrescentibus ; ad marginem granuli minuuntur mensura,

1 Trans. Mier. Soc. Lond. n. s., vol xiii., 18G5, p. 48. * '■fXo'f, a wheel.

124 THE VOYAGE OF H. M.S. CHALLENGER

augentur numero ; corona viginti-quatuor denticulorum in marginali limbo arctissime
striate In oceano Pacifico.

The centre of this magnificent disc presents a small, smooth, irregular area, to which
radiating lines of equal round granules converge on all sides. Some of these lines become
interrupted as they pass inwards, owing to the diminution of the size of the space, and
where such interruptions occur, long triangular spaces become interpolated. Near the
margin the granules suddenly diminish in size, and assume the appearance of very minute
radiating striae. Twenty-four denticules project around the broad border. Although this
Diatom does not seem to differ much from the two preceding, the small irregular umbilical
area, and the notable convexity of the valve, point to its true specific character. In
addition to the figure of the valve from the valval aspect a small ideal figure from the
zonal side has been added in order to show the notable convexity of this frustule.

Cestodiscus coronatus, n. sp. (Plate VII. fig. 9.)

Disciformis, medio granulatus, ad marginem striatus ; centrum a margine dividitur per
coronam e cellulis oblongis, grandiusculis, a qua exterius oritur denticulorum series. In
mari Pacifico.

In this Diatom, which also belongs to the group having a continuous system of granu-
lation and a small umbilical area, the beautiful granulation radiates from a small umbilical
area in the central part of the valve, while the margin is ornamented with delicate striae.
The two areas, which are thus differently sculptured, are separated by means of a corona
of beautiful large oblong cellules, from which a series of spines or denticules pass towards
the circumference. Apart from these denticules the other structural details recall those of
the genus Brightwellia, Ralfs.1

Cestodiscus gemmifer, n. sp. (Plate II. fig. 7.)

Disciformis ; granulis rotundis ; grandiusculis ; circumconfertis, radiantibus, ad aream
centralem raris et irregulariter distributis, aequalibus, ad marginem minuentibus et
adspectum zonae marginali striolatum tribuentibus ; spinae vel denticuli plures (38) mar-
sinales. In oceano Pacifico.

This large form has a disc about -fa of a millimeter in diameter. It is ornamented by
large round granules, which are crowded towards the peripherjr, but rare and scattered
about the centre. These granules gradually diminish in size near the marginal zone, and
finally assume the appearance of fine radiating striae. The spines which project around the
margin are more numerous than usual, being thirty-eight in number in the specimen figured.
The peculiar irregular distribution of the large granules in the middle area, and the laro-e
smooth central space are, however, the chief distinguishing features of the present frustule.

1 Micr. Journ., n. s., vol. i., 1861, p. 74.

REPORT ON THE 1)1 ATOM ACE/E. 125

Cestodiscus gemmifer, n. sp., var. nov. ^Plate VII. fig. 7.)

The frustulc here shown can only be regarded as a variety of Cestodiscus gemmifer,
n. sp. It differs from the latter (1.) in the character of its granules, which are smaller
and are sparsely and irregularly disposed iu the centre, from which they radiate towards
the circumference, leaving at first smooth alternating areas between the radii, but becoming
crowded more peripherally, and finally diminishing, so as to give a striated appearance
to the border ; (2.) in the greater number — forty-eight — of its denticules ; and (3.) in the
narrower condition of the border, so that the denticules are placed nearer to the margin.

Cestodiscus gemmifer, n. sp., var. decrescens, nov. (Plate XII. fig. 11.)

This variety of the same typical Cestodiscus gemmifer may be distinguished from the
latter in the following respects : (1.) The granules are large in the centre, and diminish
towards the margin ; and (2.) the periphery is ornamented with a quincuncially disposed
granulation, while the outermost belt is very delicately striated.

Cestodiscus parmula, n. sp. (Plate VII. fig. 5.)

Valvis rotundis, granulatis ; granuli potiores vel margaritas e centro irregulari radiantes,
quandoque spatiis intercedentibus, ad marginem confertse et subito in strias exeuntes ; ad
zonam marginalem latam in plures circulos distinctam pauciores spinae vel denticuli
distribuuntur. Ad Oceanum Pacificum.

This frustule is especially distinguished by the extraordinary size of the round granules.
The smooth umbilical area is absent, being replaced by one ornamented with a few large
granules, and from this radiating granulated lines pass towards the periphery. Between
these lines bare spaces may sometimes be observed, while secondary granulated lines
become interpolated towards the peripheral ends of the former, and finally the granules
become contiguous before passing abruptly into the condition of fine striae. The convexity
of the central part of the disc is moderate, and the margin — the inner part of which carries
a considerable number of denticules — is broad and divided into several distinct concentric
circles. It thus resembles one of the shields x used in warfare by the ancient Romans, a
circumstance which has suggested the specific name that has been adopted.

Cestodiscus (?) rapax, n. sp. (Plate XVIII. fig. 2.)

Forma rotunda, parva, circum convexa, medio concava ; cellulis aequalibus, medio
radiatim, in margine decussatim dispositis ; pluribus (11) apicibus circum radiantibus, et
ad instar ungularum assurgentibus.

This type differs greatly from all the previously mentioned Cestodisci, and its generic
determination must for the present remain doubtful. Its form is discoid, its centre con-

1 Parmula, a small round shield.

126 THE VOYAGE OF H.M.S. CHALLENGER.

cave, and its marginal portion notably convex. The central area is ornamented with
round radiating granules, while the broad margin has these arranged in a decussate
manner. From the border of the central concave part there extends to the peripheral
rin<r a series of eleven strongly marked radiating areas, which are raised above the plane
of the valve.

Although this singular valve differs from any of the other specific forms of Cestodiscus,
its characters still coincide with those expressed in the definition of this genus as given by
Greville. By the use of extremely oblique light, and the homogeneous immersion objec-
tive of Zeiss, I have been able to recognise that the long elevated areas project hori-
zontally from their origin, and separate afterwards like the talons of a bird of prey — an
observation which has suggested the specific name that has been applied to this inte-
resting form.

Eupodiscus, Ehrenb.

While Pritchard, in his History of Infusoria (p. 842), notifies the characteristics
of this genus, which is especially distinguished by the processes disposed around its
valves, he very justly observes that the genus A ctinocyclus of Ehrenberg is probably iden-
tical with Eupodiscus, and that therefore it is wrong to ascribe the former to the Coscinodisci.
The only notable difference between Actinocyclus and Eupodiscus consists in the number
of the pseudo-noduli or submarginal processes, of which only one occurs in the former,
while two or more are to be found in the latter.

In the Challenger Collection Eupodiscus jonesianus, Grev., is pretty frequent, having
been collected in several localities, such as the Sea of Japan, at Hong-Kong, in the Arafura
Sea, in the Southern Ocean between Kerguelen and Heard Islands, and elsewhere. At
first, however, it is difficult to determine this Diatom specifically. Notwithstanding the
well-known facility which Greville possessed of drawing the most delicate and elegant
diatomaceous frustules, his figure of this species does not agree sufficiently exactly with
its natural appearance. This will be at once admitted when a comparison is made between
plate ii. fig. 3 of the Transactions of the Microscopical Society of London, vol. x., new
series, 1862, and the description which Greville himself gives at page 22 of the same
volume. This description, however, perfectly coincides with the valves collected by the
Challenger, but it is necessary to remark that among the many specimens in my possession
not one is furnished with more than two submarginal processes, while these are disposed
as if the periphery were tripartite, the place of the third process remaining vacant.

Eupodiscus insutus, n. sp. (Plate XIX. fig. 6.)

Disciformis ; tribus processibus submarginalibus munitus ; valvis areolatis et radi-
anter punctulatis ; areolis grandiusculis, subrotundis, rariusculis. In Oceano Atlantico
meridionali.

REPORT ON THE DIATOM ACEjE. 127

This superb Diatom, from the South Atlantic, has a diameter of 186 m. Although at
first sight it seems to be identical with Eupodiscus argus, Ehrenb.,1 1 found, on comparing
it with the type of that species shown in the Typenplattcn of Moller, and with other
specimens in my possession, that it could not be confounded with the latter. Thus the
figure here given — which is correct in every way except that the punctation of the margin
is somewhat larger and thicker in the original, a character which could not be exactly
drawn without confusing the areolae — shows that these areolae are never stellate, but
irregularly circular. The three submarginal processes divide the perimeter of the valve
irregularly. The double structure, consisting of areolae and granules, observable in the
valve, as well as in that of Eupodiscus argus — a peculiarity which is also met with in the
species of other genera — seems to point to the existence in the cell walls of two strata, of
which one may sometimes become detached from the other, as if the more finely granulated
layer served as a lining for the other.

Actinoptychus, Ehrenb.

This genus was instituted by Ehrenberg in 1838, and was defined as follows:2 —
"Frustules disciform, cellulose, disc divided into equal triangular compartments by lines
or internal septa."

The term septa here employed is unfortunate, as it is used for those lines of division
which, in an almost diaphragm-like manner, cross the lumen of the diatomaceous cell.
e.g., in Striatella, Rhabdonema, &c. ; but in the present genus the lines of division
are but indications of the alternate depression and elevation of the compartments, and
are not septa in the usual sense.

Although the general aspect of the genus Actinoptychus is similar to that of Omphalo-
pelta (Ehrenb.), Cstr., the latter is readily distinguished by the possession, in each of the
triangular compartments, of an intramarginal denticule or spine.

Smith, in his Synopsis of the British Diatomaceae, has, on the other hand, confounded
the genus Actinoptychus with that of Actinocyclus — an error which resulted from his
having regarded as an Actinocyclus a species which was wrongly classed in this genus by
Kiitzing, instead of- being grouped with the species of Actinoptychus.

Actinoptychus raeanus, n. sp. (Plate VII. fig. 4.)

Forma rotunda, granulata, cellulosa; sex dissepimentis triangularibus distincta, cellulis
(areolis) subquadratis, denticuli grandiusculi plures intramarginales irregulariter distributi ;
umbilicus definitus hyalinus. Fossilis in deposito Ste Monicas in California.

Among the excellent preparations of Diatoms made from the Challenger collections by

1 Smith, Synop. Brit. Diat, vol. i. p. 24, pL iv. fig. 39 ; Rabenhorst, Flora Europaea Algaram, sect. i. p.
319, &c.

s Pritchard, op. cit., p. 839. Various species of this genus are figured in Ehrenberg's Mikrogeologie.

128 THE VOYAGE OF H.M.S. CHALLENGER

Dr James Rae, E.N., and kindly placed at my disposal in connection with my present
work, I found this singular frustule, wliick, although brought from the interesting and
rich deposits of St Monica in California, I have deemed worthy of being figured in order
to convey a clear idea of its remarkable structure. The large disc possesses six compart-
ments, and the valve is delicately granulated in a quincuncial manner. The large cellules
or areolae possess a subquadrate form — a character which is not represented in any other
type. At the margin numerous large denticules occur ; these are irregularly distributed,
and project from the inner part of the margin so as to cause the frustule to assume the
appearance of an irregularly toothed wheel.

The specific name which has been given is in honour of its discoverer, Dr James Eae.

Actinoptychus erosus, n. sp. (Plate VII. fig. 8.)

Frustulum disciforme in decern septa a tot lineis lsevibus divisum, et decussatim
granulatum ; granulis prope marginem et alibi quandoque cessantibus ; area centrali
hyalina rotunda. In Atlantico meridionali.

This form, which was collected in the South Atlantic, is divided into ten compart-
ments by ten smooth lines which do not quite reach the centre, this being occupied by an
umbilical hyaline area. The compartments are granulated in a quincuncial manner, except
in some small irregular spots, which are smooth, as if the valve had been eroded at the
places in question. The margin consists of a smooth narrow belt.

The eroded-like aspect of the compartments has suggested the specific name.

Omphalopelta (Ehrenb.), Cstr.

Much confusion has hitherto existed in connection with the genera Actinoptychus and
Omphalopelta, both of which were instituted by Ehrenberg. The definition of the latter
genus, as given by Pritchard in his History of the Infusoria (p. 841), is as follows: —
" Frustules simple, disciform ; disc cellulose or punctate, divided by imperfect septa into
cuneate rays ; centre hyaline ; spines one to each compartment."

Among the four species given by Pritchard under this genus, Omphalopelta cellulosa,

Ehrenb., and Omphalopelta areolata, Ehrenb.,1 have marginal spines in three of the

alternate compartments, while Omphalopelta punctata, Ehrenb., is said to have obsolete

spines. But it is difficult to understand how forms possessing such a structure can be

ascribed to this genus. With respect to the so-called Omphalopelta punctata, Ehrenb.,

there can be no doubt that it should rather be designated Actinoptychus punctatus, as

it coincides with all the requirements of that genus, whde in the case of the other two

above-mentioned species, either the original definition given by Ehrenberg must be

modified, or they must be enrolled in a new genus.

1 Ehrenberg, Mikrogeol., pi. xxxv. A 18, fig. 12. = Actinocyclus areolaris, Brightwell, Micr. Journ., vol.
viiL p. 93, pi. v. figs, la and 16.

REPORT ON THE DIATOMACRjE. 129

This inconsistency has not escaped the observation of Professor H. L. Smith, who, in his
synopsis of the families and genera of Diatoms, which is given in Dr van Heurck's work,
entitled, Le microscope, sa construction, son maniement et son application a l'anatomie
vegetale et aux Diatomdes, modified the definition of the genus Actinoptychus, Ehrenb.,
as follows : — " Valves with marginal spines obsolete ; or if present, few, and in alternate
compartments."

In the family of the Heliopeltece he also gives the genus Halionyx, Ehrenb.,1 to which
he attributes the following characteristics : — " Valves with a hyaline (stellate)
umbilicus, with marginal spines or teeth connected by a radial rib." But these
characters are precisely those of the genus Omphalopelta — a genus which Professor
H. L. Smith regards as synonymous with Actinoptychus, which accordingly possesses
spines in alternate compartments, or in each compartment — a circumstance overlooked by
Professor Smith. Hence it is preferable to leave the genus Actinoptychus intact, and
to modify Ehrenberg's definition of Omphalopelta, limiting the principal characteristic
to the marginal spines, whether they occur in each compartment, or in alternate compart-
ments. Moreover, the number of spines should be hrnited in number, in order to avoid
confusion between Omphalopelta and Heliopelta, Ehrenb., because in this last group the
numerous spines, in addition to its other characteristic marks, do not, when carefully
examined, present the appearance of spines, but of small spheroidal processes.

The genus Omphalopelta may accordingly be defined as follows : — Frustulum simplex,
discoidale ; valvis cellulosis vel punctatis, in cuneatis dissepimentis distinctis ; area
umbilicali hyalina ; spinis marginalibus in omnibus vel in alternis dissepimentis."

Omphalopelta antarctica, n. sp. (Plate VII. fig. 2.)

Valvis areolato-punctatis, in sex partes alternatim elevatas et depressas distinctis;
area umbilicali parva, hyalina, indefinita ; denticulo in elevatioribus dissepimentis sub-
marginali ; lato margine irregulari punctulorum corona distiucto. Ad mare Antarcticum.

This small disc, which was collected in the Antarctic Ocean, has six alternately elevated
and depressed compartments. The valve is delicately punctated in a quincuncial manner,
and also possesses distinct rounded cellules or areolae. The hyaline centre is small and
not distinctly defined, while a single intramarginal denticule occurs in each of the three
elevated segments, but is absent from the depressed segments. The margin is for the most
part smooth, but it is distinguished externally by an irregular circle of small points. The
occurrence of this circle of puncta, as well as of the rounded areolae, is especially
noteworthy.

Omphalopelta japonica, n. sp. (Plate VII. fig. 10.)

Areolato-punctata ; area umbilicali indefinita; dissepimentis pluribus (18) elevatis et
depressis, et denticulo marginali distinctis. In mari Japonico.

1 Ehrenberg, Mikrogeol., pi. xxxv. A 21, fig. 12.
(bot. chall. exp. — paet iv. — 1886.) D 17

130 THE VOYAGE OF H.M.S. CHALLENGER.

la Pritchard's History of the Infusoria (p. 841), we read: — "All the species of
Omphalopclta resemble Actinoptychus senarius," and indeed all the specific forms there
recorded have six compartments and the characteristic submarginal spines. The Diatom
now in question, however, forms a beautiful disc, with eighteen compartments, each of
which is furnished with a distinct denticule or spine. As in all the species of Omphalo-
pelta, the segments are alternately elevated and depressed, while the centre is occupied
by a smooth ill-defined area. The axis of each segment is marked by a radial line, which
extends from the central area and disappears soon after reaching the middle of the com-
partment. At the circumference, and at the middle of the outer margin of each compart-
ment, there is an evident denticule which is of the utmost importance for the generic
determination. The valve is ornamented with small decussately disposed granules, while
traces of areolation are found here and there.

The frustule presents a great resemblance to a Diatom found abundantly by Shadbolt 1
in guano from Callao, as well as in the Port Natal gathering, upon which that observer
instituted the genus Actinophcenia, and which he named Actinophcenia splendens. It is
hard to understand the reason which induced Pritchard 2 to class this form with the genus
Actinoptychus, since he described it as possessing a denticule or spine in each compart-
ment, and so should have named it Omphalopelta splendens. On comparing, however, the
description with our preseut figure, it may be observed that in Sbadbolt's species the
umbilical hyaline area, instead of being indefinite, is well defined. It is further to be
noted, on consulting the figures of Actinophcenia splendens given by Eoper,3 that the
denticulcs correspond to the jjoint of division between two compartments. Thus in
Pritchard's supposed Actinoptychus, the segments would not be alternate in the ordinary
sense, we should rather have to regard the number of the compartments as being double
what is stated, and to look upon the denticules not as placed at the middle of the segments
but as marking the extremities of the divisional lines. Moreover, in our present frustule,
the number of denticules is greater than in that of Shadbolt, and the vertices of the elevated
segments are narrower than those of the depressed segments. Finally, Shadbolt's species
is destitute of the striated margin seen in the present case. From such considerations
the specific value of the present form from the Sea of Japan can hardly be questioned.

Omphalopelta (?), sp. (?) (Plate XVI. fig. 8.)

This small irregular disc is probably a monstrous form of some species of Omphalo-
pelta. The cuneate septa are almost invisible, while the outline of the valve is entirely
anomalous. Like other species of this genus, however, it possesses a hyaline umbilical

1 See a note entitled A defence of the proposed new genus Actinophcenia, Shadb., in Micr. Journ., vol. ii.
pp. 201-203; Edwards on Diatomaceas collected in the United States, Micr. Journ., vol. vii. 1859, p. 88.

2 Pritchard, op. cit., p. 840.

3 Eoper on the Diatomaceae of the Thames, Micr. Journ., voL ii. pp. 73, 74, pi vi. fig. 2.

REPORT ON THE DIATOMACEiE. 131

area of indefinite outline and submarginal spines, one of which occurs in the middle of
each fasciculus. The granulation is fasciculate — a circumstance which prevents me from
determining the genus to which the frustule belongs, as other species belonging to this
group and possessing a similar condition of striation are unknown to me. Should similar
forms be recorded in future, they must be regarded as normal, and not as teratological, or
should a species of Omphalopelta be discovered with fasciculate striation, the present
specimen may probably be attributable to that species.

Omphalopelta parda, n. sp. (Plate XVIII. fig. 9.)

Forma magna, rotunda, in sexdecim dissepimentis alterne elevatis et depressis
divisa ; margine striato et granulorum corona cincto ; dissepimentis depressis brevi radio
ad verticem instructis, elevatis ; denticulo in media basi ; superficies celluloso punctata ;
area umbilicalis indefinita. In mari Pacifico.

This superb discoidal Diatom was procured from mud, which was brought up from a
depth of 1875 fathoms near Japan. The disc is bounded by a well-defined striated border,
which is provided on its inner side with a corona of granules. Interiorly it is divided
into sixteen cuneiform compartments, which start from a smooth central area, with
indefinite outline. These compartments are alternately elevated and depressed, the latter
appearing fiat, while the former are longitudinally ridged in a roof-like manner. Each of
the depressed segments is marked at its vertex by a radial line which soon vanishes, while
each of the elevated areas bear a denticule near the margin, and at the middle of the basal
side ; this denticule is united to the central area by means of a narrow linear space. The
valve is punctated in a decussate manner, and presents ill-defined traces of areolation,
especially in the elevated compartments. Each of the depressed segments presents a
remarkable smooth broadly linear area at its base : in this no trace of striation could be
determined.

According to the amended definition which has been given above, this very attractive
Diatom must be classed in the genus Omphalopelta ; but, apart from this emendation, it
should be ascribed to that of Actinoptychus, Ehrenb.

Omphalopelta shrubsoliana, n. sp. (Plate XXIII. fig. 2.)

In sex dissepimentis alternis triangularibus distinctum ; area umbilicali hyalina sub-
hexagonali ; dissepimentis decussatim granulatis, alterne inermibus et spina a margine
dissita instructis. Dissepimenta medio turgescunt. Ad Zebu.

This disc was collected in the Channel of Zebu, among the Philippine Islands. It is
divided into six triangular segments, of which the alternate ones are similar to each other.
These compartments are granulated in a quincuncial manner, and are slightly swollen at
the centre. Each of the three alternate segments whose apices are_transversely truncated

132 THE VOYAGE OF EM.S. CHALLENGER.

bear at their base, and at a considerable distance from the margin, a denticule which
serves to distinguish the present Diatom from other species. The apices of the non-
denticulated compartments are rounded, so that a central hyaline hexagonal area results,
which is circumscribed by straight and concave sides. The margin of the valve is
convex.

The specific name has been given in honour of Mr Shrubsole, who is well known for
his researches on the fossil Diatoms of the Lower Eocene of the London basin.

Asteromphalus, Ehrenb.1

This genus essentially consists, like that of Asterolampra, Ehrenb.,2 of an areolated
granulated disc, which bears a hyaline star and umbilical lines of division. It differs,
however, from the latter in the single important circumstance that one of the radial areas
is so much narrower than the others as to be almost obliterated. The umbilical lines and
the form of the granulated segments constitute the principal distinctive characteristics
between the specific types. The genus, which was established by Ehrenberg, has been
defined as follows : — " Frustules simple, disciform ; disc as in Asterolampra, but with
two of the punctated compartments approximate, and the interposed ray narrower than
the others." 3

Marine Diatoms belonging to this group are by no means rare, but the transparency
of the valves is such that they are often not perceived in Canada Balsam preparations.
Although the genus Asteromphalus is richer in species than that of Asterolampra, the
following new forms belongino- to the former have been observed in the Challenger
collections : —

Asteromphalus ovatus, n. sp. (Plate V. fig. 7.)

Elliptico-rotun datum ; segmentis rotundato-complanatis ; lineis umbilicalibus rectis vel
non angulariter curvatis, ab area pyriformi dimanantibus ; areis radialibus marginem non
attingentibus. In mari Antarctico.

This very small oval Diatom, which in the figure is magnified 1000 diameters, is dis-
tinguished by the circumstance that the umbilical lines do not separate from one another,
as in the others, at a central point, but originate from the circumference of the area of the
obliterated radius. This characteristic has been found, it is true, in other allied forms,
among which Brebisson constituted the genus Spatangidium ; 4 but this genus has not

1 Monatsber. d. h. Akad. d. Wiss. Berlin, 1844, p. 198.

2 See definition in Monatsber. d, k. Akad. d. Wiss. Berlin, 1844, p. 73.

3 Pritchard, op. ait., p. 836.

4 e.g., Spatangidium flabrllatum, Spatangidium peltatum, Spatangidium araclme, Spatangidium heptactis, in
Bull. Soc. Linn, de Normand., voL iii. ; Spatangidium ralfsianum, Norm., Greville on Diatomacete observed in
Californiai) Guano, Micr. Journ., vol. vii. p. 161.

REPOKT ON TEE DIATOMACI'LE. 133

been accepted, as it is not sufficiently distinct from that of Asteromphahis, to which the
forms in question accordingly remained annexed.

Asteromphahis ovatus presents, however, several points of resemblance to the Asterom-
phalus shadboUianus of Greville,1 who described this latter Diatom as an Asterolampra,
having endeavoured to join the two genera Asteromphahis and Asterolampra into one
under the latter name, because by supposing one radius of Asterolampra to remain unde-
veloped there would be left no means of distinguishing them. This supposition, however,
has failed to meet with the acceptance of other micrographers. The points of difference
which may be noted between the frustule now before us and Asteromjihalus shadboUianus,
Grev., are the following: (1.) The form of the umbilical lines which proceed from the
superior margin of the pyriform area of the obliterated radius are angularly curved and
almost broken in the Grevillian species, while in the present form they are straight, with
the exception of those separating the inferior radial areas, which are slightly curved, but by
no means bent in a zigzag manner, as in the other case. It is to be remarked, however,
that the occurrence of a bending in the umbilical lines is of little significance, because all
gradations may be found in the same species, from the most decided curves to simple
lines which are but slightly tortuous or may even be straight. (2.) The present form is
smaller than the species established by Greville; and (3.) the former, instead of being
round like the latter, is oval. From these considerations the specific value of Asteromphahis
ovatus cannot be questioned.

Asteromphahis roperianus, Grev., var. atlantica, nov. (Plate V. fig. 3.)

This frustule also belongs to the group designated Spatangidium by Brebisson. It
presents a beautiful granulated disc on which a hyaline star with six radii — 'apart from
that which is obsolete — occurs. It is very closely allied to Asteromphahis roperianus,
Grev. : " in both there is the same number of radii, of interradial areas with straight
vertices, and the central area of the obsolete radius, at which the other umbilical fines
meet, is campanulate in both. It differs from Greville's species, however, in the following
respects : (1.) The extremities of its radial areas are slightly expanded, and terminate
some distance from the margin of the frustule. (2.) The granulated segments at the
apices terminate abruptly and not in a somewhat undulating manner, as in the figure
o-iven by Greville, who represents two slightly convex curves as starting from the two
angles of the apex and meeting in a slight involution at the intermediate umbilical line.
Notwithstanding these points of difference, however, the present frustule can only be
regarded as a variety of Greville's species, and it has been named atlantica from the locality
in which it was found.

1 A monograph on the genus Asterolampra, including Asteromphalus and Spatangidium, by R. K. Greville
in Micr. Jonrn., vol. viii. p. 121, pi. iv. fig. 19, 1860.

2 Micr. Journ., vol. viii. pi. iv. fig. 14, 1860.

134 THE VOYAGE OF H.M.S. CHALLENGER

Asteromphalus challengerensis, n. sp. (Plate V. fig. 2.)

Forma rotunda ; lineis umbilicalibus tribus a centro dimanantibus, hinc bipartitis vel
tripartitis ; linear umbilicales rectce ; segmentorum vertices obtuso-rotundati ; area radii
obsoleti fere obliterata. In mari glaciali Antarctico.

This round Diatom bears a star with eight radii apart from the obsolete radius. The
umbilical lines are three in number, start from the central point, and afterwards subdivide.
The vertices of the segments are obtusely rounded, and the central area of the obliterated
radius has vanished.

Asteromphalus challengerensis (?), n. sp. (Plate IX. fig. 2.)

Forma monstrosa. In mari glaciali Antarctico.

The frustule here represented is exceedingly instructive and interesting. It is evi-
dently a monstrous form, and doubtless originates from the coalescence and simultaneous
development of two germs which have combined to form a single frustule. Although
similar teratological phenomena frequently occur in the fruits of phanerogams, I have not
met with an example hitherto among Diatoms, nor am I aware of its having been observed
by others. Although the precise specific determination of such an abnormality is a matter
of considerable difficulty, no doubt can exist as to the generic name, inasmuch as an
obsolete radius is present, and at once determines it to be an Asteromphalus. Moreover,
from the rounded form of the granulated segments, the linear outlines of the radial areas,
the locality in which it was procured, and its association with other frustules of the same
genus, it can hardly be doubted that we have here to deal with a specimen of Asterom-
phalus challengerensis.

It is to be noted that, while we are here brought face to face with the development
of a double germ or double sporule, the two germs, though the products of the same parent,
possess different characters, so that the one would apparently have given rise to a frustule
with a smaller number of radii than the other. If this be admitted, then it is clear that
no importance should be placed on the number of radii in specific determinations.

Asteromphalus wyvillii, n. sp. (Plate V. fig. 6.)

E maximis ; radiis prseter obsoletum viginti sex, intra marginem terminantibus den-
ticulo signatis ; areas granulatse intrarnarginales ad verticem truncata? ; linese umbilicales
ab area radii medii circumradiantes, simpbces vel semel et bis dichotomse, et medio
dupliciter curvatse. Diametrum arese centralis 65 /*, valvse 226 m. In mari Pacifico.

This is without doubt the most singular and superb specimen of the genus Asterom-
phalus which has up till now been recorded. Its immense size ( = 226 n in diameter),
the great number of its radii, and the relative smallness of the central disc of the star
are remarkable. From the outline of the central area of the obsolete radius proceed many

REPORT ON THE DIATOMACEiE. 135

of the umbilical Hues of the twenty-six remaining radii, and among these some remain
simple, while others bifurcate once or twice. Almost all the lines have a double curve in
the course of their length, and the hyaline radii terminate before arriving at the circum-
ference in a salient or denticulate point. The apices of the granulated segments are
sharply truncated.

The specific name of this elegant frustule has been given in honour of the late
Professor Sir Wyville Thomson, formerly Director of the Challenger Commission.

Asteromphalus antarcticus, n. sp. (Plate XVI. fig. 11.)

Forma rotunda, parva ; dissepimentis cuneatis, ad verticem truncatis ; lineis umbili-
calibus rectis ; areis radialibus linearibus et denticulo terminal: instructis. Ad mare
Antarcticum.

This small disc was collected by means of a surface net to the south of Heard Island
in the Antarctic Ocean, being found associated with many specimens of Asteromphalus
darwinii, Grev.1 It possesses five radii — apart from the obsolete radius — and the
umbibcal fines are straight and proceed from the centre. The radial areas are broadly
linear, and each terminates near the margin in a well-defined granule. The areolated
segments are bounded by internally concave margins, with the exception of those which
flank the obsolete radius, and which pass by an elegant curve into the smooth area in the
centre of the disc. The ensemble of such characters must be looked upon as sufficient
to constitute a good species.

Asterolampra, Ehrenb.

The genus Asterolampra, though bearing a great affinity to certain forms of Asterom-
phalus,2 has been too hastily united with the latter by Greville.3 It embraces some of the
most elegant forms of Diatoms, and has been defined by Ehrenberg, as noted by Pritchard
(op. cit., p. 836), in the following manner: — " Frustules simple, disciform; disc orbicular
with marginal areolated or punctated compartments, separated by smooth rays which
proceed from a hyaline central area ; central area divided by lines which radiate from the
umbilicus to the apex of each compartment, compartments and rays symmetrical." The

1 Micr. Journ., vol. viii. pi. iv. figs. 12 and 13.

2 This genus was established by Ehrenberg for certain Antarctic Diatoms brought home by Sir Joseph
Hooker, and collected during the cruise of H.M.SS. "Erebus" and "Terror," Monatsber. d. k. Alcad. d. Wiss.
Berlin, 1844, p. 198. See also Kiitzing, Spec. Alg., 1849; Araer. Journ. Sci. and Art., vol. xxii. p. 1; Ehren-
berg Mikroceologie, pL^ xxxv. A closely allied genus, Spatangidium, was proposed by Br^bisson in 1857»
Bull, de la Soc. Linn, de Normand., vol. ii.

• 3 Greville, Descriptions of Diatoms observed in Californian Guano, Micr. Journ., vol. vii. p. 157 ; Greville
on the Asterolamprae of the Barbados Deposit, Micr. Journ., n. s., vol. ii. p. 42 ; Wallich on Siliceous Organisms
found in the Digestive Cavities of the Salpas, and their relation to the Flint Nodules of the Chalk Formation,
Micr. Journ., voL viii. p. 44.

136 THE VOYAGE OF H.M.S. CHALLENGER.

rays of the hyaline star are all equal to one another, and the lines which radiate from the
umbilicus are well defined on the circular granulated area. Though several specimens of
this genus have been observed, none present any very well-marked characteristics.

Asterolampra grevillii, (Wall.) Grev., var. eximia, nov. (Plate V. fig. 5.)

This magnificent disc was procured from soundings taken at the equator in the Atlantic
Ocean. It is 175 fi in diameter, and although it does not agree in all its features with
any species hitherto recorded, it approaches most closely to Asterolampra grevillii, (Wall.)
Grev., which has been defined by Greville 1 in. the following terms : — " Areolated segments
square at the base ; rays numerous ; umbilical lines divided and arranged in parcels or
groups of from two to five lines each." la this description it is to be remarked that the
term base ought rather to be apex, as it has reference to the smaller extremity of a
pyramidally shaped area, while the word square, though applying exactly to the frustule
now in question, should be replaced by the designation subquadrate, as may be at once
observed on consulting the figures given by Greville and Wallich.

The following additional points of distinction between Asterolampra grevillii and the
Challenger frustule, may also be pointed out : — (1.) The umbilical lines in the former are,
as above indicated, in groups of from two to five, but in the latter of from three to seven ;
(2.) The number of rays in the former are from thirteen to seventeen, while in the latter
they are twenty ; and (3.) the radius of the central disc of the former exceeds one-third of
that of the entire valve, while it is less than a third in the latter.

All these distinctions, however, can hardly be regarded as of sufficient importance to
justify the establishment of a new species for the frustule now figured for the first time,
but the elegance of its ornamentation has suggested the varietal name eximia which has
been selected.

Asterolampra decora, Grev., var. nov. (Plate XVI. fig. 9.)

The frustule here shown bears a strict analogy to the Asterolampra decora of
Greville. The designation picnctiform, however, which has been applied to the latter,2
cannot be given to the cellules of the present specimen, which possesses nineteen rays
instead of the variable number — five to fourteen — recorded by Greville in the course of
his examination of twenty-five specimens of the typical species. Such distinctions seem
to indicate clearly the varietal character of the present Diatom.

Heterodictyon, Grev.

Greville established the genus Heterodictyon in the year 1863, in his interesting paper
on new and rare Diatoms published in the Transactions of the Microscopical Society of

1 Micr. Journ., vol. viii. p. 113, pi. iv. fig. 21 = Asteromphalus grevillii, Wall., Micr. Journ., vol. viii. p. 47,
pi. ii. fig. 15.

2 Micr. Journ., u. 3., voL ii. p. 45, pi. vii. figs. 4—6.

EEPOET O.N THE DIATOMACE^E. 137

London, new series, vol. ix. p. GO, and defined it in the following manner : — " Frustules
free, disciform ; disc with radiate or scattered cellules or puncta in the middle portion, and
a circle of large intramarginal cellules." The genus has hitherto comprised only the two
species Hctcrodictyon rylandsianum, Grev., and Ileterodictyon splendidum, Grev., both of
which were found by Greville ' in the Barbados deposit, Cambridge, U.S.A. ; but I am of
opinion that the former, on account of the corona of large linear or oblong cellules with
which it is provided, should be ascribed to the genus Asterolampra, and named Astero-
lampra vulgaris, Grev.,2 which is from the very same deposit, and of which other six
different forms have been given.

Heterodictyon jerrreysianum, n. sp. (Plate XXI. fig. 8.)

Disciforme ; medio inordinate cellulosum ; subtriangularium areolarum corona submar-
ginali ; spatiis intramarginalibus cellulis decrescentibus obsitis. Ad insulas Phdippinas.

This new species was collected in the neighbourhood of the Philippine Islands. Though
its elegant form might at first sight be mistaken for that of Asterolampra marginata,
Grev.,3 it may be readily distinguished from the latter, in which the central part is covered
with puncta disposed in radiating lines, by having the sculpturing arranged without order.
The corona of large cellules on the submarginal area is composed of subtriangular
compartments, two of the angles being blunted at the points of contact with adjoining-
compartments. The cellules in the intramarginal segments, circumscribed by the sub-
marginal belt, diminish towards the periphery.

The specific name has been given in honour of the well-known English naturalist, Dr
Gwyn Jeffreys.

Brightwellia, Pialfs.

This genus was established by Ralfs in honour of the well-known microscopist,
Brightwell. The Diatom, which the latter observer had named Craspedodiscus coronatus,4
was found to differ from the other members of that genus by the possession of a corona of
oblong cellules, which separated the central from the peripheral part of the disc, and Ralfs/'
who first drew attention to this important circumstance, ascribed the frustule to his new

1 Micr. Journ., n. s., vol. iii. p. 66, pi. iv. figs. 6 and 7.

2 Micr. Journ., n. s., vol. ii. p. 47, pi. vii. figs. 17-22.

3 Micr. Journ., n. s., vol. ii. p. 50, pi. viii. fig. 30.

4 Brightwell on some of the Earer or Undescribed Species of Diatomaceae, Micr. Journ., vol. viii. p. 95,
pi. v. fig. 6.

'•> Pritchard, op. cit., p. 940. " This species is very variable in size. In a dry state it is of a purplish or
brown colour, but in balsam hyaline ; the centre has the granules irregular near the umbilicus, and interrupted
by blank rays ; but near the circlet of cellules they become more regular, and form curved, moniliform lines.
The broad limb is usually brownish when dry, and marked by numerous radiating lines similar to those of
Coscinodiscus concinnus, and have in the intervals extremely minute obliquely arranged granules. The radiating
lines, although conspicuous in the dry state, nearly disappear in balsam."

(bot. chall. exp. — part iv. — 1886.) -D 18

138 THE VOYAGE OF H.M.S. CHALLENGER

genus, which he defined as being a disc with a large granulated centre, separated from a
broad punctated limb by a circle of oblong cellules.

Among the Diatoms collected by the Challenger, a beautiful specimen of this genus
has been found, but it does not exactly agree with all the characteristics indicated in the
definition given by Ealfs. The disc is notably convex, and there is a central round space,
covered with beautiful granules, which decrease in size as they approach the centre, and
are at the same time disposed in excentric curves. Similar characteristics are to be found
in Brightivellia elaborata, Grev.,1 and Brightwellia johnsoni, Kalfs, MS. ; but the markedly
round form of the large cellules, which constitute the corona and divide the central from
the outer part of the valve, readily differentiate our new frustule. The generic definition
may accordingly be modified in the following manner : — Frustulum disciforme prsestans
partem centralem cellulosam a reliqua superficie punctulata vel granulata per coronam
majorum cellularum divisam.

Brightwellia murrayi, n. sp. (Plate X. fig. 2.)

Forma rotunda, convexa ; centro cellulis decrescentibus in lineas excentricas et
concavas dispositas iusigni ; a lato margine radiatim granulato per coronam grandius-
cularum rotundarum cellularum distincta.

This elegant disc is notably convex, and possesses a broad margin, which is covered
with thickly disposed equal radiating rows of granules of medium size. The large central
area is elegantly ornamented by pearl-like granules arranged in excentric curves, and
decreasing in size as they approach the centre. The central part of the valve is separated
from the peripheral by a corona of large round cellules, which serve to render the contrast
between the two portions more conspicuous.

The specific name has been given in honour of Mr John Murray, the present Director
of the Challenger Commission.

■*&v

Porodiscus, Grev.

This genus is akin to that of Coscinodiscus, and was instituted by Greville2 to
embrace forms which agreed with the following definition : — " Frustules free, disciform,
composed of two discs united by an intermediate ring-like zone ; discs very convex,
minutely radiato-cellulate or punctate, with a conspicuous central pseudo-opening or pore."
No specimen of this interesting genus has hitherto been discovered, save in the well-
known Barbados deposit in the United States of America ; and it is therefore of the greatest
importance that the frustule now to be recorded should have been found by the Challenger
in a sounding made at the equator, as it goes to justify the belief that many, if not all,

1 Greville, Descriptions of New and Kare Diatoms, Micr. Joum., n. s., vol. i. p. 73, pi. ix. fig. 1.

2 Micr. Joum., n. s., vol. iii. p. 63.

REPORT ON THE DIATOMACEiE. 139

of the forms at present only known as occurring in geological deposits may ultimately be
discovered in the living condition in the ocean.

Porodiscus stolterfothii, n. sp. (Plate XII. fig. 8.)

Forma rhombico-elliptica ; pseudo-ostiolo subrotundo mediocri ; striis radiantibus
stipatis. In Oceano Pacifico.

This very singular Diatom at first sight recalls the Porodiscus oblongus of Greville,1
which has been characterised as being an elliptico-oblong disc with a large pseudo-open-
ing ; this, by reference to the figure, being found to be equal to one-third of the transverse
axis and perfectly round. In the present case, however, the form of the valve is rhom-
boido-elliptical, and the pseudo-opening is much smaller and somewhat oval, while the
radiating lines of granules are much more crowded than in the Grevillean species.

The association of these differences must be regarded as sufficient to constitute a new
species, which has been named in honour of the English Diatomist Dr H. Stolterfoth,
who has rendered much valuable aid in connection with the present work.

Hyalodiscus, Ehrenb.

The genus Hyalodiscus has been defined by Pritchard (History of the Infusoria, p.
814) in the following terms : — "Frustules simple, disciform; disc smooth, flat, its umbilical
portion or centre separated by a distinct suture ; " and he points out that it differs from
the genus Podosira, Ehrenb., in having a flat disc. I am, however, inclined to adopt the
view of Professor H. L. Smith, who thinks it very probable that the Hyalodisci are only
valves of Podosira — a genus which he unites to that of Mclosira, Ag. But on the other
hand, although Podosira has never yet been met with in long series of many frustules,
it seems to be desirable to retain it as a genus distinct from that of Mclosira, as we
thereby retain a distinction between two kindred series of forms.

It may here be observed, on the other hand, that it is by no means easy to understand
how O'Meara described and figured,2 under the name of Pyxidicula radiata, a form
collected by Mr Moseley at Kerguelen Island, as no Pyxidicula presents an umbilical
area which is separated by a line of suture from the rest of the surface of the valve.
O'Meara ought rather to have named his specimen Podosira (or Mclosira) radiata, as the
form of the valve, which is the only common character between his Diatom and the genus
Pyxidicula, is equally common to the two genera just named.

Hyalodiscus (Pyxidicula, O'Me.), radiatus, var. nov. (Plate X. fig. 1.)

A kindred, if not identical, frustule to that described and figured by O'Meara is here
represented. The disc has its surface divided into three well-defined zones, the wide

1 Micr. Journ., n. s., voL iii. p. 65, pL iv. fig. 5.

2 See Note on the Diatomaceous Gatherings made at Kerguelen's Land by H. N. Moseley, Journ. Linn.
Soc. Land. (Botany), vol. xv. plate i. fig. 9.

140 THE VOYAGE OF H.M.S. CHALLENGER

central area which occupies about two-thirds of the diameter being rugged rather than
granulated. The intermediate zone has very fine radiating striae, while the external circle
is ornamented by regularly disposed granules arranged in a quincuncial manner. In this
form, as well as in that described by O'Meara, dark lines of variable length radiate
outwards from the perimeter of the central zone; these, however, are not "alternately
arranged," as referred to by O'Meara, but are distributed irregularly. Moreover, here
the margin, instead of being "finely striated," is granulated in a quincuncial manner,
each granule being distinctly defined with a magnifying power amounting to only 460
diameters. The additional circumstance, which is well seen in the present frustule, but is
not referred to by O'Meara, that the edge is marked by long, irregularly disposed dark
lines of variable thickness, is noteworthy. Although both Diatoms were collected in the
Antarctic Ocean, they were procured in localities separated by upwards of 40° of longitude;
yet the distinctions which have been pointed out cannot be looked upon as sufficient to
justify the establishment of two distinct species for the two interesting organisms.

Hyalodiscus subtilis, Bail. var. japonica, nov. (Plate XVIII. fig. 4.)

The specimen here shown was collected on the coast of the Sea of Japan. It agrees
generally with Hyalodiscus subtilis, Bail.,1 in the size of the umbilical area, in the
guillocke-like disposal of the granules, and in the character of the line of suture which
limits the central area. This line, however, in the present case is somewhat more indented,
while the granules are not difficult to detect and the margin is notably wider — character-
istics which are of sufficient importance to justify us in regarding this frustule as a
variety of Bailey's typical species.

The fragment shown in Plate XXIV. fig. 4 probably belongs to the Hyalodiscus Icevis
of Ehrenberg.2

Oyclotella, Kg.

The genus Oyclotella, which was instituted by Klitzing, differs but slightly from that of
Melosira — the frustules of the latter forming; longitudinal series, while those of the former
occur singly or at most united in pairs. Though Oyclotella may thus be regarded as but
another form of the genus Melosira, the retention of the former generic name is desirable,
as tending to prevent confusion, and to facilitate the record of so many distinct types.
Of the genus Oyclotella an excellent monograph has been prepared by the well-known
micrographer Dr Albert Grunow, and his work has been incorporated in the Synopsis des
Diatome'es de Belgique of Dr van Heurck.

1 Pritchard, op. cit., p. 815, pi. v. fig. 60.

2 Ehrenberg, Mikrogeologie, pi. xxxiii. 15, fig. 17 ; = Oyclotella Icevis, Kg., and allied to Cyclotella
physoplea, Pritchard, op. cit., p. 814.

REPORT ON THE DIATOM ACE^E. 141

Cyclotella fimbriata, n. sp. (Plate XVII. fig. 16.)

Valvis discoidalibus ; area grandiuscula umbilicali bullato-rugosa, a qua stria) dense
radiantur, ad marginem fimbriam sirnulantes. In Oceano Pacifico.

This Diatom was found in a soundiug made in the Pacific Ocean at lat. 35° 41' N.,
and long. 157° 42' K, at a depth of 2300 fathoms. It is a small but elegant little disc,
adorned with very delicate radiating striae which diverge from the perimeter of a large
umbilical arch, which is rugged rather than granulated. The radiating striae form, near
the margin, a series of very small arches or elegant fimbria} — a circumstance which has
suggested the specific name of this new and interesting Diatom.

Actinocyclus, Ehrenb.

This genus was established by Ehrenberg and has been defined in the following
manner : — "Frustules simple, disciform, disc minutely and densely punctated or cellulose,
generally divided by radiating single or double dotted lines and having a small circular
hyaline intramarginal pseudo-nodule."

In the review of the genus given by Pritchard in his History of the Infusoria (p. 833),
the confusion which has been observed in connection with it is ascribed partly to Kiitzing,
who included a few species of Actinoptychus, and partly to Smith, who, in his classic
Synopsis of the British Diatomaceae, named the genus Eupodiscus, a circumstance which
led both Gregory and Roper into error. The real ground of the confusion introduced by
Kiitzing, however, consisted not so much in the fact of his having embraced some of the
Actinopychi, as in the circumstance of his having falsified the definition given by Ehrenberg,
and in having replaced it by a new and substantially different definition, which ran as
follows : — " Actinoptychus: Individua solitaria, libera; lorica bivalvis disciformis cellulosa ;
cellulae radiis septisque internis radiantibus pluribus interruptae."

It is manifest that Ehrenberg regarded the condition of the punctated or cellulose disc
with its intramarginal pseudo-nodule as the chief and essential characteristic, inasmuch as
he makes use of the word generally in speaking of the more or less punctated radiating
lines, while Kiitzing notes merely the " cellulae radiis pluribus laevibus interruptae"
without making the slightest mention of the pseudo-nodule. It cannot, however, be
denied that the pseudo-nodule constitutes a morphological factor of sufficient import-
ance to be of generic significance, so that it is necessary either to regard, like Ehren-
berg, that structure as an essential characteristic of the genus Actinocyclus, or to institute
a new genus to embrace those cellulated discs which only present radiating lines with
interruptions. In the latter case such lines should not be the means of detaching such
discoidal forms from the genus Coscinodiscus, of which Kiitzing has given the following
definition : — " Individua solitaria, libera, lorica bivalvis sdicea in latere secundario disci-
formis cribrata, sepimentis interioribus radiantibus nullis."

142 THE VOYAGE OF H.M.S. CHALLENGER

Since then the generic designation given originally by Ehrenberg must be looked upon
as valid, it may further be noted that Pritchard and Ralfs1 unite under the designation
of Actinocyclus ehrenbergii not less than one hundred and sixteen species which Ehren-
berg had established by giving special value to the radiating divisional lines, and had
named after all the gods of Olympus, afterstars, historical celebrities, gems, &c. At the
same time the species Actinocyclus moniliformis, Ralfs,2 Actinocyclus raljsii, W. Sm.,3
Actinocyclus fulvus, W. Sm.,3 and Actinocyclus crassus, W. Sm.,3 are retained, while Actino-
cyclus panhelios, Ehrenb., is regarded as a doubtful species. Actinocyclus inter punctatus,
Bright.,4 and Actinocyclus subtilis, Greg.,5 are looked upon as the representatives of a
section whose discs are generally coloured and ornamented by radiating series of points,
while Actinocyclus tessellatus, Roper,6 is regarded as the single type of a second section
embracing cellulated hexagonal discs without radiating lines. The Actinocyclus interpunc-
tatus, Bright., however, has no trace of an intramarginal pseudo-nodule, and, therefore,
cannot be regarded as belonging to the present genus, while Actinocyclus tesselatus, Roper,
possesses neither a pseudo-nodule nor any other characteristic which might authorise its
ascription to this genus.

Rabenhorst, in his Flora Europaea Algarum Aquae dulcis et submarines, follows the
example of Kutzing in omitting the important generic character above referred to, but that
work must be admitted to be greatly deficient generally in its account of marine Diatoms.

It follows, therefore, that although the presence of a pseudo-nodule is indispensable to
the conception of the genus Actinocyclus, the radiating lines or points which divide the
surface into compartments have a more or less accidental character, while among the
specimens collected by H.M.S. Challenger discoidal forms distinctly provided with intra-
marginal pseudo-nodules have been observed to be either finely and closely or sparsely
punctated, to possess discoid surfaces either with or without radiating fines, to exhibit
large radiating cells, or finally, to have a marginal pseudo-nodule and a disc minutely and
closely but irregularly punctated. Since then the essential conditions which separate one
organic form from another must be recorded, and since less essential characteristics become
more and more extended as the discovery of new and kindred forms goes on, the original
definition given by Ehrenberg may now be modified in the following manner : — Frustula
simplicia, disciformia, punctulata vel cellulosa, plerumque nonnulis lineis radiantibus
distincta, circulari pseudonodulo intramarginal i instructa.

1 Pritchard, op. cit., p. 834. 2 = Actinocyclus ternarius, Ehrenberg, Mikrogeologie, pi. xxii. fig. 9.

,. 3 = Eupodiscus ralfsii, Eupodiscus fulvus, and Eupodiscus crassus respectively. These forms, according to
Smith, probably belong to the genus Actinocyclus, Ehrenb., but as he has " limited that genus to frustules with
undulated valves," he has found it necessary to place these apart. Their position in Eupodiscus he regards,
however, as doubtful since " the process in all is rather a pseudo-nodule than a projection from the surface
of the valve." Smith, Synop. Brit. Diat., vol. i. p. 24, vol. ii. p. 86.

4 Micr. Jouni., vol. viii. p. 94, pi. vi. fig. 17.

5 = Eupodiscus subtilis, Gregory, Diatoms of the Clyde, p. 29, pi. iii. fig. 50.

6 = Eupodiscus tessellatus, Micr. Journ., vol. vi. p. 19, pi. iii. figs, la and 16.

REPORT ON TTIE DIATOMACKffl. 143

Actinocyclus clevei, n. sp. (Plate IV. fig. 6.)

E maximis ; dense et decussatim punctulatis ; centrum sex cellulis oblongis dis-
tinguitur, a quo dupliccs perspicuorum punctulorum linea? dimanant, superficiem in partes
aequales dividentes ; pseudo-nodulo intramarginali grandiusculo. Diametrum 360 n.
In mari interno Japonico.

This disc has a diameter of 360 m, and the great delicacy of its punctation contrasts
distinctly with its unusually large size. A central rosette is constituted by a few large
irregular oblong cellules, and from this region numerous double lines of more salient points
radiate outwards and terminate at the circumference, while the equal interspaces are orna-
mented by very minute puncta arranged in a decussate manner. The submarginal
pseudo-nodule is rather large, and is somewhat elevated above the surface in a subulate
manner. The specimen figured was collected in the Sea of Japan.

The specific name has been given in honour of the well-known diatomist Professor
Cleve of Upsala.

Actinocyclus japonicus, n. sp. (Plate IV. fig. 3.)

E maximis ; valvis fasciculatim punctulatis ; fasciculi punctulati lineis radiantibus dis-
septi quae in denticulo desinunt ; punctuli circum in zonam decussatim dispositi ; pseudo-
nodulum submarginale grandiusculum. Ad oras Japonicas.

This magnificent discoid form is not smaller than Actinocyclus clevei, but differs not-
ably from the latter in all its remaining characteristics. The puncta, which are dissemi-
nated over the valve, are less delicate and more sparsely disposed, and at the centre there
exists a group of small points in the middle of an irregular hyaline area. A series of
radiating punctated lines divide the surface of the valve into a number of equal segments,
which are sculptured by rows of fasciculated granules that are arranged in quincuncial
order near the circumference. The radiating lines terminate in submarginally placed denti-
cules, and the external limiting circlet is ornamented by well-defined striae. The Diatom
was found in the vicinity of the coast of Japan.

Actinocyclus ralfsii, W. Sm., var. challengerensis, nov. (Plate XXX. fig. 1.)

The frustule here shown was also collected in the neighbourhood of Japan. Although
in the course of engraving the striae have been somewhat unduly reduced in number, it is
manifest that it closely resembles the Actinocyclus ralfsii of '. W. Smith. It may be noted,
however, that the marginal denticules are more pronounced than in the latter case, while
the system of striae is fasciculate rather than radiating. Notwithstanding these distinc-
tions, it can only be looked upon as a variety of Smith's species.

144 THE VOYAGE OF H.M.S. CHALLENGER

Actinocyclus pruinosus, n. sp. (Plate IV. fig. 2.)

Valvis ad marginem dense deinde rarius punctulatis ; centrum area hyalina nonnulis
punctulis medio signata, a qua pluriinse punctulorum lineae circum distribuuntur.
Diametrum 100 m. In mari Pacifico.

This very elegant form does not entirely satisfy the generic definition given by
Ehrenberg. Thus, although a beautiful round hyaline pseudo-nodule occurs near the
circumference, the disc, which is divided into thirty-five distinct segments by salient
punctated radii, is but sparsely granulated, while, with the exception of a single fasciculus,
the granules are densely disposed near the circumference and around the centre, but
become more rare over a moderately wide subcentral or submedian area. The
centre consists of an irregular smooth area, in the middle of which a small group
of points occurs. The diameter of the valve, which was collected in the Pacific Ocean,
is 100 ft.

Actinocyclus fasciculatus, n. sp. (Plate IV. fig. 8.)

Valvis cellulosis per lineas cellularum radiantes divisis ; segmenta cellulis fasciculatim
dispositis, ad marginem in punctis decussatim distributis transeuntibus, areola umbilicali
tribus cellulis medio signatis ; pseudonodulum submarginale evidens. Fossilis ad Eich-
mond in Virginia.

The cellulated surface of this form is divided into segments by numerous radiating
moniliform lines of equal size and originating from a small central area, the middle of
which is occupied by three small cellules. The segments are ornamented by fasciculately
disposed cellules of uniform size except in an outer zone, where the margin is provided
with very minute granules arranged in a decussate manner. The submarginal pseudo-
nodule is well seen. The frustule was found in a fossilised condition at Richmond,
Virginia.

The specific name has reference to the fasciculate appearance of the sculpturing of the
component segments of the valve.

From the consideration of this and the foregoing species it is manifest that the
discoidal granulated form of the valve, combined with the presence of an intramarginal
pseudo-nodule, can alone be regarded as absolutely distinctive of the present genus, the
character of the granulation varying considerably between divergent extremes.

Actinocyclus fasciculatus, n. sp., var. nov. (Plate IV. fig. 8 bis.)

A frustule very similar to that last described is here shown. It differs, however, in
it3 smaller size, and in its locality, having been collected in the Atlantic at lat. 37° 24' N.
and long. 25° 13' W.

REPORT ON TI1E DIATOMACEdE. 145

Actinocyclus complanatus, n. sp. (Plate IV. fig. 9.)

Valvse ccllulis fasciculatim medio, propc niarginem decussatim ordinatis, hinc in
puuctulorutn lineis trauseuntes ; ad marginem rarse dentieulorum lincae distribuuntur ;
pseudonodulum intrarnargiuale evidens. In niari Japonico.

This Diatom, which was collected in the Sea of Japan, corresponds in most of its
characters to Actinocyclus fasciculatus. Its form is discoidal, and the cellules, which
are equally distributed over the entire surface of the valve, are disposed in a fasciculate
manner. Its centre, however, is not specially differentiated, and the lines which radiated
from that region in Actinocyclus fasciculatus are here absent. In this respect, therefore,
the frustule does not correspond to the definition of the genus as given by Ehrcnberg,1
but the presence of a well-defined submarginal pseudo-nodule clearly indicates its true
generic determination.

The specific name that has been given has reference to the entire absence of salient or
other differential characteristics at the centre of the valve.

Actinocyclus umbonatus, n. sp. (Plate IV. fig. 4.)

Valvse umbonatas, cellulosse ; cellulse rarius in centro, reliquum fasciculatim ordinatae ;
zona marginalis dense lineata ; pseudonodulum submarginale cylindricum lateraliter
extensum. In mari Antarctico.

This beautiful discoid umbonate Diatom was procured in the Antarctic Ocean. It
presents alternate radiating shades as in Podosira maculata, W. Sin.,2 these being due to
the distribution of the cellules in fasciculate parallel lines. Near the margin of the oblique
surface of the Diatom a small subulate cylindrical structure projects. This was observed
in many cases ; it always presented the same appearance, and from its oblique direction
it can only be the homologue of the pseudo-nodule, its position in the strongly umbonate
valve preventing it from appearing round as in other species of the genus. It is impor-
tant to note that the pseudo-nodule may assume this remarkable appearance, constituting
a well-marked process-like elevation on the surface of the valve.

Actinocyclus oliveranus, O'Me. (Plate IV. fig. 7.)

We have here represented one of the commonest and most characteristic forms that
occur in the Antarctic Ocean. It consists of a very small irregularly punctated disc
which possesses a pseudo-nodule near its margin. I had named the organism Actino-
cyclus antarcticus, but on comparing Part iv. of the Diatoms, edited by T. P.
Cleve and T. D. Moller, the same frustule was repeatedly observed by me in preparation
207 made from a sounding taken between Patagonia and the Sandwich Islands, and it
has already been designated Actinocyclus oliveranus, by the well-known Irish microscopist,

1 Pritchard, op. tit., p. 833. 2 Smith, Synop. Brit. Diat., vol. ii. p. 54, pL xlix. fig. 328.

(bot. chall. exp. — part iv. — 1886.) D 19

146 THE VOYAGE OF H.M.S. CHALLENGER.

O'Meara, in his article on the Diatoms collected at Kerguelen.1 The name which I have
suggested can accordingly only be applied in the event of some important distinctions
being ultimately determined between the two types referred to.

Actinocyclus (?) denticulatus, n. sp. (Plate IV. fig. 5.)

Granulis radiatim ordinatis, medio sequalibus, ad marginem minutissimis et stipatis ;
denticulorum corona submarginali ; pseudonodulo minimo margini proximo. In mari
Arafura, et fossilis ad Richmond in Virginia.

Great difficulty has been experienced in arriving at the generic determination of the
present form. This has arisen from the very minute size of the pseudo-nodule, which
could only be recognised after very careful examination, being found, unlike what
occurs in other species, in the immediate vicinity of the margin. Yet this peculiarity
cannot be regarded as sufficient to regard it as not intramarginal, so that the generic
definition is not infringed. It is also to be noted, although not an essential character, that
the disc is not divided into equal parts by radiating lines. The granules are of moderate
size and arranged in a radiating manner, but they become minute and crowded near the
margin. There is a submarginal corona of denticules — a circumstance which has suggested
the specific name that has been applied. The organism was found in the Arafura Sea,
and it has been observed in a fossilised condition at Richmond, Virginia.

Actinocyclus (?) anceps, n. sp. (Plate IV. fig. 1.)

Valvis granulatis ; granuli sequales ad centrum nullo certo ordine, hinc fasciculatim
distributi ; pseudonodulus minimus et ad marginem attingens. In mari Japonico.

The same difficulties as were encountered in the last case, have recurred here in so far
as the determination of the genus is concerned, namely, (l.) the small size of the pseudo-
nodule which is placed close to the margin, and (2.) the absence of radiating lines
dividing the disc into equal compartments. The granules are of equal size and disposed in
a fasciculate manner except in the central area, where they are irregular. This organism
was found in the Sea of Japan.

Actinocyclus punctulatus, n. sp. (Plate XVI. fig. 3.)

Punctulorum lineis ab area umbilicali hyalina circum radiantibus, et in strias exeun-
tibus, nonuullis denticulis distinctioribus ad marginem distributis ; pseudonodulo a
margine paulisper secedente. Ut supra.

This beautiful Diatom approaches the Actinocyclus pruinosus, above described, in

1 Journ. Linn. Soc. Land. (Botany), vol. xv. No. 82, p. 58, pi. i. fig. 7.

REPORT ON TIIE DIATOMACE^E. 147

being ornamented with punctiform granules which proceed from a smooth umbilical
areola, cover the entire surface of the valve, and towards the circumference assume the
appearance of very delicate continuous stria?. But in Actinocyclus pruinosus (Plate
IV. fig. 2) a number of radiating lines of small points pass to the circumference, and these
divide the surface of the valve into a series of segments which are ornamented with
puncta more or less crowded, and disposed in an irregular manner. Around the circum-
ference of the present frustule a number of salient denticules also occur, and the pseudo-
nodule is very distinct and removed from the margin.

The specific name that has been applied is intended to recall the peculiar character of
the punctation.

This Diatom was, like the preceding, collected in the Sea of Japan.

Actinocyclus pellucidus, n. sp. (Plate XXIX. fig 13.)

Valva vix perspicuo lineolarum circulo cincta, interius detegens coronam signoruni
duplo rariorum ; pseudonodulum submarginale distinctissimum. Ad Japonem in mari
interno.

This is without doubt the most singular form among the new species of this genus which
were collected by the Challenger. The excessive minuteness and transparency of the valve
render it almost invisible, and its discovery is entirely due to the use of a superior homo-
genous immersion lens by Zeiss. The submarginal pseudo-nodule is distinctly elevated.
The margin of the disc is surrounded by very minute striae, and somewhat nearer the
centre a corona of small linear dots which are sparsely, but approximately regularly dis-
seminated, seem to correspond to the extremities of the inconspicuous lines that radiate
from the centre. No other ornamentation of any kind could be detected. In a Canada
balsam preparation, the delicacy of the details cannot be made out even by the best
objective, and it is only by the use of monochromatic illumination that they can be
detected with certainty.

The specific name has reference to the transparency of the organism, which was
collected in an inland sea at Japan.

Actinocyclus pumilus, n. sp. (Plate XII. fig. 2.)

Forma minima ; punctulis minimis et inordinatis ; centro et margine nodulis vel
denticulis signato, processu marginali. Ad meridiem insulse Heard in mari Antarctico.

A confused and very delicate granulation ornaments this very minute disc. There is
a central nodule, and the margin of the valve is decorated by a number of salient points.
There is also present an exceedingly minute but distinct marginal process.

This interesting form, which has been named from its small size, was obtained in the
vicinity of Heard Island in the Antarctic Ocean.

148 THE VOYAGE OF H.M.S. CHALLENGER

Euodia (Bailey), Cstr.

Among the Auguliferese, Pritchard 1 enumerates the genera Euodia, Bail., and Hemi-
discus, Wall. The definition of the latter, as given by Wallich,2 is as follows: — "Frustule
free ; valve arcuate, with a marginal nodule ; cellulation hexagonal, radiate." Professor
H. L. Smith, however, according to Dr van Heurck,3 speaks of Hemidiscus as possessing
somewhat different characters, namely, " valves celluleuses, centre blanc, marge veinee."

The genus Euodia, Bail., is defined by Pritchard in these words — "Frustules cellulose
or granulate, in lateral view lunate ; " while Professor H. L. Smith,4 after describing the
Palmeria of Greville thus — " Valves with indistinct umbilicus, finely punctate with
radiating lines, dorsal and ventral margins with minute teeth or spines " — characterises
Euodia as follows — " All others, dorsal margin without spines, ventral frequently with a
small pseudo-nodule."

Thus in the case of Euodia Bailey only requires the lunate form and the cellulose
structure of the valve as distinguishing features, while Smith demands that the dorsal
margin be devoid of spines and the ventral often provided with a pseudo-nodule. In the
case of Hemidiscus, on the other hand, Wallich regards an arcuate form, a cellular radi-
ating structure, and a marginal nodule as essential ; whereas Professor H. L. Smith4
insists on " cellular valves, clear centre, and veined margin."

To avoid such confusion, I believe, from the observations which I have been enabled
to make, that the two genera, Hemidiscus and Euodia, should be united into one. This
view has also been propounded by Pritchard, who, after defining the Hemidiscus of
Wallich, adds — " We doubt whether the Hemidiscus be distinct from Euodia, since the
only distinction seems to be the marginal nodule of the former — a character perhaps
overlooked by Professor Bailey."

In the marine soundings of the Challenger the lunate forms of Euodia and Hemi-
discus have to be recorded as among the least uncommon forms of Diatoms. They
are found of all sizes and in all positions, and they are more or less arcuate in
outline, having circular or parabolic curves, the ventral line being almost plain or
possessing an inflation in the centre, and being not much less swollen than the dorsal line.
I have frequently found specimens with a more or less distinct ventral nodule, but have
at the same time observed a few devoid of any such structure. The latter I named
Euodia, and the former Hemidiscus ; but on several occasions I had the opportunity of
observing some semi-lunate valves in series of two or more frustules, and on examining

1 Pritchard, op. cii., p. 852.

2 On the Siliceous Organisms found in the Digestive Cavities of the Salpaj, and their relation to the Flint
Nodules of the Chalk Formation, Micr. Journ., vol. viii. p. 42, 1860.

3 Le Microscope, sa construction, son maniement et son application special a l'anatomie vegetale et aux
diatomees, par le Dr Henri van Heurck. — Troisieme edition, Bruxelles, 1878.

4 The Lens, vol. i. p. 18, 1872.

REPORT ON THE DIATOMACE.E. 149

the different valves the ventral line and margin alternately were found sometimes to be
provided with, but sometimes to be devoid of, a nodule ; so that in such forms the
superior valve has not the marginal nodule which exists in the inferior. This circum-
stance recalls what takes place in Cocconeis and Achnanthes, since in the former a central
nodule, and in the latter a stauros, is to be found only on the inferior valve.

Since, then, the genus Euodia, Bail., was instituted prior to that of Hemidiscus,
Wall., the name of the united genera must be Euodia, Bail., and its definition may
be amended as follows : — Frustulum cellulosum vel granulatum ; valvis lunatis ; nodulo
ventrali maranali in inferiori valva.

I have already had the opportunity of observing similar lunate valves while preparing a
report on a sounding made in the Atlantic by the "Porcupine" in 1869. * At that
time I noted the occurrence of lunate valves with and without the central nodule, and
possessing profiles and structural characteristics sufficiently distinct from those of Hemi-
discus cuneiformis, Wall.,2 by the absence of the marginal series of points and the much
oreater delicacy of the radiating granulation, to justify the establishment of a new species.
I then had a presentiment, which has now been verified by actual observation, that the
two lunate forms, the one with and the other without the nodule, should be indicated as
the inferior and superior valves of the same species ; and it may be well here to append the
definition of the species then established which is reproduced on Plate XII. fig. 1, so as
to rectify the generic name, and to note the presence of a nodule on the inferior valve : —
Euodia (Hemidiscus) inornata, Cstr. — Valvis lunatis ; linea dorsali parabolica,
ventrali late arcuata ; punctulis vel denticulis marginalibus nullis ; nodulo marginali in
inferiori valva. In mari Atlantico Septentrionali et in Oceano Pacifico.

Euodia recta, n. sp. (Plate XII. fig. 3.)

E maximis ; linea dorsali late arcuata, ventrali recta ; cellulis subradiantibus ; apicibus
acuto-rotundatis. In Oceano Atlantico meridionali.

Among the Euodise brought home by H.M.S. Challenger this form is distinguished by
its singular outline and its large dimensions. It is 132 m in diameter, and its straight
ventral line resembles the chord of an arc formed by the curved dorsal margin. The
extremities are acutely rounded, and the sculpturing is cellulate and subradiating.

The specific name of this Atlantic species has reference to the character of the ven-
tral side,

Euodia orbicularis, n. sp. (Plate XII. fig. 15.)

Forma minima, inflata ; linea dorsali curvata, ventrali arcuata ; apicibus obtusis ;
minimis punctulis conferta. In Oceano Atlantico meridionali.

1 Atti Accad. Pontif. d. nuov. Lincei, March 1871.

2 Micr. Joum., vol. viii. p. 42, pi. ii. figs 3 and 4.

150 THE VOYAGE OF H.MS. CHALLENGER.

In this small form the ventral line is almost as arcuated as the dorsal, so that the
outline is suborbicular or approximately elliptical. The extremities are very obtuse, and
the valve is delicately but irregularly punctated.

Euodia radiata, n. sp. (Plate XII. fig. 4.)

There is here represented a frustule which possesses a very different profile from the
preceding, but has the same delicate striation of radiating puncta. The apices possess a
hooked form, hence the varietal designation applied above. It was procured in the
Atlantic Ocean.

Euodia ventricosa, n. sp. (Plate XII. fig. 5.)

Parva, subradiantibus lineis punctulata ; linea dorsali admodum curvata, centrali late
convexa ; apicibus rotundatis. In Atlantico meridionali.

This small frustule is ornamented with subradiating lines of puncta, which are disposed
in excentric curves. The dorsal line is strongly arched, while the ventral is less convex.
The extremities are bluntly rounded.

The specific name has reference to the turgescence of the ventral margin.

Euodia ventricosa, n. sp., var. nov. (Plate XII. fig. 6.)

A variety which possesses a still more convex dorsal line than the typical species
just described, is here shown. The ventral line, on the other hand, is less arched, but the
character of the punctation and of the extremities is similar in both.

Systephania, Ehrenb.

This genus resembles the Stephanopyxis of Ehrenberg so strongly as to lead one
to the belief that, like Stephanopyxis, Ehrenb., and Cresivellia, Arnott,1 they are
identical. Professor H. L. Smith is inclined to regard the Systephania as simply a valve
of Stephanopyxis, and although much may be said in favour of this view, the absolute
identity has still to be demonstrated. It is indeed true that, according to Ehrenberg, the
parallelism of the rows of cellules which cover the valve is to be regarded as characteristic,
since this was visible in the two species then known, namely, Systephania corona, Ehrenb.,2
and Systephania diadema, Ehrenb. ; ' but in a preparation kindly intrusted to me by Dr
James Rae, R.N., and made from material procured at Richmond, Virginia, five specimens
occur, one of which is figured on Plate IX. fig. 11. This undoubtedly constitutes a new
and very distinct species of Systephania, although the cellules that occur on the valve are

1 Trans. Roy. Soc. Edin., vol. xxi. pt. iv. p. 538, pi. xiv. fig. 109.

2 Ehrenberg, Mikrogeologio, pi. xxxiii. 15, fig. 22, and pi. xxxiii. 18, fig. 11.

REPORT ON THE DIATOMACE/E. 151

only partially disposed in a parallel manner, so that in this particular the original defini-
tion of Ehrenberg ought not to be applied too rigidly.

Systephania raeana, n. sp. (Plate IX. fig. 11.)

Densius cellulata, ordinibus subparallelis, bina aculeorum corona decorata interna
spinis densiusculis subregularibus, altera rarioribus et irregularibus. Fossilis ad Richmond
in Virginia.

This new fossil species, which has been named in honour of Dr James Rae, R.N.,
differs from those hitherto recorded not only in not having the cellulation of the valve
always parallel, but also in being markedly more minute. Its most distinctive character-
istic, however, consists in the double corona of puncta with which it is ornamented, the
inner series being more numerous and regular than the outer. The frustule was found in
a collection made at Richmond, Virginia.

Systephania aculeata, Ehrenb., var. a nov. (Plate IX. fig. 6.)

The form here figured, which must be ascribed to the present genus, was procured in
the Sea of Japan and in the vicinity of Hong Kong by H.M.S. Challenger. It possesses
a convex outline, while the areolation is wide and hexagonal. A delicate but distinct
corona is manifest about the middle of the radii of the valve. The character of the
areolation and of the corona point to an affinity with Systephania aculeata, Ehrenb.,
although in the latter the corona is submarginal.

Systephania aculeata, Ehrenb., var. /3 nov. (Plate XXX. fig. 3.)

The valve here shown is another variety of the typical Systephania aculeata of
Ehrenberg, its corona occupying a position much nearer the margin than that of the
variety a above referred to.

Systephania (?) sp. (?) (Plate XXX. fig. 2.)

The beautiful disc here figured was collected near the Bermudas. It, like the preced-
ing, is convex and ornamented by large hexagonal areoke, while a number of irregularly
disposed submarginal puncta may readily be recognised. At the periphery a band of
very large cellules occurs and gives a somewhat remarkable appearance to this interesting

form.

As only a single specimen of this curious frustule has been observed, its generic and
specific determination must for the present remain problematical, although it may ulti-
mately be found to belong to the genus Systephania.

1 52 THE VOYAGE OF H.M.S. CHALLENGER.

Coscinodiscus, Ehrenb.

Among the numerous different forms of Diatoms which have hitherto been recorded,
those possessing discoidal outlines are without doubt the most elegant in appearance, but
they are at the same time the most rarely met with. Thus among freshwater forms they
are typified by the genus Cyclotella and a few Melosirse, while in marine collections from
the sea-shore or from the washings of sea- weeds they are only found in rare cases. To Ehren-
berg is due the credit of having first directed the attention of naturalists, geologists, and
microscopists to the so-called Infusorial Earths, tripoli and other siliceous deposits formed
by the accumulation of myriads of the siliceous remains of Diatoms that lived in bygone
geological eras. Many of these deposits manifestly represent the bottoms of ancient seas
which have been elevated by subterranean forces, and the examination of the organisms
which they contain has revealed many new diatomaceous forms, among which discoidal
frustules more beautiful and elegant than any before known, occur in great numbers. The
principal type of these discoidal forms is to be found in Coscinodiscus, Ehrenb., a genus which
has been defined by Pritchard (History of the Infusoria, p1 827) in the following manner : —
"Frustules single, discoid; disc cellular or dotted, without processes, defined border, internal
septa, or division into radiating compartments." Hence in brief any simple cellular or
punctated disc is called a Coscinodiscus, while the character of this cellulation or punctation
— whether strongly marked or minute, whether increasing from the margin to the centre or
conversely, whether regular or without order, whether in linear or curvilinear arrange-
ment, whether radiating or excentric, or forming a rosette or umbilicus — constitutes the
differential characteristic of many species.

That the extension of research in this department of marine biology should result in the
continued increase of the number of species of this important genus is not to be wondered at,
and the collection procured by the Challenger Expedition is a large and interesting one.
Three magnificent new species which were brought home by this Expedition have been
already described by the well-known Irish microscopist, Rev. E. O'Meara, M.A., and
to the first of these the name of Coscinodiscus craspedodiscus, O'Me.,1 has been given
on account of its very great size (Plate III. fig. 5). It may readily be seen with the
naked eye, presenting the appearance of a hoop or ring, one millimetre in diameter. Its
hoop-like form is due to the circumstance that the outline is somewhat convex and strongly
siliceous, with large hexagonal areolae, while the central part is very thin and so trans-
parent that an accurately adjusted illumination is required to reveal its sculpturing and to
discover the form of its smooth central areola. The second has been named Coscinodiscus

1 This interesting Diatom has been defined as follows: "Diameter 0O22", centre large, free from arcola-
tion. Areoles radiate, at the margin large, hexagonal, thence somewhat compressed, decreasing in size towards
the centre, and somewhat elongated. Towards the centre some of the radiate lines of areoles are somewhat
shorter than others, in consequence of which the free centre has somewhat of a star-like appearance." — Quart.
J own. Mia: Sci., vol. xvii. p. 5G1

REPORT ON THE DIATOMACE.E. 153

arafuscensis, O'Me., the specific name being probably intended for arafurensis — since the
sea which extends from the Aru Islands to Torres Strait, where the specimen was found,
is called the Arafura Sea. The frustule is a little smaller than the preceding, from
which it may also be distinguished by the following circumstances, namely (1.) the sub-
hexagonal areola) decrease from the margin towards the centre and are stronger than in
Coscinodiscus craspedodiscus; and (2.) the central areola is smaller, loss stellate, and
irregular.

The third form which was brought from Kerguelen is also an enormous disc. It has
been named Coscinodiscus moseleyi, O'Me.1 It is very convex, and its central rosette
is formed by eight large unequal cellules or areolae and minute subquadrate granules
arranged in small radiating groups.

In examining the numerous discoidal frustules which have now to be recorded, it is by
no means easy to determine the limits that are to be set to the genus Coscinodiscus.
Some of the granular or cellular discs correspond perfectly to the generic definition above
quoted, but several other forms possess so extremely delicate punctations as to surpass in
this respect all known species of Coscinodiscus, and to render it highly improbable that
they could have been observed by Ehrenberg with his comparatively imperfect microscope
when he established that genus. This view is confirmed by W. Smith, who, to prove that
his Coscinodiscus concinnus could not be confounded with the Coscinodiscus centralis of
Ehrenberg, remarks 2 that the cellules of the former could not have been detected by means of
the instrument used by Ehrenberg. But the punctations of the forms now in question
are even more minute than those of Coscinodiscus concinnus, so that they cannot be
regarded as conforming to the definition of that genus, but must be looked upon as
belonging to a new genus which I shall name Ethmodiscus3 on account of the ex-
ceedingly fine condition of the granulation.

Coscinodiscus arafurensis, O'Me., var. nov. (Plate II. fig. 4.)

The frustule here figured is one of the largest members of the genus Coscinodiscus
that has to be recorded here. Its diameter is 349 M-, and it is ornamented with large
radiating cellules which become smaller towards the centre, where a smooth irregular
areola occurs.

Though possessing the large size just noted the present valve is smaller than that of
Coscinodiscus arafurensis, O'Me., and much less than that of Coscinodiscus craspedo-
discus, O'Me. With the last-named frustule, however, it has been found to be associated
in a collection made in the Arafura Sea, although Coscinodiscus craspedodiscus, O'Me.,
was also obtained in great numbers in a sounding of great depth taken in the Pacific Ocean

1 Journ. Linn. Soc. Lond. (Botany), vol. xv. No. 82, p. 57, pi. i. fig. 6.

2 Synopsis of the British Diatomaceae, vol. ii. p. 85.

3 rfin,ti, a filter for liquids.

(bot. chall. exp. — part iv. — 1886.) D 20

154 THE VOYAGE OF H.M.S. CHALLENGER.

near the coast of South America. Between Coscinodiscus craspedodiscus, O'Me. — which
is represented on Plate III. fig. 5, and in which the areolation corresponds exactly to
that of the original except in the case of the margin, where the details could not be
adequately pourtrayed — and the present frustule there is no risk of confusion, while the
latter may be distinguished from Coscinodiscus arafurensis, O'Me.,1 in the following
respects : — (1.) It is of smaller size ; (2.) its radiating rows of cellules regularly diminish
from the circumference to the centre, where there is (3.) a smooth area somewhat smaller
than that of Coscinodiscus craspedodiscus, and terminated less irregularly than that of
Coscinodiscus arafurensis. Notwithstanding, however, the difference in size, in the char-
acter of the areolation, and in the condition of the central areola, I am of opinion that we
are here dealing with nothing more than a variety of O'Meara's typical species.

Coscinodiscus mirificus, n. sp. (Plate III. fig. 6.)

E maximis ; granulorum lineis, radiautibus ; area centrali, irregulari, grandiuscula ;
cellulae punctulorum lineis circumducuntur. Diametrum = 326 m. Ad Hong-Kong in
mari Sinensi.

This singular species is closely allied to the above-mentioned Coscinodiscus arafurensis,
O'Me. Its diameter is 326 m., and its large central areola has a very irregular outline.
The granulation is radiating, but the granules are at the same time disposed in exceutric
curves which resemble the guilloche of a watch. When examined with a homogenous
immersion lens and accurately adjusted light each cellule or areole is found to be bounded by
a hexagonal margin of extremely minute punctiform granules (Plate III. fig. 6 a). This
curious frustule is from the neighbourhood of Hong-Kong.

Coscinodiscus papuanus, n. sp. (Plate III. fig. 3.)

E maximis ; granulis minimis radianter per nonnullas lineas distinctions in totidem
denticulos submarginales exeuntes divisis ; centrum nonnullis rarioribus granulis notatur.
Diametrum = 152 /x. In mari Arafura.

This large disc (152 /m. in diameter) is covered with radiating lines of small granules.
These are separated by rows of very minute granules which pass centripetally from as
many submarginal points or denticules, but disappear towards the centre. Here a few
less crowded granules occur, and this circumstance serves to distinguish the present form

1 The original observations published on Coscinodiscus arafurensis, O'Me., are as follows : " The form is
large, diam. O015", however, considerably smaller than the very striking species exhibited ... by Mr. O'Meara
under the name of Coscinodiscus craspedodiscus, a comparison with the leading features of which would best
pourtray the characteristics of the present. Here the broad margin so remarkable in the former is absent. In
the present form as in it the radiate lines of areoles terminate some distance from the centre ; the central blank
space, however, is rrmch smaller, and the lines of areoles are of more equal length. Areoles of margin sub-
hexagonal, diminishing in size towards the ends ; they are shorter, broader, and much more robust than in
Coscinodiscus craspedodiscus." — Quart. Journ. Micr. Set., vol. xvii. p. 463.

KEPOKT ON THE DIATOMACEiE. , 155

from the Coscinodiscus concinnus of W. Smith, with which it agrees in its other charac-
teristics. Although such a difference is a small one, its specific value must still be
admitted. Coscinodiscus papuanus was collected in the neighbourhood of New Guinea.

Coscinodiscus denticulatus, n. sp. (Plate III. iig. 8.)

E maximis ; striis aequalibus, radiantibus ; superficie denticulis sparsim uotata ; mar-
gine striato distincto. Diametrum=130 fx. In mari Pacifico.

This large Diatom is covered with uniformly radiating granules, while the surface of
the valve also bears at intervals more prominent denticules. The margin is formed by a
distinct band, which is delicately fluted and separated from the valve by a well-defined
line. The specific name which has been applied has reference to the prominent valval
denticules.

Coscinodiscus^stellaris, Roper, var. nov. (Plate III. fig. 2.)

The frustule here delineated was found near the ice-barrier of the Antarctic. It consists
of a very small disc which is ornamented by lines of minute radiating puncta of so great
delicacy that it has been found impossible to represent them adequately in the figure.
The centre is marked by a cross formed of four oblong granules analogous to those found
in Coscinodiscus stellaris, Eoper,1 of which the present form must be regarded as a variety.

Coscinodiscus centralis, Ehrenb., var. nov. (Plate II. fig. 3.)

A variety of Coscinodiscus centralis, Ehrenb.,2 is here shown. It was collected in the
Sea of Japan, and only differs from the typical specimen in possessing a fluted margin
which is entirely absent in the latter.

Coscinodiscus variolatus, n. sp. (Plate II. fig. 5.)

Granulis parvulis aequabbus, fasciculatim radiantibus ; superficies granulis elevatioribus
maculata. Ad insulas Philippinas.

The present small but very elegant disc from the neighbourhood of the Philippine
Islands is ornamented with small but distinct granules, which are disposed in fascicules
consisting of parallel lines. The surface is spotted and the spots result from small groups
of granules which rise beside one another in the form of denticules. The specific name
has reference to the pitted appearance shown by the surface of the valve.

Coscinodiscus patera, n. sp. (Plate II. fig. 6.)

Forma umbonata; area umbilicali nonnullis punctulis notata, a qua punctulorum linear
radiantur, quandoque prope marginem cessantes. In mari Pacifico.

1 Micr. Journ., vol. vi. p. 21, pi. iii. fig. 3 ; Pritchard, pp. cit., p. 828, pi. v. fig. 83.

2 Ehrenberg, Mikrogeologie, pL xviii. fig. 39 ; Gregory, Diatoms of the Clyde, p. 28, pi. iiL fig. 49.

156 THE VOYAGE OF H.M.S. CHALLENGER.

This Diatom was obtained in the middle of the Pacific in a depth of 2900 fathom3.
As indicated in the outline adjoining the figure, it presents, when viewed from its zonal
side, a hat-like appearance. From the central areola, which is marked by a group of small
puncta, radiating lines of similar puncta proceed towards the periphery, although in some
cases they do not quite reach the circumference of the valve. The specific name has
reference to the form of the valve.

Ooscinodiscus umbonatus, n. sp. (Plate II. fig. 8.)

Valvis in centro depressis, hinc elevatis, dein depresso-complanatis ; puuctulis radiato-
fasciculatis ; fasciculi totidem denticulis ad marginem signantur. In mari Pacifico.

This form, which is very closely allied to that last referred to, was found in the same
collection. Like the former, it possesses an umbonate form, but the centre of its elevated
part — as may be observed in the outline accompanying the figure — is notably low. The
valve is punctated in a radial manner, but the punctation differs from that of the preceding
in being fasciculate. The central area is ornamented with a small group of minute points in
the middle. At the circumference the fasciculi are separated by means of a short series of
small, closer, and more salient puncta.

Ooscinodiscus (?) bifrons, n. sp. (Plate II. fig. 1.)

Frustulum valvis dissimilibus ; striis egre conspicuis, et denticulorum lineis radianter
signatis, in una crebrioribus et ad marginem cessantibus, in altera rarioribus et marginem
attingentibus. In mari glaciali Antarctico.

On examining a preparation made near the ice-barrier of the Antarctic on 24th February
1874, a small delicate disc was recognised. This is marked by irregular radiating lines,
which are sometimes interrupted by small but very salient puncta or denticules, and dis-
appear at a short distance from the circumference. By the use of strong oblique illumination
and the superior homogeneous immersion objective of Zeiss, the bottom of the valve is
found to be ornamented by striae of extraordinary deUcacy, which could only be adequately
represented in the figure by the use of a greater magnifying power.

A second valve, also shown on Plate II. fig. 1, was found, which exactly corresponded
with the former in position and in size, but which presented a very distinct type of orna-
mentation, its surface being marked by sparsely disposed lines of granules or minute puncta.
These all originate much nearer the margin, and some go to the centre, while others disap-
pear sooner or later before reaching that point. It can hardly be doubted that, from the
perfect coincidence of the perimeters of the two valves, we have here to deal with a species
which, like Cocconeis, possesses dissimilar valves ; yet this supposition remains to be verified
by actual observation ; and it may, indeed, ultimately be discovered that they are the
representatives of distinct species.

"Whether these valves belong to the genus Ooscinodiscus, or to another new genus,

[REPORT ON THE BIATOMACEjE. 157

must for the present remain a matter of doubt, although the dissimilarity of the valves,
the very delicate striation, and the radiating rows of denticules seem to point to the latter
alternative. It may, however, in the meantime be treated as a species which provisionally
ranks among the Coscinodisci.

Coscinodiscus (?) janus, n. sp. (Plate II. fig. 2.)

Valvis dissimilibus, vix conspicue striatis et denticulorum lineis a margine excuntibus,
in una valva subdensioribus, in altera rarioribus centrum attingentibus vel prius aut
postea cessantibus ; ad marginem ambae decern distinctioribus granulis notantur. Ut supra.

The two dissimilar valves here figured must be regarded as forming a single species,
which is, however, to be looked upon as a doubtful member of the present genus. The
valves were found in the same preparation as that in which Coscinodiscus (?) bifrons
occurred, and they possess the same characteristic radiating lines of small, salient puncta,
which proceed from the margin towards the centre, while they also exhibit two distinct
systems of striation. The margin of each valve, however, is ornamented by eight equi-
distant points, or, perhaps, little protuberances, and, as in Coscinodiscus (?) bifrons, the
perimeters of the two exactly coincide.

Coscinodiscus (?) dimorphus, n. sp. (Plate XVII. fig. 6.)

Minimus ; valvis irregulariter, una rarius, altera crebrius punctulata (denticulata).
In Atlantico meridionali.

This very minute disc also possesses valves which are dissimilarly granulated, and might
therefore be regarded as belonging to distinct species, or at least to varieties of one species.
They, however, belong to a single frustule which was collected in the South Atlantic.

Coscinodiscus comptus, n. sp. (Plate XIII. fig. 9.)

Forma discoidalis plurimis punctulorum lineis circumradiantibus, quorum nonnuUae
prope centrum, reliquse ad medium radii vel circiter evanescunt. In mari Antarctico.

This elegant small disc is probably closely allied to the preceding. It is surrounded
at the periphery by numerous radiating striae of small points, the majority of which proceed
only a short distance towards the centre, while a few reach the margin of a smooth but
small central areola. Although the aspect of the granular lines and the delicacy of the
striation present obstacles in the way of regarding this form as a Coscinodiscus, it may
for the present be enrolled in that genus. Its specific name has reference to the elegance
of its sculpturing.

Coscinodiscus antarcticus, n. sp. (Plate XII. fig. 10.)

Forma parva, inordinate punctata, et raris spinulis circumfusa. Ad meridiem insula?
Heard.

158 THE VOYAGE OF H.M.S. CHALLENGER.

This small disciform valve lias some affinity to Coscinodiscus denticulatus (Plate III.
fig. 8), inasmuch as its surface is, like that of the latter, sparsely covered with spines
or denticules. Its granulation, however, instead of being disposed in a radiating manner,
presents no distinct order, while its margin is not striated, but plain. The two frustules
also differ greatly in point of size, and their specific values cannot be doubted.

Coscinodiscus atlanticus, n. sp. (Plate V. fig. 8.)

Cellulis aequalibus fasciculatim radiautibus a margine ad medium radii, in centro
autem inordinate dispositis. In mari Atlantico meridionali.

In this elegant disc the radiating lines of granules are disposed in fascicules around
the margin, but the arrangement in the central part is irregular and more or less lax. As
the character of the granulation is of the greatest importance in the determination of the
species of the present genus, the specific value of the valve now before us is at once
obvious. It was collected in the South Atlantic.

Coscinodiscus atlanticus (?) n. sp., var. nov. (Plate III. fig. 7.)

The valve here represented differs from the typical specimen of Coscinodiscus atlan-
ticus in the following respects: — (1.) The part occupied by the radiating fasciculately
disposed lines of granules is considerably greater than that ornamented by irregularly
arranged granules ; and (2.) the sculpturing passes into the condition of very debcate
striation at the margin, where a well-marked rins; of considerable breadth occurs. As
it is not possible to determine the precise importance of these distinctions, I regard the
present form provisionally as a variety.

Coscinodiscus stellaris, Roper, var. fasciculata, nov. (Plate V. fig. 9.)

The valve here represented was obtained near the ice-barrier of the Antarctic, and
must be regarded as a variety of Coscinodiscus stellaris, Roper. It is distinguished from
the latter by its very delicate fasciculated granulation — a circumstance which is not men-
tioned by Roper in the case of his typical species, nor indicated in the figure given by
that observer, and reproduced by Pritchard (see page 155).

Coscinodiscus (?) pacificus, n. sp. (Plate VIII. fig. 5 ; and Plate XXII. fig. 1.)

Valvis cellulosis ; cellulis subhexagonalibus, insequalibus ; margine late striate In
mari Pacifico.

The valves here shown are ornamented with large subhexagonal cellules, and each pos-
sesses a wide striated border, which serves to recall the genus Endictya of Ehrenberg.1
Whether they must be ascribed to the latter genus, however, cannot be determined till

1 Ehrenberg, Mikrogeologie, pi. xxxv. A 18, figs. 6 and 7; Pritchard, op. cit., p. 831, pi. v. fig. 70.*

REPOKT ON THE DIATOMACE/E. 159

the frustule has been observed from its zonal side; they have accordingly been provi-
sionally assigned to the genus Coscinodiscus, with which there at least exist marked
affinities.

Coscinodiscuslafricanus, Janisch, var. rotunda, nov. (Plate XXIV. fig. 3.)

The type of which the present frustule is regarded as a variety is represented on
plate fix. figs. 24 and 25, of A. Schmidt's Atlas of the Diatomacese.1 The character of the
granulation, which radiates from an excentric point, is the same in both, but the figures
given by Schmidt are elliptical, while the valve here shown is round, and bears radiating
lines which are not equidistant. These differences, however, cannot be looked upon as of
more than varietal significance.

&

Coscinodiscus decrescens, n. sp. (Plate XII. fig. 14.)

Valvis striato-cellulosis ; cellulis grandiusculis a ccntro ad marginem minuentibus ;
margine nonnullis punctulis (denticulis) signato. Ad mare Philippinarum.

This minute disc, from the Philippine Sea, is distinguished by the large hexagonal
cellules which are found on its central part, but which decrease in size towards the
periphery. The margin is ornamented with a few more salient points which render it
more distinct and prominent.

Coscinodiscus ebulliens, A. S., var. nov. (Plate V. fig. 1.)

The frustule here delineated is adorned with irregular cellules, which are irregularly
distributed, the larger usually occupying a well-marked area about half-way between the
centre and the periphery. Although presenting marked affinities with Coscinodiscus
ebidliens, A. S., which is shown on plate lxi. figs. 11 and 12 of Schmidt's Atlas, it may
be readily distinguished from the latter by the possession of a neatly striated margin,
which is not represented in the typical species.

Coscinodiscus undulatus, n. sp. (Plate VIII. fig. 3.)

E maximis ; cellulis subradiantibus hexagonis ; valvis concentrice undulatis ; centrum
oranulo vel cellula vacat. Diametrum = 390 n. In Oceano Pacifico.

This superb disc, which measures not less than 390 y". in diameter, is ornamented with
large hexagonal cellules, except at the centre, where a small smooth space occurs. Its
surface is broadly undulated, so that under the microscope the cellules seem to occur in
alternate zones — a circumstance which has suggested the specific name.

1 In a note appended to the explanation of his figure of the typical species Schmidt says : " Peripherie ellip-
tisch, Mitte der Sculptur stets zur Seite geschohen."

160 THE VOYAGE OF H.M.S. CHALLENGER.

Coscinodiscus obovatus, u. sp. (Plate VIII. fig. 4 ; Plate XVIII. fig. 7 ; and Plate
XXII. fig. 9.)

Valvis obovatis ; cellulis sequalibus ad marginem radiantibus, medio decussatis. In
mari Pacifico.

The valves represented in the present figures all possess an oval outhne. The cellules
are arranged in rows which proceed from the periphery towards the centre, but around
this point the ornamentation varies in character, although it often assumes a simple
linear and decussate arrangement. In the form shown in Plate XVIII. fig. 7, which may
be regarded as the typical form, the margin is provided with an undulating line. The
frustules were all found in the Pacific Ocean.

Coscinodiscus curvatulus, Grun., var. nov. (Plate III. fig. 10.)

In this figure there is represented a very small disc which is richly covered with
granules, disposed in seventeen radiating lines, passing from the periphery to the
centre. Each line originates at a marginal indentation, and the granulation between the
lines is arrarjged parallel to them in each intermediate area. This valve can only be
regarded as a variety of Coscinodiscus curvatulus, Grun.,1 in which the lines, instead of
being straight, are slightly curved — a distinction which is of little importance.

Coscinodiscus reniformis, n. sp. (Plate XII. fig. 12.)

Frustulum reniforme ; striis radiantibus ; cellulis grandiusculis, ad centrum minuen-
tibus.

This novel valve possesses a reniform outline, and is ornamented with radiating cellules
which decrease in size towards the centre. That it is a normal and not a teratolosical
form is shown by the circumstance that several specimens have been observed either
entire or in fragments from different and widely separated localities.

Coscinodiscus lentiginosus, Janisch. (Plate V. fig. 4.)

Mediocris, punctulorum lineis radiantibus, rariusculis, interruptis. In mari Antarctico.

At Station 146, off Marion Island, lat. 46° 46' S., long. 45° 31' E., in a depth of 1375
fathoms, the valve here figured, along with many other discoidal forms, was obtained. The
entire collection has already been reported on under No. 207 of the interesting series of
Diatoms edited by Cleve and Moller, but I am not aware that the present species has
been referred to by Janisch in any other publication. Although the naming of a species
in a preparation is not regarded as equivalent to the publication of the form, yet the
designation lentiginosus given by Janisch is so apposite that it has been retained here.

1 A. Schmidt's Atlas, pi. lvii. fig. 33 : " Die radialen Krumniungen haben in den beiden Schalen eine
entgegengesetzte Eichtung."

REPORT ON THE DIATOMACE^E. 161

The frustule is of moderate size, and bears radiating but interrupted lines of thinly
disposed granules.

Coscinodiscus polyradiatus, n. sp. (Plate III. fig. 4.)

Granulis sequalibus in lineis parallelis distributis, superficiem in pluribus triangularibus
dissepimentis dividentibus ; margine lato, striato. Ad mare Antarcticum.

This elegant disc was collected in the Antarctic Ocean. It is divided by several
radiating lines into a number of triangular areas. These lines are made up of granules of
uniform size, and parallel rows of similar granules are disposed in the intermediate areas
until the whole of the central part of the disc is covered. This is inclosed by a large
striated border — a circumstance which especially distinguishes it from Coscinodiscus fasci-
culatus, A. S.,1 while a narrow peripheral margin bearing concentric rings is also present.
That this frustule constitutes a good species cannot be doubted.

Coscinodiscus gemmatulus, n. sp. (Plate XVII. fig. 9.)

E minimis ; margine lato, punctulato ; medio raris granulis vel margaritis in irregulares
et raros ordines radiatim dispositis. In mari Indico.

This small but elegant disc is surrounded by a large striated and punctated marginal
belt. The centre is ornamented with sparsely disposed bines of rare granules which do
not reach the centre.

Coscinodiscus cycloteres, n. sp. (Plate XXII. fig. 8.)

E minimis; disciformis, valvis margine hyalino et zona granulata distinctis, a qua
aequaliuni granulorum linese procedunt, nonnullse ad centrum, alia3 plus minus breviores ;
linese centrales ab elevatiori granulo vel denticulo oriuntur. In mari glaciali Antarctico.

This elegant small disc possesses a hyaline margin within which a densely granulated
belt runs round the valve. From this belt a series of granulated lines proceed towards
the centre, and of these a few almost reach that point, while the others are shorter and of
different lengths. The lines which pass almost to the centre originate at the inner border
of the hyaline margin in somewhat salient granules or denticules.

Coscinodiscus (?) polygonus, n. sp. (Plate XXII. fig. 6.)

Valvis polygonis finissime striolatis; superficies plurimis denticulorum lineis irregu-
lariter distributis et interruptis ornata. Ut supra.

Among the collections made in the Antarctic Ocean polygonal discs fringed by a very
delicate dentation are not unfrequently met with. The surface of the valve in the present
case is ornamented by numerous radiating lines of small points or denticules — the lines

! * Compare A. Schmidt's Atlas, pL lvii. figs. 9 and 10.
(bot. chall. exp. — part iv. — 1886.) D 21

162 THE VOYAGE OF H.M.S. CHALLENGER.

being irregular and irregularly interrupted. The centre of the valve is for the most part
smooth and bears only a small group of granules. Although it is not possible to deter-
mine the cause of the polygonal form it may be conjectured that it has resulted from the
detachment of its outer border. The generic and specific determinations which have been
given are to be looked upon as provisional — (l.) because the surface of the valve, like
those of several of the species above enumerated, is very delicately striated, and (2.) because
the outline here presented is probably abnormal.

Coscinodiscus megacoccus, n. sp. (Plate XVII. fig. 2.)

E minimis; cellulis vel areolis subhexagonalibus arrandiusculis; mar seine cellulis
subquadratis constituto. In Oceano Pacifico.

This very minute organism from the Pacific Ocean is singular in being covered by a
limited number of large cellules and by possessing a distinct border of cellules or sub-
quadrate granules. In these respects it presents affinities with the similarly ornamented
Coscinodiscus subconcavus, Grun., which is figured in Schmidt's Atlas, plate lxii. fig. 7.
It may, however, be readily distinguished from the latter — (1.) by its much smaller
diameter, and (2.) by the fact that the cellules are notably larger than in any other
known species.

Coscinodiscus (?) rudis, n. sp. (Plate XXII. fig. 4.)

Forma distinctis granulis aequalibus elevatioribus grandiusculis instructa, rudem super-
ficiem efficientibus. Ad insulas Philippinas.

The generic determination of this disc, from the Philippine Islands, must be looked
upon as provisional, as it may ultimately be found to belong to the genus Pyxidicula,
Ehrenb., its valve being sufficiently convex to entitle us to express such a conviction. It
is distinguished by its large and distinctly elevated papillae or granules, which are hexa-
gonal at the base, but orbicular at the summit. On account of these papillae the external
surface, when seen from the zonal aspect, is tuberculated in appearance — a circumstance
which has suggested the specific name.

Coscinodiscus" (?) venulosus, n. sp. (Plate XVII. fig. 1.)

Forma minima, centro lineolis sinuosis radiantibus distincto. Ad meridiem insulae
Kerguelen.

This small hyaline disc was observed in a collection made to the south of Kerguelen.
It only shows a corona of tortuous and ill-defined lines around the centre, where a smooth
irregularly bordered area occurs. Although I have provisionally placed this curious valve
among the Coscinodisci, its generic determination is by no means free from doubt.

REPORT ON THE DIATOMACE^E. 163

Coscinodiscus (?) sp. (?) (Plate X. fig. 10.)

The form which is represented in this figure is up to the present time unique and must
be looked upon as an organism of a very uncertain nature. The valve is oval, and pro-
vided with a wide hyaline border. The surface is covered with uniform, densely, but
irregularly disposed granules, and the boundary between the hyaline and granulated
surfaces is irregidar in form. The specimen figured is the only one which has been
observed, and occurs in a preparation made by Dr Rae, and kindly placed by him at my
disposal. This being so, the exact systematic position which should be assigned to it must
in the meantime remain a matter of doubt, though it may be provisionally assigned to the
genus Coscinodiscus. It was collected in the neighbourhood of Zebu, one of the Philip-
pine Islands.

Coscinodiscus diophthalmus, n. sp. (Plate XVI. fig. 4.)

Valvis cellulosis, duplicem prsebentibus areolam ovalem depressam ; cellulis rotundis
grandiusculis ad centrum decrescentibus. In mari Pacifico.

Among the many interesting preparations made by Dr Pae and forwarded to me
by Mr Johu Murray, were some labelled Coscinodiscus excavatus, which could not
be reconciled with the Diatom so named by Greville. The latter has been defined
in the following manner by Pritchard : 1 — " Disc large, with hexagonal cellules decreas-
ing in size towards the centre, which has three conspicuous depressions alternating
with the same number of elevations." On the other hand, the valve now in question
possesses round cellules, while, in place of the three alternating elevations and depres-
sions, there exists two large depressed oval hyaline areas, which appear as perfora-
tions, and an equal number of slight elevations. Unless these structures can be
compared with the two processes of the genus Auliscus, Ehrenb., they find no analogy
among other known Diatoms. The depressions do not certainly show that they
are closed by a siliceous plate, since in the interior and at a lower level no other
cellulated wall could be discovered, yet the valve cannot be considered as an abnormal
or teratological one, because it has been recognised in three collections which were made
in widely separated localities in the Pacific Ocean. The true systematic position of this
Diatom must accordingly for the present remain a matter of doubt. The specimens
observed have been of different dimensions, and the large round cellules ornamenting the
valves were observed to decrease slightly in size towards the centre.

Coscinodiscus diophthalmus, n. sp., var. monojihthalma, nov. (Plate XVI. fig. 7.)

This variety is provided with a single perforation-like depression and a single elevation.
The general sculpturing of the valve resembles that of the preceding species, and, like the
latter, the present form was collected in the Pacific Ocean.

1 Pritchard, op. cit., p. 829 pi. viii. fig. 26 ; Schmidt's Atlas, pL lxv. fig. 1.

wooc

164 THE VOYAGE OF H.M.S. CHALLENGER

Coscinodiscus rhombicus, n. sp. (Plate XXII. fig. 11.)

Forma rhomboidalis, apicibus rotundatis ; granulis ad centrum rariusculis nullo certo
ordine, ad marginem crebescentibus et sensim in strias radiantes transeuntibus. In mari
Japonico. . •

This elegant valve from the Sea of Japan possesses a rhomboiclal outline. The central
space is ornamented by well-defined granules which are free from one another, and
arranged in no definite order. They decrease in size, but become more numerous towards
the margin, and finally pass into the condition of delicate radiating striae. In its general
characters this Diatom recalls the genus Cestodiscus, but the absence of rare distinct and
salient granules around the margin prevent it from being ascribed to that genus. The
specific name has reference to the form of the outline of the valve.

Coscinodiscus lanceolatus, n. sp. (Plate XVII. fig. 19.)

Forma parva, elliptico-lanceolata, granulis vel cellulis stipata a centro radiantibus et
decrescentibus. Ad meridiem Australia.

This elliptico-lanceolate form was procured in the neighbourhood of Sydney, South
Australia. It is densely covered with irregular granules or cellules, which decrease in size
as they pass from the centre towards the circumference.

Coscinodiscus ovalis, Roper. (Plate XVII. fig. 18.)

This form was found in a sounding made near Yedo, in the Sea of Japan. The frustule
exactly corresponds to those in a collection of Coscinodiscus ovalis, Roper,1 which was given
to me by the well-known French microscopist Alphonse de Bre'bisson, and it shows marked
differences from any of the preceding types.

Coscinodiscus margaritaceus, n. sp. (Plate XVIII. fig. 3.)

Mediocris ; valvis margaritarum subeequalium seriebus ab area centrali radiantibus
distinctis, quae abrupte ad marginem in punctulorum lineolas transeunt. In mari
Antarctico.

In its general characters this Antarctic valve is closely allied to the genus Cestodiscus.
Its form is circular, and it is ornamented by rows of beautiful cellules, which are disposed
in radiating lines. The central area is smooth and irregular. Towards the periphery the
granules suddenly become small, so that the border is formed by thickly disposed lines of
minute points. It differs, however, from the Cestodisci in possessing no prominent
points or processes around its circumference, and it must accordingly be looked upon as a
specific form of the genus Coscinodiscus.

1 Micr. Journ., vol. vi. p. 22, pi. iii. fig. 1, 1858 ; Pritchard, op. cit., p. 831, pi. v. fig. 78.

REPORT ON THE DIATOMACE/E. 165

Coscinodiscus radiatus, Ehrenb., var. abyssalis, nov. (Plate XXIX. figs. 2, 11, and 15.)

Those figures represent a Coscinodiscus which was found in the interior of two
Echini procured from a depth of 1340 fathoms at Station 47, lat. 41° 14' N., long.
65° 45' W. The differences observable between the frustules tend, however, to lead one
to believe that one has to deal with forms having no relation to one another. The
drawings have been made from the appearance presented by one valve mounted in a
dry air cell instead of Canada balsam, and in such circumstances when examined under
the most perfect homogeneous immersion objectives, constructed by Zeiss of Jena, or
by Eeichert of Vienna, and others, a different image is obtained at each slightest
focal change, so that it is not easy to determine which should be regarded as typical.
These different appearances, however — apart from those caused by diffraction — reveal the
minute details of the ornamentation of the organism. Thus, although I regard the very
small granules which limit the hexagonal areolae in fig. 2 to be a real structural peculiarity,
they cannot be observed in Canada balsam preparations. In the latter preparations the
surface is found to be covered with areolae, which radiate from the centre to the peri-
phery, where they gradually diminish in size, so that it presents affinities to Coscinodiscus
radiatus, Ehrenb.1 On comparing it, however, with the specimen of that species given
in the Typenplatten of Moller, the areolae are found to differ in size in the two cases,
being in the new form about one-third larger. Thus it must be regarded as a variety of
Ehrenberg's species, its varietal name having reference to the fact that it was procured in
deep water in the Atlantic. It is desirable that the two forms should be accurately com-
pared when both mounted in the dry state.

Willemoesia, n. gen.

Several of the curious and interesting forms shown on Plate VIII. figs. 8, 8 a, and
8 b, have been observed from different localities. The three frustules represented are all
elongated, and are closely related to one another. All are granulated, but in the first
the granules are disposed in an indefinite manner, while in the second and third the
punctations are uniformly distributed over the valves, being in the second decussate,
but in the third irregular. The form of the first is long and cuneate, while that of the
others is sublinear, but in all three one extremity is cuueately, while the other is simply,
rounded.

The fact that several specimens of these forms have been observed from different
localities is opposed to the belief that they are teratological or anomalous forms, and as
they all present some common characteristics, while each retains its own special and
distinctive marks, they may well be regarded as three types of a new genus.

1 Ehrenberg, Mikrogeologie, pi. xxL fig. 1 ; Smith, Synop. Brit. Diat., pi. iii. fig. 37 ; Pritchard, op. ciL,
p. 830, pi. xi. figs. 39 and 40.

166 THE VOYAGE OE II. M.S. CHALLENGES.

This genus has been named in honour of Dr Rudolph von Willemoes-Suhm, one of
the naturalists on hoard H.M.S. Challenger, who died during the Expedition.

A more precise statement of generic and specific characters may in the meantime
be omitted pending a further series of observations on this very interesting group of
organisms.

Ethmodiscus,1 n. gen.

Frustula solitaria, discoidalia ; valvis tenuissime et inconspicue striolatis ; forma plus
minus convexa, quandoque diversimode denticulata ; zona connectiva punctulata.

Under the genus Coscinodiscus, which is well adapted for the reception of species
having conspicuously granulated markings, the desirability of establishing a new genus
for certain forms possessing an almost invisible sculpturing was alluded to. For this new
genus the name of Ethmodiscus is here adopted, a word which has reference to the
presence on the valves of extremely delicate points, which may be compared to the fine
pores of a filter. Although such forms could not have been accurately observed by the
use of the relatively imperfect microscopes employed by Ehrenberg, and although it is
with the greatest difficulty and by the use of the greatest obliquity of illumination that
the best instruments of the present day can reveal the exquisite delicacy of their
ornamentation, yet the opportuneness of establishing the genus has been proved by the
discovery in surface gatherings of certain discoid forms of enormous size and extreme
fragility, which are characterised by their very minute granulations.

The form of the valves of the specimens bearing these delicate markings — which cannot
be made out when they are mounted in balsam, but become visible in dry preparations —
is found, when they are broken by the action of the flame of a spirit-lamp, to be notably
convex.

The existence of large fragments of frustules, which must also have belonged to
Diatoms of extraordinary size, and which bear characteristic punctations (Plate XIV. figs.
4 a, 4 b, 4 c), had often been revealed by the examination of organic siliceous forms derived
chiefly from soundings. These fragments, however, remained undetermined, and their
true significance has only recently been disclosed by preparations made on board the
Challenger. One of these preparations includes three large cylindrical Diatoms terminated
by two notably convex surfaces (Plate XXII. fig. 10), and provided with hoops or con-
necting zones showing punctations arranged in a quadrate manner. This would lead one
to believe that the fragments in question belonged to the connecting zones of Diatoms of
this genus ; and this belief has now been proved by the recognition in one of these
fragments of converging lines of small points, along with a similar quadrate sculpturing
on a portion adhering to its margin. The difficulty of finding any converging lines or any

1 qt)/j.o(, a sieve.

REPORT ON TITE IUATnMACE.E. 1G7

parts connected with these fragments may be explained by the long cylindrical forms of
the frustules.

The following are the essential characteristics of the genus Ethmodiscus : — (1.) The
extremely delicate granulation or striation of the frustule ; (2.) the discoidal form and
notable convexity of the valves ; (3.) the great development of the connecting zone, which
may sometimes cause the axial line to exceed the length of the transverse diameter; (4.)
the presence of small points arranged in a quadrate manner on that zone; and (5.) the
frequent occurrence of distinct elevated granules, which rise upon the valves in a coronal
or radial manner.

The following new species are included in this genus : —

Ethmodiscus punctiger, n. sp. (Plate III. fig. 1.)

Valvis convexissimis, finissime radiato-punctulatis, denticulorum corona marginali
decoratis, et ad marginem granulorum ordine signatis. Diametrum = 143 m. Prope
Yedo, in mari Japonico.

This large and beautiful disc comes from a very interesting collection of pelagic
Diatoms made in the Bay of Yedo. It possesses a very convex form, and is ornamented
with delicate radiating punctated striae. It has a submarginal corona of distinct points
or denticules, and its border is marked by a row of very small granules.

Ethmodiscus convexus, n. sp. (Plate III. fig. 9.)

Valvis admodum convexis subinconspicue punctulatis ; denticulis medio raris, ad mar-
ginem dense ordinatis. Diametrum= 123 n. In mari Arafura.

This beautiful species was found in the Arafura Sea. Its disc is markedly convex,
bears an almost invisible striation, and is ornamented with very minute denticules, while
near its periphery a circle of similar denticules is also found. This species cannot be eon-
founded with any Coscinodiscus on account of the extreme delicacy of its ornamentation
and the decided convexity of its valves, which is equal in extent to two-thirds of their
radius. The specific name has reference to the form of the valves.

EthmodiscuS;radiatus, n. sp. (Plate XII. fig. 9.)

Forma grandiuscula finissime striolata ; zona lata marginali inordinate granulata, a
qua granulorum radii plus minus ad centrum pergunt ; media superficies nonnullis granulis
sparsa ; distinctiorum granulorum corona marginali. In mari Arafura.

This species is also from the Arafura Sea. It is of discoidal form and bears extremely
delicate strise. It exhibits, towards the margin, a large belt of small granules confusedly
disposed, and from this belt there run towards the centre several straight rows of granules.

168 THE VOYAGE OF H.M.S. CHALLENGER.

These rows, none of which reach the centre, are of different lengths. The central space
bears a number of irregularly scattered granules, while the periphery is marked by minute
strijB and is adorned by a beautiful corona of distinct granules.

Ethmodiscus japonicus, n. sp. (Plate XXII. fig. 2.)

E maximis ; valvis vix perspicue radianter striatis ; margine finissime striolato, et
brevium distinctiarum lineolarum corona decorate Ad sinum Yedo, in mari Japonico.

A rich gathering of pelagic Diatoms made with surface nets in the Bay of Yedo con-
tained this magnificent hyaline disc, which, in a dry preparation, measured 145 m. The
very delicate marginal hoop, under oblique illumination, reveals a very fine striation. The
corona consists of beautiful small radiating lines, and the valve is covered by very delicate,
almost invisible, radiating bands.

The specific name has been derived from the locality in which it was first recorded.

Ethmodiscus japonicus, n. sp., var. nov. (Plate XVI. fig. 1.)

We have here represented a form which resembles Ethmodiscus japonicus. It may
be distinguished from the latter, however, by having the intervals between the short
small lines which adorn the perimeter of the disc somewhat narrower and by being of
smaller size.

Ethmodiscus coronatus, n. sp. (Plate XXII. fig. 7.)

Mediocris ; valvis tenuissime striolatis, duplici alterno granulorum ordine marginali
coronatis, et raris granulis sparsis. Ad mare Arafura.

This species is rather small but elegant. It possesses a discoidal surface which is
delicately striated, and bears a few irregularly disposed round granules. There is also
present a beautiful corona formed by two alternate rows of granules which are thinly
scattered around the margin of the valve.

*o*

Ethmodiscus coronatus, n. sp., var. a and /3. (Plate XII. figs. 7, 13.)

Two varietal forms of the preceding species are here figured. They differ from the
typical frustrule in having a marginal corona which, instead of being regular and formed
by two alternate lines, is irregular and of several rows.

Ethmodiscus humilis, n. sp. (Plate XVII. fig. 4.)

Forma minima discoidalis, granulo centrali et granulorum ordine eircum signata. In
mari Antarctico.

This species is from a depth of 325 fathoms. It occurred among mud in a sounding
taken among the ice-masses of the Antarctic. Its form is that of a small disc which
bears a single granule in the centre, and a marginal line of similar granules.

The simplicity of the ornamentation of this form has suggested its specific name.

REPORT ON THE DIATOMACE/E. 169

Ethrnodiscus obovatus, n. sp. (Plate XVII. fig. 5.)

Forma minima late obovata, nodulo grandiuseulo umbilicali, raris puuetulis circum
sparsis, et margine distinctioribus granulis signato, iustructa. Ad mare Arafura.

This type has some resemblance to the last, but its form is oval, and it exhibits a
slight deviation in the regularity of its outbne at one point. It possesses a large granule
in its centre and its margin is also marked by a number of somewhat smaller though
distinct granules. Between the marginal corona and the centre of the disc a small
number of irregularly disposed rounded points occur. .

Ethmodiscus perichantinos, n. sp. (Plate XXII. fig. 3.)

Forma suborbicularis, vix conspicue striolata, irregulari punctulorum agmine umbili-
cali, raris punctulis ad marginem et varise magnitudinis sparsa, et punctulorum eorona
spiuulosa cincta. In mari Arafura.

This species, of which several specimens were obtained in the Arafura Sea, has a
rounded and irregular outbne, and exhibits a radiating striation which is very difficult to
recognise. There is an irregular group of small granules at its centre, its margin is strewn
with thiuly scattered puncta, and its border is granulated and thorny.

The specific name of this form is derived from the condition of its border.

Ethmodiscus diadema, n. sp. (Plate XVIII. fig. 1.)

Forma mediocris, convexiuscula, radianter striolata, margine variorum processuum
corona insignita. In Antarctico, ad meridiem insulse Heard.

This type, which was found in the icy Antarctic Sea to the south of Heard Island,
possesses a disciform convex frustule. It has conspicuous radiating striae and its margin
is adorned by several (20) large granules or more salient points, which are disposed at
somewhat irregular intervals.

The specific name has reference to the presence of salient marginal puncta.

The examination of several fragments of connecting zones belonging to the extremely
large Diatoms already alluded to, revealed the existence of several species, of which the
following are noteworthy : —

Ethmodiscus gigas, n. sp. (Plate XIV. fig. 5.)

Forma rotunda, maxima ; valvis convexis insequalibus quarum una cum area umbili-
cali (?) ; lineolis radiantibus (?) ; cingulo cylindrico punctulato. Diametrum = 1633 //. In
Atlantico, ad insulas Capo Verde.

This is a very large form provided with valves which are unequal, one being more
convex than the other. It possesses an umbilical areola which may be of varying size,

(bot. chall. exp. — part iv. — 1886.; D 22

170 THE VOYAGE OF H.M.S. CHALLENGER.

thereby showing the uncertainty which must often exist in the definition of species when
these have not been established upon living forms. Its valves and connecting zone
exhibit a slight striation.

Ethmodiscus wyvilleanus, n. sp. (Plate XIV. fig. 6.)

E maximis ; subcylindricum ; valvis convexis medio complanatis, radiolatis ; zona
connectenti punctulata. Diametrum tequatoriale = 1000 m. : polare = 1457 m. In Oceano
Atlantico.

Although smaller than Ethmodiscus gigas, this species is still of colossal size. Its valves
are convex, and each has a flat slightly compressed centre. The connecting zone is greatly
developed, so that the frustule is cylindroidal, the ratio of the equatorial to the polar axis
being approximately as 2 to 3. This zone is clearly punctated in a quadrate manner, and
the valves are ornamented with radiating strise or rows of small points.

Ethmodiscus tympanum, n. sp. (Plate XIV. fig. 3.)

E maximis ; exacte cylindricus ; valvis annulatis cingulo inclusis. Ad superficiem
maris Japonici.

This curious type is not so large as the two preceding. Its form is exactly cylin-
drical. It possesses, like Biddulphia, a distinct belt superposed to the hoops of the two
valves.

The specific name has reference to its cylindrical outline.

Ethmodiscus (?) sphaeroidalis, n. sp. (Plate XXII. fig. 10.)

Frustulis spherico-compressis, binatim conjunctis; valvis convexis inaequalibus, una
scilicet convexo-complanata, altera convexa et medio concava. Diametrum = 882 At. In
mari Pacifico.

This species is interesting not only on account of its size, but also because two simi-
lar frustules were found associated together — a circumstance which renders the generic
determination doubtful. Both forms are of a compressed spheroidal outline and are
united by a common belt, which exhibits a very delicate line of suture in the middle.
This belt is somewhat analogous to that which occurs in some Melosine and Podosirie.
On each frustule two other lines, which probably indicate the incapsuling of the two
valves, occur.

It is to be noted that, although the character of the belt seems to indicate a relation-
ship to the Melosirse and Podosirse, its size and fragile nature point to an affinity with
Ethmodiscus.

Plate XIV. figs. 4, 4a, ib, 4c, represent fragments of various species.

INDEX.

The chief references are given in Roman type. Synonyms and incidental references are in Italics. The more important pages

are indicated by darker figures.

Plate

Figure

Page

Actinoptychus — continued.

Plate

Figure

IV'.'

Achnanthes, Eory.

40, 58

, 149

raeanus, n. sp. .

vii.

4

127

kerguelenensis, n. ep.

XX.

15

41

senarius

130

longipes, Ag.

92

Alloioneis, Schumann

24, 35

parallela, n. sp.

six.

11

41

antillarum, CI. et Grun.

35, 36

rhomboides, Ehrenb.

41

I XV.

5)

vcntricosa, Kg. .
ACTINISCI'S scxfurcatus, Ehrenb.

41
82

» ii
var. nov.

< XX.

' xxviii.

14

14)

35

bioctonarius, Ehrenb.

82

japonica, n. sp. .

XX.

12, bis.

36

list pt) narius, Ehrenb.

82

Amphiprora, Ehrenb. Kg. .

8, 39, 66

Actinoctclus, Ehrenb.

:.\

126, 127
142

141,

alata, Ehrenb. Kg.
fimbriata, n. sp.

xxvii.

15

40
40

(?) anceps, n. sp.

iv.

i

146

plicata, Greg.

40

antarcticus
areolatus, Bright.

145
128

„ var. japo- (
nica, nov. . . 1

XXX.

8

40

clevei, n. sp.
complanatus, n. sp.

iv.
iv.

6
9

143
145

Amphitetras, Ehrenb.

-1

101, 102, 103,
105

crassus, W. Sm.

142

antediluviana, Ehrenb.

102

(?) denticulatus, n. sp.

iv.

5

146

ornata, Shadb. .

105

ehrenbergii

142

v-Uk'Sii, Bright. .

107

fasciculatus, n. sp.

iv.

8

144

145

Amphora, Ehrenb.

4, 17, 32

» » v»r- |
nov. . . .1

iv.

8, bis.

144

coarctata, Grun.
coffeaeformis, Kg.

19
17

fulvus, W. Sm. .

142

crassa, Greg.

19

interpunctatus, Bright.

142

decora, n. sp.

xxvii.

14

18

japonicus, n. sp.

iv.

3

143

egregia, Ehrenb.

19

moniliformis, Ralfs .

142

meneghiuiana, n. sp.

xxvii.

16

17

oliveranus, O'Me.

iv.

7

145

oblonga, Greg. .

...

17

panhelios, Ehrenb.

142

obtusa, Greg.

20

pellucidus, n. sp.

xxix.

13

147

oceanica, n. sp. .

xxvii.

20

20

pruinosus, n. sp.

iv.

2

144, 146

147

philippinica, n. sp.

xxvii.

2

19

pumilus, n. sp.

xii.

2

147

polyzonata, n. sp.

xxvii.

18

18

punctulatus, n. ap.

xvi.

3

146

scalaris, n. sp. .

xxvii.

19

18

ralfsii, W. Sm. .

142

143

speciosa, n. sp. .

xxvii.

1

17

,, „ var. chal- 1
lengerensis, nov. $

XXX.

1

143

staurophora, n. sp.

xxvii.

6

20

thaitiana, n. sp.

xxvii.

15

19

subtilis, Greg. .

...

142

ANARAPHIDIEM .

6

iernarius, Ehrenb.

142

Anguliferse ....

...

106, 148

tessellatus, Roper

142

ARAPEIDIE^E .

6

umbonatus, n. sp.

iv.

4

145

Asterionella, Hassall .

60, 51, 54

Actinophsenia

130

bleakeleyii, W. Sm. .

50, 51

spkndcns, Shadb.

130

formosa, Hass. .

20, 51, 55

Actinoptychus, Ehrenb.

Z|

127, 128,
130, 131,

129,
141

frauenfeldii, Grun.
glacialis, n. sp. .

xiv.

1

50, 54
50, 51

dives ....

...

112

gracillima(Hantzsch.), [
Heib. . . . $

XXV.

6

51

erosus, n. sp. . ■

vii.

8

128

punctatus . . .

128

spiralis

50

172

THE VOYAGE OF a M.S. CHALLENGER.

Plate

Fjgure

Page

Biddulphia — continued.

Plate

Figure

Page

Asterolampra, Ehrenb.

...j

116, 132, 133,
135, 137

quinquelocularis
reticulata, Roper

100
102

decora, Grev.

136

reticulata, Roper, var. )
inermis, nov. )

xxvi.

9

102

„ „ var. nov.

xvi.

9

136

grevillii (Wall.), Grev.

136

roperiana, Grev.

xxvi.

4

106

» » " * (

v.

5

136

septcmlocularis .

100

var. eximia, nov. . )

triloculare

100

niarginata, Grev.

137

tuomeyi, Bail. .

xxvi.

10

106

vulgaris, Grev.

137

„ var. |
pacifica, nov. \

XXX.

6

106

Asteeomphalus, Ehrenb.

I|

116

132,

133,

134,

turgida

100

antarcticus, n. sp.

xvi.

11

135

weissflogii, Gruu.

xxvi.

2

104, 105

challengerensis, n. sp.

v.

2

134

sp. (?)...

xxiii.

13

104

(?) n. sp.

ix.

2

134

sp. (!) . . .

xxvi.

1

104

darwinii, Grev.
ovatus, n. sp.

v.

7

132

135
,133

(?)sp. (?) . . .

1 xix.

( XXX.

l\

105

roperianus, Grev.

133

Brightwellia, Ralfs .

124, 137

„ „ var. |
atlantica, nov. • )

v.

3

133

elaborata, Grev.

138

johusoni, Ralfs .

138

shadboltianus, Grev. .

133

murrayi, n. sp. .

X.

2

138

wyvillii, n. sp.

V.

6

134

Camptlodiscus, Ehrenb., Men.

62, 66, 91

Aulacodiscus

122

anceps, u. sp.

xvi.

2

66

Auliscus, Ehrenb.

163

bicinctus, n. sp.

xi.

12

63

Bacillaria, Gmel.

51, 68, 69

erosus, n. sp.

xi.

3

63

paradoxa, Gmel.

68

„ var. nov.

xi.

5

63

socialis, Greg.

69

humilis, n. sp. .

xi.

8

64

„ var. in- )
dica, nov. . . )

XXV.

9, 10

69

japonicus, u. sp.

xi.

1

62

lepidus, n. sp. .

xi.

7

63, 64, 65

Bacteriastrum, Shadb.

... 82, 85, 98

105

nitens, n. sp.

xi.

6

65

brevispinum, n. sp.

XV.

8

83

oceanicus, n. sp.

xi.

4

65

,, ,, var. nov. xv.

6

83

orbicularis, n. sp.

xvi.

10

64

curvatum, Shadb.

82

philippinarum, n. sp.

xi.

9

64

furcatum, Shadb.

82

wallichianus, Grev. .

66

spirillum, n. sp.

\ xix.
( xxix.

"l\

83

„ var. )

xvi.

6

65

thaitiensis, nov. !

varians, Lauder

82, 84

zebuauus, n. sp.

xi.

10

62, 64

„ var. )
princeps, nov. >

xiv.

":i

84

Oerataulus, Ehrenb. .

100, 101

xxix.

la;vis, Ehrenb. .

•

100

„ „ var. nov.

xxiii.

i

84

turgidus, Ehrenb.

100, 101

walohii, Hall's

83

var. |
polyceros, nov. )
Cera toneis arcus, Kg.

xxvi.

6, 8

101

„ „ var. his- 1
pida, nov. . . 1

xxiii.

3

83

50, 55

(

90,96,98,100,

Cestodiscus, Grev.

\ 122, 123, 125,

Biddulphia, Gray

...

101,

102,

103,

(

126, 104

(

105,

170

convexus, n. sp. ' .

vii.

6

123

aurita (Lyngb.), Breb.

103

104

coronatus, n. sp.

vii.

9

124

australis .

100

gemmifer, n, sp.

ii.

7

124, 125

japonica, u. sp.

xxiii.

14

104

ii i> Tar- (

xii.

11

125

lavis ....

100

decrescens, nov. . '

(Amphitetras) ornata, i

- I

vii.

7

125

Shadb., var. hir- >

xxiii.

9

105

nov. . . 1

suta, nov. . . )

parmula, n. sp. .

vii.

5

125

parallel.!, n. sp. (?) .

( xxiii.
( xxvi.

1\

103

(?) rapax, n. sp.
trochus, n. sp. .

xviii.
vii.

2

1

125

123

pellucida, n. sp.

xxvi.

5

103

,, ,, var. nov.

vii.

3

123

pulchella, Gray .

100

102

(75,

... J81,
(98,

76, 77, 80,

var. )

Cmtoceros, Ehrenb. .

82, 85, 89,

major, nov. )

xxiii.

C

102

99, 105

pumila, n. sp. .

xxiii.

12

103

104

bacteriastrum, AVallich

83

KITORT ON THE DIATOM ACEiE.

173

Cn.ETOCEROS — continued.
convolutum, n. sp.
criophilum, n. sp.
curvatuni, n. sp.
decipiens, Cleve

dioladia, n. sp. .

dispar, n. sp.
incurvum, Bail

„ „ var.

umbonatum, nov.
janischianum, n. sp. .
protuberans, Lauder .

„ „ var. "nov.

radiculum, u. sp.

,, var. a

,. ii .> P

sp. (?) nov.
Chrctocerotidse gregarise
solitaria?

Cocconeis

cyclophora, Gruu.
Cocconerna, Ehrenb.
Conferva ...

armillaris, 0. F. Mull

biddulphiana

obliquata .
Cokethkon, n. gen.

criophilum, n. sp.

» » var,

hispidum, n. sp.

murrayanum, n. sp.

sp. .

(?) sp. (?) . .

Cosckodiscus, Ehrenb.

Plate Figure

vm.
xix.
viii.

xxi.

xxi.
xxi.

XV.

xxi.

1 I

7,8 i
C

xxix. 10,16

Page
78
78
77
76

82

76

80

80

africauus, Janisch, var
rotunda, nov.

antarcticus, n. sp.

arafurensis, O'Me.

„ var

atlanticus, n. sp.

(?) u. sp.,
var. nov.

(?) bifrons, n. sp.

centralis, Ehrenb.
„ „ var. i

comptus, n. sp. .

concinnus, W. Sm.

craspedodiscus, O'Me

curvatulus, Grim.

curvatus, Grun., var.

cycloteres, n. sp.

decrescens, n. sp.

denticulatus, n. sp.

(?) dimorphus, n. sp.

XXIV.

xii.

in.
ii.

ii.

xiii.

ill.
xxii.j
xii.

iii.
xvii.

14

12, 15

3, 5

4
7
6

10

4

S

10

14

77

76

76

79, 80

79

79

80

80

80

34, 49, 58, 91,

99, 149, 156

34

81

92

92

100

96

85, 86, 87

85

85

86

86

86

87

66, 126, 138,

141, 152, 153,

156, 157, 159,

162, 163, 164,

165, 166, 167

159

157

153, 154

153

158

158

156, 157

153, 155

155

157

137, 153, 155

152, 153, 154

■ 160

160

161

159

155, 158

157

?!

3

6
6

4

xxix. 2, 11, li

obovatus, n. sp. .

ovalis, Roper

(?) pacificus, n. sp.

papuanus, u. sp.
patera, n. sp.
polygonus, n. sp.
polyradiatuB, n. sp.
radiatus, Ehrenb.

, , , , var

abyssalis, nov.
reuiformis, n. sp. . xii. 12

rhombicus, n. sp. . xxii. 11

rudis, n. sp. . . xxii. 4

stellaris, Roper .

„ „ var.

fasciculata, nov
») i»

subconcavus, Grun.
umbonatus, n. sp. . ii. 8

undulatus, n. sp. . viii. 3

variolatus, n. sp. . ii. 5

venulosus, n. sp. . xvii. 1

(?) sp. (?) . . . x. 10

CraspedodiBcus coronatus, )

Bright.
Cresswellia, Arnott
turgida, Grev.
t arris, Grev.
CRYPTORAPHIDIE^E
Cyclophora, n. gen.
tenuis, Cstr.

„ ,, var. tro

pica, Grun. .
>• » ., nov,

Cyclotella, Kg.

fimbriata, u. sp. . xvii. 10

Icevis, Kg. .

physoplea

rota .

rolula

COSOIHODISOOS — continued. Plate Figure

diophthalmus, n. sp. . xvi. 4

>i "

var. ruonophthalma,

nov.
ebulliens, A. S. .

„ „ var. nov,

excavatus, Grev.
fasciculatus, A. S.
gemmatuluB, n. sp.
(?) janus, n. sp. .
lanceolatus, n. sp.
lentigiuosus, Janisch
margaritaceus, n. sp.
megacoccus, n. sp.
mirificus, n. sp. .
moseleyi, O'Me. .

ii.
xvii.

v.
xviii.
xvii.

iii.

viii.
xviii.
xxii.
xvii.

I viii.
xxii.
iii.
ii.

19
4
3
o

Page
163

163

159
159
163
161
161
157
164
160
164
162
154
153

160

164
158

154, 155
155
161
161

165

165

160
164
162

155, 158

158

155
162
156
159
155
162
163

137

87, 150

87

88

6, 69

57, 58

58

58

58

116, 140, 152

141

140
140
112
112

174

Cymbella, Ag. Kg.

(?) arcus, Haas. .

criophila, n. sp.

marina, n. sp.

pelagica, n. sp. .
Dactyliosolen, n. gen.

antarcticus, n. sp.
Denticella aurita, Ehrenb.

gracilis, Ehrenb.
Penticula .

noiia, Greg.
Diatoma, De CandoUV

auritum, Lyngb.

biddulphianum ■

gracillimum, Hautzsch

hyalinum, Kg. .

rhonibicum, O'Me.
Dicladia .

capreolus, Ehrenb. .
DlMEEEGBAMMA, Ralfs .

nanum, Greg. .

M „ var. thai-

tiensis, nov.

Diploneis .

entomon, Ehrenb.

proserpincr, Ehrenb. .
Discoplea (?) rota, Ehrenb. .

rotula, Ehrenb. .
Endictya, Ehrenb.

Ethmodiscus, n. gen. .

convexus, n. sp.
coronatus, n. sp.

o and /3
diadema, n. sp. .
gigas, n. sp.
humilis, n. sp. .
japonicus, n. sp.

» var. u
obvatus, n. sp. .
periaehantiuos, u. sp.
punctiger, n. sp.
radiatus, n. sp. .
(?) sphieroidalis, n. Bp.
tympanum, n. sp.
wyvilleanus, n. sp.
Eucampia, Ehrenb.

balaustium, n. sp.

THE VOYAGE OF H.M.S. CHALLENGER.

Plate Figure

XXVU.

xxvii.
xxvii.

xxv. 22

xix. 7, 8

m.
xxii.

xii.

xviii.
xiv.

xvii.
xxii.
xvi.
xvii.
xxii.

iii.

xii.
xxii.

xiv.

xiv.

minor, nov. • '

zodiacus, Ehrenb.
Euuotia .

arcus, Ehrenb. .

fJ W. Sm. .

formica, Ehrenb.

EUODIA (Bail.), Cstr. .

(Hemidiscus) inoruata, Cstr.
orbicularis, n. sp.

9

7

7,13

Page

5, 20

50

20, 21

20, 21

21

75

75

103

103

46

46

57, 68

103

100

51

58

67

76, 80, 81

81, 82

44, 46

46

46

27

28

28

112

112

158

153, 166, 167,

170

167
168

4
2
1
5
3
1
9

10
3
6

XVUl. •>

xviii. 6

168

169

169, 170

168

168

168

169

169

167

167

170

170

170

97, 98

97

98

15

20, 50, 55

20

50

47

148, 149

149

149

Edodia — continued.

radiata, n. sp. .
recta, n. Bp.
ventricosa, n. sp.

>, » ™

Eupleuria, Aruott

Eupodiscus, Ehrenb. .

argus, Ehrenb. .
crassus
fulvus

insutus, n. sp.
jonesianus, Grev.
ralfsii

subtilis, Greg. .
tesscllatus .
Fragilabia (Lyngb.), Agardh.
antarctica, n. sp.
linearis, n. sp. .
(?) pacifica, Grun.
(?) schwarzii, Grun
GALLIONELLA, Ehrenb.
sol, Ehrenb.
sp. (?) Bail.
Gephybia, Arnott

dyerana, O'Me .
gigantea, Grev.
n. sp. (?) •
Gltphodesmis (Grev.), Cstr,
challengerensis, n. sp
eximia, Grev.
margaritacea, n: sp.
murrayana, u. sp.
(?) an Dimeregramma
(?) sp. (?)
Goniothecium
Gbammatophoea, Ehrenb
stricta, Ehrenb.

,» var- n0Vl
Gyrosigma, Hassall
transvcrsalc
Halionyx, Ehrenb.
Heliopelta, Ehrenb.
Heliopeltea; .
Hemiaultjs, Ehrenb. .

antarcticus, Ehrenb.
glacialis, n. sp. .

sp. (?) • •

Hemidiscus, Wall.

cuneiformis, Wall.

inomata
Hetekodictyon, Grev.

jeffreysianum, n. sp.

rylandsianum, Grev

spleudidum, Grev.
Hyalodisods, Ehrenb.

lsvis, Ehrenb. .

(Pyxidicula, O'Me.) ra
diatus, var. nov.

Bubtilis, Bail.

Plate
xii.
xii.
xii.
xii.

Figure
4
3
5
6

XXV.

xix.

XV.
XXV.

XVUl.

xviii.

XXV.

xxi.

12

9

10
19

13

10
12

10

12

4
9,13

Page

150

149

150

150

41, 42

122, 126,141,

142

127
142
142
126
126
142
142
142
47, 55, 56
56
56, 57
56
56
92
93
100
41, 42
42
42
42
43, 44, 45
44
43, 44
44
43

44

76, 80, 81

57

57

57

36

37

129

129

129

98, 99

99

100

99

148, 149

149

149

136

137

137

137

139

140

139

140

REPORT ON THE DIATOM AC K.E.

175

Plate
xviii.

Hyalodiscus — continued.

subtilis, Bail., var. japo- )
nica, nov. . . 1

ISTHMIA, Ag.

enervis, Ehrenb.

,, ,, var.

japonica, nov.

obliquata, var. tenuior

polymorphs, Mont.
Laidema, Clevo

annulata, Cleve . . viii.

elongata, n. sp. . ix.

(?) moseleyana, n. sp. xxiv.

pumila, n. sp. . . ix.

Mastogloia, Thwaites

apiculata, W. Sm.

kerguelenensis, n. sp . xv.

thaitiana, n. sp. . xxvi.

Mklosira, Ag.

arenaria, Moore .
clavigera, Grun.
costata, Grev.

„ ,, var. nov. xxiii.

glomus, n. sp. . . xxi.

1
major, nov.
hyalina, n. sp.
labuensis, Cleve
radio ta
sol, Ehrenb.

J "

< xvn.

' xxi.
xxi.

Figure
4

tbaitiensis, n. sp.

thirmulii, Meneg.

varians, Ag.

westii, W. Sm. . xxi.

sp. (?) . . . . xxi.

Molleria (Cleve), Cstr.

artarctica, n. sp. . xviii.

cornuta, Cleve . . xxv.

Monogramma ventricosa, Ehrenb.~ ...

Nayictla, Bory

abnormis (?) n. sp.
arcus, Ehrenb.
bisulcata, Lagerstedt
brasiliensis, Grun.
bullata, Norman
„ „ var.

carinata, nov.
,, Norman, var.

obtusa, nov.
„ Norman, var.

/ AAA.

rhomboidea, nov. )
cyclophora, n. sp. . xxviii.
decipiens, n. sp. . . xxvii.
directa

11
11

5
10

16
11

1, 3

10

18
17

Page
140

96

90

96

96

100

86, 89, 90

89, 90

89, 90

90

89

8, 22

22

22

22

S, 92, 93,139,

140, 152, 170

93

116

93

93

95

95

94

95

139

93

93

94

100

92

94

95

97, 98

98

97, 98, 99

41

58, 18, 23, 25,

26, 27, 28, 32,

33, 34, 35, 53

27

50

31

30

29, 30

29
29

30

34
32

26

Navicula — continued.

cntomon, Ehrenb.
entomon, Ehrenb., var,

thaitiana, nov.

„ „ var. ('

firma, var. subundu

lata, Grun. .
grunowii, Rabenh. (?)
jamaicenBis, Grev.
janischii, n. sp. .
(?) jejuna, A. S. .

„ „ var,

nov. .
kerguelenensis, n. sp.
kutzingii, Gruu. .
liber, Sm. .
longa .
lyra, Ehrenb.
„ „ var. sig

nata, A. S. .
mammalis, n. sp.
mirabilis, n. sp.
ostrearia, Kg.
oxeia, n. sp.
parallela, n. sp. .
phocnicentcron, Ehrent
pristiophora, Janisch
proserpinao, Ehrenb.
smithii, Breb.
spectabilis, Grev.
subrhomboidea, n. sp
thaitiana, n. sp.
zanzibarica, Grev.

> > ,i var

zebuana, nov.
sp. (?)

Plate Figure

Nitzschia, Hassall, W. Sm.

mammalis, n. sp.
obesa, n. sp.

„ ,, var. nov
plana, W. Sm. .

„ „ var. ze

buana, nov.
socialis, Greg. .
vermiculata, n. sp.
Odontella, Ag.
aurita, Ag.
polymorphs.

Omphalopelta (Ehrenb.), Cstr.

autarctica, n. sp.
areolata, Ehrenb.
cellulosa, Ehrenb.
japonica, n. sp.

xx.

XX.

XXX.

XX.

xxviii.
xxviii.

17
10

XX.
XXX.

XX.

xxviii.

xxvin.

XX.
XX.

xxviii.

xxviii.
xxviii.
xxviii.

xx.
xxvii.
xxvii.

xxix.
xiii.
xiii.

12

11
16

13

10

12

13
15

17

6

3, 7

10

5
11
13

10
12

in

Page
96

28
28

31

28, 29
29
29
33

33

33

28

35

26

33, 34

33

30
34

32
31
31
81
• 28
28
34
32
30
27, 2S
31

31

32
32
32
34
35
35
40, 66,67,68,
69, 90

68
67
67
67

67

69

68

100, 101

103

100

127, 128, 129,

130, 131

129
128
128
129

176

THE VOYAGE OF H.M.S. CHALLENGER.

Omphalopelta — continued.
parda, n. sp.
punctata, Ehrenb.
shrubsoliana, n. sp.
splendens .

(?)BP. (?) .

Orthosira

dickieii
Palmeria, Grev. .
Pinntjlabia, Ehrenb.. .
criopLila, n. sp.
divergens, W. Sm.
entomon, Ehrenb.
major, Rabenh.
peregrina, Ehrenb.
raeana, n. sp.
sp.(?). . .
Plagiogramma, Grev. .
jamaicense, Grev.
margaritaceum, n. sj
tessellatum, Greg,
thaitiense, n. sp.
PLErjROSIGMA, W. Sm.
arafurense, n. sp.
balticum, W. Sm.
decorum, W. Sm.
delicatulum, W. Sm.
elegantissimum, n. sp
formosum, W. Sm.
japonicum, d. sp.
naviculaceum, Breb.
»» j»

var. nov.
smithianum, n. sp.
Bpeciosum, W. Sm.
„ „ var.

nov. .
thaitiense, n. sp.
Iramsversale, W. Sm.
Podosira, Ehrenb.

maculata, W. Sm.
(Melosira) radiata
Podosphemia
Polyceratium
Pokowscus, Grew

oblongus, Grev. .
stolterfothii, n. sp.
Porpeia, Bail.
PSEUDORAPHIDIE^E
Pseudostictodiscua angulatu
Pyxidicula, Ehrenb.

radiata, O'Me. .
RAPHIDIE.E .
Rhabdouema
RhaPHONEIS, Ehrenb.

australis, H. L. Sm.
elliptica, n. sp. .
fasciolata, Ehrenb.
japonica, n. pp. .
mammalia, n. sp.

Plate Figure
xviii. 9

Page
131
128
131
130
130
93

148

... 5,23,25,27,33

xv. 2 26

24

28

5

25

xv. 3 25

xv. 1 26

45

45

xix. 13 45

45

xix. 4 45

8, 36, 37, 39

xxviii. 5 38

37

37

37

xxviii. 1 37

37

xxix. 14 38
37

XXVUl.

xxvm.
xxviii.

XIX.

xxvi.

13

12
3

37

38

38

37

38

37

87, 139, 170

145

139

8

106

138

139

139

101, 102

6, 41

112

139, 162

139

6, 17

3, 70, 127

48, 49

49

49

49

49

48, 49

Rhizosolenia, Ehrenb.

arafurensis, n. sp.
eriensis, H. L. Sm.
(?) flaccida, n. sp.
gracilis, H. L. Sm.
imbricata, Bright,
insequalis, n. sp.

inermis, n. sp. .

japonica, n. sp.

murrayana, n. sp.

polydactyla, n. sp.

robusta, Norman
, , , , v;

nov. .

sima, n. sp.

,, ,, var. nov.

styliformis, Bright.
Rutilaeia, Grev.

edentula, n. sp. .

epsilon, Grev.

recens, Cleve

tulkii, n. sp
Spatangidium, Br<5b.

arachne

flabellatum

heptactis .

pellatum .

ralfsianum, Norm.
Stauboneis, Ehrenb., Kg.

brebissonii, n. sp.

glacialis, n. sp.

gracilis, Ehrenb.

oblonga, Bail.

pacifica, n. sp. .

phccnicenteron, Ehrenb.

pygmea, n. sp. .

salina, W. Sm. .
,, ., var. i,

RaJjenh.
,, ,, var. c, nov.

thaitiana, n. sp. .
Stauroptera, Ehrenb.

aspera, Ehrenb. .
Stephanopyxis, Ehrenb.

apiculata, Ehrenb.

campana, n. sp. .

kittoniana, n. sp.

rapax, n. sp.

turris

Stictodiscds, Grev.

affinis, n. sp.

„ „ var. late

zonata, nov.
,, .. var. nov,

Plate Figure Page

i 69, 70, 72, 73,
74, 75, 85, 80
xxx. 12 74, 75

70
xxix. 4 74

70
xxiv. 1, 1, bis. 73

xxiv. 15 71

xxiv

xxiii.
xxiv.
xxiv.

xxiv.

I 10, 13 )

7
12

XXIV.

xxix.

XVUl.

XV111.

XIX.

ix.
iz.

l.

xvii.

71

72
72
71

70, 73

73

11 71

9 72

72

90, 91

14 92

91

91

11 91, 92

... 132, 133, 135

132

132

132

132

132

23,24,27

xv. 4 24

xxvii. 11 24, 25

7, SI

xx. 7, 11 24

xx. 9 23

7, 81

xxix. 7 25

23

13
16

14
5
9

11

6

8

24

23
23

23

24

87, 150

88

88

87

88

88

106, 112, 113,

114, 115, 116,

117, 118, 120

119

119

119
119

REPORT ON TIIK MAToMAlT.l.

177

3ti< i sci a continued

Plate

Pi 'iir

Page

anceps, n. sp.

i.

5

116, 121

.•mgulatiis, Grim.

112

argus, A. S.

115

bicoronatus, n. sp.

vi.

5

115, 120

» var. )
punctigera, nov. . 1

xiii.

2

121

buryanus, Grev.

113

californicus, Grev.

115, 117

■i „ var. )
areolata, Grim. . 1

11.-.

,, „ (
ecostata, Grim. . >

115

.. ., j
nankoorensis.Grim. 1

115

crozieri, Kitton .

112

elegans, d. sp. .

xiii.

3

121

eulensteinii, Cstr.

i.

'I1*

117, 118.

1.
xvii.

xiii.
xvii.

xxv.

XXV.

xxv.
xxv.

xxv.

17
14

12

1
10

gallapagensis

hardmanianus, Grev,

harrisonianus

hexagonus, n. sp. . xvii.

,, ,, var. nov. xvii.

inflatus

japonicus, n. sp.

jeremianuB

johnsonianus, Grev.

kittonianus, Grev.

margaritaceus, n. sp.

margaritiferus .

morsianus, A. S.

multiplex

novarse, Cleve .

parallelus .

polygonius

quadriornis

radiatus, n. sp. .

radfordianus, n. sp.

retieulatus, n. sp.

simplex, A. S.

trigonus, n. sp. .

varians, n. sp.
Striatella

unipunctata
Syndendrium
Synedba, Ehrenb.

atlarjtica, n. sp. .

capitulata, n. sp.

fimbriata, n. sp.

lanceolata, n. sp.

„ „ var.

thaitiensis, nov.

pbilippinarum, n. sp.

thalassiotbrix, Cleve
Systephania, Ebrenb. .

aculeata, Ehrenb.
„ „ var. a

„ „ var.

/S nov.
(BOT. CHALL. EXP. — PART IV. — 1886.)

121

115

115

116

122

122

114

119

116

117

115

120

115

116

116

112

114

114

115

117

118

120

115

1 122

7 120

3, 70, 127

4

76, 80, 81

... 51,52,67,68

16 53

13 52

14 52
20 52, 53

IS

15

IX.
XXX.

53

52

54

150, 151

151

151

151

Systephania mntinwil.
Corolla, I'.liivnb,

diadems, Ehrenb.
racana, n. sp.

(?) sp. (?) . . .
SuniRKU.A, Turp., Ehrenb.,
W. Sm. .

arctissima, A. S.

argus, n. sp.

dives, n. sp.

fastuosa, Ehrenb., var.
abludens, Grun. .

grandiusottla, n. sp. .

intercedens, Grim.

japonica, n. sp. .

linearis, W. Sm.

multicostata, n. sp.

ocellata, n. sp. .

thaitiana, n. sp. .
Tabellaria . . . .

Terebbabia, Grev.

barbadensis, Grev.

kerguelensis, O'Me. .

(?)sp. (?) . . .
Thalassiosika, Cleve .

nordenskibldii, Cleve .

var. nov.
Thalassiothrix (Grun.), Crtr.

curvata, n. sp. .

fraueufeldii (Grim.), Cstr.
Toxonidea, Donkin .

challengeriensis, n. sp.

„ \

var. nov. . . )

Trachisphenia australis, Cleve

Triceraticm, Ehrenb. .

abyssale, n. sp. .
arctieum, Bright.
„ „ var.

kerguelenensis, nov.

kerguelenensis /3

1

Plate

Figure Page

150

150

ix.

11

151

XXX.

2

151

59, 60, 62, 66
59

X.

9

60

X.

4

59, 60

59

X.

5

61

59

X.

8

60

59

X.

6

61

X.

7

60

xix.

3

61

4, 57, 58
46, 47
46, 47
46, 47

xxv.

2

47

95

95, 96

XXX.

4

96

... 50,51,52,53,54

xxiv.

6

55

xiv.

7,8

50, 54, 55
39

xxvi.

15

39

xxvi.

14

39

49
102, 106, 111,
113, 115, 116,

121

xiii.

6

111

105, 107

xiii.

7

107

107

kerguelenensis y,

xiii.

5

107

nov.

)

armatum, Roper

110

„ var. o,

nov.

vi.

2

109

atlanticum

xvii.

3

111

calvescens, n. sp.

ix.

1

107

cariosum, n. sp. .

vi.

6

108

coroDatum, n. sp.

vi.

7

108

eulensteinii, Grun.

106,

117

118

favus, Ehrenb. .

:,'i

103, 109,
111

110,

var
late-areolata

1

ix.

3

109

D 23

178

THE VOYAGE OF H.M.S. CHALLENGER.

RATIl'M — contain i

Tlate

Figure

Page

favus, Ehrenb., var. i
pacifica, nov. . 1

vi.

1

109

ferox, n. sp.

vi.

4

107

fimbriatum, Wall.

108,

109

110

,, ,, var. )

nov. . . . )

ix.

12

110

gallapagense, Cleve

115

grande, Bright. .

110

grunowianum, n. sp. .

xvi.

5

110

111

harrisonianum, Grev.

113,

114

11«

incrassatum, n. sp.

is.

10

110

injlatutn

114

insutum, n. sp. .

XXV.

7

111

jcremianum, A. S.

116

liiuolatiim, Grev.

114

margaritaceum, Ralfs

114

margaritiferum, Cleve

115

megastomium, Bright.

108

multiplex, Janisch (?).

116

Triceratitjm — continued.

orientate, Harv. et Bail

parallelum, Grev.

pavimentosura, n. sp

polygonium, Grev.

pulvillus, n. sp.

punctigerum, n. sp.

quadratum, Grev.

quadricornc, Grev.

sarcophagus, n. sp.

thaitieuse, n. sp.

tumescens, n. sp.

vnlheaii, var. |3 .
Willemoesu, n. gen. .
Zygoceros, Ehrenb.

mobilensis, Bail.

rhombus, Ehrenb.

surirella, Ehrenb.

tuomeyi, Bail.

Date

Figure

Tage
llii
114

xiii.

8

108

114

vi.

8

106

xiii.

4

108

114
114

vi.

3

109

xiii.

14

106

VI.

9

109

107
165

... 100,

101, 102
101
101
101
106

PRINTED BY DALI.ANTTNE, HANSON AND CO., EDINBURGH AND LONDON.

PLATE I.

(bot. chall. exp.— part iv.— 1886.)

Da

PLATE I.

Figure 1. Stictodiscus radiatus, n. sp. .

2. Stictodiscus japonicus, n. sp. .

3. Stictodiscus radiatus, n. sp., var. no v.

4. Stictodiscus affinis, n. sp.

5. Stictodiscus anceps, u. sp.

6. Stictodiscus affinis, n. sp., var. nov.

7. Stictodiscus adensteinii (Grun.), Cstr.

8. Stictodiscus reticularis, n. sp.

PAGE

117

119

117

119

116

119
117

120

The Voyage of H.M.S:Challen§er

Duttomaci <r I'll

■

PLATE II.

PLATE II.

Figure 1. Coscinodiscus (?) bifrons, u. sp.

2. Coscinodiscus (V)janus, n. sp.

3. Coscinodiscus centralis, Ehrenb., var. now

4. Coscinodiscus arafurensis, O'Me., var. nor,

5. Coscinodiscus variolatus, n. sp.

6. Coscinodiscus patera, n. sp. .

7. Cestodiscus gemmifer, n. sp. .

8. Coscinodiscus umbonatus, u. sp.

PAGE

156
157
155
153
155
155
124
156

l)l,ll(JIII,l, , ■ I 'I . II

PLATE III.

(bot. CHALL. EXP. — PART IV. — 1886.) D b

PLATE III.

Figure 1. Ethmodiscus, punctiger, n. sp.

2. Coscinodiscus stellaris, Roper, var. nov.

3. Coscinodiscus papuanus, n. sp.

4. Coscinodiscus poly radiatus, n. sp.

5. Coscinodiscus crasj^edodiscus, O'Me. .

6. Coscinodiscus mirificus, n. sp. .

7. Coscinodiscus atlanticus (?), n. sp., var. nov.

8. Coscinodiscus denticulatus, n. sp.

9. Ethmodiscus convexus, n. sp.
10. Coscinodiscus curvatulus, Grun., var. nov.

PAGE

167
155
154
161
152
154
158
155
167
160

I 'I . Ill,

-

'

PLATE IV.

PLATE IV.

Figure 1. Actinocyclus (?) anceps, n. sp.

„ 2. Actinocyclus pruinosus, n. sp. .

,, 3. Actinocyclus japonicus, n. sp. . *

„ 4. Actinocyclus umbonatus, n. sp.

,, 5. Actinocyclus (?) denticulatus, n. sp.

„ 6. Actinocyclus clcvei, n. sp.

„ 7. Actinocyclus oliveranus, O'Me.

,, 8. Actinocyclus fasciculatus, n. sp.

„ 8 6ts. Actinocyclus fasciculatus, n. sp., var. nov.

,, 9. Actinocyclus complanatus, n. sp.

PAGE

146
144
143
145
146
143
145
144
144
145

Diai0mac<x& I* I I V

w

8 bis

PLATE V.

(bot. chall. exp.— part rv.— 1886 )

D c

PLATE V.

Figure 1. Coscinodiscus ebulliens, A. S., var. nov.

2. Asteromphalus challengerensis, n. sp.

3. Asteromphalus roperianus, Grev., var. atlantica, nov.

4. Coscinodiscus lentiginosus, Janisch

5. Asterolampra grevillei (Wall.), Grev., var. eximia, nov.

6. Asteromphalus wyvillii, n. sp. ...

7. Asteromphalus ovatus, n. sp. ...

8. Coscinodiscus atlanticus, n. sp.

9. Coscinodiscus stellaris, Roper, var. fasciculata, nov. .

PAGE

159
134

133
160
136
134
132
158
158

ager'

Ihn/, in i, /,,■,/■ P] V

PLATE VI.

PLATE VI.

Figure 1. Triceratiumfavus, Ekrenb., var. pacijica, nov

2. Triceratium armatum, Roper, var. S, nov

3. Triceratium sarcofagus, n. sp.

4. Triceratium ferox, n. sp.

5. Stictodiscus bicoronatus, n. sp.
G. Triceratium cariosum, n. sp. .

7. Triceratium coronatum, n. sp.

8. Triceratium pulvillus, n. sp. .

9. Triceratium tumescens, n. sp.

PAGE

109
109
109
107
120
108
108
106
109

The Vqya I i mger"

Ihii/rillltlt rn- I'l \ I

■

■<■'■

WBSP

■

Caslxacane direst

MrKfJ-

PLATE VII.

(bot. CHALL. EXP. — PART IV. — 1886.) D d

PLATE VII.

Figure 1. Cestodiscus troclms, n. sp.

2. Omphalopelta antarctica, u. sp.

3. Cestodiscus trochus, u. sp., var. uov.

4. Actinoptychus raeanus, n. sp.

5. Cestodiscus parmida, n. sp.

6. Cestodiscus convexus, n. sp. .

7. Cestodiscus gemmifer, n. sp., var. nov

8. Actinoptychus erosus, n. sp. .

9. Cestodiscus coronatus, n. sp. .
10. Omphalopelta japonica, n. sp.

PAGE

123
129
123
127
125
123
125
128
124
129

"r ■ i

PLATE VIII.

PLATE VIII.

Figure 1. Chsetoceros dicladia, n. sp.

2. Cliwtoceros protuberans, Lauder, var. nov.

3. Coscinodiscus undulatus, n. sp.

4. Coscinodiscus obovatus, n. sp. .

5. Coscinodiscus (?) pacijiciis, n. sp.

6. Chsetoceros dispar, n. sp.

7. Lauderia annulata, Cleve

8. Willemoesia, sp.
8 a. T'Ti'Wemoesia, sp.
8b. Willemoesia, sp.

PAGE

82

76

159

160

158

76

89

165

165

165

'ilenger

dialomacecB IM VIII

i

8*

-680

-

■

PLATE IX.

(bot. chall. bxp. — pabt iv. — 1886.) D e

PLATE IX.

Figure 1. Triceratium calvescens, n. sp. .

2. Asteromphahis challengerensis (?) n. sp., forma monstrosa, nov

3. Triceratium favus, Ehrenb., var. late-areolata, nov.

4. Lauderia elongata, n. sp.

5. Stephanopyxis hittoniana, n. sp.

6. Systephania aculeata, Ehrenb., var. a, nov.

7. Dactyliosolen antarcticus, n. sp.

8. Lauderia pumila, n. sp.

9. Stephanopyxis rapax, n. sp. .

10. Triceratium incrassatum, n. sp.

1 1. Systephania raeana, n. sp.

12. Triceratium Jimbriatum, Wall., var. nov.

PAGE

107

134

109

89

87

151

75

89

88

110

151

110

I) minium nr I'l IX

i K*

t <

i

v :i%

J&r,

■V

V

era del

. ■

.

PLATE X.

PLATE X.

Figure 1. Hyalodiscus (Pyxidicula, O'Me.) radiatus, var. nov

,, 2. Brightwellia murrayi, n. sp. .

„ 3. Melosira sol, Ehrenb., var. nov.

,, 4. Surirella dives, n. sp.

,, 5. Surirella grandiuscula, n. sp.

,, 6. Surirella multicostata, n. sp. .

„ 7. Surirella ocellata, n. sp.

,, 8. Surirella japonica, n. sp.

,, 9. Surirella argus, n. sp.
„ 10. Coscinodiscus (1), sp. (?)

FADE

ar. nov. .

139

138

93

59

61

61

.

60

.

60

.

60

.

163

I

Vinlemtwjxe I' I X

m

A

■^'

.^7-

:>'

,

■

•

PLATE XI.

(bot. CHALL. EXP. — PART IV. — 1886.) -D f

PLATE XI.

Figure 1. Campylodiscus japonicus, n. sp.

,, 2. Campylodiscus bicinctus, u. sp.

„ 3. Campylodiscus erosus, n. sp. .

„ 4. Campylodiscus oceanicus, n. sp.

„ 5. Campylodiscus erosus, n. sp., var. nov

„ 6. Campylodiscus nitens, n. sp. .

,, 7. Campylodiscus lepidus, n. sp.

„ 8. Campylodiscus humilis, n. sp.

,, 9. Campylodiscus philippinarum, u. sp.

,, 10. Campylodiscus zebuanus, n. sp.

PAGE

62
63
63
65
63
65
63
64
64
62

. . .

HhiIiiiiihi til' I'l XI

'" " ' "-&-,

■

PLATE XII.

PLATE XII.

Figure 1. Euodia (Hemidiscus) inornata, Cstr.

,, 2. Actinocyclus pumilus, n. sp. .

,, 3. Euodia recta, n. sp. .

„ 4. Euodia radiata, n. sp. .

',, 5. Euodia ventricosa, n. sp.

„ 6. Euodia ventricosa, n. sp., var. nov.

,. 7. Ethmodiscus coronatus, n. sp., var. a. nov.

,, 8. Porodiscus stolterforthii, n. sp.

„ 9. Ethmodiscus radiatus, n. sp. .

,, 10. Coscinodiscus antarcticus, n. sp.

,, 11. Cestodiscus gemmifer, n. sp., var. decrescens, nov.

,, 12. Coscinodiscus reniformis, n. sp.

,, 13. Ethmodiscus coronatus, n. sp., var. B. nov.

,, 14. Coscinodiscus decrescens, n. sp.

,, 15. Euodia orbicularis, n. sp.

PAGE

149

147

149

150

150

150

168

139

167

157
125

160

168

159

149

I) I, i /„/,,,,, ; .'■ I 'I. Ml

'

•

^

9

.

■-■;.'

PLATE XIII.

(BOT. CHALL. EXP. — PAST IV. — 1886.) D g

PLATE XIII.

Figure 1. Stictodiscus trigonus, n. sp. .

2. Stictodiscus bicoronatus, n. sp., var. punctigera, nov.

3. Stictodiscus elegans, n. sp.

4. Triceratium punctigerum, n. sp. .

5. Triceratium arcticum, Bright., var. kerguelenensis j, nov.

6. Triceratium abyssale, n. sp. .

7. Triceratium arcticum, Bright., var. kerguelenensis, nov.

8. Triceratium pavimentosum, n. sp. .
,, . 9. Coscinodiscus comptus, n. sp. .
„ 10. Nitzschia plana, W. Sm., var. zebuana, nov. .
,, 11. Nitzschia obesa, n. sp.
„ 12. Nitzschia vermiculata, n. sp. .
„ 13. Nitzschia obesa, n. sp., var. nov.
,, 14. Triceratium thaitiense, n. sp. .

FAOB

122
121
121

108

107

111

107

108

157

67

67

68

67

106

I ' I . MIL

PLATE XIV.

PLATE XIV.

Figure 1. Asterionella glacialis, n. sp. .

„ 2. Bacteriastrum varians, Lauder, var. princeps, nov.

„ 3. Ethmodiscus tympanum, n. sp.
Figures 4, 4 a, 4 &, and 4 c. Ethmodiscus, sp. (Fragmeuta)
Figure 5. Ethmodiscus gigas, n. sp. .

,, 6. Ethmodiscus ivyvilleanus, n. sp.
Figures 7 and 8. Thalassiothrix frauenfeldii (Grun.), Csfcr.

PAGE

50
84
170
170
169
170
54

- !'l . XIV.

PLATE XV.

(bot. CHALL. EXP. — PART IV. — 1S86.)

Dh

PLATE XV.

Figure 1. Pinnularia sp. (?) .

„ 2. Pinnularia criophila, n. sp. .

„ 3. Pinnularia raeana, n. sp.

,, 4. Stauroneis brebissonii, n. sp. .

,, 5. Alloioneis antillarum, CI. et Gruu., var. nov.

„ 6. Bacteriastrum brevispinum, n. sp., var. nov.

„ 7- Corethron, sp.

8. Bacteriastrum brevispinum, n. sp.

,, 9. Navicula grunowii, Rabenh. (?) .

„ 10. Gephyria gigantea, Grev.

,, 11. Mastogloia kergtielenensis, n. sp.

PAGE

26

26

25

24

35

83

86

83

28

42

22

,t I'l . \\

I

PLATE XVI.

PLATE XVI.

Figure 1. Ethmodiscus japonicus, n. sp., var. do v.

2. Campylodiscus anceps, n. sp. .

3. Actinocyclus punctulatus, n. sp.

4. Coscinodiscus diophthalmus, n. sp.

5. Triceratium grunoivianum, n. sp.

6. Campylodiscus wallichianus, Grev., var. thaitiensis, nov.

7. Coscinodiscus diophthalmus, n. sp., var. monophthalma, nov.

8. Omphalopclta (?), sp. (?)

9. Asterolampra decora, Grev., var. nov. . .

10. Campylodiscus orbicularis, n. sp.

11. Asteromphalus antarcticus, n. sp.

PAGE

•

168

•

66

•

146

•

163

•

110

65

163

130

136

64

135

Dia/oiru ■ ea" I 'I. XT]

■

PLATE XVII.

(bot. chall. bxp. — pabt iv. — 1S86.) D i

PLATE XVII.

Figure 1. Coscinodiscus (?) venulosus, n. sp.
,, 2. Coscinodiscus megacoccus, n. sp.
,, 3. Triceratuim atlanticum, n. sp.
,, 4. Ethmodiscus humilis, n. sp.
,, 5. Ethmodiscus obovatus, n. sp. .
,, 6. Coscinodiscus (?) dimorphus, n. sp.
,, 7. Stictodiscus varians, n. sp.
„ 8. Stictodiscus affinis, n. sp., var. nov.
,, 9. Coscinodiscus gemmatulus, u. sp.

„ 10. Stictodiscus radfovdianus, n. sp.

,, 11. Stictodiscus affinis, n. sp., var. late-zonata, nov

,, 12. Stictodiscus margaritaceus, n. sp.

„ 13. Melosira sol, Ehrenb., var. nov.

,, 14. Stictodiscus hexagonus, n. sp., var. nov

„ 15. Amphiprora fimbriata, n. sp. .

,, 16. Cyclotella fimbriata, n. sp.

., 17. Stictodiscus hexagonus, n. sp.

„ 18. Coscinodiscus ovalis, Roper

,, 19. Coscinodiscus lanceolatus, n. sp.

PAGK

162
162
111
168
169
157
120
119
161
118
119
120

93
122

40
141
122
164
164

I 'I. Ml

: c o ° c 0 ° ° ^

0

■■'■-. • ' o

fr° °

PLATE XVIII.

PLATE XVIII.

Figure 1. Ethmodiscus diadema, n. sp. .

2. Cestodiscus (?) rapax, u. sp. .

3. Coscinodiscus margaritaceus, n. sp.

4. Hyalodiscus subtilis, Bailey, var. japonica, nov.

5. Eucampia balaustmm, n. sp.

6. Eucampia balaustmm, n. sp., var. minor, uov.

7. Coscinodiscus obovatus, n. sp., forma typica

8. Molleria antarctica, n. sp. .

9. Omplialopelta parda, n. sp. .

10. Glyphodesmis maryaritacea, n. sp.

11. Rutilaria tulkii, n. sp.

12. Glyphodesmis murrayana, n. sp.

13. Glyphodesmis challcngerensis, n. sp. .

14. Rutilaria edentula, n. sp. .

PAGE

169

125

164

140

97

98

160

98

131

44

91

43

44

92

'

Diaiomacea N. XVIII

Caslrac.' ■

PLATE XIX.

Dk

(BOT. CHALL. EXP. — PART IV. — 1886.)

PLATE XIX.

Figure 1. Biddulphia (?) (an Triceratium (?) ) sp. (?)

2. Bacteriastrum spirillum, n. sp.

3. Surirella thaitiana, n. sp. .

4. Plagiogramma thaitiense, n. sp.

5. Dimeregramma nanum, Greg., var. thaitiensis, nov.
[6. Eupodiscus insutus, n. sp. .

Figures 7 and 8. CJitetoceros dicladia, n. sp. .
Figure 9. Fragilaria linearis, n. sp.

„ 10. Glyphodesmis (?) (an Dimeregramma (?) ) sp. (?)

„ 11. Achnanthes parallela, n. sp. .

,, 12. Rhaphoneis japonica, n. sp.

,, 13. Plagiogramma margaritaceum, n. sp.

„ 14. Stephanopyxis campana, n. sp.

PAGE

105
83

61
45
46
126
82
56
44
41
49
45
88

'• ' ! . .\ 1 \ .

. ..

■ - * - -

I!

PLATE XX.

PLATE XX.

Figures 1 and 3. Navicula brasiliensis, Grun. .
Figure 2. Navicula mammalis, n. sp.

„ 4. Navicula subrhomboidea, n. sp.

,, 5. Navicula thaitiana, n. sp.

,, 6. Navicula, sp. (?) .

Figures 7 and 11. Stauroneis oblonga, Bailey
Figure 8. Navicula oxeia, n. sp.

,, 9. Stauroneis pacifica, n. sp.

,, 10. Navicula entomon, Ehrenb., var. (?)

,, 12. Navicula (?) jejuna, A. S.

,, 12 fo's. Alloioneis japonica, n. sp. .

,, 13. Stauroneis salina, W. Sm., var. c, nov.

„ 14. Alloioneis antillarum, CI. et Grun., var. nov.
„ 15. Achnanthes kerguelenensis, n. sp.
„ 16. Stauroneis thaitiana, n. sp.

,, 17. Navicula entomon, Ehrenb., var. thaitiana, nov.

„ 18. Gephyria (?) sp. ....

PAGE

30
30
30
27
34
24
31
23
28
33
36
23

35

41
23

28

42

This is the same form as figured on Plate XXV. fig. 19.

Duztemacea PI. !0i

.

i

•

' ■.

I

A '

■

l". Iv

PLATE XXI.

(BOT. CHALL. EXP. — FART IV. — 1886.) ^ 1

PLATE XXL

Figure 1. Melosira hyalina, n. sp.

„ 2. Melosira thaitiensis, n. sp.

Figures 3 and 5. Corethron hispidum, n. sp.

Figure 4. Corethron murrayanum, n. sp.

„ 6. Corethron (?) sp. (?), forma juvenis

,, 7. Melosira sol, Ehrenb., var. nov.

„ 8. Heterodictyon jeffreysianum, u. sp.

,, 9. Hemiaulus, sp. (?), valva terminalis (?)

10. Melosira glomus, n. sp.

,, 11. Melosira, sp. (?)

,, 12. Corethron criophilum, n. sp., var nov.

,, 13. Hemiaulus, sp. (?), valva terminalis (?)

,, 14. Corethron criophilum, n. sp. .

,. 15. Corethron criophilum, n. sp., var. nov.

„ 16. Melosira ivestii, W. Sm.

PAGE

94
94
86
86
87
93
137
99
95
95
85
99
85
85
94

lh,i/nni<rri'<f

XXI

I

PLATE XXII.

PLATE XXII.

Figure 1. Coscinodiscus (?) pacificus, n. sp.

„ 2. Ethmodisctts japonicus, n. sp. ...

,, 3. Ethmodiscus perichantinos, n. sp.

„ 4. Coscinodiscus (?) rudis, n. sp. .

„ 5. Triceratium arcticum, Bright., var. kerguelenensis (S, nov.

6. Coscinodiscus (?) polygonus, n. sp.

7. Ethmodiscus coronatus, n. sp.

8. Coscinodiscus cyclotcres, n. sp.
„ 9. Coscinodiscus obovatus, n. sp. .
„ 10. Ethmodiscus (?) sphceroidalis, n. sp. .
,, 11. Coscinodiscus rhombicus, n. sp.

PAGE

158
168
169
162
107
161
168
161
160
170
164

■

lh,,/<,,„.,,<;r l*J. \.\||

-

PLATE XXIII.

(bot. CHALL. EXP. — PART IV. — 1886.)

PLATE XXIII.

Figure 1. Bacteriastrum varians, Lauder, var. nov.
,, 2. Omphalopelta shrubsoliana, n. sp.
,, 3. Bacteriastrum wattichii, Ralfs, var. hispida, nov.
„ 4. Melosira glomus, n. sp., var. major, nov.
„ 5. Melosira costata, Grev., var. nov.
„ 6. Biddulphia pulchella, Gray, var. major, nov.
„ 7. Rhizosolenia japonica, n. sp. .
„ 8. Rhizosolenia (?)....
,, 9. Biddulphia (Amphitetras) omata, Shadb., var. hirsuta, nov,

,, 10. Biddulphia parallela, n. sp. (?)

,, 11. Biddtdphia parallela, n. sp., var. nov.

„ 12. Biddulphia pumila, n. sp. .

„ 13. Biddtdphia, sp. (?)

„ 14. Biddulphia japonica, n. sp. .

PAGE

84

131

83

95

93

102

72

72

la, nov. .

105

103

103

•

103

104

104

Vuu - II. Will.

k

PLATE XXIV.

PLATE XXIV.

Figures 1 and 1 bis. Rhizosolenia imbricata, Bright.

Figure 2. Rhizosolenia poly dactyla, n. sp.

„ 3. Coscinodiscus africanus, Janisch, var. rotunda, nov.

„ 4. Hyalodiscus Icevis (?) Ehrenb.

,, 5. Rhizosolenia 7'obusta, Norman, var. nov.

„ 6. Tlialassiothrix curvata, n. sp.
Figures 7, 8, 10; and 13. Rhizosolenia inermis, n. sp.

Figure 9. Lauderia (?) moseleyana, n. sp.

„ 11. Rhizosolenia sima, n. sp.

,, 12. Rhizosolenia, murrayana, n. sp.

„ 14. Rhizosolenia, sp., vel forma monstrosa (?)

„ 15. Rhizosolenia iniequalis, n. sp. .

PAGE

73
71
159
140
73
55
71
90
71
72
73
71

ViOtOllU ,,,■,/■ I'l WIV

? — — ;

t> <

t> <i

P <

i —

> — <

i >)

1

PLATE XXV.

(BOT. CHALL. EXP.— PART IV.— 1886.)

PLATE XXV.

Figure 1. Fragilaria (?) (an Terebraria (?) ), sp. (?)

2. Terebraria (?), sp. (?) .

3. Cyclophora tenuis, Cstr., var. no v.

4. Hemiaulus glacialis, n. sp.

5. Isthmia enervis, Ehrenb., var. japonica, nov. .

6. Asterionella gracillima (Hantzsch.), Heib.

7. Triceratium insutum, n. sp. .

8. Molleria cornuta, Cleve.
Figures 9 and 10. Bacillaria socialis, Greg., var. indica, nov.
Figure 11. Fragilaria linearis, n. sp., et Fragilaria, sp. (?)

12. Fragilaria antarctica, n. sp.

13. Synedra capitulata, n. sp.

14. Synedra Jimbriata, n. sp. .

15. Synedra philippinarum, n. sp..

16. Synedra atlantica, n. sp. .

17. Fragilaria, sp. .

18. Synedra lanceolata, n. sp., var. thaitiensis, nov.

19. Gephyria, n. sp. (?) .

20. Synedra lanceolata, n. sp.

21. Fragilaria (?)

The determination of this form is somewhat doubtful, and its detailed descrip
tion has been omitted from the text.

,, 22. Diatoma rhombicum, O'Me. .

PAGE

47
47
58

100
96
51

111
98
69
57
56
52
. 52
52
53
57
53
42
52

57

'

kw

'■

11

225
1

1(1

1

S53

12

;

13

14

16

17

18

20

21

19

fr»0

1

22

560

PLATE XXVI.

PLATE XXVI.

PAGE

Figure 1. Biddulphia, sp. (?)

„ 2. Biddulphia weissflogii, Grun. .

„ 3. Rhaphoneis mammalis, n. sp.

„ 4. Biddulphia ropcriana, Grev. .

„ 5. Biddulphia pellucida, n. sp. .
Figures 6 and 8. Cerataulus turgidus, Ehrenb., var. polyceros, uov.
Figure 7. Biddulphia parallela, n. sp. (?)

„ 9. Biddulphia reticulata, Roper, var. inermis, no v.

„ 10. Biddulphia tuomeyi, Bail. ....

„ 11. Mastogloia thaitiana, n. sp. .

„ 12. (?) Undetermined form

,, 13. Rhaphoneis clUptica, n. sp.

,, 14. Toxonidea challengeriensis, n. sp., var. uov. .

,, 15. Toxonidea challengeriensis, n. sp. .

•

104

•

104

♦

48

106

103

101

103

102

106

22

49
39
39

• I ' I . X.WI

.

1

+

!

10

It

1

'

12

14

'

I.',

PLATE XXVII.

(BOT. CHALL. EXP. — PART IV. — 1886.) D 0

PLATE XXVII.

Figure 1.

„ 2.

Figures 3

Figure 4.

„ 5.

' „ 6.

„ 8.

„ 9.

„ 10.

„ 11-

„ 12.

, 13.

, 14.

, 15.

, 16.

, 17.

, 18.

, 19.

, 20.

Amphora speciosa, n. sp.

Amphora philippinica, n. sp.

and 7. Navicula, sp. (?)

Cymbclla pelagica, n. sp.

Cymbella criophila, n. sp.

Amphora staurophora,, n. sp.

Amphora polyzonata, n. sp., var. nov.

Stauroneis glacialis, n. sp., var. nov. .

Navicula, sp. (?)

Stauroneis glacialis, n. sp.

Ceratoneis arcus, K.g.=Eunotia Arcus, W. Sm.

The form here shown was found not unfrequently in the gathering from the
port of Tahiti. It approaches in its character the genus Amphora, and may
prohably belong to this genus.

Cymbella marina, n. sp.
Amphora decora, n. sp.
Amphora thaitiana, n. sp.
Amphora meneghiniana, n. sp.
Navicula decipiens, n. sp.
Amphora polyzonata, n. sp. .
Amphora scalaris, n. sp.
Amphora oceanica, n. sp.

PAGE

17
19
35
21

21
20
18
25
35
24
50

21

18
19
17
32
18
18
20

""7.-/VVC V I X V \ II

•

.

.

10

d

a

is

f - if l/l

i

16

Y

ia

I!)

\_ , I i -

■

•20

PLATE XXVIII.

PLATE XXVIII.

Figure 1. Pleurosigma elegantissimum, n. sp.

2. Pleurosigma speciosum, W. Sm., var. nov.
,, 3. Pleurosigma naviculaceum, Breb., var. nov.
,, 4. Pleurosigma thaitiense, n. sp. .
„ 5. Pleurosigma arafurense, u sp.

6. Pleurosigma smithianum, n. sp.
„ 7. Navicula bullata, Norman, var. carinata, nov,
,, 8. Navicula zanzibarica, Grev., var. zebuana, nov
,, 9. Navicula spectabilis, Grev., var. nov. .
„ 10. Navicula bullata, Norman, var. obtusa, nov.
„ 11. Navicula (V) jejuna, A. S., var. nov. .
„ 12. Navicula parallela, n. sp.
„ 13. Navicula, sp. (?)

„ 14. Alloioneis antillarum, CI. et Grun., var. nov
,, 15. Navicula, sp.
,, 16. Navicula kerguelenensis, n. sp.
,, 17. Navicula, sp. (?)
„ 18. Navicula cyclophora, n. sp.
,, 19. Navicula abnor mis (?), n. sp. .

PAGE

37
37
37
38
38
38
29
31
32
29
33
31
32
35
32
33
32
34
27

Viauniicuxuz PI. VWI

r

10

17

560
I

19

PLATE XXIX.

(bot. CHALL. EXP. — PART IV. — 1886.) D P

PLATE XXIX.

PAGE

Figure 1. Bacteriastrum spirillum, n. sp. . . . . . 83

Figures 2, 11, and 15. Coscinodiscus radiatus, Ehrenb., var. abyssalis, nov. . 165
Figure 3. Bacteriastrum varians, Lauder, var. princeps, nov., frustulum terminale 84

4. Rhizosolenia (?) flaccida, n. sp. . . . . .74

5. Nitzsehia mammalis, n. sp. . . . . . 68

6. Bacteriastrum tvallichii, Ralfs, var. hispida, nov. . . . 83

7. Stauroneis pygmea, n. sp. . . . . .25

8. Rhizosolenia (?) forma sporangialis (?) . . . . 74

9. Rhizosolenia sima, n. sp., var. nov. ..... 72

Figures 10 and 16. Chcetoceros incurvum, Bail., var. umbonatum, nov. . . 80

Figure 12. Grammatophora stricta, Ehrenb., var. nov. . . . .57

„ 13. Actinocyclus pellucidus, n. sp. . . . .147

,, 14. Pleurosigma japonicum, n. sp. .... 38

.

Diatomaceee Pi. XXIX

PLATE XXX.

PLATE XXX.

Figure 1. Actinocyclus ralfsii, W. Sm., var. challenger ensis, nov.

2. Systephania (?), sp. (?) .

3. Systephania aculeata, Ehrenb., var. /3, nov. .

4. Thalassiosira nordenskioldii, CI., var. nov.
4 bis. Thalassiosira nordenskioldii, CI. (fig. given by Grunow)

5. Navicula janischii, n. sp.

6. Biddulphia tuomeyi, Bail., var. pacifica, nov.

7. Navicula bullata, Norman, var. rhomboidea, nov.

8. Amphiprora plicata, Greg., var. japonica, nov.

9. Biddulphia (?), sp. (?)
10. Navicula mirabilis, n. sp.
LI. Rhizosolenia, sp., fragmentum

12. Rhizosolenia arafurensis, n. sp., fragmentum

13. Navicula lyra, Ehrenb., var. signata, A. S.

14. Rhizosolenia, sp., fragmentum

PAGE

143

151

151

96

96

29

106

30

40

105

34

74

74

33

74

!h. ,/,.,, ,„.,-.,■ Pi. XH

•

ill

'v, ■ : !■' ,: i'fi'v' i'i).: .