iWHW-liUwi/ll

fmaTTUUrm

^'^f'

-.'l*

m^m.

■■•,■■■ ■■ r-y^,^ '-•!*■-

y „ »» -'

■1 -'

THE

VOYAGE OF H.M.S. CHALLHNGER.

ZOOLOGY-VOL XXI.
TEXT.

REPORT

ON THE

SCIENTIFIC RESULTS

OF THE

VOYAGE OF H.M.S. CHALLENGER

DURING THE YEARS i 8 7 3-7 6

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.

KF.r.nrs professor of natural history in thf. university of edinburch

DIRECTOR of THE CIVILIAN SCIENTIFIC STAFF ON BOARD
AND NOW OF

JOHN MURRAY

ONE OF THE NATURALISTS OF THE EXPEDITION

Zoology— Vol. XXI.

TEXT
Ptibltst)fti b]) 0i:Xitt of I^er iWajestp's (Soliernment

PRINTED FOR HER MAJESTY'S STATIONERY OFFICE

AND SOLD liY

LONDON :— EYRE & SPOTTISWOODE, EAST HARDING STREET, FETTER LANE

EDINBURGH :— ADAM & CHARLES BLACK

DUBLIN :— HODGES, FIGGIS, &: CO.

1887 A^

Price (including Volume of Plates) Seventy SInllings. I '»v^

U \

M

A

PRINTED BY NEILL AND COMPANY, EDINBURGH,
FOR HER majesty's STATIONERY OFFICE.

CONTENTS.

Report on the Hexactinellida collected by H.M.S. Challen-ger during the years

1873-1876.

By Dr. F. E. Schulze, Professor of Zoology iu the University of Berlin.

TEXT.

9093

EDITORIAL NOTE.

The Report on the Hexactinellida, by Professor F. E. Scliulzc of the

University of Berlin, occupies the whole of the present Volume, the Text and

Plates being bound up in separate parts. This Report forms Part LIII. of

the Zoological Series of Reports on the Scientific Results of the Expedition ;

it extends to over 500 pages and is illustrated by lO-i Lithographic Plates
and many woodcuts.

Sir C. Wyville Thomson gave special attention to this group of Sponges
during the Expedition, and it was his intention to describe them in detail
in these Official Reports. Owing, however, to ill health and other circum-
stances he was never able seriously to commence the work, and at the time
of his death there remained only a few Plates which had been prepared
under his directions.

It must be regarded as a singularly fortunate circumstance that I was, in
1882, able to induce so eminent a Spongiologist as Professor Schulze to
undertake the examination and description of this most characteristic deep-
sea group of animals, to which he had already given special attention.
The result of Professor Schulze's labours is presented in this magnificent
Monograph, which will be welcomed by naturalists in all parts of the world
as a substantial and brilliant addition to our scientific knowledge.

Some portions of the German manuscript were translated by Mr. John
Rattray, M.A., B.Sc, F.R.S.E., but by far the larger part was translated by
Mr. Arthur Thomson, M.A., F.R.S.E., and the whole was revised by him.

The Manuscript was received by me in instalments between the 10th
August 1886 and the 30th May 1887.

John Murray.

Challenger Office, 32 Queen Street,
Edinburgh, 1st Aurjust 1S87.

THE

VOYAGE OF H.M.S. CHALLENOEE.

ZOOLOGY.

KEPORT on the Hexactinellida collected by H.M.S. Challenger during
the Years 1873-76. By Dr. F. E. Schulze, Professor of Zoology in
the University of Berlin.

PEEFACE AND HISTORY OF THE WORK.

In December 1879, I was asked by Sir WyviUe Thomson to examine some specimens of
Euplectella aspergillum, which had been procured on the Challenger Expedition, and
which had been preserved with special care for histological purposes. The results of
my investigation were to be published as part of the Report of the Challenger Expedi-
tion. Though at that time engaged in the study of the Adriatic Sponges I did not
hesitate to concur with the request, as I was strongly attracted by the prospect of
being able to study the hitherto almost unknown soft structure of the Hexactinellida.
I soon received, from the Challenger office in Edinburgh, an entire specimen of
Euplectella aspergillum (from oif Zebu), preserved with its soft parts in methylated
spirit, and six bottles containing fragments of the same species preserved in absolute
alcohol, or in other preservative media, such as chromic acid, picric acid, acetate of
potash, glycerine, &c.

After the results of my investigation of this, in part, well preserved material had
been published in the Transactions of the Royal Society of Edinburgh (1880, vol. xxix.
p. 661, Tab. A), Sir Wyville Thomson asked me to work over with him all the
Hexactinellida obtained during the expedition. He was to undertake the study of the
skeletal structures, whUe the investigation of the soft parts was to be entrusted to me.
To this I readily agreed, and began at once with the examination of some samples, which
had been previously sent to me at Graz.

(ZOOL. CHALL KXP. — PART LIII. 1886.) Ggg 1

r

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

This pleasant arrangement for joint research was, however, frustrated by the illness
and death of Sir Wyville Thomson. I had, accordingly, to lay aside the whole work for
a prolonged period, till, in March 1882, Mr. John Murray asked me to undertake the
complete investigation of the Challenger Hexactiuellida, and for this purpose, in the first
instance, personally to undertake the separation of the latter from the rich collection of
Sponges obtained during the expedition and preserved in Edinburgh. Having com-
pleted this in April of the same year, I received, in June, most of the Hexactinellida
collected by the Challenger Expedition, which were forwarded to me at Graz in a well-
packed condition. Subsequently, at my request, a few additional Hexactinellida, which
were collected before by Sir Wyville Thomson and Mr. John Murray in the eastern part of
the Atlantic, during the expeditions of the " Lightning," " Porcupine," " Knight Errant,"
and " Triton," were handed over to me for purjDoses of comparison, and also a number
of microscopic preparations which had been made by Sir Wyville Thomson.

My material was afterwards very agreeably increased by the acquisition of a not
unimportant collection of Hexactinellida, partly dry and partly preserved in alcohol,
which Dr. Doderlein, Director of the Zoological Museum in Strassburg, had accumulated
during a prolonged stay in Japan, partly as the result of his own collecting on the
small island of Enoshima. This collection he very kindly placed at my disposal for
investigation.

At my request the Challenger Commission agreed that I should include this valuable
material in my work on the Challenger Hexactinellida, and approved of what thus
necessarily involved an increase of plates. I therefore accepted the offer of Dr.
Doderlein, and that all the more willingly since the increase of material considerably
facilitated the systematic work. On the other hand, I was enabled to compare minutely
a collection zealously and systematically made for several years from a definite locality
remarkably rich in Hexactinellida, with that collected in the same locality by the
Challenger Expedition.

As to the ways and means of acquiring this collection, which consisted of about
twenty-five dried forms and three bottles of preserved specimens, Dr. Doderlein
communicated to me the following notes : — " With the exception of a Euj)lectella
(pweni) from Simonoseki on Kiushiu, the whole collection was gathered in Sagami
Bay, near Enoshima. Most of the dried specimens were obtained from a shop in
Enoshima, where they were collected for me by the owner. The three glasses with
spirit contain specimens which I myself obtained with a trawling apparatus from depths
varying from 80 to 240 fathoms, between Enoshima and Misaki. A few dead specimens
which I obtained there have been sent dry. I have added a single specimen of
Hyalonema (sieboldii) in which the sponge body is magnificently developed, but it is by
no means the largest specimen which I brought with me."

Professor Oscar Schmidt of Sti'assburg was kind enough to hand over to me, for

EEPORT ON THE HEXACTINELLIDA. 8

a prolonged period, for purposes of comparison, portions of most of the Hexactinellida
which he has described, and also a very large nuniber of microscopical preparations.
From Dr. Marshall of Leipzig I likewise obtained the use of portions and preparations
of all the specimens which formed the basis of his own Hexactinellidan ol)servations.
I have been similarly assisted by Hofrath Director A. B. Meyer of Dresden,
Professor Semper of Wiirzburg, Professor von Martens of Berlin, and Professor Moseley
of Oxford. In the important work of determining forms, a visit which I paid to London
was of much value, since, through the Challenger Expedition Commission, I was enabled
to study the original specimens of several Hexactinellida described by Owen, Gray,
Bowerbank, Carter, and others. During my work in London, not only were the rich
treasures of the Museum made accessible to me in the most liberal way by Dr. Giinther,
Director of the Zoological Department of the British Museum, but I was also accommo-
dated with a private work-room in the new building of the British Museum in South
Kensington. There I enjoyed the most attentive and friendly assistance of Mr.
S. 0. Ridley.

I had valuable assistance of another kind from Dr. Vosmaer of Naples, who had the
great kindness to place at my disposal the manuscript of that portion of his great
Spongiological Bibliography (in course of ^preparation) which dealt with the
Hexactinellida.

In the preparation of the numerous plates I have been ably helped by my assistant,
Dr. von Miihrenthal, who has photographed a number of objects, and by Mr. G.
Firtsch, one of my students, who has executed numerous drawings from my sketches
and according to my directions.

It is my pleasant duty to return "my best thanks to all the gentlemen whom I have
named, but especially to the Director of the Challenger Office, Mr. John Murray.

With regard to the figures given in the plates, I have still to note the following
points. From the scientific papers belonging to Sir Wyville Thomson, sixteen
lithographed plates have been handed over to me, all of which have been incorporated
in the present work, viz.. Pis. I., 11. , V., VIL, VIII., IX., X., XV., XXXI. , XXXV.,
XL., XLV., XLVL, LI., LXVL, and LXVII. As these had not yet been printed ofi"
when I received them, I was able to make several corrections, though I have pre-
ferred to leave most of them unaltered. Only from Pis. XXXI. and LXVII. a
few figures have been removed, which represented spicules not really belonging to the
Sponges figured. Every authority will admire the beautiful and life-like figures both
of the entire Sponges and of the separate preparations which are found in Sir WyviUe
Thomson's plates. I was not able, however, to follow the method illustrated in these
plates, since I could not confine myself merely to the general form of the Sponge, nor to
the structure of the skeletal elements, but had also to give representations of the soft
parts. It seemed to me, further, especially desirable to display the internal architecture

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

of the different species in a clear way, so that it might be possible, at a glance, to survey
and compare the different forms. With this view I have not only given figures of the
entire Sponges, of special fragments, and of the skeletal elements, but also numerous
ideal (less highly magnified) sections through the body-wall. In these diagrammatic
figures, which are composed from a number of microscopic sections, the skeletal elements
are indicated in blue. If I had attempted to copy the individual sections exactly as
they appeared, the essential and typical could not, as a rule, have been distinguished
from the unessential and accidental, except, of course, by giving a larger number of
illustrations than seemed justifiable for such a slight possible advantage.

Since I had repeated occasion to restudy the rich material, even after the plates were
printed off (i.e., at a time when it was no longer possible to introduce corrections), I have
been forced sevei'al times to indicate a change in my original opinion, by a correction in
the text, or by a change in the specific designation of the plates.

GENERAL HISTORICAL INTRODUCTION,

All siliceous Sponges in which the spicules belong to the triaxial type were, in 1870,
designated by Oscar Schmidt " Hexactinellidse." This title, which I would only modify
into Hexactiuellida, since the termination idas should be exclusively reserved for the
designations of ftimilies, soon found general acceptance, although Wyville Thomson had
previously established under the name of Vitrea an order of siliceous Sponges in which the
characteristic peculiarity was defined by the reference of all the spicules to the hexradiate
type. The older title proposed by Wyville Thomson has not gained currency, since
among the representatives of Vitrea enumerated by him forms occur in which the skeletal
elements are not referable to the hexradiate type, and which accordingly belong to another
order.

Confining myself at this stage to a short review of the historical development of our
knowledge of the Hexactiuellida in general, I purpose subsequently to preface the
description of each genus with an accurate account of the relevant literature.

First of all, there are a few isolated contributions of comparatively ancient date, which
give some account of certain structures which differed essentially in their peculiar form
and siliceous framework from any marine forms then known, but which, when subjected
to careful criticism, can be referred to certain now fairly familiar Hexactiuellida. Thus
Rozier's Journal de Physique for 1780, and a transcript from that work in the Magazin
fiir das Neueste aus der Physik und Naturgeschichte (published by Lichtenberg, Bd. 1,
Gotha), 1781, contain a description and a very characteristic representation of a form
belonging to the genus Dactylocalyx. This contribution, for an acquaintance with
which I am indebted to Dr. W. Marshall of Leipzig, is probably the earliest notice of a
Hexactinellid. In the Synopsis of the Contents of the British Museum for 1832, Gray
described a peculiar " glass rope like " structure preserved in the British Museum. This
he named Hyalonema, and described it carefully, though without recognising its real
nature as the basal tuft of a Hexactinellid.

In the Voyage de "I'Astrolabe," 1833, Quoy and Gaimard figured and described, under
the name Alcyoncellum speciosum, a sponge form which undoubtedly belonged to the
Hexactiuellida.

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

Similar individual descriptions became gradually more numerous, and up to 1860
the following may be noted as most important : — Dactylocalyx pumiceus, Stuchbury,*
1841, Euplectella asjKrgilluin, Owen,^ 1841, Farrea sp., Owen,* 1857, Aphrocallistes
beatrix, Gray,* 1858, and Myliusia callocyathus, Gray,* 1859.

A more profound study of the skeletal structure of Hyalonema sieholdii was made in
1860 by Max Schultze.® He also discovered, in those spicules which did not externally
exhibit a cruciate or stellate, but merely a simple rod-like form, an intersection of the
axial canals in a median swelling, whieli indicated the fundamental stellate type of all
the spicules. He was also the first to discover the close affinity of Hyalonema and
Etiplectella, which, on account of the common character of the spicular tuft, he united in
the group " Lophospongite."

Bowerbank'^ (1862) was less fortunate in hia perception of the affinities of the
Hexactinellid genera known to him, viz., Alcyoncellum [Euplectella, Owen), Quoy and
Gaimard, Hyalonema, Gray, Dactylocalyx, Stuchbury, and Farrea, Bowerbank. For
while he placed the genus Alcyoncellum, Quoy and Gaimard (with Euplectella, Owen, in
parenthesis), in his suborder Silicea with sjyiculo-radiate skeletons, between Ecionema,
Bowerbank, and Polymastia, Bowerbank, he referred the genus Hyalonema, Gray, to
another quite different suborder, characterised by spiculo-reticulate skeletons, between
Halichondria, Flemming, and Isodictya, Bowerbank. Of each of the two genera,
Dactylocalyx, Stuchbury [ = Iphiteon, Mus. Paris), and Farrea, Bowerbank, he
made, on the other hand, a special suborder, of which the former was characterised
chiefly by solid siliceo-Jihrous, and the second [Farrea) by canaliculated siliceo-fibrous
skeletons.

In Gray's System of Sponges,^ which appeared in 1867, the Hexactinellida then
known were not yet united into a common group. For while Gray placed the family of
the Euplectellidse, consisting of Alcyoncellum and the closely allied Euplectella, with his
Esperiadse and Tethydse, in the order of the Acanthospongise (with spicules of more than
one form or kind in the same Sponge) and within the subsection Sjjiculospongise (with free
spicules), on the other hand he united the family of the Aphrocallistidpe, consisting of
the genus Aphrocallistes, with the family of the Dactylocalycidse, including Dactylocalyx,
Stuchbury, Myliusia, Gray, MacAndrewia, Gray, and Farrea, Bowerbank, in a special
order, " Corallispongise," within the subsection " Dictyospongise " (in which the skeleton is
formed of a continuous siliceous or horny network). The Corallispongias were charac-
terised by Gray as : — " Hard, coral-like Sponges, entirely formed of siliceous spicules,
anchylosed together by siliceous matter into a network. Mass covered with a thin coat
of sarcode when alive."

' Proc. Zool. Soc. Land., vol. ix. pp. 86, 87. ^ Proc. Zool. Soc. Load., vol. ix. pp. 3-5.

3 Trans. Linn. Soc. Land., vol. xxii. pp. 117-124. '' Proc. Zool. Soc. Land., vol. xxvi. pp. 114, ll.";.

5 Proc. Zool. Soc. Land., vol. xxvii. pp. 437-440. '' Die Hyalonemen, 1860, 4.

7 Phil. Trans., vol. clii. 2 pp. 747, 830, 1087. » Proc. Zool. Soc. Land., 1867, pp. 117, 492, :001.

REPORT ON THE HEXACTINELLIDA.

Glaus ^ now pronounced against any close systematic union of Euplectella and
Hyalonema.

On the otlier hand, in 1868 Wyville Thomson first recognised the common relation-
ship of all Hexactinellida, though he was not able to separate them sufficiently sharply
from the Lithistida. The characters of his order Vitrea were given in these words : - —
" Sarcode in small quantity, very soft, never containing formed horny matter, either
fibrous, membranous or granular. The skeleton consists entirely of siliceous spicules,
either separate (in fascicles or scattered) or anastomosing and combined into a continuous
siliceous network. The sarcode contains small spicules of a different character from the
general spicules of the skeleton, and of complicated forms. The spicules, whether of the
skeleton or of the sarcode, may all be referred to the hexradiate stellate type," and in
another place (Phil. Trans., 1869, p. 713) : — "In all the known genera all the spicules
are modifications of the hexradiate stellate type."

In the order Vitrea, Wyville Thomson noted {loc. cit., p. 713) the following genera and
species : —

Genus 1. Euplectella, Owen,

,, 2. Habrodictyon, Wyv. Th.,

,, 3. Aphrocallistes, Gray,

., 4. Dactylocalyx, Stuchbury,

5. Farrea, Bowerbank,

6. Holtenia, Wyv. Thom.,

7. Hyalonema, Gray (in part).

8. Adrasta (n. gen.),

Euplectella aspergillum, Owen.
Habrodictyon speciosum, Quoy et Gaimard.
Habrodictyon corbicula, Valcncienne.
Aphrocallistes beatrix, Gray.
Aphrocallistes bocagei, Wright.
Dactylocalyx pumicea, Stuchbury.
Dactyloccdyx subglobosa, Gray.
Dactylocalyx prattii, Bowerbank.
Dactylocalyx callocyathes, Gray.
Dactyloccdyx azorica, Gray.
Dactylocalyx{?)torva,D'a.chiissa.mget'M.iche\otti.
Farrea occa, Bowerbank.
Holtenia carpentcri, Wyville Thomson.
Hyalonema sieboldii, Gray.
Hyalonema lusitanicum, Gray.
Hyalonema loveni, n. sp.
Adrasta infundibulum, n. sp.

If we except the genus Adrasta, which, though named, has remained undescribed,
and even undiagnosed, and further, the above-mentioned Hyalunema loveni, which is not
sufficiently defined, and finally the doubtful Dactylocalyx torva, Duch. and Mich.,
there are only two of the above-cited species quoted which are not Hexactinellida, but
belong to the Lithistida, namely, Dactylocalyx prattii, Bowerbank, and Dactylocalyx

1 Uber Euplectella aspergillum, 1868, p. 4. ^ ^„„_ and. May. Nat. Hist, 1868, vol. i. pp. 11-4-132.

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

azorica, Gray. The young forms described by Wy^Tlle Thomson in his memoir on
Holtenia, are in several points suggestive of Hyalonema.

Among the numerous siliceo-Jibrous sponges, which Bowerbank* described at the same
time in the Proceedings of the Zoological Society, there are also many forms, such as
Caliapsis cidaris, Bowerbank, Dactylocalyx heteroformis, Bowerbank, Dactylocalyx
macandreivii, Bowerbank, Dactylocalyx prattii, Bowerbank, Dactylocalyx masoni,
Bowerbank, Dcoctylocalyx hoiverhankii, Johnson, Dactylocalyx polydtscus, Bowerbank,
which belong not to the Hexactinellida, but to the Lithistida. With regard to Bowerbank's
remarks on the structure of the Hexactinellida, his description of a special skin or
cortical layer with peculiar skeletal elements deserves to be noted.

In addition to further descriptions of some new Hexactinellid species, — such as
Eurete simplicissima, Semper, Enpdectella oweni, Herklots and Marshall, Pheronema
annsB, Leidy, Semj^erella schultzii, Semper, &c., — protracted discussions on the systematic
position of Hyalonema and Euplectella, and on their mode of attachment, were continued
for several years.

Some new genera and species from the Atlantic area were described in 1870 by Oscar
Schmidt,^ who, after very accurate study of the axial relations of the siliceous spicules, was
the first, as above noted, sharply to distinguish the Hexactinellida from all other Sponges,
and especially from the Lithistida, with which they had been hitherto more or less united.

While the genera Lanuginella and Sympagella, which were then established by 0.
Schmidt, are without doubt true Hexactinellida, I cannot say the same of Placodictyum
cucumaria, 0. Schmidt. Preparations of the skeleton which exhibit broken plates and bent
knotted rods, were kindly given to me for examination by Prof. 0. Schmidt himself, and
from these I have been able to convince myself that by the action of acids the elements
are dissolved with the liberation of gas, so that they are evidently not of silica, but lime
carbonate. It seems probable, in fact, that we have to deal with a Holothurian — Thyone
— and this supposition is confirmed by the general form of the organism as figured.

By means of microscopic examination — which had been but little applied to the study
of fossil Sponges — certain forms belonging to the genera Scyphia and Ventriculites were
indubitably referred by 0. Schmidt to the Hexactinellida.

About the same time Saville Kent,^ who had found on the coasts of Sj^ain and
Portugal several little known Hexactinellida, including the new genera Asconema, Kent,
and Fieldingia, Kent, gave a systematic review of all the then known Hexactinellid
genera, and distributed them in two suborders, as follows : —

I. CoEALLiosPONGiJS, Gray. Sponge body supported by an anastomosing or con-
tinuous reticulate skeleton. Reproductive gemmules entirely membranous, aspiculous.

1 Pwc. Zool. Soc. Land., 1868, pp. 118-137; 18G9, pp. 66, 323, 389.

2 O. Schmidt, Grundziige einer Spongienfauna des atlant. Gebietes, 1870.

3 Monthly Micr. Journ., 1S70, p. 241.

REPORT ON THE HEXACTINELLIDA. 9

Genera : — Euplectella, Owen, Hahrodictyon, Wyville Thomson, Aphrocallistes,

Gray, Farrea, Bowerbank, Aulodictyon, Kent, MacAndrewia, Gray, Dacty-

localyx, Stuchbury, Fieldingia, Kent.

II. Callicispongi^e, S. Kent. Sponge body supported by an interlacing or isolated

spicular skeleton ; never by a reticulate and continuous one. Reproductive gemmules

membranous, furnished with protective spicules.

Genera : — Pheronema, Leidy, Hyalonema, Gray, Asconema, Kent, Sympagella,
0. Schmidt, LanugineUa, 0. Schmidt, Vazella, Gray.
In his Grundziige der Zoologie, Glaus maintained the union of the spicules into a
firm network to be an important character of all Hexactinellida (which constitute his
suborder of Hyalospongite), while Carter, 1873, like Saville Kent, only emphasised the
differences in the modes of union of the spicules as the main basis of his classification.
Carter ^ established the three following families : —

I. VlTREOHEXACTINELLIDA.

Spicules held together by silicified fibre.

Dactyloccdyx, Myliusia, Euplectella aspergillum, Aphrocallistes, Atdodictyon,
Farrea, Sympagella.

1. Patulina. Example — Dactylocalyx immiceus.
Gray.

2. Tubulina. Example — Euplectella aspergillum,
Owen.

3. Scopulifera. Example — Aphrocallistes hocagei,
Wrio;ht.

Including the three groups <|

II. Sarcohexactinellida.

Spicides held together by amorphous sarcode.

Asconema, Crateromorpha, Rossella, Hahrodictyon, Hyalonema, Pheronema,
Meyerina.

1. Eosettifera. Example — Rossella, Carter, and Cratero-
morpha meyeri, Gray.

2. Birotulifera. Example — Hyalonema sieholdi, Gray;
Holtenia, Wyv. Thomson; Meyerina, Gray, and
Labaria, Gray.

Including the two groups

III. Sarco-vitreohexactinellida.

Spicules held together in one part by vitrified fibre, in the other by amorphous

sarcode.
Euplectella cucumer, Owen.

1 Ann. and Mag. Nat. Hist, ser. 4, vol. xii. y. 349 ; vol. xvi. p. 199.
(ZOOL. CHAIX. EXP. — PAET LIII. — 1886.) "S8 "^

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

These siliceous spicules which are united iuto a continuous network, or which form, iu
virtue of their large size and mutual apposition, a supporting framework for the entire
sponge body, were termed by Carter " skeleton spicules," — in opposition to the far smaller
" flesh spicules " which are loosely embedded in the soft tissue. In the new genera
RosseUa, Carter, and Crateromoiyha, Gray, Carter was able to distinguish several species.

In 1875 Marshall published his researches on the skeleton of several sponge forms,
some imperfectly known, and others newly discovered, e.g., the genera Sclerothamnus,
Marshall, and Pervphragella, Marshall. The affinities of the Hexactinellida were
discussed by Marshall in a special work. To start with, he distinguished Synauloidse
and Asynauloidge. In the former the entire lattice network is said to be penetrated by
a continuous system of axial canals, while in the latter the canals of the spicules which
fuse to form the network do not communicate. To the Synauloidas Marshall referred
only the genus Sclerothamnus. The Asynauloidse he divided into (1) Monacid*, with
only one form of spicule ; (2) Pleionacidfe with forks and rosettes, in addition to
the six-rayed forms ; and (3) Pollacida3, with numerous distinct forms of spicules, a
special dermal skeleton, and an inner covering for the gastral cavities. While Marshall
placed in the division Monacidfe only the genus Eurete, he assigned to the Pleionacidse
the genera Lanuglnella, Schmidt, Asconema, Kent, Farrea, Bowerbauk, Periphragella,
Marshall, Aulodictyon, Kent, Fieldingia, Kent, and A2)hrocaUistes, Gray ; to the
PoUacidse, on the other hand, he ascribed the family of the Holteuiadge with Holtenia,
Wyv. Thomson, Crateromorpha, Gray, RosseUa, Carter, Sympagella, Schmidt, Placo-
dictyum, Schmidt, the family of the Euplectellidse with Euplectella, Owen, and Hahro-
dictymn, AV. Thomson, and the family of the Hyalonematidse, with Labaria, Gray,
Pheronema, Leidy, Semperella, Gray, and Hyalonema, Gray.

In 1877, SoUas^ described with great thoroughness a new fossil Hexactinellid genus
with two species. Both in the dermal layer, or "oscular plate" as he termed it, and
in the thick body mass, SoUas noted a framework of siliceous strands intersecting at
right angles. The usual axial canals were present, but the nodes were not penetrated
by them, i.e., they did not exhibit any octahedral or lantern-like form. To indicate
the systematic position of this new genus, SoUas elaborated the following classification
of the Vitreohexactinellids, according to the characters of the skeletal network : —

I. Sexradiate skeleton spicules, always rectangular. Stauronemata.

(a) Skeletal network, with simple nodes.

1. One layer in thickness, . . . Farrea.

2. Several layers thick, . . . Stauronema.

(b) Skeletal network having the nodes complicated by

the presence of an octahedral lantern about

each node, ..... VentriculitidEe, including Myliusia

grayi.
' Ann. and Mag. Nat. Hist., ser. 2, vol. xix. p. 1.

REPORT ON THE HEXACTINELLIDA. 11

II. Sexradiate skeleton spicules, with rays making any angle

■with each other. Aphrocallistidse, . . Aphrocallistes, Dadylocalyx, Ip-

Jiifeon, Siromafojiora (Callo-
dictyon, Sollas, n. gen.) con-
centrica.
III. Skeleton spicules, cemented into ladder-like fibre. Eu-

plectellidre, ...... Eapledella, Sympagella.

Zittel's epoch-making work on fossil Sponges' contains very accurate descriptions,
not only of the form of the body and nature of the surfaces, but of the system of canals
that penetrate the body, and especially of the fine microscopic structure of the siliceous
framework. The flesh spicules which lay loosely in the soft tissues, and were thus, for
the most part lost, could not of course be so closely studied.

As the main basis of his classification Zittel emphasised the differences in the modes
of union exhibited by the skeletal spicules, a basis of division which had been already
employed by Saville Kent and Carter. He distinguished those forms " in which the
skeletal spicules usually I'emain isolated and are only united by sarcode," from those
" in which the skeletal spicules are fused in a regular manner and form a continuous
lattice-work with cubical or polyhedral meshes." The former he named Lyssacina,
the latter Dictyonina.

That intercommunication of the lumina of the axial canals throughout all the
spicules fused into the lattice-like framework, which had been observed by Marshall in
Sclerothamnus, was not corroborated by Zittel, either in any fossil Hexactinellid or even
in Sclerothamnus itself. It seemed to him, further, that the formation of a special
group of Monacid*, in Marshall's sense, was unwarranted, at least as regards the
division of the Dictyoninse, but he himself formed, within the Lyssacina, from certain
fossil genera, a similar group, and ranged alongside of it the Pleionacidse and
and PoUacidse.

I will here cite the fundamental principles of Zittel's Hexactinellid system of
1878 :—

Class Spongia.
Order HEXACTINELLIDA, 0. Schmidt.

Siliceous Sponges with six-rayed spicules, isolated or fused into a lattice-work of a
hexradiate pattern. All the siliceous elements exhibit the same fundamental structure,
with an axial cross formed by three central canals intersecting at right angles. In
addition to the peculiar skeletal needles there are numerous isolated flesh spicules, mostly
very delicate in form.

^Abhandl. d. II. CI. h. barer. Akad. d. Wi^s., xiii., 1878.

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

Suborder I. DICTYONINA, Zittel.

Skeletal spicules fused in such a way that every arm of a six-rayed spicule is applied
to the corresponding arm of a neighbouring spicule, so that both become surrounded
by a common siliceous covering. The continuous skeleton consists of a framework,
with cubical or irregular meshes. Flesh spicules present or absent.

Family 1. Astylospongid^, Zittel.

Sponge body very thick walled, unstalked, free (occasionally fixed by a broadly
expanded base). Water vascular system consisting of radial canals extending from the
surface to the centre, besides which vertical tubes disposed in radial rows to the number
of eight or ten, are generally present. Lattice framework tolerably irregular with
thick nodes of intersection.

Only fossil forms from the Silurian.

Family 2. Euretid^.

Fixed sponge bodies beaker-like, cylindrical, top -like or branched. Skeleton lattice-
like ; the intersection nodes of the fused hexradiate spicules non-perforated. Surface
naked or protected by a thickening of the outer skeletal layer, sometimes covered with
a very delicate network of fused spicules, which, in theii* form, differ but slightly from
those of the rest of the skeleton. This outer dermal meshwork also surrounds the
ostia. Structure of the root resembling that of the rest of the sponge body. Flesh
spicules absent or present.

(a) Canal system well developed. Ostia of the blind radial canals occur variably on
either surface. In addition to fossil genera belonging to different strata, this suljfamily
is represented by the living genus Sclerothamniis, Marshall.

(b) Canal system absent or scarcely developed. Besides the fossil genus Verruco-
ccelia, the living genera Farrea, Bowerbank, Eu7-ete, Marshall, and Aulodictyon, S. Kent.

Family 3. Coscinopori 'dm, Zittel.

Sponge body beaker-like, stellate or branched, more frequently compressed. Eadial
canals very numerous, simple, straight, and blind. Ostia small. Skeleton finely meshed,
dense, stony. The numerous radial canals exhibit the regular formation of cubical
meshes. The intersection nodes of the six-rayed spicules are thick and seldom per-
forated. Dermal layer usually absent or only formed hj a thickening of the outermost
skeletal layer.

Examples. — Coscinospora and other fossil genera.

REPORT ON THE HEXACTINELLIDA. 13

Family 4. Mellitionid^, Zittel.

Sponge body branched, spherical or flat. Body-wall completely perforated by nume-
rous tubular water canals, and thus divided into honeycomb-like cells. Skeletal spicules
with thick nodes of intersection. Surface (naked or) covered by a delicate, meshed or
porous, siliceous skin, which also covers the openings of the canals. Root tuft absent.

Examples. — The genus Aphrocallistes, Gray, which is known both in the living and
fossil form, the living genus (X) Fieldingia, Saville Kent, and the fossil genus Stauronema,
Sollas.

Family 5. Ventriculiti d^, Zittel.

Sponge body simple or polyzoic, beaker-, funnel-, or top-like, cylindrical or branched.
Wall irregularly folded. Lattice framework with octahedral perforated nodes of inter-
section. Canal system usually well developed. Radial canals blind. Both surfaces
with ostia or longitudinal furrows. Dermal layer seldom absent, and usually formed by
thickening of the outer skeletal layer. Roots consisting of prolonged siliceous fibres
united by transverse bridges and without axial canals.

Examples. — Ventriculites and other fossil genera.

Family 6. Staurodermi dm, Zittel.

Sponge body top- or funnel-like, seldom branched. Lattice skeleton more or less
reo'ular. Intersection nodes thick or octahecbal, perforated. The outer, or both surfaces
of the wall provided with stellate spicules, which difi"er in form from those of the rest of
the skeleton, and are either but loosely cemented to one another, or lie embedded in a
continuous siliceous skin.

Examples. — Stam-oderma and other fossil genera.

Family 7. M^andro3pongid.e, Zittel.

Sponge body consisting of intricately labyrinthine and anastomosing thin-walled tubes
or foUse. Canal system absent or scarcely developed. Intercanalicular system always
present. Dermal layer absent or represented by a continuous siliceous superficial skin.

Examples. — Besides Cystispongia and other fossil genera, the living genera Dactylo-
calyx, Stuchbury, Periphragella, Marshall, and Myliusia, Gray {pro parte).

Family 8. Callodictyonid.e, Zittel.

Sponge body beaker-llke. Wall consisting of a regular wide-meshed lattice-like
framework with octahedral intersection nodes ; canal system absent or confined to the
sometimes very thick dermal layer of the lattice-like skeleton.

Examples. — Callodictyon, Zittel, and other fossil genera.

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

Family 9. Cceloptychid.e, Zittel.

Sponge body umbel-like, stalked. Wall thin, deeply folded ; the central cavity
divided into radial chambers. Upper surface shallow or deepened, entirely enveloped
by a continuous dermal layer which usually consists of variable coarse and finely porous
strands. Canalicular ostia only on the under side of the umljel, on the backs of the
folds, sometimes also on the stalk. Lattice framework with large, regular, cubical
meshes. The intersection nodes of the amalgamated six-rayed spicules octahedral,
perforated. The radii of the six-rayed spicules bear thorny and root-like protuberances.

Example. — The single fossil genus Cveloj^tychium.

Suborder II. LYSSACINA.

The whole skeleton consisting of spicules which are bound together only by sarcode
(exceptionally also by a smooth siliceous substance in iiTegular ways). Flesh spicules
usually present in rich abundance and much differentiated.

Family 1. M on acid. ^, Marshall.

Whole sponge body composed of similar spicules.

Examples. — Acanthospongia, Salt, and Stauractinella, Zittel.

Family 2. Pleionacice, Marshall.

Main portion of the skeleton composed of six-rayed spicules and also of forks and
rosettes.

Examples. — The living genera Asconema, Kent, and Lanuginella, Schmidt.

Family 3. Pollacid^, Marshall.

Form of skeleton and flesh sj)icules very variable. Siaeeial dermal skeleton and
inner lining of the gastral cavities present. Base usually forming a root-tuft of long
siliceous spicules.

Examples. — The living genera Holtenia, Schmidt, Pheronema, Leidy, Cratero-
morplia, Gray, Rossella, Carter, Sympagella, Schmidt, Placodictyon, Schmidt, Eiqjlec-
tella, Owen, Hahrodictyum, Wyv. Thomson, Laharia, Gray, Semperella, Marshall
{Meyerina, Gray), Hyalonema (Carteria), Gray, and 1 Acestra, Eoem, from the
Silurian.

In the report which Marshall and Meyer' made in 1877 on the Philippine Hexac-

' Miitheilungen aus d. Konigl. Zool. Museum in Dresden, 1877, p. 263.

REPORT ON THE HEXACTINELLIDA. 15

tinellida, tliey described, iu additiou to Sume new species of the genera Myliusia and
Aulodictyon, a new genus Ilyalocaulus, and gave an explanation of the hollow octa-
hedral nodes which occur so abundantly in fossil Hexactinellida, but less frequently
in living forms.

To the communications which Wyville Thomson ' made in 1877 on some specially
noteworthy Hexactinellida from the Challenger material, — such as Euplectella
suherea, Hyalonema toxeres, Folhpogon amadou, Lefroyella decora, and others, — •
I will refer at greater length when noting the literature of the genera and species in
question.

An important addition to our knowledge of the Hexactinellida was made in 1880 by
0. Schmidt.^ He gave a description of the forms collected by Agassiz in the Gulf of
Mexico, and entered into a discussion of their aiiinities and other general questions,
without, however, undertaking an independent elaboration of the system.

Schmidt does not deny the merit of Zittel's systematic grouping of all known fossil
and living species into genera, families, and orders, but while admitting its utility
for the practical purposes of the geologist, expresses his doubt whether Zittel's
system represents even approximately the natural affinities. He does not therefore
feel himself warranted in arranging the genera described in any completely articulated
system.

In the living Dictyoniua, Schmidt sees only "representatives of the incompletely or
altogether unknown fossil Sponges, whose affinities will for ever remain concealed"; the
Lyssacina, on the other hand, are " for the most part so closely related to one another,
that the boundaries between the genera may be moved quite arbitrarily." The
Euplectellidse and Hyalonematid^ appear to him genuinely natural families, but he does
not give the same credit to Marshall's Holteniadse.

It is a great pity that some of the genera established by 0. Schmidt have been
described from fragments but slightly characteristic, and sometimes so shortly that it is
difficult to obtain any sufficient conception of their characters. This is especially the
case with the new genera Diaretida, CyatheUa, Diplacodium, Pachaulidium, Rhahdo-
stauridium, and Leiobolidium.

The three genera Farrea, Bowerbank, Eurete, Marshall, and Aulodictyon, Kent, are
united by Schmidt into a new genus Farrea, of which he found but one species, Farrea
facunda, represented by numerous and certainly very variable examples. A new genus,
Syringidium, Schmidt is inclined to refer to the Lefroyella decora, Wyville Thomson,
figured and briefly described in Wyville Thomson's Atlantic.

Of the genus Cystispongia, hitherto only known in the fossil condition, he describes
a living representative as Cystispongia superstes. The genus Margaritella, Schmidt, is
supposed to be closely allied to Ccdo2)tychium, but it does not possess the perforated

1 The Atlantic, 1877. '- Die Spongien des Meerbusens von Mexico, ii. \\ 33.

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

octahedral nodes. Joannella, Schmidt, is distinguished by its narrow meshed network
and the marked fissures associated with this character. In one species of the new genus
Scleroplegma [Scleroiolegma laterna) there are perforated octahedral nodes, while in the
other species of the same genus the nodes are solid. In Volvulina sigsbei, 0. Schmidt,
which is distinguished by the knotted cords of the siliceous network of beams, Schmidt
thinks that he can prove the non-fixity of certain characters of the Dictyonine lattice-like
tissue, characters which have been principall)^ used by Zittel in the classification and
determination of fossil Hexactinellida. While, namely, in some parts of the lattice
network smooth rays and rough knots appear, in other regions of the same specimen the
rays are also rough. Many specimens of the same species exhibit rough rays and smooth
knots, the exact reverse of what has just been stated above. Schmidt has also found
that the form of the meshes is very variable and uncertain, they are sometimes cubical,
sometimes predominantly polyhedral, sometimes quite irregular.

Of Euplectellidge Schmidt mentions, besides Eiqylectella jovis, which is a species
nearly related to JEiq^lectella suberea, Wyv. Thomson, a new genus Regadrella, including
the single species Regadrella phoenix, which is devoid of the siliceous tuft and is
fixed to the rocky substratum by a firm base. Though in this form a very constant
character of the Lyssacina is evidently absent, it must be noted that other forms occur
which are quite firm and compact beneath, lattice-like in the middle, and loosely
constituted above, such as, for example, Hertwigia falciformis, Schmidt, and Rhahdo-
pectella tintinnus, Schmidt, and these were direct transitional forms between Dictyonina
and Lyssacina.

Of the genus Hyalonema Schmidt has described only fragmentary pieces. The form
described by him as Asconema hentii is certainly a species of Hyalonema.

The then little known structure of the soft body of the Hexactinellida was, in 1880,
elucidated by my ^ investigation of some well-preserved specimens of Eiiplectella aspei'-
gillum, which were given to me by Sir Wyville Thomson from the treasures of the
Challenger expedition.

In a subsequent study of the Hexactinellid material from the Gulf of Mexico, which
had been already utilised by 0. Schmidt, but which was re-examined by Weltner^ in
1882, the structure of the peculiar skin and covering layers was especially considered.
This had, it is true, been previously described by Zittel in several fossil forms, but in
living forms it had hitherto been insufiiciently known, and had been investigated, in
fact, only in a few cases, such as in Aphrocallistes by Zittel, in Myliusia by Marshall, and
in Dactylocalyx pumiceus by SoUas. Weltner was able to demonstrate the presence of
these structures in numerous modern Dictyonina, e.g., in Farrea, Syringidium, Aphro-
callistes, Volimlina, Joannella, Margaritella, Scleroplegma, and Cystispongia. Like
Zittel he distinguished " covering layers, dependent and independent of the lattice-

1 Trans. Roy. Soc. Edin., vol. xxix., 2, p. 661. " Weltner, Beitriige zur Kenntniss des Spongien, 1882.

REPOET ON THE HEXACTINELLIDA. 17

skeleton." The former are flat thickenings of the lattice-skeleton on those regions where
the sponge was in contact with, or lay against some solid body ; while the " independent
covering layers " surround the external and also the gastral surface, in the form of a
skin-like web of fine free or united spicules.

The remarkable rigid balls, first observed by Saville Kent in his Fieldingia lagettoides,
and later byO. Schmidt in Cystispongia superstes, which seemed to be formed of condensed
portions of the lattice-like network, were observed by Weltuer in several other Dictyonina,
such as Scleroplegma kmterna, Myliusia zittelii, and Margaritdla ccelopti/chioides.

From the preliminary reports on the Hexactinellida collected oft" the coasts of France
and West Africa by the French deep-sea expedition of the " Travailleur " and
" Talisman," it appears that not only were most of the species collected by the
Challenger discovered in the localities in question, but some new sjDccies were added.

In a popular account of the expeditions of the " Travailleur " and " Talisman," La vie
au fond des mers, 1885, Filhol notes some of the Hexactinellid forms, e.g., " Euplectella
suberea, Wyv. Thomson, cj[ui est largement repandue dans I'Atlantique nord. Pendant
la croisi^re du Talisman nous les avons dragudes &, diverses reprises par des fonds variant
entre 900 et 2300 metres. En certains points elles ^taient d'une extreme abondance et
devaient couvrir d'assez vastes espaces."

A new form is reported {loc. cit, p. 284), and figured on pi. viii., viz., " Trichaptella
elegans, H. Filhol, fixe sur des Coraux (Lophohelia) sur les c6tes du Maroc par 865
metres de profondeur. Sa base est formee de spicules siliceux agglutinees les uns avec
les autres et formant ainsi un r^seau d'une grande solidite. Le restant du corps de
I'eponge, qui s'dlargit dans sa partie moyenne, est souple comme chez les Euplectelles.
L'oscule, ferme par un treillage a mailles grandes et irregulieres est entoure par une
collerette de longs spicules d'une extreme delicatesse."

" Les Pheronema paraissent etre repandus dans tout I'Atlantique, dont elles habitent
en certains points de tres grandes profondeurs. Communes sur la c6te du Portugal, elles
apparaissent encore plus nombreuses au large des cotes du Maroc et du Senegal, en pro-
fondeur de 600 metres jusqua 2200 metres."

" Certaines d'entre elles sont remarquables par un enorme developpement, alors que
d'autres, telles que Pheronema Parfaiti se font remarquer par leur transparence et
I'absence de collerette de spicules autour de l'oscule."

" L'Asconema setubalense n'avait 6t6 trouve, jusqu'au voyage du Talisman, que sur
les c6tes du Portugal. Lors de la campagne de ce dernier bateau, nous I'avons
reeueiUi sur les c6tes du Maroc, au voisinage du Cap Bojador, par 410 metres."

" Les Aphrocallistes sont communes sur les c6tes du Portugal, du Maroc, du Senegal
et s'etendant aux parages des iles du Cap Vert, des Canaries, des Azores. Les Aphro-
callistes Bocagei constituent, au fond de certaines portions de I'Atlantique nord, des

('ZOOL. CHALL. EXP. PART LIII.— 1886.') Ggg 3

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

colonies quelquefois assez distantes les unes des autres. Nous les avons obsei'vdes a,
des profoudeurs bien definies et cela h. partir de 860 metres jusquh, 2200 metres."

" Les Aseonema ont et^ trouvees avec les Aphrocallistes ; les Hyalonema et les
Euplectella etaient presque tousjours associ^es."

A Catalogue of the Fossil Sponges in the Geological Department of the British
Museum, which appeared in 1883, contains a thorough account, by G. J. Hindc, of the
fossil Sponges in the above museum. The Hexactinellida are thoroughly discussed, and
the whole work closely adheres to Zittel's system. In my general survey I. may
subsequently refer to the fossil Hexactinellids, and shall then rely on Zittel's pioneer
work and on Hinde's excellent memoir.

In 1884' Zittel demonstrated that the family of Astylospongidas (with the genera
Astylospongia, F. Rcemer, Palieomanon, F. Eoemer, ProtachiUeum, Zittel, and Eospongia,
Billings) belonged not to the Hexactinellida, as had been hitherto supposed, but to the
Lithistida, and, in fact, to the group Anomocladina.

In a collection of marine Sponges gathered in Japan by Dr. T. Anderson, Mr. H.
T. Carter ^ notes the occurrence of four Japanese Hexactinellids, namely, Hyalonema
sieboldii. Gray, Farrea occa, Bowerbank, Periphragella elisse, Marshall, and Hexac-
tinella ventilabrum, new species. Carter. Of these accurate descriptions are given,
based partly on dried specimens. Especially important in this report, as it appears to
me, is a clear and comprehensive description of the skeletal parts of Farrea occa,
Bowerbank, accompanied with excellent figures. A definite and reliable conception of
this, hitherto somewhat indistinct, species has thus been established.

' Jahrb. f. Miii-, Bd. ii. p. To, 1884. ■ Ann. and Mag. Nat. Hist, 1885, ser. 5, vol. xv. p. 387.

GENERAL REMARKS ON THE FORM AND STRUCTURE OF THE

HEXACTINELLIDA.

Although the different species of the Hexaetiuellida vary greatly in form and
structure, they nevertheless exhibit so essential an agreement in the fundamental
features of their organisation, that it is easy to reduce all known forms to a common
type. Apart altogether, for the present, from the skeletal portions, this common type
resembles a simple sac, in which the outer surface is formed of a thin skin rich in pores
(the dermal membrane), through which water enters into a space (the subdermal
trdbecidar space) crossed by fine trabeculse. Within this there is a single layer of
closely arranged chambers shaped like the finger of a glove, with wide internal openings,
and with fine walls (the memhrana reticidaris), supported by an elegant quadrate net-
work, and penetrated by small round pores {chamber pores) for the passage of water. A
connecting membrane extends between the internal openings of the chambers, and is in
direct continuation with the walls of the chambers. This membrane generally agrees in
sti'ucture with that bounding the chambers, and closes internally the clefts and apertures
which occur between the latter. Through the pores of the chamber walls and the
connecting membrane, the water reaches an inner or siibgastral trabecular space,
which is likewdse crossed by fine trabeculse, and is separated from the wide gastral cavity
by a porous or net-hke internal limiting skin, or gastral membrane.

From the exterior inwards the following successive layers occur : — (1) the outer hmit-
ing skin or dermal membrane ; (2) the outer or subdermal trabecular framework ; (3) the
cAa?n6ers with their connecting membrane ; (4) the inner or subgastral framework ; (5)
the inner uniting or gastral membrane.

The body of every Ilexactinellid is invariably made up of these five layers in the
same order, but the designation of the layers as outer and inner refers not so much to
their relation to the centre of the individual sponge, as to the direction of the stream of
water. Thus the dermal membrane always refers to that through which the water enters
the sponge body, and the gastral membrane, on the other hand, to that through wliidi it
escapes from the body-wall into the gastral space, or directly to the exterior.

In the attempt to trace the manifold modifications of the numerous species of

20

THE VOYAGE OF H.M.S. CHALLENGER.

HexactincUida from tlie fundamental typo, I will begin with those forms which, like
Bathydonts fimhriatus, are only slightly removed from the original saccular form, and
which by a simple elongation have assumed a tube-like outline. In a section through
the lamelliform smooth wall of Bathydorus Jimbriatus, such as is somewhat diagram-
matically represented in PI. LVIII. fig. 2, the only essential modification of the funda-
mental ty|ie is a peculiar folding of the chamber layer. This layer exhibits a series of
closely disposed broad protrusions of approximately similar form and ec^ual size, which

Fig. 1. — Section of the wall of Walteriafleininingii, n. geu. et sp.

raise the smooth external skin in such a way that diverticula, traversed and divided by
trabeculse, extend inwards from the subdermal lacunse of the outer trabecular space.
The lumina of the protrusions which open by a wide round aperture into the inner
trabecular space are at first destitute of a trabecular framework.

In all the numerous Lyssacina, which, like Acanthascus (Pis. LVI.-LVIL), Rossella
(PI. LV.), &c., resemble a thick- walled beaker or cup, the folding of the chamber layer is
continued by the formation of successive protrusions, so that branched eflferent passages
of roundish section are formed, between which corresponding complex incurrent passages
penetrate inwards from the outer trabecular spaces. This development of a more or less
richly branched system of afferent and efferent passages, which are, however, completely
separated by the chamber layer, remains essentially unchanged, even with such

REPORT ON THE HEXACTINELLIDA. 21

oomplicated folding of the clinmber layer as occurs in the very thick-walled species
Pheronema (Pis. XLII.-XLVI.) and Poliopogon (Pis. XLVII.-L.). The efferent canals do
not, however, by any means, always remain free from the trabecular framework which is
so abundantly developed in the afferent spaces. This is very frequently illustrated, as,
for example, in Malacosaccus (Pis. XVIIL, XIX.), Crateromorpha (Pis. LXI.-LXIII.),
Hyalonema (Pis. XXVII.-XLJ.), Pohjloplms (PI. LIV.), Rossella (PL LV.), and others,
where a reticular lining penetrates from the subgastral trabecular space into the outer-
most diverticulum of the efferent canal system, without, however, at any time crossing
the lumen of the efferent passages, or entering the cavity of the chambers. The gastral
membrane extends smoothly over all the (excurrent) openings of the efferent canals,
in the form of a sieve-like net, and thus forms the boundary of a simple gastral space,
as in Rossella (PI. LV.), Pheronema (Pis. XLII.-XLVI.), &c. ; or it lines the niche-
like depressions of the wall which may be of simple or complex form, as in Mala-
cosaccus (Pis. XVIIL, XIX.); or finally it passes through the wide excurrent openings
of greatly branched efferent canals which open directly into the gastral space, and
extends within them on to the terminal branches, as in Hyalonema (Stylocalyx)
depressum (PI. XXXVI. fig. 1) and others.

Further variations of the simple saccular form are exhibited by many Hexactinellida
in the formation of a terminal sieve plate covering the wide opening at the extremity
{Euplectella, Pis. I., V., Holascus, Pis. XV.-XVIL, Hyalonema sieboldii, PI. XXVIL,
and others). In many cases, too, the body -wall exhibits a more or less regulai-ly arranged
set of gaps, by means of which a direct communication is established between the gastral
cavity and the external medium. While these gaps in the walls of Euplectella
(PI. I.) and Taegeria (PL VII.) are circular, and bounded by an iris-like membrane'
which is capable of contraction, in Walteria (PL IX.) they occur as irregularly
angular meshes, like the lattice-work of a basket. Striking modifications may arise
by the formation of a stalk, which is, indeed, always in the form of a simple continua-
tion of the lower portion of the body-wall, from which, however, it often differs
widely in diameter. It varies greatly in the length attained, and is more or less
sharply truncated. The stalk is generally quite round and smooth, but frequently with
characteristic curvature [Gaulophacus elegans, PL XXV.), or prominent protuberances
[Crateromorpha murrayi, PL LXIIL), and is sometimes even branched {Symjxtgella mix,
PL XXII. fig. 4). It is either solid or tubular, generally the latter in long stalks (Canlo-
pthacus), and its lumen either opens into the gastral cavity, or is in communication with,
and so belongs to the efferent canal system. By the expansion of the upper oscular
margin many species, e.g., Rhahdocalyptus mollis (PL LXIV. fig. 1) and others, acquire
a funnel-like shape. A further widening and flattening leads to the formation of a flat
saucer-like body, while a more unilateral growth results in an ear or shell-like form
{Euryplegma auricidaris, PL OIL), or even in certain circumstances in a simple perpen-

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

diciilar plate-like form {Chonelasma, Pis. LXXXVII. -XCL). The wall of the cup may
be complicated by thimble-like sacculations, as in Aphrocallistes hocagei (PI. LXXXIII.
fig. 1), and, in such cases, the gastral cavity may be separated by several transverse
net-like diaphragms into a series of partitions. If the outer margin of a stalked,
originally cup-shaped sponge, becomes folded outwards and downwards through great
development of the median portion, a fungoid form arises which, in the genus Caulo-
phacus (Pis. XXIV.-XXVL), exhibits several varieties of outline. In this way then, as
the gastral cavity and osculum have thus been lost, what was originally the internal
gastral has become the upper and outer surface, so that the water enters the body
from below, and escapes again from the upper outer wall. In Aulochone (Pis. LXVL,
LXVIIL) the originally upper portion of the gastral membrane has, on account of the
folding of the oscular waU, been turned towards the outside, and thus forms the outer
wa.ll of the cylindrical or approximately hemispherical body, while the under portion
of the gastral cavity along with the lumen of the tube-like stalk connected with it, has
remained unchanged.

In many Dictyonina the elongation of the sac-like body, without any marked
thickening of the waU, results in the formation of more or less thin-walled tubes
in which the lumen remains approximately the same. These tubes often branch
in a tree-like fashion, frequently dichotomously [Aphrocallistes ramosus, PI. LXXXVI.
fig. 1), while the multiplication and union of branches may form an anastomosing network
of tubes, from which numerous terminal branches arise, each provided with an osculum.
The latter is the sta,te of the case, e.g., in the genera Fan-ea (Pis. LXXL, LXXII.) and
Eurete (Pis. LXXVII.-LXXIX.). In Farrea the young tube-wall begins on the outer-
most terminal branches as a very thin plate with a simply folded chamber layer, and
the whole wall is gradually somewhat thickened, with the increasing folding of the
chamber layer ; while in Eurete the ends of the tubes are continuous outgrowths of
the entire thickness of the wall.

The main tubes in expanding into a funnel-shape sometimes give ofi" lateral branch-
tubes, which have a tendency to branch and anastomose, as in Peripliragella (PI. LXXX.)
and Aulocalyx (PI. LX.). In some species, which consist, for the most part, of a net-
like system of anastomosing tubes, with terminal and lateral oscular openings, a special
covering layer may occur which envelops the whole body, and which, as an independent
plate, not only spreads out laterally from the oscular walls as a fine porous skin or
net-like sieve for the inflowing water, but also extends over the oscula as a sieve plate,
with wider apertures. This remarkable structure, which may be termed a cover, is
seen in Aulocystis (PL CIV.), and also, though in quite diiferent form, in Semperella
(Pis. LL, LIL), where the oscula appear, not so much as round apertures, but rather as
irregular longitudinal clefts on the sides of the body. The covers of these genera difi"er
also in this : in Aulocystis the cover appears as a direct continuation of the entire wall

REPORT ON THE HEXACTINELLIDA. 23

of the tube (though without the chambers) ; in Semperella, on the other hand, the cover
consists, in that portion which is spread over the afferent passages, simply of the uplifted
external skin, while the sieve plates extending over the oscular clefts appear as a
chamberless continuation of the entire wall of the tube.

HISTOLOGY.

Thoiigh the material placed at my disposal for investigation was not very well suited
for histological research, some facts of interest were established. It ought to be observed,
in the first place, that the histological structure is so uniform throughout the entire
group, that the modifications to be noted are hardly of au important character.

As I pointed out in my communication On the Structure and Arrangement of the Soft
Parts in Euplectella aspergillum} the Hexactinellida exhibit, like all other Sponges, three
histological layers, viz., two distinct layers of epithelium, and an intermediate connective
tissue with various substances enclosed within it. As to the delicate single layer of
flat epithelum which covers the entire surface exposed to the water, I have not been able
to detect the contours of the cells, but simply their characteristic, small, clear, spherical
nuclei. These, and the small, shining, granule-like nucleolus, were distinctly recognis-
able on well-preserved portions of the surface when strongly stained with hsematoxylin
or picrocarmine. The nuclei are distributed with tolerable uniformity, and project a
little above the general surface of the cell, as is indisputably demonstrated on profile
view. They thus lie in the very outermost portion of the bounding surface. This can
be best seen on the dermal and gastral limiting membranes, but these epithelial nuclei
are also recognisable as slightly projecting elevations, on the netted strands of the sub-
dermal and subgastral trabecular spaces, and also on the trabeculse and membranes
extending between the ciliated chambers (PI. IV. fig. 8).

A peculiar character, which differs not inconsiderably from the known relations of
other Sponge groups, is exhibited by the epithelium which lines the inner surface
of the chambers. One could not, of course, expect that in these epithelial cells,
structures so delicate as the flagellum and collar should be preserved. Although I have
taken every trouble, I have failed to detect these structures which are so constant in the
chamber-cells of all known Sponges. It cannot be doubted, however, that they were
really present. In my variously stained preparations, the chamber-cells appear as small
bodies, projecting for a variable distance into the lumen of the chamber, and provided
with small spherical nuclei and usually with a single shining nucleolus, which is
sometimes of a cylindrical, but usually of a kidney-shaped roundish form. The
remarkably regular disposition of these cells, and their connection Ijy flat lateral
bands, which extend over the membranous chamber- wall, are worth noting. Since

' Trans. Eoy. Soc. Edin., vol. xxix., 1880.

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

these bands are disposed iu rows, intersecting approximately at right angles, at an
average distance of 0'06 mm., and united further by stronger basal strands, a lattice-like
pattern results, in which the meshes are approximately rectangular or slightly rhombic
in form. As to the nature of the somewhat strongly refracting and slightly stained
connecting strands, of which the stronger cross at right angles, while the finer branches
are irregularly reticulate (PI. LXII. fig. 8), I have not been able to come to a perfectly
clear decision. I have little doubt, however, that we have here to do with the
anastomoses of protoplasmic processes from adjacent cells. Each cell seems to be
connected with its four neighbours by one main connecting process and several fine
lateral branches (PL LXII. figs. 7, 8).

The connective substance, which forms the main framework of the soft body, does not
differ essentially from that of most other siliceous Sponges. In the hyaline, somewhat soft
matrix, two kinds of cells can be distinguished, namely, (a) simple stellate or spindle-
shaped connective-tissue cells with little protoplasm, and a small oval nucleus, and (6)
larger cells with clear vesicular nucleus, and with a more or less abundant accumulation
of refracting, intensely stained granules of various sizes. As a rule, these granules lie
loosely beside one another, and may be included in lappet -like processes of the cell
(PI. IV". fig. 8, and PI. LXII. fig. 8). Sometimes, however, they are united in firm
clumps. Occasionally I found them brownish or yellow in colour. As I have previously
noted, I regard these bodies as reserve nutriment, analogous to fat or starch. Small
groups of round cells occasionally occur, but their import is not known (PI. LXII. fig. 8).

All the skeletal structures belong exclusively to the connective-tissue. On the
surface of the spicules there is a more or less distinct thin layer of hyaline matrix, the
so-called spicular sheath, which, on specially thick needles, is seen as a finely fibrous
membrane. In the connective substance, finally, the genital products occur, the sperm
masses and ova, in more or less abundance, and usually in the same individual. The
sperm masses, both in young and mature stages, are exactly like those of other siliceous
Sponges, such as Reniera. In their immature form the ova are indistinguishable from
connective-tissue cells. They subsequently increase in size and develop refi"acting yolk
granules, and exhiint a very characteristic aspect owing to the enlargement of the nucleus.
It is remarkable that in the adult (0"3 mm. in diameter), irregularly roundish ova of
Euplectella asjiergillum, along with which ripe sperm masses also occurred, the nucleus
was situated not in the ovum itself, but lay freely in a superficial depression into which
it had been squeezed. This expulsion of the nucleus was probably the result of the
drastic preservative treatment.

It is curious that I have never been able to discover any distinct segmentation stages. It
would not, however, be justifiable to jump to the conclusion that the ova leave the body of
the Sponge as such, and undergo subsequent development outside the mother organism.

As to the developmental history I am only able to communicate a few observations.

REPORT ON THE HEXACTINELLIDA. 25

Some species undergo multiplication by budding. This process can be studied
extremely well in Polyloplms 'philippinensis, where the same specimen frequently exhibits
numerous stages in the development of buds, from small papilla-like elevations to
adult individuals ready to be separated off (PI. LIV. fig. 1). After the numerous
conical protuberances of the lateral surface, which are apically equipped with a
tuft of slightly divergent needles, have grown out to a greater length, and have
been drawn out into nipple-like structures, they become constricted, and the outer
portion expands into a pear-like form. The bud continues to grow gradually larger,
expanding especially on its free terminal portion ; a roundish opening is formed
at the distal pole, while on the lateral surface a number of conical projecting tufts of
needles appear irregularly disposed. The portion connecting the bud with the mother
Sponge becomes longer, and at the same time thinner ; finally the union between the
two is wholly severed, and the bud thus separated is attached only by a weak bundle of
siliceous needles. Strain and friction soon effect complete separation, and the bud at
length falls to the ground, where it becomes independently attached by its own tuft of
spicules (PL LIV. fig. 1).

When, in such a process of gemmation, the undeveloped individuals remain attached
by an elongated stalk, ramified colonies with terminal individuals result, as for example
in Sympagella nux (PI. XXII. fig. 4).

These young buds are in their young stages admirably suited for investigation,
affording a convenient view of the whole organisation of the form in question. This
may be illustrated by glancing at fig. 2 on PL LIIL, where a longitudinal section through
Poli/lophus philippinensis is represented. The relation of the entire chamber layer to
the general canal system and water stream becomes at once distinct. It may be clearly
seen that, however much the layer of chambers is folded, it always forms a continuous
intermediate layer between the afferent and efferent lacunae or canals, forms in fact a
filtering layer, through the pores of which the stream of water has to pass.

The above gemmation which results from proliferations of the sponge- waU, must be
of course distinguished from a formation of colonies very frequent among Dictyonina,
which occurs, however, as the result of the folding of the exuberant margin of a cup or
tube-like rudiment. The dichotomous or more complex tubes, which frequently exhibit
lateral branches, in FaiTea, Eurete, Periphgragella, Myliusia, Sec, result wholly from
the folds of the growing portion becoming gradually closed into complete tubes
(PL LXXII. fig. 3 ; PL CIII. figs. 1, 2) as may be readily demonstrated from a careful
comparison of the different stages in the development of the colony.

In some cases, as for example in the young specimens of Lanuginella pupa (PL LIII.
figs. 4, 5) which were found isolated, I was of the opinion that I had before me very
young forms developed from the ovum. I inferred that chiefly from the fact that both
the youngest, simply spherical specimen (2 "3 mm. in diameter) (PL LIII. fig. 4), and a

(ZOOL. CHALL. EXP. — PABT LIII. 1886.) Ggg 4

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

somewhat larger oval form, were surrounded by a perfectly continuous skin, and exhibited
no trace of a connecting stalk. The chamber layer in the former case was still simjjly a
closed sphere, while in the second a rupture had taken place towards the pointed pole of
the oval body, at the point apparently where the osculum would be developed. A
beautiful post-embryonal series, with certain modifications of form, was observed by
Wj'ville Thomson* in Pheronema [Holtenia) carjxnteri.

Generally speaking, however, no important modification in the form seems to occur
during the development. In those Lyssacina in which the spicides are never soldered
together, the growth may apparently continue until the death of the animal ; so that
giant forms of 50 cm. in diameter and more may arise, as Poliopogon gigas, Mala-
cosaccus vastus, &c. In many Lyssacina, however, which, as they grow, typically
exhibit a soldering of the principal needles, there appears to be a definite limit of
growth (e.g. in Euplectella aspergillum). It is readily intelligible that when the
soldering of the needles has progressed from the median portion of the tube to
the basal tuft on the one hand, and to the firm terminal sieve-plate on the other,
any further extension of the sponge-body is really impossible. It is difierent with the
Dictyonina, in which the principal needles (Dictyonalia) are immediately after their
formation united with one another into a connected framework. Here, however, both on
the free margin of the cup- or tube-like body, and on the whole dermal and gastral
surface, there is a persistent continuance of growth through the laying down of fresh
portions of the framework and simultaneous displacement of the loose dermal and gastral
skeleton. It is only when the dermal skeleton becomes itself rigid, through the union of
its spicules in a reticulate framework, or by a continuous siliceous membrane, that
an absolute check is placed upon any further lateral growth. This seems also to occur in
individual cases, e.g., in Fieldingia, and with apparent regularity in several fossil forms.

In many species the whole body dies and falls to pieces at once, while in others
dissolution is more gradual and begins at the base. Thus many specimens, especially
of Dictyonina, and also many Lyssacina, are dead in their basal portion, which consists
sim2>ly of a perfectly macerated skeleton, representing either the loose fibrous tuft, as in
PoJiopogon amadou, or a dead portion of the dictyonal skeleton. The latter case is
illustrated in PI. GIL fig. 1, in reference to Euryplegma aurictdaris, where the lower
boundary of the persistent, somewhat darker soft body is sharply marked ofi". On a large
beautifully developed specimen (40 cm. in height) of Aphrocallistes vastus, which Dr.
Gotsche brought from Japan, the whole basal portion for about a hand's length has been
killed. The siliceous elements rooted in the mud are not, however, always wholly dead.
The long tuft needles of the Hyalonematidge and the anchor needles of the Euplectellid* un-
doubtedl}^ retain their life, being probably nourished through the strand of their axial canals.

In all skeletal elements which have been exposed, for a lengthened period after

' Phil. Trans., vol. clix. p. 70.

REPORT ON THE HEXACTINELLIDA. 27

death, to the influence of sea-water, one always notices a more or less marked widening
of the axial canal. This is due to the fact that the central siliceous layers are slightly
more soluble than the outer. The time that has elapsed since death can thus be
approximately estimated, and the differences have, of course, no systematic importance
as was formerly supposed.

THE SKELETON.

Under the title skeleton I include all the solid parts of the Sponge, whether these
are bound together into a united framework, or lie isolated in the soft tissues. I
would avoid the expressions, "skeletal" and "flesh" spicules, which are much used by
Carter, Zittcl, and others to distinguish the hard parts into two leading categories ;
since, on the one hand, these expressions are ambiguously used by diff'erent authors,
and even by one and the same author, and, on the other hand, since they in no way
indicate any essential distinction. By "spicules of the skeleton" Carter originally under-
stood "large spicules, which are only concerned in the formation of the supporting
structure or skeleton," but he subsequently applied the term only to the larger free
spicules, in contrast to the smaller free spicules or "flesh spicules," and to the vitreous
fibre of the continuous lattice framework. Zittel used the phrase " skeletal spicules,"
for the most part in reference to the skeletal parts concerned in the formation of a
united lattice-like framework, but he occasionally applies it also to the larger free
spicules which form the chief support of the soft parts.

Most spicules present a structure similar to. those which Max Schultze and Claus
first discovered in the larger spicules of Hyalonema sieboldii and Euplectella asp)er-
gillum. A fine central canal, corresponding to the typical axes of the skeletal
body, is surrounded by numerous concentrically arranged layers of a solid substance.
In its appearance and other physical jaroperties the latter resembles glass so closely
that it has been often, without the least hesitation, spoken of as vitreous fibre. For the
same reason the entire Hexactinellid group has been designated " vitreous sponges," or
more shortly "Vitrea." As to the nature of the. soft, finely granular mass composing
the axial thread or cord which fills the central canal, but little is as yet known.
During the growth of the spicule it appears to be connected with the surrounding soft
parts through an opening which is present at the end of every ray. After the ray has
ceased to grow in length, the terminal opening is closed by an expansion of the layers
of the glassy substance. That all spicules have a central canal is very probable, but not
quite certain, since it has not been possible to recognise it in all cases. Especially in
very thin and small spicules it is often very difticult or impossible to detect it. I have
never found it in lateral prickles, prongs, scales, and the like, nor in the secondary
terminal rays of the rosettes and scopulas, nor in the short transverse connecting beams.

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

the so-called " synajsticula," wliich stretch so frequently between neighbouring spicules
of many Lyssacina, binding them into a solid framework, nor finally in those remark-
able lattice networks which occur in many Hexactinellida on regions in contact with
foreign bodies, but especially where the Sponge has grown on a solid substratum.
In individual cases, as, e.g., in the thickened extremities of many anchor spicules,
the central canal exhibits a brush-like division into several diverging, blind, terminal
branches (PI. III. fig. 29 ; PL XIV. fig. 5).

The innermost layer immediately surrounding the central canal is called by Claus
the " axis cylinder." It is generally distinguishable from the usually many layered
outer cortex by its somewhat feebler refractive power and by the absence of lamination.
From the behaviour of the spicules when heated, and when examined in polarized light,
Max Schultze determined that the individual lameUse were separated from each other
by thin layers of an organic substance. After cautious heating, fine brown carbon
streaks were seen between the adjacent hyaline layers. By the use of the polarising
apparatus double refraction could be demonstrated in the thin intermediate layers, but
not in the substance of the lamellae.

Professor Maly of Graz was kind enough to analyse a number of spicules
from the root tufts of a Poliopogon amadou, and has summed up the result in the
following note, with which he has favoured me: — "The spicules after being placed in a
desiccator, and dried at 105°, still contained 7 "16 per cent, of water in chemical union.
They are, therefore, not silicic acid in the mineralogical sense (quartz substance), but a
hydrated silicic acid, and therefore resemble opal, in which the amount of water very
frequently varies from 6 to 8 per cent."

In no other group of Sponges is there so great a variety in the form of the skeletal
elements as in the Hexactinellida, yet nowhere are the numerous individual forms of the
spicules so readily referable to a common fundamental type.

As was first recognised by Wyville Thomson, and subsequently abundantly confirmed
by Oscar Schmidt, Marshall, and others, a system of three equal axes intersecting at
right angles is the fundamental structure of the skeletal parts in all Hexactinellida, and
that both in regard to the spicules united into a continuous framework, and those which
lie isolated. Though the disposition of the axes is thus in fundamental agreement with
that of the regular crystallographic system, the structure of the spicules is by no means
the result of crystallisation. The silicic acid occurs here in an absolutely amorphous
condition like that of the opals with which they are, chemically, so closely related.
Only in a few fossil specimens does the existence of double refraction indicate a secondary
transformation into the crystallised state.

The axes of the individual skeletal elements are usually quite distinctly indicated by
the disposition of the principal rays in relation to the point of intersection, but frequently
they can only be demonstrated by examination of the central canals.

REPORT ON THE HEXACTINELLIDA. 29

The fimdamental form of all the spicules is the simple, regular, six-rayed form. The
manifold deviatious from this form may be understood as the results of the following
modifications : — (1) the unequal development of the rays, which may lead to the complete
suppression of one or more, so that in extreme cases only a single ray attains full
development ; (2) division of the rays into terminal branches, varying in number, form,
and direction ; (3) the development of local thickenings or unilateral swellings in the
form of knobs, thorns, prickles, knots, and similar ornamental protuberances ; (4) the
curvature of the chief rays or their branches.

The isolated spicules which sometimes occur, in which more than six principal rays
seem to run out from a nodular point, may be usually referred to a very deep
division of one or several of the principal rays, as the result of which the secondary or
terminal rays have been closely approximated to the point of intersection, and thus
simulate the principals (PI. XCVII. fig. 3).

In many cases in the continuous frameworks, more than six beams are seen running
out from a nodular point, but of these, as is well known, only six belong to one spicule,
while the others belong to neighbouring hexacts, and have become fused to the former
spicule at the node of intersection.

In our survey of the different forms of spicules which occur in the HexactineUida,
it will be convenient to divide them into six main groups, according to the number of
perfectly developed rays. These six forms may be termed Hexacts, Pentacts, Tetracts,
Triacts, Diacts, and Monacts, and they will be treated consecutively in that order.

Hexacts.

Regular Hexacts are aU spicules in which the rays lie at right angles to one another,
and are of equal length and similar form. One of the most simple and at the same time
most frequent of the regular hexacts possesses straight, perfectly round and smooth rays,
in which the diameter becomes uniformly less from the point of intersection to the
extreme tips (PL III. fig. 15; PI. XXVII. fig. 20). Even in these simple forms,
however, there are manifold difi"erences in the size of the radii. Besides the gradual
running to a point, the rays frequently exhibit an irregular decrease of the diameter
towards the outer extremity. The latter may thus be conicaUj^ sharpened, rounded in
various ways, or transversely truncated. The ray may also exhibit a terminal swelling
of varying form, a sharply truncated terminal knob (PL XCV. figs. 3, 4), or a
hemispherically arched, transverse, terminal disc provided with marginal prongs (PL XL
fig. 3; PL LV. fig. 8). In the latter cases there is no manifest decrease in the diameter
of the ray from within outwards.

The rays may be roughened by small elevations thickly studded throughout their
whole length (PL LV. fig. 5), or in particular regions (PL LII. fig. 5). They are often

30 THE YOYAGE OF H.M.S. CHALLENGER.

also beset with thorns and prongs, varying in form, size, and direction (PI. LVIII. fig. 6 ;
PI. XL fig. 2 ; Ph XXVII. fig. 13 ; PI. XVI. fig. 8).

In certain cases, as Marshall has shown, individual thorns become greatly developed,
are directed obliquely inwards, and unite with the opposite thorns of other ra)"s on the
same hexact. This results in the remarkable formation of octahedral margins round
the intersections of the lattice-like framework in many fossil, and also in some living
Hexactinellida (Pi. CIV. fig. 3).

The rays of most regular hexacts are straight, but curved forms often occur. The
curvature may be quite irregular and undulating, but it has in most cases a definite form
and direction, varying in diflerent species of Sponges, and in many instances
characteristic. The curved rays may be smooth or rough, and are sometimes also beset
with prickles (PL XXVII. fig. 10). When the rays run out to a pointed extremity, the
name " oxyhexact" may be applied. When a knob or disc-like thickening is formed at
the end of each ray the term " discohexact " may be conveniently used.

In many regular hexacts secondary rays appear, either as diverging and similar
terminal branches on the princij)al rays, or arranged in a circle or a disc-like transverse
expansion of the latter, or finally they may be uniformly distributed close to one
another. To such hexacts, provided with secondary or terminal rays. Carter has applied
the term " rosette." The number of terminal rays on each principal is generally constant
in one and the same kind of rosette, though it may occasionally vary in different rays,
even on the same rosette, and on different rosettes otherwise similar. Variations in
the form, direction, and dimensions of the terminal rays result in numerous kinds of
rosettes.

The length of the terminal rays is generally in inverse ratio to that of the basal
principals, which are, as a rule, cylindrical and smooth, and less frequently rough or
beset with prongs (PL LXXXVIII. figs. 8, 9 ; PL XCI. fig. 7). The terminal rays are
sometimes smooth, sometimes rough (PL XXIV. fig. 6), or covered with prongs (PL XXV.
fig. 6), sometimes straight, sometimes bent in different ways, but usually in such a way
that the planes of curvature in all the terminal rays are directed radially towards the
axis of the principal. If the curvature is simple, either its concavity or its convexity
may be directed towards the principal axis ; if it is S-like the terminal ray bulges in its
proximal or inner portion towards the exterior, while the distal outer portion has its
convexity turned towards the axis of the principal, and bends outwards like the petal of
a lily. The terminal rays are seldom quite cylindrical, in many cases their diameter
decreases towards the free ends, in other cases the converse is observed. The very end
may be pointed, truncated, rounded ofi", or terminated by a sharply truncated terminal
expansion of a knob-, cylinder-, disc-, or bell-like form.

As to the abundant rosettes, it is convenient to distinguish in the first place those
forms in which the principal rays are wholly or partly divided into two or more uniformly

REPORT ON THE HEXACTINELLIDA. 31

pointed, outwardly directed, and divergent terminals. Such forms I have designated
" o.ryhexasters." All the six principal rays are usually divided into an equal numljcr
(two to five or more) of terminals, but the number of the latter may vary in tlie dilierent
principals even in one and the same oxyhexaster ; and it may even happen that indivi-
dual principals remain undivided. In the extreme case (PI. LVI. fig. 8), only one of the
six principals is forked, while the other five run out to simple points. It is noteworthy
that, in a division of a principal ray, the divisional planes of the two principals which are
directly opposite to one another, and therefore belong to one axis are mutually disposed
at right angles (PI. LVI. fig. 7). The terminal rays of the oxyhexaster are usually
straight (PI. III. fig. 1 ; PL XVII. fig. 8), but slight curvature frequently occurs. In
such cases the rays are either simply convex internally or externally (PI, XXI. fig. 6 ;
PI. XIII. fig. 6), or they are S-shaped (PI. LXII. fig. 5). The terminal rays may be
sometimes quite irregular and wavy, or else hook-like (PI. XIV. fig. 13; PI. XV. fig. 9),
or even sharply bent (PI. XXVI. fig. 7).

Striking forms which occur in many Euplectellidse and here and there in Cratero-
morpha, may be termed brush-rosettes, " graphiohexasters." The principal rays are much
broadened and bear a bundle of long, straight, thin, terminal rays in parallel or slightly
diverging disposition (PL XV. fig. 19; PL XII. fig. 5). A closely related form,
distinguished, however, by the slightly waved curvature of the delicate terminals, is
represented in PL CIV. fig. 4. In many rosettes with numerous S-like terminal rays,
disposed in concentric circles, there is a certain resemblance between the tuft of rays,
and a down-feather or pluma. I have, therefore, called these forms " plumicomes."
Their individual terminal rays may run to a point at the outer extremity, or becoming
gradually thickened towards the curved ends, be rounded ofi" terminally (PL LIV.
figs. 4, 6).

A perfectly uniform increase in the thickness of the straight terminal rays, on to the
broad, rounded, free extremity, is exhibited by an unusually large form of rosette, which
is also further characterised by a fringe of strongly bent booklets (PL LX. fig. 3).
Another form, represented in fig. 2 on PL LXX. , is characterised by the sharply truncated
cylindrical thickening of the outer part of the straight terminal rays, which thus exhibit
a certain resemblance to the spikes of a Typha.

In the rosettes with rounded terminal knobs, " sphasrohexasters," the terminal rays
occur in varied form. They may be quite straight (PL CI. fig. 7), simply curved, S-shapcd
(PL XCT. fig. 7), or finally irregularly curved in a wave-like fashion (Pi. LXXVIII.
fig. 12). They are of equal thickness throughout their whole length.

Where transverse terminal discs are developed in the discohexasters, the stalks
seldom retain a cylindrical form (PL XII. fig. 4), but are as a rule thickened either
externally or internally (PL XII. fig. 8 ; PL XIII. fig. 3). The terminal discs are fixed
transversely on the corresponding terminal ray by their centre, or they may form an over-

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

hanging expansion of the curved outer end of the ray. In the former case they are
radially symmetrical, in the latter bilaterally symmetrical, with a plane of symmetry,
indicated by the S-shaped curvature of the terminal ray.

Although the radial terminal discs vary greatly in form, they all agree in this that
the outer terminal surface is convexly arched or conical, while the lateral margin is
provided with small teeth or longer protuberances. If the marginal teeth are large in
relation to the central body of the disc a many toothed anchor form results (PL XXV.
fig. 6). If the middle portion of the anchor, on the other hand, is more strongly
developed, the result is a hemispherical form with a toothed margin, or a campanulate form
with long marginal protuberances (PI. XII. fig. 4). The number and form of the marginal
teeth varies in different rosettes. It is frequently four, but in other cases six, eight,
twelve, or indefinitely more.

To all rosettes in which the secondary rays bear on their ends a transversely directed,
radially symmetrical terminal disc, I would apply the term " discohexasters." Those
forms, however, in which S-shaped terminal rays bear on their ends a bilaterally
symmetrical disc-like thickening, I would call "Jloricomes " — a term used in reference to
the resemblance of the bundle of rays to a flower perianth, and first employed by
Bowerbank in the description of Euj^flecteUa aspergillum. In these floricomes the
formation of small marginal teeth is usually confined to the terminal portion of the disc-
like thickening that projects or overhangs externally (PL III. fig. 11; PL XIX. fig. 5),
so that a hand or claw-like form arises. In other instances smaU teeth are formed over
the whole margin of the disc-like thickening (PL XVIII. figs. 6, 7).

With these regular hexacts numerous irregular forms are contrasted. The latter are
characterised by the variable formation of one or more rays ; the distinctions are
associated with difference in length of ray, or with modification of form. To the first
category belong the sword-like hexacts, which are found beneath the skin of many
Eujjlectellidse, such as Euplectella aspergillum, (PL III. fig. 10), Euplectella crassistellata
(PI XII. fig. 4), and others. In these the five outer rays are simple, smooth, round, and
run out to a point ; they are either altogether or almost exactly of equal length, while
the sixth ray, directed inwards, is also smooth, round, and pointed, but is fully twice as
long as the others. Frequently, too, besides the ray directed inwards, the outer radius
on the same axis is longer than the tangentials, so that the resemblance to a cross-hilted
sword may become very obvious, as in Tasgeria pulchra (PL XL fig. 5). This lengthen-
ing of the rays, in one or in two axes, is quite frequent, e.g., in those hexacts with
long thread-like rays, which are found so abundantly in the parenchyma of Malac-
cosaccus vastus. In this form the two rays on the radial axis are usually shorter than
the four tangentials. The inverse relation occurs in the hexacts which Lie in the gastral
membrane of Asconema setuhalense, where the two rays of the radial axis are longer than
the four tangentials (PL XXI. fig. 5).

REPORT ON THE HEXACTINELLIDA. 33

The shortening of one of the six otherwise similar rays is very frequent, and leads
finally to the formation of pentacts.

Much more frequent, however, than mere differences in length, are modifications of
form. A ray may, in its entire configuration, more or less markedly differ from its five
neighbours. This is the case, e.g., in many hexacts with fir-tree-like distal rays, pinuli,
such as occur in the skin oi Aulascus johnstoni (PI. XXH. fig. 3), Caulophacus elegans
(PL XXV. figs. 4, 5), and many other Asconematida^, and also in Aphrocallistes
(PI. LXXXIV. fig. 8 ; PL LXXXV. fig. 4). More frequently, however, the diflferentia-
tion affects two radial rays generally on one and the same axis, as is usually the case in
the hexacts which occur in the skin of the Asconematidse, and are provided with
one prominent, scaly, fir-tree-like, pronged ray (PL XXII. fig. 9 ; PL XXVI. fig. 9).
Similar forms occur also in many EuplecteUidse (PL XII. fig. 3 ; PL XIII. fig. 2).
Apart from the differences already noted, in regard to the rays of many rosettes,
some other hexacts, with secondary rays, exhilait further irregularities. Thus, e.g.,
Aphrocallistes heatrix. Gray, is markedly distinguished by the character of the spicules
which lie separately in the soft parts. In these, two of the principal rays, not however
on the same axis, are elongated and curved, and occasionally divided into four pointed
terminals, while the four other principals remain simple and short (PL LXXXIV.
figs. 9, 10).

Pentacts.

In support of the theory that pentacts have phylogenetically arisen from hexacts by
the atrophy of one ray, the existence of numberless transitional forms, with a more or
less manifest rudiment of the sixth ray, may be adduced ; while it ought to be noted
that the disposition of the five well-developed rays is in thorough agreement with
that of the corresponding rays in hexacts and, finally, that pentacts occur almost
exclusively close to the bounding layer, where the development of one of the two rays
standing at right angles to the limiting surface, is either impeded or specially favoured.
Either the distal or the proximal ray may thus undergo atrophy, with the associated
increase of the other.

The rays may be smooth (PL LVI. fig. 6), or rough (PL LV. fig. 3), cylindrical or
attenuated tovrards the exterior, and pointed, rounded, or even thickened at the
extremity.

The simple case in which all the five rays are of ec^ual form and size is of frequent
occurrence (PL LV. fig. 3), but the unpaii-ed ray usually difiers in some point from the
four others. It may project freely beyond the bounding surface, bearing lateral
prickles which overlap one another like scales. The resemblance to a fir-tree, thus pro-
duced, has earned for the entire spicule the designation pinulus. Such pinuli, in which

(ZOOL. CHALL. EXP. — PART LHI. 1886.) Ggg 5

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

the sixth ray may also appear, occur tjrpically in the two families of the Asconematidse
and Hyalonematidse. Further differences between the unpaired fifth ray and the other
four occasionally occur.

As in hexacts, so here curved rays often occur, and the ray may be curved through-
out its whole length, or only in a particular portion. The curvature is frequently
exhibited only by the four rays which form the cross, and these are usually curved
towards the unpaired straight ray (PI. LXXIV. figs. 1, 2). In many of the pentacts,
which project from the sponge-body, this curvature of the four cruciate rays has
been so effected that they have assumed an anchor form, and have, in fact, the function
of an anchor (PI. III. fig. 23 ; PI. XXXIII. fig. 10).

Peculiar curvatures of a different kind are exhibited by the four cruciate and
tangential rays of many pentact pinuli, which adhere closely to large siliceous beams
of hypodermal spicules (PI. LI I. fig. 6).

The four rays, which lie at right angles to the two radials, form an acute angle with
each other, and are sharply curved just at their origin in those large pentacts of Rossella
antarctica (PI. LV. fig. 9, 13), which have been protruded from the outer skin of the
lateral waU.

In the fork- or broom-like spicules {scopulie), which frequently possess four outwardly
directed teeth in addition to the long stalk, I was, like 0. Schmidt, unable to trace
into the teeth the fine axial canals, though their cross of intersection is often very plainly
visible just below the forking. It seems to me improbable, therefore, that the teeth
can be regarded as. principal rays, and all the more since their number is by no means
always four or five, but sometimes six or more (PI. XCII. figs. 4, 6 ; PI. XCIV. fig. 5).
I should be more inclined to compare them with the terminal rays of the rosettes. The
invariably simple and straight stalk of the scopulse either ends in a point, or is truncated,
or exhibits a knob-like thickening. It is usually smooth, being but rarely provided,
terminally or throughout, with tubercles or transversely directed prongs. The teeth, on
the other hand, exhibit manifold variations in number, length, form, and position. They
usually arise in whorls from, or just above the knot-like swelling on the stalk which
contains the axial cross. They usually vary in number from four to six, but in
individual cases more may be present. Sometimes they diverge but slightly at their
point of origin (PI. LXXXIV. fig. 5), and, in other cases, somewhat widely (PI. XCVIII.
fig. 8). After then- original divergence they may also become parallel to one anothei",
and to the chief axis (PL XCVIII. fig. 9). They are frequently simple in the gastral
part, but exhibit in the dorsal part an S-like curvature (PI. XCII. fig. 6), or are sharply
bent somewhat above their origin, so that the outer portion is considerably divergent
(PL LXXVII. fig. 10 ; PL XCIV. fig. 5). Some are pointed (PL LXXVIII. figs. 3, 5),
and others truncated (PL XCII. figs. 5, 7), but most of them bear a knob-like or even
spherical terminal swelling, which is usually provided with numerous fine backwardly

REPORT ON THE HEXACTINELLIDA. 35

bent prongs (PI. LXXIX. figs. 4, 5 ; PI. LXXXIV. fig. 4). These finely pointed prongs
often occur over the whole branch (PL XCII. fig. 5), or on special portions of the same
(PI. LXXXLV. figs. 3, 5). I am, on the whole, inclined to regard the scopulse not as
pentacts, but rather as diacts or monacts.

Tetracts.

If two of the six rays of a hexact on the same axis are not fully developed, simple
cruciform spicules arise, with four rays in the same plane. Such regular tetracts occur in
the outer, and in the inner (gastral) membrane of many Rossellidse ; as also on the lower
end of the body of Hyalonema, in the quadrate latticework of the main supporting
framework in many Euplectellidse, and in many other species. Rudiments of the two rays
which have not been fully developed can be detected at the nodes of intersection. Such
regular tetracts are either quite smooth, or uniformly rough (PL LVI. fig. 5), or beset
wdth knobs. Many, however, are only rough or knobbed terminally. Some are more or
less uniformly pointed, while others are truncated, rounded, or provided with a knob-like
thickening. The four rays are not always equally long (PL III. fig. 28 ; PL XXXV. fig. 8).
In many tetracts they are wholly or partially curved, and that either in the plane of
the cross (PL III. fig. 27; PL XXXI. fig. 15) or in a spherical form (PL LIV. fig. 7).
A few cases occur, which seem to have arisen, not by the abortion of two opposite
hexact rays, but of two rays at right angles. In such forms only one of the three typical
axes is fully developed, and of the two others only one ray of each has remained
(PL III. fig. 20).

Triacts.

The three rays of a triact usually lie in the same plane, and consist of two rays
belonging to one axis with a third at right angles to these. They are either straight or
slightly bent, quite smooth or terminally rough, besides being frequently swollen and
knobbed. While the rays of the main axis are generally equal, the unpaired third ray is
usually distinctly shorter (PL III. figs. 12, 25), and seldom larger (PL III. fig. 26) than
the others. The insertion of this unpaired ray is often opposite a prong, which doubtless
represents an abortive fourth ray (PL III. figs. 19, 28). If the unpaired ray is very long,
and the two others are bent towards it, a peculiar anchor form results, such as may be
seen among the prominent lateral, and basal-tuft spicules of many Hyalonematidse, such
as Pheronema, PoUopogon, and Semperella (PL XLVIII. fig. 14a).

Rarely it happens that all the rays are markedly curved (PL XVI. figs. 3, 4), and
correspond in position to three of the edges of a cube.

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

DiACTS.

The two rays of a diact belong either to the same or to different axes, and may
resemble or differ from one another. The two rays usually form together a straight or
curved rod. Only very rarely are they disposed at right angles (PI. LIX. fig. 16).

Many diacts show traces of undeveloped rays in the form of knobs (PI. V. fig. 7), or
as a ring-like thickening (PI. XXI. figs. 7, 8, 10), or in the presence of two or four canals
crossing the main axial canal at right angles. In many cases every trace of their deriva-
tion from hexact forms has been lost (PI LXII. fig. 6). Where only two knots occur
on the boundary between the two developed rays, or where two opposite cross canals
occur, we have to deal with the derivation of the diact from a tetract form, in which
the tliird axis of the original hexact has been entirely lost.

Many variations occur in the terminal development of the two rays, which are some-
times simply rounded, sometimes pointed, thickened, or knob-like, and often also pro-
vided with button or umbel-like structures of the most diverse kind. The two rays
may be uniform, or most variably differentiated. They may be smooth or rough, wholly
or partly beset with points, prongs, or spines, varying extremely in number, form,
size, and direction. The direction of the prongs, with which many diacts are beset
aU round and throughout their whole length, usually remains the same from one end
to the other (PL XLV. fig. 6, and PI. LXXI. fig. 4), but the dii'cction may be reversed
on either of the two rays, so that the prongs have their points turned to either end of
the diact (PI. XVI. fig. 12), or, on the other hand, towards the centre (PI. XL. fig. 5).

The obliquely directed prongs occasionally assume a flat form, like certain leaf buds
(PI. XL. fig. 8), or the scales of the fir cone (PI. XXXVI. fig. 7).

A very peculiar and typical diact structure, characteristic of the whole family of the
Hyalouematidte, is found in the so-called " Amphidiscs," in which a terminal expansion of
a disc-like, or spherical form, always curved towards the centre, is borne on the end of •
each ray. The disc is prolonged into several (six to twelve) tooth or shovel-like marginal
protuberances, which in the peculiarities of their length, form, and direction, contribute
essentially to the characteristic features of the different genera and species.

Among the asymmetrical diacts some forms occur, which exhibit a certain similarity to
the terminal umbels of the Amphidiscs. Certain anchor-like forms exhibit curved,
hook-like, more or less flat teeth, which suggest the main rays of a pentact; but the
anchors exhibit, at the same time, the marginal prongs characteristic of the thickened
terminal portion of a diact, which has been pulled out to a great length, and ends in a
point or in a simple knob (PI. III. fig. 29 ; PI. XVI. fig. II). The anchor teeth of such
a diact do not, of course, possess any central canal, and the axial cross of the central
canal is frequently clearly seen at a considerable distance from the whole terminal thicken-
ing of the anchor stalk (PL XVL figs. 11, 13).

REPORT ON THE HEXACTINELLIDA. 37

I liuvc applied the term " Uncinata " to the rod-like forms which run out to a point
at both ends, and are provided with barbs pointed in the same direction. For such forms
Carter^ has lately suggested the name " Barhula," Init this term would not be generally
imderstood. In some species these more or less thickly placed prongs lie close to the
body of the rod, while in others they are obliquely directed. They are sometimes delicate
and narrow, sometimes broad and scale-like.

Certain Hexactinellidan families have typical and regular Uncinata, while in others
they are absent. They are usually directed at right angles to the surface, with their
external points in the skin (PI. LXXIV. fig. 1, PL LXXXIV. fig. 1) ; occasionally they lie
obliquely to the surface or quite irregularly (PI. LXXVIII. fig. 2 ; PL XCV. fig. 2).

Among the asymmetrical diacts a form occurs in which one of the two rays bears
numerous obliquely and outwardly directed teeth or scales, whilst the other remains
smooth (PL XXX. fig. 7), or exhibits only small knobs (PL XL. fig. G).

Feeble, irregular, undulating curvature is exhibited especially on the frequently
numerous tuft-like, long and filiform diacts (PL V. fig. 14). The simple arc-Hke form
occurs, on the other hand, in shorter and stronger diacts (PL III. fig. 21 ; PL LXII.
fig. 6 ; PL LXIII. fig. 4). A. spiral cork-screw-like form is illustrated by the diacts of
Hyalostylus dives, which are rough on one side (PL LXX. figs. 5, 8). More marked
curvature of both rays in the same plane is exhibited by small diacts in Holascus stellatus
(PL XIV. fig. 12). In other cases the rays are bent towards each other in a hook-like
fashion, but on opposite sides (PL XVI. figs. 5-7).

MONACTS.

While the derived nature of a monact spicule is in many cases determinable by the
presence of rudimentary abortive rays (PL LXV. fig. 8), or by the persistent intersection
of the corresponding axial canals at one end of the spicule; this becomes difficult when
neither rudiments of other rays, nor traces of their axial canals persist. The presence of
a knob or disc-like expansion at one end does not of itself determine the monact
character of the spicule in question, since, as we have seen in the anchor-shaped diacts,
the axial cross of the central canal, which is the decisive character, lies at some distance
from the thickened end, and may thus demonstrate the diact character of the simple
spicule.

It seems to me, however, that those spicules, called by Carter " Clavulw," which run
to a point at one end, and bear a knob or terminal toothed umbel at the other
(PL LXXI.-LXXV.) as characteristically seen in the genus Farrea, are really monacts ;
although I have. Like 0. Schmidt, been unable to detect an axial canal in their terminal
umbel, or in the swollen portion l)elow. This opinion is, however, the more probable,

' Ann. and Mag. Nat. Hist., 1885.

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

since 0. Schmidt has discovered and figured such an axial canal cross, just in the middle
of four teeth on the little anchor spicules (of the same genus Farrea), which belong to
the same system as the " ClamdaB."^

Though I agree on this point with 0. Schmidt, I cannot accept his opinion that the
monact umbel and anchor spicules of Farrea are homologous with the terminal rays
of certain discoliexasters to which they have some resemblance. These tenninal rays
I regard simply as prongs without axial canals ; the umbel and anchor spicules, on the
other hand, I regard as true monacts.

Mode of Union of the Spicules.

After this general review of the most important forms of spicules, I pass to describe
their modes of union. In many Hexactinellida there is no distinct union between the
individual spicules, which either lie c|uite isolated in the soft tissue, or exhibit only a
slight connection or mutual support by being closely disposed side by side, or by being
interwoven or entangled with each other (PI. XVII. fig. 6). In other cases, however,
numerous spicules are bound into a firm framework by a laminated siliceous substance.
It may be that parallel or closely opposed rays become surrounded by a common
concentrically layered sheath of siliceous lamellae and so become united into a beam
(PI. LXXVI. fig. 5 ; PI. C. fig. 2), or it may be that the ends of the rays of one spicule
are opposed to the intersection nodes of another, and become fixed as if soldered ; or
further, it may be that the rays of adjoining spicules crossed in any dii-ection are bound
together by lamellae of silex, so laid down that the interspaces are filled by web-like
layers of siliceous substance (PL XX. fig. 3). Smaller hexacts frequently occur in which
the end of one ray is soldered transversely to a larger beam (PI. XXVII. fig. 8).
If there be no immediate contact of the . adjoining spicules, then boss or cone-like
elevations may grow out from the sides of two opposite beams, meet one another, and,
becoming surrounded with laminated siliceous material, form transverse bridges or
Synapticula, — which have not, of course, axial canals (PI. XX. figs. 2-4). Finally, in
certain conditions, both adjoining spicules, and the branches of one and the same spicule,
may become connected by a fine delicate lattice-work, with quadrate or rounded meshes.
The beams of the network arise at right angles to the lateral borders of the spicular
rays, and always lie in the same plane, while they are further united by transverse
anastomoses. Here again there is no axial canal (PI. LXIV. fig. 3). When this lattice-
work is greatly developed, the gaps may be quite filled up, so that finally a thin siliceous
plate results. It is remarkable that such lattice-work and plates are found only in
bounding surfaces which come into contact with solid bodies, especially where the Sponge
has grown on a solid substratum, Ijut also round about foreign bodies which have

1 Spongien des Meerbusens von Mexiko, ii. p. 38, Taf. v. fig. 9.

REPORT ON THE HEXACTINELLIDA. 89

penetrated into the Sponge parenchyma, and on the surface of Sponges which are set in
cavities of stones or are surrounded by sand and graveh

Several attempts have been made to utilise for systematic purposes the various modes
in which the spicules are more or less united into a framework. Thus, in contrast to
the Coralliospongia, which possess a firmly united spicular framework, Saville Kent has
named the HexactineUida in which spicules remain disconnected, Calicispongia, and
Carter has distinguished the following three great groups: — (l) spicules united ])y
silicified fibre ; (2) spicules united by amorphous sarcode ; (3) spicules united partly by
vitrified fibre, and partly by amorphous sarcode.

Marshall, on the other hand, maintains that the free or united condition of the
spicules is less significant in the classification of HexactineUida than the manner in
which the union actually occurs. His distinction of Synauloidse, with open communica.
tion between the axial canals of all dictyoualia, and Asyuauloida^, without such a union
of the axial canals, did not, however, find acceptance, for it was soon shown that there
are no Synauloidse in this sense. Zittel accepted Marshall's ideas, however, to this
extent, that he also based his classification mainly on the nature of the union between
the dictyonalia, and distinguished two great divisions, Lyssacina and Dictyonina.
The Lyssacina of Zittel embrace, besides forms with disconnected spicules, those in
which there is simply a cementing of the spicules, that is to say, such an external
union that the spicules do not seem to be checked either in their free arrangement
or in their perfect development. In the Dictyonina the spicules of the lattice-
framework, the dictyonalia, are, on the other hand, normally fused in such a way that
the corresponding rays of neighbouring spicules are closely apposed, and become so
completely united by a uniform coating of silex, that their original independence is
revealed only in the presence of two separate but closely approximated axial canals.
Zittel has also drawn attention to the fact, that Dictyonina frequently possess spicules
which are apposed and bound together in more ii-regular ways ; and 0. Schmidt has
confidently maintained the occurrence of transitional forms between Lyssacina and
Dictyonina. He insists that this dictyonal character is manifested by many forms,
as, e.g., in his genus Hertwigia, in the firmly united inferior portion, while the loose and
irregular union of the spicules in the middle, and the entire absence of fusion in upper
and outer portions, relates the form to the Lyssacina.

It seems to me indeed diificult to distinguish, in many cases, whether a form belongs
to the Lyssacina or Dictyonina. I have, in fact, found forms in which, in a few places,
regular fusion was to be observed, while, in other respects, almost all the spicules were
cemented together in the irregular way of the Dictyonina. I have, therefore, long
endeavoured to find other characters which might serve as sure points of distinction
between the two groups, but I have been able to discover only one fact, which might,
indeed, justify a separation of the groups, though only at best a distinction of degree,

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

aud incapable of practical applicatiou in many cases. While in the Lyssacina the
cementing of the spicules occurs at a relatively very late stage, and generally only
after the Sponge has attained its final form and size, the continuous framework of
the Dictyonina is formed at once in every newly developed portion, and constitutes from
the very beginning an important factor.

The spicules of the Lyssacina are, during the entire period of growth, easily
separable from one another ; but when they become united by the cement into a
rigid framework, the growth of the sponge ceases. In the Dictyonina, on the other
hand, on account of the early development of the rigid framework, growth can only
continue through increase on the surface or at the ends, but there is as little definite
limit to this superficial growth as to the general growth of those Lyssacina in which the
spicules are never firmly united.

It seems to me worthy of notice that, in certain divisions of the Lyssacina, as, e.g.,
in the great family of the Hyalonematidse, and in the subfamily of the Holascinae, a
firm union of the spicules never occurs. It is certainly no accidental fact that it is in
these very divisions that the largest species occur, such as Poliopogon gigas, Poliopogon
amadou, Pheronema giganteum, Malacosaccus vastus.

Position and Arrangement of the Spicules.

The position and arrangement of the spicules now remain to be considered. A
definite grouping and naming of the spicules is of obvious advantage for description even
in those cases where no sharp distinctions exist between the various categories. I would,
therefore, endeavour to complete the classification and nomenclature already adopted by
Marshall.

Prostalia.

The more or less prominent spicules which occur over the outer surface of the Sponge
I call prostalia. They occur only in Lyssacina, and may, from their position, be more
intimately defined and differentiated as hasalia, pleuralia, and marginalia.

Basalia. — The basalia are prominent spicules at the lower end of the Sponge,
grouped together like bunches of long hairs, and forming the root-tuft which serves for
anchoring the animal in the mud. This is one of the characteristic family peculiarities of
the Hyalonematidse, but it also occurs in many Euplectellidse and some Rossellidfe.

When the knowledge of the Hexactinellida was still limited to a few forms, such as
Euplectella aspergiUum, Hyalonema sieholdii, and some Dictyonina, it was proposed to
utilise the presence or absence of a root-tuft as a leading principle of classification, and
to erect a special group of " LophospongiaB." In this, the importance of what is merely
an adaption to the nature of the ground, was over estimated. We now know, in fact,

REPORT ON THE HEXACTINELLIDA.

41

very closely related forms, perhaps even referable to the same genus, one of which occurs
on soft ground, and possesses a completely developed root-tuft, while the other, which

Dlcfyonalia.

Fig. 2. — Diagram of the arrangement of the Spicules.

grows on a firm substratum, exhibits no trace of such a structure. It is, indeed, readily
conceivable that one and the same species might, in different circumstances, produce a
root-tuft or not.

(ZOOL. CHALL. EXP. — PART LIU. — 188G.)

Ggg6

*2 THE VOYAGE OF H.M.S. CHALLENGER.

The elongated spicules wliich form the root-tuft are either smooth or beset with
barbs, disposed iu various ways. While the upper end, which is concealed in the
parenchyma, always runs out to a simple point, the free inferior extremity usually bears
an anchor-like structure, w^hich varies greatly in form and morphological significance. A
knob-like terminal swelling may occur, from the sides of which a few prongs, arranged
in a whorl, project obliquely upwards and outwards ; or again, four cruciate rays may be
present, lying in two typical axes, or rarely two rays in one transverse axis. These rays
are bent upwards in hook-like fashion. In the first case we have usually to deal with diacts,
iu which the axial-canal cross is usually recognisable on the stalk, at some distance
above the anchors, while the anchor-teeth do not exhibit any axial canal (PI. XIV.
fig. 5). In the other cases, however, we have to do with pentacts and triacts, in which
four cruciate or two curved transverse rays lie in the same plane, with axial canals
usually distinctly recognisable (PL III. fig. 23, and PI. LIV. fig. 9).

The arrangement of the barbs on the anchor-stalk, and the form of the anchor-teeth,
are often generically characteristic. In many root-tufts besides the anchor-spicules,
simple pointed diacts frequently occur.

Pleuralia. — The spicules that project markedly from the lateral walls of the Sponge
sometimes exhibit an essential similarity^ to the basalia, from which they cannot, indeed,
be sharply separated even in regard to position. In Pheronema giganteum, for instance
(Pis. XLV., XLVL), bundles of long spicules with round swellings are radially
distributed, with tolerable regularity, over the whole outer surface ; and those directed
downwards (basalia) differ from those placed in a more lateral position (pleuralia), only
in their greater length, slight curvature, and union in a basal tuft. In the Polylophus
philippinensis (PL LIV. fig. 1) numerous pleural spicules assist in the formation
of a local tuft, by bending down and becoming approximated to the basalia. The
outermost termination of these pleuralia exhibit the same anchor-structures as are present
in the basalia of the same form. Thus, both in pleuralia and basalia, in Pheronema
giganteum, for instance, two slightly bent hooks, lying in the same transverse axis, occur
(PL XLV. fig. 9), and in Polylophus philippinensis four intersecting transverse rays
with a gentle curvature (PL LIV. fig. 9).

In other cases the prominent radial pleuralia, which occur either in bundles or isolated,
have the form of simple, terminally pointed diacts. This is the case, e.g., in Acanthascus
(PL LVI. fig. 2). In Euplectella suberea the long, slightly bent radial rays of the large
jirincipal pentacts of the quadrate lattice-work extend radially beyond the side walls
(PL V. figs. 1, 15). In Rossella velata, on the other hand, the external end of the
large pentact is that at which the four tangential rays intersect at right angles.
All these pentact pleuralia project for nearly equal distances from the surface, and
almost touch each other laterally, so that a delicate veil is formed, enveloping the whole
Sponge.

J

REPORT ON THE HEXACTINELLIDA. 43

The pentact pleuralia of Rossella antarctica (PI. LV. figs. 1, 7, 9) deserve special
notice. They arise from small conical elevations and project radially in bundles from
the outer surface of the Sponge, while their four tangential rays, which arise transversely
to the radials, extend at tolerably equal distances over the surface of the Sponge. The
tangential rays, which are provided with small prongs and a uniformly rough cortical
layer, do not intersect at right angles, but are so disposed at acute angles that the four
together enclose a right angle (PI. LV. figs. 9, 13). Between these pcntacts, which
also form a remarkable veil-Hke structure, other longer simple pointed radially pro-
jecting diacts occur.

Marginalia. — In the oscular margin of numerous Lyssacina there is a circle of more
or less wddely projecting spicules — marginalia — which usually consist of elongated
diacts, in which the axial canal cross, which is often distinctly recognisable, or even
indicated externally by boss-like swellings, usually lies almost in the plane of the
outer skin. The projecting distal ray is for the most part thickly beset with outw^ardly
directed prickles and prongs ; it is less frequently quite smooth, and usually terminates
in a point, though sometimes in a small knob-like thickening (PI. L. fig. 4). The
internal, usually much shorter ray of the diact exhibits in some cases small proxi-
mally directed prongs (PI. XL. fig. 6), but is usually smooth and uniformly pointed.
The marginalia include those spicules which project freely in a cufi"-like fashion from
the margin of the terminal sieve-plate in many Euplectellidse. These difi"er from the
above chiefly in this, that their four transverse rays are not abortive, but remain more or
less long, so that the spicules are not diacts but hexacts. At the oscular aperture of
Tsegeria a peculiar form occurs in which the distal rays are specially long and peculiarly
bent (PL VIL);

Dermalia.

As to the spicules of the dermal skeleton, which all deserve the title dermalia, some
belong wholly or at least specially to the outer bounding skin, and have their axial cross
and transverse rays within the latter, while others lie for the most part under the
dermal membrane, with a more or less specially developed proximal ray extending for a
variable distance towards the interior, and with the axial cross and transverse rays
either lying immediately below the inner side of the dermal membrane, or even some-
what removed towards the interior. Although these tw^o forms of dermalia are not by
any means sharply separable from one another, it may be convenient to distinguish
them by the special designations autodermalia and hypodermalia, especially where they
occur close to one another.

As examples of autodermalia, which are exclusively confined to the dermal
membrane, I may cite the dermal tetracts of Lanuginella pupa (PL LIIL figs. 4, 5) and

1

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

Polylophus 2^hiUppinensis (PI. LIV. fig. 2), and also the dermal diacts of Bathydorus
haculifer (PI. LIX. fig. 11). A ray projecting towards the exterior is found in the
pentact dermal pinuli of all Hyalouematidte and many Asconematidse, while a ray
extendino- inwards occurs in the dermal peutacts of Rossella antarctica (PI. LV.
figs. 2, 3) and other Eossellidae. Autodermalia with rays projecting both inwards and
outwards from the dermal membrane occur in Balanites (PL XXIII. figs. 13, 14),
Aulascus (PL XXII. figs. 2, 3), and as amphidiscs in all Hyalonematidse.

As hypodermalia I would note, in the first place, those sword-like hexacts in the
Euplectellidge, which, with their short distal rays, raise the dermal membrane into small
peaks, and are usually surmounted by an attached floricome. Hypodermalia are
well illustrated, too, by those strong pentacts which, in many Hexactinellida, lie with
their four tangential rays closely under the dermal membrane, and bear a greatly
prolonged proximal, penetrating like a strong peg at right angles to the surface, more or
less deeply into the subjacent parenchyma. The tendency of these hypodermal pentacts
to grow inwards may be recognised where the axis cross occurs in the dermal membrane,
from the fact that the tangential rays frequently extend obhquely inwards from their
points of intersection, and always lie beneath any other dermalia which may be present
(PL XXXVI. fig. 1).

To the dermal skeleton I refer finally those spicules which I have noted as clavulse
and scopulse. These forms are disposed at right angles to the outer surface, have the
greater part of their elongated uniaxial body embedded in the parenchyma, usually,
however, reaching the skin or even extending beyond it with their broadened terminal
portion, which contains the axial cross of the central canal. Here, too, we include those
fine raphides which sometimes occur in bundles close to the radial rays of other dermalia
(PL XCII. fig. 2 ; PL XCIII. fig. 2).

Where the skin is raised above the rest of the parenchyma as an independent plate,
the spicules of the dermal skeleton are either confined to this plate without being con-
tinued on the outer side of the subjacent parenchyma, e.g., in Semjoerella schuUzei
(PL LII. fig. 3), and in Hexactinella lata (PL XCV. figs. 1, 2), or they occur not only
on the independent skin plate, but also on the outer surface of the parenchyma, which
then exhibits a special layer of skin, as in Euryplegma auriculare (PL CII. fig. 3), and
Myliusia zittelii.

Gastralia.

Eelations similar to the above are exhibited by the gastral skeleton which supports
the gastral surface and inner side of many efi'erent canals, and which, in some cases,
appears at the oscular border as a direct, and but slightly altered continuation of the
dermal (PL XXIII. fig. 14 ; PL LXVIII. fig. 1). The diff"erence between gastralia and
dermalia is, however, usually distinct enough (PL XXV. fig. 3), and the boundary

REPORT ON THE HEXACTINELLIDA. 45

between the two is generally sharply defined by the peculiar spicules — marginalia — of
the oscular orifice (PI. XXXVIII. fig. 1).

As in the dermal skeleton, so here, under the more superficial, deeper spicules occur,
that is, spicules further removed from the inner bounding surface, and more embedded in
the parenchyma. These are strong hexacts or pentacts with their radial rays disposed
at right angles to the inner surface, and with the four cruciate transverse rays parallel to
the same (PL XXII. figs. 2, 5). In harmony with the term hypodermalia, these may be
designated hypogastralia. I must, however, note that hypogastralia are often wanting
where hypodermalia are present, and the same is true of other typical spicules like
floricomes, amphidiscs, clavulse, scopulse, &c. It may be laid down as a rule that
the spicules of the gastral skeleton resemble the dermalia of the same Sponge in
general characters, but not in their special development, dimensions, and the like.
Thus, for example, the radial axis in one or two rays in the gastralia is frequently
well developed, while it is absent in the dermalia, and so hexacts in the gastral
membrane are often contrasted with pentacts in the dermal membrane (PI. LVIII.
fig. 2). In other cases the free ray is .short and broad in the dermal pinuli, but long
and thin in the gastral (PL XXV. fig, 3). Where the dermal clavulse exhibit a knob-like
extremity, the corresponding gastralia have long anchor-teeth (PL LXXV. fig. 2).
Dermal scopulae with pointed teeth are contrasted with gastral scopulse with button-
like teeth (PL LXXVII. fig. 2), and so on.

As the dermalia of the outer skin do not usually pass into the aS"erent subdermal
spaces and canals, so the gastralia do not, as a rule, pass from the inner skin bounding
the gastral cavity into the efierent canals (PL XXI. fig. 2 ; PL LVI. fig. 2). In not a
few Hexactinellida, however, there is a development of peculiar canalaria, which lie on
the inner surface of the efi'erent canals, and appear to be a continuation of the gastralia ;
whether it is that, in the absence of a special continuous gastral skin, the efl:erent canals
open directly with large orifices into the gastral space, and the gastral skeleton simply
enters into the eflferent passages (PL XXXV. fig. 2 ; PL XXXVI. fig. 1), or that a
special gastral skin encloses the gastral space with a well-developed gastral skeleton,
while the efi'erent passages are, in addition, provided with similar spicules — canalaria
(PL XLVI. fig. 1). The canalaria are usually present only in the main stems and large
branches of the efi'erent canal system, but are not continued into the ultimate blind ends
(PL XXXVIII. fig. 1).

Parenchymalia.

While in the dermal and gastral skeleton there is always a very distinct and tj-jjical

agreement in the position and arrangement of the spicules, this is not the case everywhere

with the parenchymal skeleton. I think, however, that here also certain simple relations

may be recognised as original, from which the less regular have been secondarily developed.

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

On the supposition that the arrangement of the parenchymalia will be least modified
where the original form of the Sponge is a thin-walled sac, and the typical six-rayed
condition of the spicules most distinctly persist, I will start from such conditions as
occur in Holascus and Farvea. Here the chief supporting framework of the parenchyma
is formed of large regular hexacts, which are disposed at right angles or parallel to the
bounding surfaces. These forms, which may be termed directalia, simply lie in the one
case with their corresponding rays apposed to one another (PL XVI. fig. 2 ; PI. XVII.
fig. 2), while in the other they are closely united into a rigid framework (PL LXXIII.
fig. 2 ; PL LXXVL fig. 5).

The invariably six-rayed sjaicules, which are concerned in the formation of the
continuous dictyonal-skeleton, and which I call dictyonalia, are always distinctly
recognisable as such, for even when they have not become connected into a regular cubical
meshwork, but only in an irregular manner, they are readily recognised by their axial canals.
This is not the case with the spicules indicated as principalia, which form the chief
supporting framework of the parenchyma in the Lyssacina, for these, like the dictyonalia,
do not retain their typical position, and further, by no means always exhibit the six rays,
being frequently reduced to pentacts, tetracts, or even indeed to simple diacts. Thus,
in Euplectella aspergillum, for example, tetracts occur in the formation of the quadrate
lattice-work on the inner side (PL II. figs. 2, 5), while in Eup>lectella suherea (PL V.
fig. 15) and in Euplectella nodosa (PL XIV. fig. 2) pentacts occur for the same purpose
and in the same positions ; in Crateromoipha, Rhahdocalyiitus, and Aidochone,
numerous strong diacts are found both in the body and in the stalk.

Closely apposed to the strong rays of the principalia slender elongated spicules
frequently occur, with two or three rays, but seldom more (PL XVII. fig. 6). These I
would call comitalia. They are sometimes straight and parallel to the corresponding
ray of the principal spicule, but they usually extend in a winding and wave-like manner
on or round the outer surface of the principal.

In regard to the position of the parenchymalia, which are found more or less
abundantly between the dictyonalia or the principalia and their attendant comitalia,
few general points can be noted. While elongated spicules of this category frequently
run at right angles to the surface (as the imcinata of many Dictyonina), or are disposed
in any other way with complete regularity, very numerous, small, crowded spicules
often occur, like the small hexacts, the rosettes and their derivatives, which are almost
always scattered quite irregularly in the parenchyma.

REPORT ON THE HEXACTINELLIDA. 47

NOMENCLATURE AND TECHNICAL EXPRESSIONS.

Dermal membrane. — The external limiting membrane, through the pores of which the

water enters the body.
Dermal pores. — The large and small pores which perforate the dermal membrane.
Suhdermal traheculas. — The delicate strands of tissue which form an irregular framework

extending between the dermal membrane and the chamber layer.
Subdermal trabecular space. — The space between the dermal membrane and the

chamber layer, and partly traversed by the subdermal trabecular framework.
Chamber layer.— The more or less folded layer of adjacent ciliated chambers.
Connecting membrane. — The continuation of the chamber wall, stretched between the

terminal openings of the chambers.
Chamber pores. — The small round apertures in the chamber wall.
Gastral membrane. — ^The internal limiting membrane directly surrounding the gastral

space.
Gastral pores. — The pores of very varied size which perforate the gastral membrane.
Subgastral trabeculse. — The delicate strands of tissue which are united into an irregular

framework, extending between the chamber layer and the gastral membrane, and

also frequently into the efferent canals.
Subgastral trabecular space. — The space between the chamber layer and the gastral

membrane, which is partly traversed by the subgastral trabecular framework.
Terminal sieve-plate. — A sieve-like perforated plate, which extends over the broad

terminal opening of many tubular or cup-shaped Hexactinellida.
Parietal gaps. — Apertures over the whole external waU of the Sponge, through which

the space within communicates directly with the surrounding medium.
Membrane of the parietal gaps. — An iris-like, circular membrane stretched across the

gaps, with circular bands of muscular fibres, by means of which the orifice may be

narrowed or entirely shut.
Covering plate. — A porous plate which surrounds like a capsule the body of some forms

with tubular framework, and which is united only to the terminal oscular opening

of the tube.
Spiculum. — Every independent and originally isolated skeletal element.
Principal ray. — The primary ray which springs directly from the central nodal point of

a spicule.
Terminal ray. — The branch or secondary ray springing from the outer end of a

principal.
Hexact, pentad, tetrad, triad, diad, monact. — Nouns and adjectives used to designate

the spicules according to the number of their principal rays.

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

Regular hexact. — A hexact with six rays at right angles to one another, of equal length

and of similar form.
Oxyhexact. — Hexact Viiih rays running out to a point.
Sphaerohexact. — Hexact with spherical terminal knobs.
Discohexact. — Hexact with a transverse disc at the end of each ray.
Rosette or hexaster. — A hexact with equal terminal rays.
Oxyhexaster. — A rosette with straight or bent terminal rays, running out to a

point.
Graphiohexaster. — A rosette where the ends of the principal rays bear a bundle of long

fine terminal rays in a brush-like manner.
Sphierohexaster. — A rosette with spherical knobs at the ends of the terminal rays.
Discohexaster. — A rosette with transversely disposed, radially symmetrical discs on the

outer ends of the terminal rays.
Floricome. — A rosette with S-shaped terminal rays, which are arranged in a whorl

like the petals of a lily, and which pass at their outer end into a thickened

terminal plate, arched outwards, and provided with external marginal teeth or

claws.
Plumicome. — A rosette with numerous S-shaped terminal rays, whose external curved

ends form several tiers, one upon another.
Pinulus. — A pentact or hexact in which one ray bears oblic^ue lateral teeth or prickles,

presenting a resemblance to a fir tree or fir cone.
Scapula. — A fork-like spicule with a long straight stalk passing at one end into two or

more teeth.
Amphidisc. — A diact, at each end of which a convex expansion occurs, which bears six

or more backwardly bent marginal teeth.
Ancora. — A rod-like spicule at one end of which two, or several whorled, transverse

spikes occur, which are bent backwards like hooks.
Uneinatum. — A straight rod, pointed at both ends, and beset all over with barbs point-
ing in the same direction.
Clavula. — A rod which bears at one end a club-shaped or transverse discoidal exj^ansion.
Prostalia. — All the larger spicules which project far beyond the outer surface of the

body, and which may be readily seen with the naked eye.
Basalia. — The prostalia which project downwards from the lower end of the body, and

which form the basal root-tuft by which the Sponge is fixed in the mud.
Pleuralia. — The prostalia which project beyond the outer surface of the lateral portion

of the body.
Marginalia.— The prostalia which project in wreath-Hke arrangement round the oscular

margin.
Dermalia. — All the spicules which stand in a definite relation to the external skin.

REPORT ON THE HEXACTINELLIDA. 49

Autodermalia. — Dermalia which lie either completely, or at least with their axial cross,

in the dermal membrane.
Hypodermalia. — Dermalia which lie with their tangential rays more or less closely

beneath the dermal membrane.
Gastralia — All the spicules which stand in a definite relation to the gastral mem-
brane.
Autogastralia. — Gastralia which lie either completely, or at least with their axial cross,

in the gastral membrane.
Hypogastralia. — Gastralia which lie with their tangential rays close beneath the gastral

membrane.
Canalaria. — Spicules whose axial cross lies in the membrane which lines the eff"erent

canals. '

Parenchymalia. — All spicules which are confined to the parenchyma of the soft parts,

and which belonoj neither to the dermal nor to the gastral membrane.
Dictyonalia. — The parenchymalia which become fused to form the continuous skeletal

framework of the Dictyonina.
Principalia. — The more or less strong parenchymalia which, in certain circumstances,

become fused by siliceous matter, and which constitute the supporting framework

of the Lyssacina.
Comitalia. — -Thin parenchymalia which are very closely apposed to the principalia.
Intermedia. — Parenchymalia situated between the principalia or dictyonalia.

(ZOOL. CHALL. ESP. — PART Lin.--1886.) Ggg

DESCRIPTION OF GENERA AND SPECIES.

Order HEXACTINELLIDA, 0. Schmidt, sen TRIAXONIA.

Sponges with very loose soft tissue, in which the spicules are either isolated or united
by a siliceous cement into a connected sUiceous skeleton. The spicules belong to the
triaxial type, or are readily derivable from it.

Suborder I. LYSSACINA, Zittel (Pis. I.-LXX.; PI. GIL).

Hexactinellida in which the needles either remain always isolated, or are partly
subsequently united in an irregular fashion, often forming strands bound together by
sUiceous cement, or ladder-like trabeculae, by means of numerous synapticula.

Tribe I. HEXASTEEOPHOEA, F. E. Schulze (Pis. I.-XXVL; Pis. LIII.-LXX.; PI. GIL).

Hexasters are always found iu the parenchyma. The chambers are clearly marked
off from one another, and are thimble-shaped.

Family I. Euplectellid^, Gray (Pis. I.-XX. ; PI. LXX.).

Saccular or tubular Lyssacina, in which the inferior blind extremity is either rooted
in the mud by means of a tuft of fibres, or fixed by a compact base on a firm sub-
stratum. The relatively thin lateral wall is in some genera perforated by round or
irregular apertures or gaps, more or less regularly arranged, while in others it is non-
perforated. The transversely truncated or dome-like upper end is generally (everywhere V)
covered by a perforated sieve-plate, and is bordered by a wreath of freely projecting
marginal spicules (marginalia), or by a cufi"-like fringe. ■ The outer surface of the lateral
wall — apart from the gaps — is either uniformly smooth, or exhibits ridge-like elevations ;
sometimes it is richly furnished with radially projecting spicules. On the inner surface
there may be observed — apart again from the gaps — furrow-like grooves in more or less
regular arrangement.

The parts of the skeleton are either entirely isolated, or partly united in an irregular

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

manner into a continuous framework. This fusion is effected by a process of cementing
(soldering), or by means of synapticula.

Tlie dermal skeleton is formed of sword-like hypodermal hexacts, of which the pro-
longed proximal rays penetrate the parenchyma at right angles to the surface, while the
shorter distal rays raise the outer skin into a conical point, and either bear on their tips a
protruding floricome, or are surrounded by several freely projecting pointed diacts. The
four tangential rays of the hypodermalia, which stand at right angles to one another, lie
just below the sieve-like perforated skin, and form by regular apposition of the corre-
sponding rays a quadrate network.

The gastral skeleton is similarly formed of hypogastral hexacts or pentacts. The
prolonged distal ray enters the parenchyma at right angles to the gastral surface, while the
opposite proximal ray, when fully developed, raises the gastral skin in an internal point,
and also, in some cases, bears on its end a floricome. By the closely apposed tangential
rays of the hypodermalia a quadrate lattice-work is formed for the support of the gastral
skin.

In. the parenchyma of the body, in addition to simple hexacts and their derivatives
w^th a smaller number of rays, numerous rosettes occur, chiefly oxyhexasters and
discohexasters.

The chambers, which lie close to one another in a very much folded single layer, have
a simple saccular or beehive form. In their thin walls can be seen the elegant quadrate
network formed from the anastomoses of the ciliated cells, and here and there a round
chamber pore. The delicate trabecular framework, which extends on the one hand between
the outer skin and the chamber layer, and on the other, between the latter and the gastral
membrane, is abundantly penetrated by the lacunse and canal-like apertures of the in-
current and excurrent canal system ; but besides this, there extends everywhere, between
the openings of the chamliers, a thin net-like layer, in which numerous oxyhexasters Lie
embedded.

According to the presence or absence of parietal gaps, I have established two sub-
families, namely, the Euplectellinse with, and the Holascinae without gaps.

Subfamily 1. Euplectellin^ (Pis. I.-VL; PL XIII. ; PI. XIV. figs. 1-5).

Euplectellidse, in which the lateral wall is perforated either by more or less regularly
arranged circular gaps, with a membranous margin and circular muscles, or by iri-egular
angular apertures. The sword-like hexacthypodermalia bear a floricome at the end of
their distal rays. The hypogastralia are, as a rule, simple pentacts without a proximal
ray, but here and there hexacth}q30gastralia also occur, with a floricome at the end of
the principal ray. The central tj^a and best starting-point for the study of this
subfamily is the genus Eujilectella.

REPORT ON THE HEXACTINELLIDA. 53

Geuus 1. Euplectdla, Oweu (Pis. I.-VL; PI. XIII. figs. 5-7; PI. XIV. figs. 1-5).

1841. Owen, Proc. Zool. Soc. Lond., vol. ix. pp. 3-5.

1843. Owen, Trans. ZooL Soc. Lond., vol. iii. (2) pp. 203-206, tab. xiii.

1847. Toiilmin-Smith, Ann. and Mag. ISTat. Hist., ser. 2, vol. xx. p. 176.

1857. Owen, Trans. Linn. Soc. Lond., vol. x\ii. (2) p. 117.

1858. Bowerbank, Phil. Trans., vol. cxlviii. (2) p. 279.

1861. Ehrenberg, Monatsber. d. k. preuss Akad. d. Wiss. Berlin, 1, p. 448.

1862. Bowerbank, PliiL Trans., vol. clii. (2) p. 747.
1864. Bowerbank, Monograph Brit. Spongiadre, vol. i.

1 866. Gray, Ann. and Mag. Nat. Hist., ser. 3, vol. xviii. p. 487.

1867. Bowerbank, Proc. Zool. Soc. Lond., pp. 18, 351.
1867. Gray, Proc. Zool. Soc. Lond., p. 492.

1867. Semper, Archiv f. Naturgesch. vol. xxxiii. 1, j). 81.

1868. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. i. p. 161.
1868. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. ii. p. 373.
1868. Marshall and Herklots, Archives NeerlandaLses, iii. p. 435.

1868. Glaus, Ueber EuplecteUa aspergiUum.

1869. Moore, Ann. and Mag. Nat. Hist., ser. 4, vol. iii. p. 196.
1869. Bowerbank, Proc. Zool. Soc. Lond., p. 66.

1871. Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. viii. p. 330.

1872. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. ix. p. 442: vol. x. p. 134.

1873. Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. xii. p. 349.

1873. "Wyville Thomson, Good Words, July, p. 510.

1874. Th. Higgin, Ann. and Mag. Nat. Hist., ser. 4, vol. xiii. p. 44.

1875. Bowerbank, Proc. Zool. Soc. Lond., p. 272.

1875. Marshall, Zeitschr. f. wiss. Zool., Bd. xxv., Suppl., p. 142.

1876. Willemoes-Suhm, Zeitschr. f. wiss. Zool., Bd. xxvi. p. Ixxiii.

1876. Marshall, Zeitschr. f. wiss. Zool., Bd. xxvii. p. 113.

1877. Wyville Thomson, The Atlantic, p. 138.

1878. Chimmo, Natural History of EuplecteUa aspergillum.

1879. Moseley, Notes by a Naturalist on the Challenger.

1879. Agassiz, Bull. ilus. Comp. Zool., voL v.

1880. 0. Schmidt, Spongien des Meerbusens von Mexico, vol. ii. p. 60.

1880. E. E. Schuke, Trans. Roy. Soc. Edin., vol. xxix, (2) p. 661.

1881. Witefield, Araer. Journ. Sci. and Arts, July and August.
1881. Milne-Edwards, Comptes rendus, t. xciii. pp. 876, 931.
1881. Walcott, Amer. Journ. Sci. and Arts, vol. xxii. pp. 394, 395.
1885. Filhol, La vie au fond des mers, p. 282.

History. — After a preliminary communication which appeared in 1841,^ Owen pub-
lished in 1843 the first accurate description of EuplecteUa aspergillum^ In this some-
what brief, but precise and clear description of a well-preserved skeleton from the Philip-
pines, the structure, as seen by the naked eye, or under the magnifying power of a simple
lens, is so thoroughly investigated that later observers have had but little to add to it.

As the most essential diff"erence between his EuplecteUa aspergiUum and the
Alcyoncellum speciosum, Quoy and Gaimard, Owen emphasises the fact that in EuplecteUa

1 Proc. Zool. Soc. Lond., vol. ix. pp. 3-5.

2 Trans. Zool. Soc. Lond., vol. iii. pp. 203-205, pi. xiii., 1843.

5-1 THE VOYAGE OF H.M.S. CHALLENGER.

a transverse, slightly convex sieve-plate closes that end which he regarded as the inferior,
while Alcyoncclhun speciosum appeared to exhibit a sim^ile inferior extremity. Further,
as he had inadvertently read the word Alcyoncellum, Alcyonellum, he maintained that,
even if there was a generic agreement between his Sponge and that described by the
French investigators, yet the name Alcyonellum must be changed since it had been
already bestowed by Lamarck on a Bryozoon. Owen also erred in regarding Alcyonellum
gelatinosum, Blainville, as synonymous with Alcyoncellum speciosum, Quoy and Gaimard.
In a paper which appeared in 1857,' Owen described a second species of the same
genus under the title Euplectella cucumer. This he distinguished from Euplectella
aspergillum by its barrel-like form, by the absence of the oblique ridge and of the cuff-
like collar at the extremity, which in this species is truncated and closed by a sieve-
plate. The single specimen u23on which the description was based was presented to
Captain Etheridge by the King of the Seychelles.

The generic characters of Euplectella, which were more definitely formulated in this
than in the previous memoir, are as follows : — "A cjdindroid hollow form of body, closed
at the wider end Ijy an irregular network, and at tlie narrow end by the terminal tuft of
finer filaments into which the parietal fibres are there resolved. The parietal fibres, or
those that constitute the wall of the cylinder, are regularly disposed, and intersect each
other at definite and nearly equal distances throughout its extent. They consist of
longitudinal, transverse, and oblique fibres, the latter being of two kinds, winding spirally
round the cylinder, but in ojDposite directions. The longitudinal and transverse fibres
are the thickest ; they are arranged at intervals of from one to two lines, averaging one
line and a half apart, and divide the cylinder wall into square spaces of about the latter
diameter. The longitudinal fibres are external to the transverse ones, to which they are
bound by the oblique or spiral fibres ; these are, some external, some internal, to the
others, and they close by their decussation alternate quadrate intervals between the
longitudinal and transverse fibres. The angles of the alternate open squares are
intersected by finer and less regular oblique fibres, which reduce their area more or less
to a circular form."

The fact that, in the specimens on which the description of EuplecteUa cucumer was
based, the tuft of long siliceous spicules included a number of foreign bodies, led Owen to
believe that the fixing of this sponge, and also of Euplectella asiyergillum, was not
efiected by means of a sieve-plate, but by the long hair-like tuft ; and accordingly, that
the natural position was the inverse of what he had formerly described in regard to
Euplectella aspergillum,.

A treatise by Bowerbank, which appeared in 1858,^ contained a detailed descri^ation
of sponge spicules. In this memoir several of the beautiful microscopical spicules which

1 Trans. Linn. Soc. Land., vol. xxii. (2), pp. 117-124, pi. xxi.

2 On the Anatomy find Physiology of the SjjongiaJa;, Pliil. Trans., vol. cxlviii. p. 279, pis. xxii.-xxvi.

REPORT ON THE HEXACTINELLIDA. 55

lie scattered in the soft tissues of Euplectella aspergillum and Euplectella cucumer are
described and figured.

In Max Schultze's well-known research, " Ueber Ilyalonema," published in 1860,
there are also some observations on Eu2)lectella spicules. It is interesting to note
that Schultze regarded Alcyoncellwn speciosum, Quoy and Gaimard, as identical with
Euplectella aspergillum, Owen. The same view was held by Bowerbank, who in 1862,
in his third communication On the Anatomy and Physiology of the Spongiadse,'
regarded it as indisputable that the older generic name Alcyoncellum of Quoy and
Gaimard should replace the name of Euplectella given by Owen. He himself, however,
inconsistently employed not the specific designation speciosum but aspergillum.
Instead of adopting the generic diagnosis proposed by the French authors, Bowerbank
preferred the following summary of the characters of the genus Alcyoncellum : — ■' Sponge
fistulate; fistula single, elongate, without a massive base. Skeleton: primary fasciculi
radiating from the base in parallel, straight, or slightly spiral lines ; secondary fasciculi
at right angles to the primary ones. Oscula congregated, with or without a marginal
boundary to their area."

In his British SpongiadiB (1865), Bowerbank figured and described difi"erent parts of
the skeleton of Alcyoncellum aspergillum, Alcyoncellum corbicula, and a new but not
minutely described species, Alcyoncellum 7'ohustum, Bowerbank (strictly Alcyoncellum
Tohusta)}

In a short historico-critical paper on Venus' Flower Basket, published in 1868,
Gray maintained the identity of Alcyoncellum speciosum, Quoy and Gaimard, and
EuplecteUa aspergillum, Owen.^ Since the designation Alcyoncellum was first employed
as a generic name for a calcareous Sponge {Alcyoncellum gelatinosum, Blainville)
by Blainville, Owen's generic name Euplectella might be preferred ; but the original
specific designation speciosa of Quoy and Gaimard should be restored, so that this
beautiful siliceous Sponge should in future be called Euplectella speciosa, Quoy and
Gaimard.

In the following year, 1867, Gray described the structure of the skeleton of a young
specimen of the same species.* He refers the bent form of the tubes to the presence of
Crustacea which appear constantly in the interior. Euplectella cucumer, Owen, is not
regarded as a distinct species, but only as a form- variety of Euplectella aspergillum,
Owen.

During his stay in the Philippines Semper had opportunity of examining numerous
specimens of Euplectella aspergillum. He questioned the identity of Alcyoncellum
speciosum, Quoy and Gaimard, and Euplectella aspergillum, Owen, and wished to

> Phil. Trans., vol. clii. p. 110.3. = Loc. cit., figs. 257, 258.

3 Ann. and Mag. Nat. Hist., ser. 3, vol. xviii. p. 487-490.
Ann. and Mag. Nat. Hist., ser. 3, vol. xL\. pp. 44, 138.

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

restore the latter designation.^ Tlic Crustacea that constantly occur in the large lumen
of the tube, Semper identified as a Palseraonid, along with which there frequently occurs
an isolated Aega, named by him Aega spongioi^hila.

In the same year (1867), in his attempt to form a general system of the sponges,^
Gray erected an order of "Acanthospongise" in which "spicules of more than one form or
kind" appear " in the same sponge," and to this he referred, among others, a special family
of the Euplectellidas with the following characteristics : — " Sponge tubular, skeleton
composed of longitudinal, transverse, and oblique bundles of spicules, intersecting each
other and forming a network. Sarcode mucilaginous, studded with many rayed stellate
spicules." To this family Gray also referred, in addition to the genus Eiqylectella, two
other new genera, CorhiteUa and Heterotella, which differ from Euplectella chiefl}^ in the
absence of the regular longitudinal and circular disposition of the fibrous skeletal
strands. The diagnosis of the genus Euplectella is given by Gray ^ as follows : — •
"The tubes regular, gradually wider above, formed of regular longitudinal and transverse
bundles of filiform spicules, which are crossed in an oblique direction with more slender
fascicles or separate filiform spicules, and strengthened externally with transverse or
obliquely raised ridges ; the upper ridge forming a fringe at the top of the tubes, between
the edge of the tubes and the irregularly netted lid. Sarcode thin, studded with many
rayed stellate spicules, with long simple or trifid rays, or with short rays divided at the
end into several converging rays, forming a bell-shaped series."

As specific examples Euplectella aspergillum, Owen, and Euplectella cucumer, Owen,
are cited.

In 1868, Glaus published a detailed research on the architecture, formation and
structure of the skeleton of Euplectella aspergillum, Owen.* The different forms
of spicules are intimately described, and the fusion of certain groups by the deposition of
siliceous lamellfe is demonstrated.

In Bowerbank's critique on Gray's ari-angement of the Sponges,* the separation of the
family of the Euplectellidse from the " siliceo-fibrous sponges " is censured, since their
skeleton is " truly siliceo-fibrous."

A new Japanese species belonging to the genus Eiq^lectella has been described by
Herklots and Marshall,* under the name of Euplectella oweni. It is characterised both
by its saccular form, which gradually diminishes upwards, and by a peculiar arrangement
of the round parietal pores, with which variations in the skeletal structure are associated.
In this species there is no cementing of the long siliceous spicules into longitudinal
fibrous strands, and in place of the outwardly directed oblique ridges there are downy
siliceous hairs covering the surface.

1 Archivf. Naturgesch., vol. i. pp. 84-89, 1867. ^ Proc. Zool. Soc. Land., pp. 492-558.

2 Loc. cit., p. 528. ^ Ueber Euplectella aspergillum, p. 4, 1868.

^ Proc. Zool. Soc. Lond., p. 118, 1868. ^ Archives nderland Sc. exact, etnatur., vol. iii. p. 458.

REPORT ON THE HEXACTINELLIDA. 57

In an article on Semperella (Hyalonema) schultzii (1869), Gray also expressed an
opinion on the relation of Etij)lectella cucumer, Owen, and Euplectella asjoergillum.^ He
said : — " The different shape of the body of Etiplectella aspergillum and of Euplectella
cucumer may indicate that they grow in different situations and circumstances. Euplec-
tella cucumer most probably grows in the mud, kept in its place by a mooring of stones,
as figured in Professor Owen's plate. The curved form of Euplectella aspergillum would
lead one to believe that it most probably grows on the side of a perpendicular rock, but
I have no proof that this is the case, except the form. If it grew from a horizontal
surface, the top of the tube or cloaca would not be uppermost, and the Euplectella would
not be in the upright position natural to all Sponges and other animals and plants that
live on the bottom of the sea."

With regard to the habitat of Euplectella aspergillum, Owen, a communication pub-
lished in the same year, 1869, by Thomas J. Moore,^ is of interest. He communicates
the account given by a certain Captain Morgan, who observed at Zebu (which is at a
distance of three nautical miles from the village of Talisay), the capturing of " Regadera"
by the natives. On the two extremities of the transverse beam of a large T-shaped iron
apparatus, a long bamboo rod with numerous iron hooks is fixed horizontally on each
side. This apparatus is let down upon the muddy ground at a depth of 130 or 135
fathoms, and is dragged over the bottom by the boat, which is carried along by the ebbing
tide, until the fisherman observes that something has been caught. On pulling it up,
some Regaderas are usually found to have been caught by the hooks. These are at
first dirty and of a yellow colour, but after maceration in fresh water they acquire a
beautiful appearance. The root tuft of the Sponge remained in the soft muddy or sandy
bottom, while the entire upper tubular portion of the body projected freely, and its
truncated terminal surface " turned itself towards the setting sun." Another verbal
communication made by Captain Morgan to Moore is also noteworthy, since, according
to it, the natives only capture the Regadera when the boat is carried onwards in a given
direction, but never when it moves in the opposite course.

In the change which Gray ^ proposed in 1872 upon his above-mentioned Sponge System
of 1867, he placed within the order Coralliospougiae, and in that section which possesses
" hexradiate spines in the sarcode," the family of the Euplectellidte, including the single
genus Euplectella, and he referred them to that subdivision in which " the sponge is free,
attached to the mud by numerous elongated filamentous spicules surrounding its base,"
and has " small recurved spines at the end"; the "skeleton being formed of elongated
spicules more or less united by siliceous secretion."

The family, and at the same time the somewhat altered generic characters, are now
stated as follows : — " Sponge tubular, free, formed of bundles of elongated thread-like

1 Ann. and Mag. Kat. Hist., ser. 4, vol. ii. p. 377. ^ ^i„„. and May. Xat. Hist., ser. 4, vol. iii. p. 196.

^ A7in. and Mag. Nat. Hist., ser. 4, vol. ix. pp. 442-4G1.
(ZOOL. CHALL, EXP. PART LIU. 1886.) Ggg 8

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

spicules placed in horizontal, transverse, and oblique directions, often crossing each other,
forming a more or less irregular network, and often closed at the top by a netted lid
formed of shorter spicules ; the base with elongated free spicules terminating in three or
four short spines, by which it is fixed to the mud. The sarcode mucilaginous, studded
with differently shaped spines, some of which are many rayed, stellate, with clavate
arms."

In the system wliich Carter proposed ' in 1873 for the Hexactinellida he did not class
together the well-known forms hitherto united in the family of the Euplectellidse, but
referred Euplectella aspergillum to one of his three chief divisions, where the " spicules"
were "held together by silicified fibre "; the genus Habi'odictymi, Wyville Thomson, he
relegated to the groujD whose " spicules " were "held together by amorphous sarcode";
while for Owen's Euplectella cuciimer, whose spicules are only united in the lower portion
of the sponge by being cemented with siliceous matter into a rigid framework, while they
remain isolated above, he instituted, because of this character, a special third division.

As characteristic of Euplectella aspergillum, he noted that the body was " tubular,
unbranched, and closed at the extremity." This specific character, however, applies
equally well to Euplectella cuciimer, Owen, and to Habrodictyon, "Wyville Thomson.
The two species of the latter, Habrodictyon sp>eciosum and Hcdirodictyon corbicula,
although separated by Wyville Thomson, Carter thought it necessary to unite in one
species, Habrodictyon or Alcyoncellum speciosum.

In 1874, Higgin ^ described the skeletal structure of a specimen of Euplectella
aspergillum preserved in the Liverpool Free Museum, which had already, on an earlier
occasion, been inspected by Wyville Thomson during his stay in Liverpool, and was
referred to in a letter from the Challenger in Good Words, July 1873, p. 510. That
communication by Wyville Thomson ran as follows: — "Several samples of Euplectella
very closely allied to the Philipjiine species, if not identical with it, came up in the trawl
off Cape St. Vincent, and gave us an opportunity for the first time of seeing this Sponge
alive. Dr. J. E. Gray writes to the Annals and Magazine of Natural History that
specimens have been received of Euplectella aspergillum in spirit, and that in these the
glassy framework is entirely masked by a soft browni corky coating of sarcode. Our fresh
specimens entirely bear out Dr. Gray's description. It would be diflicult to imagine that
the thick, somewhat clumsy, brown tube, perforated with irregular openings, contained
any arrangement of support so delicate and symmetrical."

" Although the forms of all the spicules, down to the most minute and complicated,
are identical, the wall of the tube in the European specimens of Euplectella is not
coherent as in most of the Philippine examples. The original spicules of the skeleton
remain separate from one another, and do not become soldered together. One would
think that this would be at all events a perfect specific distinction, but one or two of the

1 Ann. and Mag. Nat. Hist., ser. 4, vol. xii. p. 349. ■ Ann. and Mag. Nat. Hist., ser. 4, vul. xiii. pp. 44-48.

REPORT ON THE HEXACTINELLIDA. 59

specimens of Euplectella aspergillum, particularly one in the Museum of Liverpool, are
in this condition, and I am not yet prepared to say whether all may not be thus soft at a
particular stage of growth."

Higgin found that a rigid, firmly united, siliceous framework was only present on the
under portion of the sponge, while the upper part appeared C[uite pliable. In the inferior
portion the skeletal spicules were provided with a siliceous sheath, and were firmly
connected by means of this ; above, on the other hand, this union of the spicules was
entirely absent.

According to Higgin, large four-rayed spicules form a fundamental quadrate network.
The arms of these spicules lie in exactly longitudinal and transverse directions, and are
apposed to one another. The arms of the cross spicules, which are about 2 cm. long,
stretch across three or four meshes, and the longitudinally directed arms are still longer.
Close to the four-rayed spicules (and united with their arms in a bundle-like manner),
are simple, long, rod-like needles with thorny ends. The long arms of the three-rayed,
and less frequently also of the four- or six -rayed spicules, are employed in the formation
of the oblique fibrous bands. In these spicules two approximately equal arms extend in
one straight direction, while the third, or the remaining two or four arise transversely,
either straight and at right angles, or slightly bent. The spindle-like swollen extremities
of the arms are thickly covered with fine spines. In the description of the simple spicules
Higgin does not add anything to the researches of Clans and Carter, but corroborates the
opinion already expressed by Semper, that the entire skeleton of Euplectella aspergillum
always consists in the young state of isolated spicules, which, only at a later stage,
become partially fused, and more or less firmly united by the deposition of layers of
siliceous substance. This process of fusion begins in the inferior portion of the lattice-
like tube.

A communication by Bowerbank ^ gives a more minute account of the arrangement of
the loose spicules in the tissue of Euplectella aspergillum. There is here to be noted an
important advance in our knowledge of the dermal skeleton. Bowerbank first recognised
that, for the support of the dermal layer, the whole surface is provided with a system of
radially directed, floret-like, six-rayed spicules, and that each of the centrally directed
longest rays of these is apposed to a more deeply seated six-rayed spicule, while the four
lateral rays are applied to the corresponding lateral rays of the four adjoining spicules of
similar formation, and thus form a regular network of approximately equal quadrate
meshes, while the distal ray, which is of about the same length as the lateral rays,
projects radially outwards. Each of these freely projecting, radial, distal rays, bears
terminally one of these elegant structures which are designated by Bowerbank "floricomo-
hexradiate spicules." These ap23ear in no other region of the soft body of this sponge,
and are regarded as protective weapons against small worms, &c. The quadrate meshes

1 Ptoc. Zool. Soc. Land., p. 503, 1875.

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

of the skin skeleton are said to be completely closed by a thin transparent membrane,
and in the latter, groups of small " quadrifurcate hexradiate spicules" occasionally occur.
In the interior of the soft body Bowerbank thought he could perceive numerous
" gemmulfe" surrounded by a simple membrane.

In Marshall's research on the Hexactinellida, published in 1875,Mhere is a detailed
account of the Euplectella oireni, Marshall and Herklots, from Japan, which had been
already shortly described by Marshall and Herklots. This species is accurately com-
pared with the Philippine EuplecteUa aspergillum, which Marshall had also the oppor-
tunity of studying in a young specimen, with as yet entirely unfused spicules.

While there is a great general resemblance in the forms and position of the spicules,
as is particularly obvious in comparing EuplecteUa oweni with EuplecteUa aspergillum,
there never occurs that fusion of the main spicular bands which occurs in Evp>lectella
aspergillum when it becomes old, and which leads to the formation of the elegant
lattice-like framework. While, moreover, Euplectella aspergillum, which is always
much bent, presents an approximately round tube, continually increasing in diameter
from the base to the free extremity, and is provided laterally with ridge-like, oblique,
outwardly directed elevations, and at the extremity with a cuff, bounding the terminal
sieve-plate, Euplectella oweni consists of a perfectly straight tube which is oval in
transverse section, and without external ridges or a terminal circular cuif. From the
broadest part of the tube, which is situated about the boundary between the inferior and
middle third, the diameter diminishes very gradually upwards. The closure is efi'ected
by a sieve-plate which is somewhat strongly arched outwards.

Sir C. Wyville Thomson^ pubhshed in 1877, a short description of a new species of
Euplectella {Euplectella suberea, Wyville Thomson), of which three more or less injured
specimens, figured in a woodcut {loc. cit. p. 29), were collected to the west of Gibraltar.
It may be well here to repeat the words of the highly respected leader of the expedi-
tion : — " The fine species for which I propose the name Euplectella suherea, of which three
specimens, all unfortunately more or less injured, were taken in the trawl, forms a hollow
cylinder about 25 cm. in length by 5 cm. in diameter. The walls are composed, as in
Euplectella aspergillum, of a fundamental, square meshed, siliceous network, bands of
spicules running longitudinally from end to end of the sponge, and transverse bands
intersecting these at right angles. The spicules are in some cases straight and smooth,
frequently four projecting knobs ranged round the centre of the shaft of the spicule
show that, in essential form, the spicule is six-rayed, and often one of the side rays is
strongly developed and projects to a distance of half an inch or more from the surface
of the sponge. The spicules are all free from one another, and those composing the
bands can easily be teased asunder w\i\\ a pair of needles. In this species, as in
Eupjlectella aspergillum,, the corners of the square meshes are filled up, a pale brown

1 Zeitschr.f. wiss. Zool, Bd. sxv. Suppl., p. 142. 2 xiie Atlantic, vol. i. pp. 138-140.

REPOKT ON THE HEXACTINELLIDA. 61

corky-looking substance reducing them to round tube-like holes and rising into spirally
arranged ridges between them ; but the ridges, instead of having a continuous glassy
skeleton, have their soft substance supported by a multitude of delicate six-rayed
separate spicules interspersed with the usual minute sUiceous stars and rosettes. The
sponge is hirsute, with sheaves of feathered spicules which project from the crests of the
spiral ridges, and a series of like sheaves of great length replace round the mouth the
fretted frill of the Phihppine Islands form. The mouth is closed by a very delicate
network of a gelatinous substance supported by sheaves of fine needles. The corre-
spondence in form between its ultimate spicules and those of Euplectella aspergillum
appeared to be so close, that when I first saw this sponge I suspected that it might turn
out to be the same thing under different conditions. I am now, however, convinced
that the two species are entirely distinct." Of importance, too, is Sir Wyville Thomson's
further observation :-^-" No commensal Crustacean has been found in any of the Atlantic
specimens of Eiqylectella."

On Euplectella aspergillum, Dr. v. Willemoes-Suhm ' also made some notes during
the Challenger Expedition. He writes : — " A single specimen of the water-pot-like sponge
was first accidentally dredged seventy years ago, and this, about thirty years ago (1841)
fell into the hands of Owen. High offers for further specimens were then made and the
second was purchased at a high price. Eight or ten years ago they were still dear, when
suddenly the fishermen, induced by this circumstance, discovered in the immediate
vicinity of the city of Zebu, a place from which they captured Euplectella in abundance
by means of an apparatus, constructed of bamboo rods and provided mth hooks, which
was pulled along the sea bottom. The sponge Uves at this place at a depth of 100
fathoms in blackish mud. During our stay in Zebu the ship proceeded one day to the
place in question for the purpose of procuring specimens. A bamboo apparatus from a
fishing boat and a small dredge from the ship were simultaneously lowered. While,
however, the former procured an abundant supply, we got nothing, and only the force
of one of the large dredges was suSicient to tear up the sponges, which were evidently in
great numbers, but very firmly implanted in the mud." A detailed account of the
capture of Euplectella asjxirgillmn by the fishermen of Zebu, and of the apparatus used
for the purpose, has been given by Chimmo in a paper '^ which appeared independently,
and is accompanied by a plate.

With Chimmo's account, the report given by Moseley ^ of the method and apparatus
of capture employed by the fishermen of Zebu, entirely coincides.

Agassiz* mentions a specimen of Eupilectella from the collection of Sponges made by the
expedition in the Carribean Sea. It is probably the same specimen which Oscar Schmidt *

1 Zeitschr. f. vdss. ZooL, Bd. xxvi. p. Ixxiii. * Natural History of Euplectella aspergilliun, 1878.

3 Notes by a Naturalist on the Challenger, 1879. ♦ Bull. Mus. Conip. Zobl., vol. v., 1879.

* Spongien des Meerbuseus von Mexico, p. 60.

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

described in the following year (1880), as having been dredged at St. Lucia in 42.3
fathoms, and at Granada in 416 fathoms, and which he named Euplectella jovin. This
species is closely allied to Wyville Thomson's Euplectella siiberea, but differs from it in
its large prickles, which project freely outwards, and which, to the number of four or five,
form a ring round each of the circular parietal pores. It differs further in the possession
of a remarkable spicule with a fibula-like double hook. 0. Schmidt also describes a
transversely expanded inferior sieve-plate at the lower end of the tube, which differs
from the superior terminal plate only in being less firm.

To the Euplectellidse 0. Schmidt also refers his Rcgadrella phoenix, Hertwigia
falcifera, and Rhabdopectella tintinnus. In all three the inferior extremity does not,
as in the genus Euplectella, run out into a basal tuft, but presents a tolerably firm basal
portion, which either consists, as in Regadrella, of a dense mass growing out into knobs
and lobes, or, as in Hertivigia, of irregularly branched protuberances, or finally, as in
Rhabdopectella, of a simple stalk with a disc-like terminal plate, which is frilled at the
margin.

While the siliceous spicules of Regadrella phoenix do not differ essentially in form
from those of Euplectella asperg ilium, the general shape and the entire architecture of
the sponge is quite distinct. From the massive base a cup arises which is composed at
first of a much perforated plate, and further upwards of a flexible lattice-like network of
obliquely intersecting beams with round meshes. On the irregularly shaped uj)per
margin of the cup a spicular wreath projects like a cuff. The terminal aperture is, as in
Euplectella, closed by a sieve-plate. It is remarkable that within the cup of older dead
specimens younger forms had settled, so that two, or sometimes even three, individuals
appeared as if fixed into one another.

On the branched basis of Hertivigia fcdcifera is seated an " irregular labyrinth of
cavities with thin membranous walls, which are supported by lattice-like plates of
obliquely crossed rods and fibres." On account of the deficient preservation of the
obviously very brittle upper portion, Oscar Schmidt was not able to ol^tain any definite
idea of the form of the entire sponge. Among the siliceous spicules which lie scattered
in the soft parts, there are, besides six- to three-rayed spicules and the typical Euplectella
floricomes, remarkable hexradiate rosettes with four-toothed terminal umbels attached
to the individual arms of the rays, and also rosettes with long backwardly bent teeth on
the terminal umbels, and especially the structures called by 0. Schmidt " sickle rosettes,"
in which each of the six principal rays bears either four simple sickle-like teeth, or a
hemispherical terminal disc with several whorls of sickle-like teeth. Eods occur here and
there with numerous oblique lateral prongs at one end. 0. Schmidt mentions also,
delicate siliceous nets on whose exceedingly fine filaments small terminal hooks and
terminal umbels are found, but these nets seem to me to be fragments of Eadiolaria.

The stalk of Rhabdopectella tintinnus, in older specimens, expands towards its upper

REPORT ON THE HEXACTINELLIDA. 63

end into a very irregular network with large meshes, so that a labyrinth of hollow spaces
is formed. Among the isolated siliceous spicules, in addition to the well-known
Euplectella-&OTicome&, the branches of whose rays have a very fine basal portion and an
unbent plate with a toothed margin at the extremity, 0. Schmidt found rosettes with fine
hook-like teeth, or with terminal umbels bearing four or more prongs, and other forms
in which each of the six principal rays runs out into five teeth which bear on their
extremities hemispherical discs with numerous strong, recurved, marginal rays. Of these
five teeth with their terminal discs, the middle one is always distinctly smaller than the
four surrounding it. A remarkable spiral bow-shaped spicule with transverse furrows is
also mentioned by 0. Schmidt, and finally the presence of many rayed stars is noted.

A small sponge with similar siliceous parts, but consisting merely of a simple tube
open above, with a loose stalk, and a basal plate which resolves itself at the margin into
spicules and spicular bundles, is described by 0. Schmidt as a young Rhahdopectella
tintinnus, and he has accordingly conjectured that the larger (older) specimens are only
the upper portions of similar tubes which had been lost in the dredge.

The first reports of fossd Euplectellidse have been made during the last few yeai-s.
R. P. Witefield ^ first drew attention to the fact that the structures which were described
from the Chemung group of New York, the Waverley beds of Ohio and of several other
places, under the name of Dictyophyton, and which were mentioned in the Sixteenth
Repoi't of the State Cabinet of Natural History of New York, p. 84, as "algse of a peculiar
form and mode of growth," present a certain resemblance to Euplectellid skeletons, since
in the wall of the tube longitudinal and transverse bands of fibres are seen to be united
into a network with rectangular meshes. Soon after, Witefield '■ confirmed this view in
greater detail, and supported it by reference to a Uj^hantwnia dawsoni, which had been
described by T. W. Dawson, pointing out that the beams of the rectangular network
consist of "fascicles of slender cylindrical rods or spicules, quite cylindrical and smooth,"
and that the " spaces between the bands and threads are covered by a thin fibre, which
is alternately elevated or depressed in the adjoining spaces."

By this communication of Witefield, Walcott^ was induced after closer examination to
announce, as a fossil Euplectellid, an organism which he had formerly described, under
the name of Cyathophycus, as a fossil Alga, of a tube-like form, and with an enveloping
coat of retiform structure. He described "horizontal and perpendicular series of narrow
bands crossing each other at right angles, so as to form a network with rectangular
interspaces, the narrow bands being formed of thread-like spicula resting on or one
against the other." Of the form described as Cyathophycus, Walcott said: — "The
striking resemblance to UuplectcUa is seen at a glance, although the convex summit of

1 Aracr. Joum. Sci. and Arts, July 1881; Aim. and Mag. Nat. Hist., ser. 5, vol. viii. p. 167.

2 Amer. Joum. Sci. and Arts, August 1881; Aim. and Mag. Nat. Hist., ser. 5, vol viii. p. 237.

3 Amer. Joum. Sci. and Arts, 1881, vol. xiii. p. 394, 395.

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

the latter genus is absent, and the margin curves over and downward on the inside to a
considerable distance at least, how far is yet unknown." Moreover, he erects a species
with the name CyatJiophycus suhsphgericus, and adds that " each species preserves the
rounded rim of the circular aperture at the summit."

Character of the Genus. — ^The more or less regularly disposed circular parietal gaps
are surrounded by a soft membrane provided with circular muscle fibres. The tubular
or slightly swollen body runs out into a tuft of siliceous fibres at the narrowed inferior
extremity, while the upper, transversely truncated end is closed by a flat or outwardly
convex sieve, and bounded by a thickened border, which is continued into a wreath of
spicules, or into a cuff-like fringe. The larger parenchyma spicules have in the later age
a tendency to fuse together, and in adult specimens may unite to form a continuous
framework.

The groundwork of the whole skeleton is a system of intersecting longitudinal and
circular siliceous strands, which consist of the prolonged rays of strong hexacts, pentacts,
or tetracts, and of the apposed thread-like, thin, comital tetracts, triacts, and diacts. The
intersections of the fibrous strands form quadrate meshes. The rosettes scattered in the
parenchyma are oxyhexasters. The terminal plates of the typical protuberant floricomes
have from six to eight teeth.

1. Euplectella aspergillum, R. Owen (Pis. I. -IV.).

In general form Euplectella aspergillum resembles a thin-walled tube of approxi-
mately circular section ; the somewhat narrowed lower end is continued downwards into
a bundle of longitudinally directed or slightly diverging siliceous fibres, while the
superior transversely truncated extremity is closed by a watch-glass-shaped, outwardly
convex sieve-plate. In its upper two-thirds the tube is straight or only slightly bent,
while the inferior third usually exhibits a more marked, often almost semicircular
curvature (PL I. figs. 1, 2). If there are two curves, they usually lie in the same jilane,
but run in opposite directions, so that a form resembling a slightly bent, S-shaped cornu-
copia results. The length of the full-grown specimens, exclusive of the root-tuft, which
measures from 4 to 8 cm., is on an average 24 cm.; the breadth, in the upper portion,
varies from 4 to 5 cm.; the lower portion, just above the basal tuft, measures from 2 '5 to
3 cm. The dimensions of the individual specimens vary, however, very considerably.

The whole wall of the tube is perforated by numei'ous parietal gaps, circular apertures
of tolerably regular form, and varying from 1 to 2 mm. in diameter. These gaps lie at
the bottom of furrow-like grooves, and each gajJ is surrounded by a stretched, soft, iris-
like, circular membrane, provided with a circular muscle. The arrangement of the
parietal pores is such, that two intei-secting systems of parallel right and left-handed
spirals, ascending at an angle of 45°, may be readily recognised. This regularity.

REPORT ON THE HEXACTINELLIDA. 65

however, is here and there disturbed by the iuterpolatiou of separate gaps, and it
ceases immediately beneath the superior cuff, in the formation of a perfectly circular
row.

The distance between two adjoining parietal apertures in the same row is, in the
upper portions of fully-developed specimens, about 4 mm.; at the lower end, on the other
hand, and in young forms the distance is only 3 mm. or less. In young specimens, and
on the inferior portions of older forms, a slightly arched protuberance is formed in the
middle of every four adjacent gaps. By the fusion of the elevations which lie between
two adjacent spiral rows, raised bands are formed, which, by further increase, become
prominent ledges or ridges. The arrangement and development of these ridges vary
greatly on the walls of the lower part of the tube, they are always but feebly developed,
but they gradually increase in height towards the upper end. In some specimens they
hardly attain a height of 2 or 3 mm.; in other cases, however, they rise to a height of
10 mm. or more. Though they run, for the most part, parallel to the spiral rows of gaps,
they seldom continue in the same direction, beyond a semirevolution. They often bend
round at right angles, just at the end of half a spiral turn, and extend in the opposite
direction down the other side of the tube. Sometimes again they assume an angular or
undulating course, and here and there they may even form a meshwork. In specimens
bent in an S-like curve, I usually observed several parallel ridges extending in an
ascending spiral on both sides, from the lower concavity of the tube to the upper
concavity on the opposite side (PL I. fig. 1).

Among the ridge-like formations must also be ranked that ring-like smooth ledge
which borders the terminal sieve-plate, and which, though in many cases merely
suggested, attains in others a height of 10 mm. Usually, however, the "cuff" is not
in direct connection with the rest of the ridge system, but is separated from it by a
concave circular zone varying from 5 to 10 mm. in breadth. It is distinguished from the
other ridges by being thinner, and by possessing sharper edges.

On the inner surface of the wall of the tube may be observed a tolerably regular
system of circular and longitudinal ledges, which together form a quadrate lattice-work.
The circular ledges, however, extend somewhat further inwards than the longitudinal.
The quadrate meshes which are thus formed have a breadth of 3 or 4 mm. They difler
essentially from one another, inasmuch as the one series exhibit in their centre
the circular parietal gaps already mentioned, with several minute grooves at the
circumference, while the others only show one or more furrow-like grooves of various
breadth and depth. The areas which exhibit deep furrows correspond to those ridges
which are prominent externally, while the shallower grooves underlie the simple
elevations that occur between every four gaps, in those regions which are not provided
with ridges. The arrangement of the parietal apertures in oblique spkal rows, and the
rhombic form of the areas occurring between every four adjoining gaps, are conditioned

(ZOOL. CHALL. EXP. — PART LIII. — 1886.) Grgg 9

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

by the fact that the perforated meshes alternate (with tolerable regularity) both
longitudinally and transversely with those which are closed (PI. IV. fig. 2).

The number of transverse ridges projecting inwards amounts to sixty or eighty in a
full-grown specimen. I have counted thirty longitudinal ridges on the ujjper portion,
and about twenty in the lower, which agrees tolerably well with the numbers given by
Marshall. The increase in the number of longitudinal ridges in the upper part of the
tube is due to the splitting which here and there occurs.

The watch-glass-like, arched, terminal sieve-plate consists of a lattice-work of laterally
compressed ridges of various thickness, which, though exhibiting no very regular
arrangement, yet suggest a wheel-like reticular structure. One can distinguish, at least,
three or four main beams which are approximately circular and several which extend
radially. These form the primary meshes, which are again divided by narrower and less
prominent ridges. Here and there a broader plate is formed in the network, as if by the
confluence of the stronger beams.

At the lower end of the body the longitudinal bundles of siliceous fibres gradually
emerge on the surface, and breaking up into separate spicules, form the basal tuft. This
has a length of from 4 to 8 cm., is tube-like in its upper portion, but towards the lower
end becomes brush-like through divergence of the component fibres. This tuft accord-
ingly encloses a central inversely conical cavity, into which the extreme lower end of the
lattice-like skeleton of the tube-wall extends downwards for a variable distance. In all
full-grown specimens I found that this extreme end of the tube was dead, and at a
distance of several centimetres from the terminal opening, which is from 1 to 2 cm.
broad, the end of the tube was devoid of aU soft tissue, in fact macerated and generally
filled with a firm stojiper of mud. The younger the specimen examined, the better was
the preservation of the lower end of the tube, and the narrower the terminal opening. I
was, however, unable to discover, in any of the specimens at my command, any "pointed
terminal cone, formed from the longitudinal and spiral strands of the parietal tissue,"
such as Marshall has observed in a very young specimen, and has designated the " inferior
sieve-plate."

In uninjured specimens whose soft parts had been well hardened by being preserved
in absolute alcohol, no external openings except the parietal gaps could be seen with the
naked eye. The sponge was of a pale yellowish-grey colour. The consistence of the
soft tissue which covered the skeletal framework in a somewhat thin layer resembled
that of hread crumbs, while Wyville Thomson in The Atlantic, p. 136, observed: —
" In fresh specimens of Euplectella aspergillum the crystal framework is covered and
entirely masked by a layer of grey-hroimi gelatinous matter."

The perforated dermal membrane, which is beset with numerous, small, conical pro-
tuberances, extends smoothly over the much folded chamber layer, and is connected with
it only by the outer trabecular framework, which is much riddled by the subdermal spaces

REPORT ON THE HEXACTINELLIDA. 67

and by the incurrent canals (PI. IV. fig. 4). The saccuhir chambers which lie adjaceut
to ouc another have either grown together laterally, at the points of contact, so as to form
narrow longitudinal stripes, or are connected by means of short beams (PI. IV. figs. 4, 6).

The interspaces and canals extending from the apertures of the outer trabecular
framework, and passing in between the chamber's, are terminated by a connecting
membrane which is stretched, as a direct continuation of the chamber walls, between the
margins of the chamber orifices.

The inner trabecular framework, on the other hand, together with the internal
perforated skin, which is known as the gastral miemhrane, enters the excurrent canals
and lines them as far as the orifices of the chambers (PI. IV. fig. 4).

Without entering here more minutely into the histological relations of the soft parts
thus generally referred to, I would merely note, that in some cases I found, both in the
inner and in the outer trabecular framework, numerous sperm-balls, and also ova of
different sizes (up to 0"3 mm. in diameter), filled with round yolk-granules. In all these
ova it was peculiarly remarkable that the germinal vesicle which contained a large
nucleolus was protruded outwards, and lay on the surface in a round hollow pit-like
depression of the surface.

The Skeleton. — The filagree-like lattice framework, which is readily obtained by
macerating older specimens, as also the spicules which lie freely in the soft tissue, are so
well and so accurately described by such earlier observers as Owen, Glaus, Marshall,
and Carter, that I content myself w4th referring to their excellent works, and wiU here
only briefly note the more important points in connection mth the architecture of the
skeleton, characterise the different kinds of spicules according to their form and
arrangement, and discuss, at greater length, some points which have hitherto been but
little noted.

The groundwork of the entire continuous tube-skeleton consists of two layers of
beams crossed at right angles. One of these layers, the inner, is composed of circular
strands of fibres, while those of the outer have a longitudinal direction. Both together
form a lattice-work which is more clearly recognisable on the inner side, and which
exhibits quadrate meshes, varying from 3 to 5 mm. in breadth. Outside these two
layers of beams, and partly also between them, are two other systems of intersecting
fibres, which run diagonally to the former, and surround the tube in opposite directions
in oblique spirals. They are particularly obvious on the outer side of these quadrate
meshes of the main lattice-work which neither have parietal pores, nor are covered by
external ridges. These externally protruding elevations or ridges are supported by two
layers of steeply-set beams, which lie just beneath the two lateral surfaces, and which, like
the opposite rafters of a very steep roof, have their outer ends united at an acute angle
corresponding to the sharp edge of the ridges. They are also firmly fused laterally to
one another, and to the strong lattice-work of the tube (PI. IV. fig 3). These rafter-like

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

beams are crossed on the one hand by long spiral fibres, which run parallel to the edge
of the ridges, and on the other hand by small short beams, which run through the
ridges transversely. The ridges are doubtless to be regarded as folds of that part of
the tube-wall which extends over those meshes of the inner quadrate lattice-work
not occupied by parietal gaps, and which is supported by spiral fibres crossing one
another obliquely. The long ridge-fibres running parallel to the margin are to be
looked upon as a direct continuation of one of the two systems of spiral fibres, which
cross one another obliquely, and further, the rafter-like beams which intersect the
former at right angles are to be ascribed to the other system of those spiral bands.
The latter appear in the ridges as if laid in a fold sharply involuted on the outer
extremity (PI. II. fig. 8).

The strong circular and longitudinal fibrous bauds, which are so manifest on the
inner side of the macerated tube, have for their groundwork the much drawn-out, strong
rays of the regular, simple and cruciform tetracts. These are interwoven in a peculiar
manner, and become subsequently firmly united by cementing matter and synapticula.
While all the circularly arranged rays of these tetracts run along the inner side of the
tube, the longitudinal rays on the other hand lie transversely across the outside, and all
the rays extend over several adjacent tetracts, thus producing the interlacing which is
suo-o-ested in PI. II. fig. 2. The long siliceous fibres which form the basal tuft join the
longitudinal beams of the quadrate network externally, from the commencement of the
inferior third portion of the tube.

The narrow, much prolonged comital spicules, which are closely applied by the two
long rays of their principal axis, partly to the circular rays and partly to the longi-
tudinal rays of the large tetracts, and which at a later period become fused together, are
for the most part triacts (PI. II. fig. 4 ; PI. III. fig. 12), whose unpaired ray is continued
into the oblique spiral bands of the tube-wall, or contributes to the formation of the
ridges. More rarely the comital spicules are diacts or irregular tetracts, in which the
longer rays lie in one and the same axis, while the two other shorter rays are placed at
right angles to one another (PI. III. fig. 20).

The oblique spiral bands of fibres which extend over the covered meshes of the
quacbate network, and the firm margins of the parietal gaps, are chiefly composed of the
long principal axes of triacts. Here and there irregular tetracts also occur, and diacts
whose atrophied transverse axes are usually more or less prominent owing to the
presence of prongs.

Pentacts and hexacts are less frequently found among the thread-like elongated
spicules (PL III. fig. 17).

The extremities of all those needles, which are subsequently firmly united, exhibit
remarkable modifications, a few of which are figured on PL III. figs. 2-10. The rays
seldom run out to a point, with a gradual decrease of diameter (PL III. fig. 2) ; in most

REPORT ON THE HEXACTINELLIDA. 69

cases tlie extremity is conical and sharp (PI. III. figs. 3-5) or simply rounded (PI. III.
figs. 7-9). It is either smooth or provided with small knobs. These prongs or knobs,
which 2iroject obliquely outwards, give the ends the appearance of club-like thickenings
(PI. III. figs. 9, 12, 17, 20, 25), but this may be also exhibited by smooth extremities
(PI. III. figs. 10, 14).

The tuft of siliceous spicules, rooted in the mud, is formed of two essentially diflferent
kinds of spicules, the main difference between which may be shortly noted by the
designations " pronged diacts " and " smooth pentacts." Both l)egiu in an upper,
pointed, smooth extremity, and run out into an anchor. But, while in the comparatively
rare smooth pentacts the intersection of the axial canals lies in the centre of the four
anchor teeth, which are always crossed at right angles, and while not only the long
smooth anchor stalk, but also each of the four upwardly bent anchor teeth is traversed
longitudinally by an axial canal (PI. III. figs. 22, 23), the intersection of the axial canals
in the pronged diacts does not occur in the greatly thickened inferior terminal knobs but
somewhat above this in the spinose stalk (PL III. fig. 29). Inferiorly the axial canal runs
out generally in a penicillate fashion into a variable number of diverging branches. The
smooth anchor teeth of the prong-bearing diacts, which (to the number of three to eight or
more) stand in a whorl on the side of the terminal knob, and project outwards or upwards,
possess no axial canal (PI. III. fig. 29), and are accordingly to be considered not as true
rays, but only as lateral outgrowths like the prongs of the stalk. The tuft spicules may
in full-grown specimens attain a length of 10 cm. or more. During growth they project
downwards and outwards from the longitudinal bundles of fibres in the tube-walls, and
spread out in a brush-like manner. A bending back of the fibres against the lateral walls
of the tube does not normally occur. Predominant among the spicules which form the
upper sieve-plate, and which are distinguished by their particularly firm union, are strong
triacts with distorted angles, and diacts whose two rays either form a plain arch or an
obtuse angle. Eegular cruciform tetracts, or pentacts and hexacts, seldom occur here.

In addition to the spicules above referred to, which become for the most part firmly
united, there are numerous isolated spicules in the parenchyma of the soft parts, as well
as in both the limiting membranes. These compose the so-called flake-like tissue
(" Flockengewebe " of authors). This designation is due to the resemblance which the
heaps of such needles, liberated by maceration and dried, bear to flakes of snow. As the
age of the sponge increases some of them become cemented together, while others remain
isolated throughout life. Among the former are numerous delicate spicules with a variable
number of rays, but for the most part triacts, which in their whole appearance resemble
the long and thin comital spicules above referred to, but which possess outwardly bent
rays not so long as those of the comitalia (PI. III. fig 25). The spicules which always
remain isolated may be classed in the following five groups : — (1) parenchyma spicules,
which are distinguished by thick, short, pointed rays, uniformly conical towards the outside ;

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

(2) liexact liypodermalia ; (3) pen tact hypogastralia ; (4) oxyliexasters scattered among
the parenchyma ; (5) protuberant floricomes.

The spicules which belong to the first category, and are provided with thick, short,
conical rays, are scattered in the inner portion of the walls of the tube. They are
particularly abundant, and are arranged in a cruciform manner in the circular membrane
surrounding the parietal gaps. In this situation peutacts chiefly occur, in which the
unpaired ray penetrates the parenchyma radially outwards, while the four rectangularly
crossed rays of the two other axes lie parallel to the bounding surface. Often, however,
(especially in the inner thinner marginal portion of the circular membrane), only one
of the two latter pairs of rays is fully developed, so that triacts arise whose paired rays,
lying in the same axis, extend tangentially to the free margin of the parietal gaps, while
the unpaired ray is directed radially. If, again, the latter remain undeveloped, diacts
arise which have been designated by Marshall " compass-spicules " on account of their
resemblance to a compass needle, being much swollen in the middle. Completely formed
hexacts belonging to this category of spicules are less frequent ; when they occur they
lie in the parenchyma at some distance from the gastral surface, or from the free margins
of the parietal gaps. They are usually regularly developed (PI. III. fig. 18), more rarely
shortened in one ray (PI. II. fig. 1, a.b.c). While in the lattice-like network of the
outer skin dermalia proper are absent, regularly arranged liypodermalia occur
underneath as slender hexacts of equal form and size, and provided with a greatly
prolonged proximal ray, four times longer than the other five rays, which are of
approximately equal length, and like the former run gradually out to a point (PL III.
fig. 16 ; PI. IV. figs. 3, 4, 5). The long principal axis is at right angles, the two
transverse axes are parallel to the surface of the body.

In well-preserved portions the axial cross of the hypodermalia lies about O'l mm.
beneath the skin. Their distal ray extends into a point-like elevation of the skin, which
it seems to push out. On its outer end it carries a floricome which extends over the
skin (PI. IV. fig. 4).

Since the corresponding tangential rays of the neighbouring hypodermalia are apposed
to one another laterally for half of their length or even more, quadrate or rectangular
meshes are formed, over the centre of which the skin is, as a rule, somewhat depressed
(PI. IV. fig. 4). If this involution extends still further, the skin finally becomes applied
closely against the tangential rays of the hypodermalia, and it may even seem as if the
latter occurred in the skin. On the summit of the outer ridge there is usually a row of
particularly strong and long hypodermalia.

To the system of the hexact hypodermalia there corresponds, on the inner side of the
entire tube and of the large excurrent passages, a system of similar slender pentact
hypogastralia with pointed extremities (PL III. fig. 13 ; PL IV. fig. 3). These occupy
the same relative position to one another and to the gastral limiting membrane of the

REPORT ON THE HEXACTINELLIDA. 71

inner trabecular framework as the hypodermalia do to the outside skin. The long unpaired
distal ray extends into the parenchyma, while the four short rays of the two other axes
form a cross which lies close beneath and parallel to the gastral membrane. Though this
system of hypogastral pentacts also extends for a considerable distance into the excuxrent
passages, it does not extend as far as their terminal diverticula (PI. IV. fig. 3).

The numerous rosettes present in the parenchyma are all oxyhexasters whose short
compressed principal rays are continued with a slight terminal thickening into the
strongly diverging, straight terminal rays which are six times longer. The number of
terminal rays belonging to a principal ray varies from two to five, and it is the same
in most cases on all the six arms of a rosette, though it may difier in individual arms.
The most common case is for every arm to have three terminal rays (PI. III. fig. 1).

Such oxyhexasters occur on the one hand in the outer trabecular framework, with
which they are compressed between the convex outer sides of the chambers ; and on
the other hand, in the inner trabecular framework, with which they extend to the final
expansions of the excurrent canal system, between and in front of the terminal openings
of the chambers. They do not occur, however, either in the chambers themselves or
in their delicate walls. The dermal and gastral membranes are also free from them.

The other rosette, much more striking on account of its elegant form and exposed
position — the floricome — is regularly distributed over the whole outer skin, and is scattered
on the tips of the conical elevations of the skin into which the distal extremity of a
hypodermal hexact extends (PL IV. figs. 3, 4, 5).

Each of the six short, narrow, principal rays becomes divided into seven or eight
terminal rays, which are bent in an S-like manner, are arranged in a whorl like the
petals of a lily, and consist of a very thin basal portion or stalk, shghtly bent and con-
vex towards the outside, and of an outer portion which gradually increases in thickness.
The latter bends outwards in a stronger curve, and ends in a firm, almost hemispherical
plate, whose sharp semicircular outer border is prolonged into six or eight claw-like
prongs (PL II. fig. 9 ; PL III. fig. 11).

Since these terminal rays of the floricome usually project freely from the pointed
elevations of the skin, either entirely or with their claw-bearing extremities, they
penetrate any soft body which may come in contact with the sponge. The delicate stalk
will then readily break ofi", and the terminal portion remain in the foreign body. In
this manner the floricomes of the Euplectella serve as weapons of defence against the
attacks of soft skinned animals such as fishes, molluscs, worms, and the like.

I observed that most floricomes had one of their axes so applied to the distal ray of
the corresponding hypodermal hexact that they lay at right angles to the surface of the
skin. Not unfrequently, however, I found them also lying obhquely as if misplaced.
Though floricomes are not to be found on many, and especially on prominent portions of
the skin, it may be reasonably supposed that they were originally present here as well as

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

in the adjoining and more protected regions, where they occur in a perfectly regular
arrangement. Even on the outer side of the beams of the sieve-plate I have frequently
found them intact over large areas. On the borders of the parietal gaps, as well as on
the entire inner surface of the sponge and the excurrent passages, they are entirely
absent, nor have I found them in the parenchyma.

Among the many examples of this beautiful and interesting species which the
Challenger Expedition collected from the same locality, namely, near the island of Zebu,
at a depth of 100 fathoms, some young specimens, unfortunately much injured, of
which two seemed bound together by an oblique tube-like anastomosis, deserve special
mention. The larger of these two tubes, which stand almost parallel to one another at a
distance of 2-5 cm., is bent in a hook-like manner, and just above the connecting bridge,
on the side turned towards the other specimen, is somewhat flattened and partly involuted.
Not including the basal tuft, it is 20 cm. long, and from 2-5 to 3 cm. broad. The
distance between longitudinal and transverse spicular bundles is from 2 or 3 mm., the
breadth of the tolerably regular parietal gaps is about 1 mm. The outer ridges and the
cuff are well developed, and they attain a height of 2 or 3 mm. The terminal sieve-
• plate is remarkably highly arched. Somewhat above the middle there is an acutely
angular division of the tube into two similar tubular branches. The one branch
continues in the direction of the principal portion to the highly-arched terminal sieve-
plate with an approximately equal diameter; the other bends obliquely to the side,
becomes somewhat narrower, and is continued directly into the wall of the second shorter
and narrower straight tube, of 10 cm. in length, and 1 to 5 cm. in transverse diameter.
This extends for 2-5 cm. to the inferior funnel-like narrowed extremity, which runs out
into a delicate basal tuft, while the upper very delicate and soft portion is about 6 cm.
in length, and is provided with a terminal sieve-plate. This small specimen was
unfortunately much damaged in its delicate upper part, yet it may still be clearly
recognised that longitudinal bundles of fibres pass over directly from the larger
EuphcteUa through the obliquely ascending bridge into the longitudinal strands of the
smaller portion, and extend both upwards and downwards. Since the portion of the
short tube which lies beneath the connecting bridge, and especially that part towards
the larger tube, agrees thoroughly in the character of its wall with the side branch of the
larger tube, it apparently represents a direct continuation of the latter; the upper
portion, on the other hand, up to its superior sieve-plate, becomes gradually narrower in
its meshes and enfeebled in all its parts, and is, moreover, manifestly younger than the
under portion, so that I do not think we have here to deal with two independent
specimens which stood close to one another, and which, at a subsequent period become
united by a binding-tube, but am rather of opinion that originally only the larger
existed. A portion of the upper half of its lateral wall was separated from the upper
half in such a way that it hung down to the ground in an arch. This arch thereupon

EEPORT ON THE HEXACTINELLIDA. 73

united so as to form a tube, which constituted the present connecting bridge and the
under extremity of the small tube, while from its upper convex side a new young tube
grew upwards.

Another young specimen, in which the amalgamation of the spicules has not yet
commenced, is from 2'5 to 3 cm. broad, possesses a highly arched sieve-plate, and a very
well developed system of ridges, together with a completely formed cuff. The ridges have
already attained a height of 3 mm. or more. The inferior extremity is unfortunately
absent.

2. Euplectella suberea, Wyville Thomson (PI. V. ; PI. VI. fig. 3).

Though the Challenger specimens of the new species which W5rville Thomson named
Euplectella suberea were not, as a whole, well preserved, the discoverer was still able
by combining the various fragments, to obtain so perfect a conception of the size, form,
and structure of the whole sponge that he was able to publish, with the help of the
artist, the restoration presented in PL V. fig. 1. As is evident indeed from this
excellent fio-ure, and from Wyville Thomson's own words, which have been quoted
verbatim above on p. 60, the sponge in question is a straight, round, slightly
swollen tube, varying from 20 to 25 cm. in length, and measuring about 5 cm. in
its greatest breadth. Its walls, it is true, present a general similarity to those of
Euplectella aspergillum, but it may be readily distinguished from the latter by the
greater regularity of its structure, by the absence of the high external ridges and of the
cuff, as well as by the numerous separate projecting radial spicules.

The parietal gaps lie at a distance of 6 to 8 mm. from one another, within trench-like
pits, which are connected by intersecting systems of external, oblique, and spiral furrows,
so that the regular spiral arrangement of the parietal apertures becomes all the more
manifest. Between every four adjoining parietal gaj)s there is a rhombic area, bounded
by the connecting furrows, and provided with a flat, slightly convex, projecting elevation.

The upper transversely truncated extremity of the tube is bordered by a delicate
marginal wreath, from which isolated spicules project upwards and outwards in groups,
without forming a continuous fringe. Close beneath the margin there is a wreath
of densely placed parietal apertures, while inside the margin a wide-meshed lattice-like
network is spread out transversely, the delicate strands of which are arranged partly
in a circular, and partly in a radial manner.

The inferior extremity of the tube, which has become narrowed to about the half
of the greatest transverse diameter, is devoid of soft parts, and runs out into an incom-
pletely preserved basal tuft of long, thin, siliceous fibres.

An examination of the inner surface of the wall shows that here, as in Euplectella
aspergillum, a quadrate, lattice-like network of longitudinal and circular ridges projects
inwards. The meshes are seen to be occupied alternately with a parietal gap, and with a

(ZOOL. CHALL. EXP. — PAET LIII. — 1886.) Ggg 10

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

I'ouud exhalcut orifice of the effereut canal system which lies underneath every outwardly
arched elevation of the walls of the tube {PI. V. fig. 12 ; PI. VI. fig. 3). On the projecting
ledges there are also numerous smaller furrows, and the orifices of shorter efi'erent canals.
Although the preservation of tlie soft parts of the specimens handed over to me for
examination is not so good as that of many specimens of Eujilectella aspergillum, I
have still succeeded in definitely determining the essential structural relations.

The much folded layer of saccular chambers is, as in Ewplectella aspergillum, united
with the external skin by means of the outer trabecular framework, and with the gastral
membrane by means of the similar internal trabeculte. In this also, the latter is con-
tinued into the terminal diverticula of the efferent passages, forming a continuous
internal lining (PI. VI. fig. 3).

The thickness of the tube-wall amounts, in the arched portions, to about 3 mm., and
through the inwardly projecting ledges to between 3 '5 and 4 mm.

The chief skeletal framework is formed of strong pentacts, in which the four cruciately
disposed rays, which vary from 1 to 3 cm. in length, are arranged longitudinally and
circularly so that they are closely apposed to each other longitudinally, or even cross one
another transversely, and so that the longitudinal rays always lie externally to the
circular. The breadth of the quadrate meshes thus formed amounts to 3 or 4 mm. The
radial unpaired ray, which projects freely outwards in a straight or slightly curved course,
attains a length of 1 '5 to 2 cm., and, like the four other rays, runs out to a simple point.
A radial ray does not, however, arise from every point of intersection of the longitudinal
and circular fibrous strands ; on the contrary, both in the longitudinal and transverse
directions, a simple point of intersection without an independent pentact always alternates
with one which is occupied by the intersecting nodes and the radial rays of a pentact
(PI. V. fig. 15).

The intersections of these longitudinal and transverse fibres correspond, not to the
middle of an elevation lying between four parietal apertures, but to those external
furrows which unite the parietal apertures in spiral lines (PI. V. figs. 1, 12).

The longitudinal, as well as the circular rays of the large pentacts, are closely
surrounded by a mantle of delicate comitalia, whose greatly prolonged rays apply
themselves either parallel to, or in gentle windings round the thick round pentact-rays.
As a rule the comitalia consist of triacts, wdiose unpaired ray stands at right angles to
the two long rays which lie in one line, and serves for the support of the adjacent soft
parenchyma. Long diacts with central intersecting nodes are also frequent (PI. V.
fig. 14). The extreme ends of these comitalia form blunt points, and exhibit a slight
roughness, while the rest of the ray is smooth. With the comitalia are associated those
numerous parenchymalia, with from two to six shghtly bent or straight rays (PI. V.
figs. 16-20), which serve for the support of the parenchyma of the soft parts between the
principal strands of fibres. They are somewhat thicker than the comitalia, and their

REPOET ON THE HEXACTINELLIDA. 75

rays are never so much prolonged. Hexacts seldom occur among the parenchymal
spicules, and like the somewhat more frequent pentacts are irregularly developed. The
numerous tetracts which are present, exhibit, indeed, for the most part, rectangularly
intersecting axes, but these seldom lie entirely in a plane, being for the most part
uniformly bent towards the surface ; it may be that the two rays belonging to one axis
run out in the same direction and straight, while the two others are somewhat bent, or
are disposed at an obtuse angle to one another. Sometimes tetracts occur with one of
the rays disposed at right angles to the surface of the three others. The majority of
these parenchymalia have only two or three rays. Of the triacts the two rays belonging
to the same axis are usually bent in a slight curve, from about the middle of the convex
or concave side of which the third straight ray sjarings (PI. V. fig. 16). The diacts are
sometimes straight, sometimes slightly bent, pointed at both ends, or more or less
rounded. They exhibit in the middle of their axial cross, four, two, or one tubercle,
and are frequently roughened, not only near their ends, but throughout.

Peculiar small, straight, rough diacts, with truncated or rounded extremities, and
with four tubercles projecting markedly from the axial cross and often exhibiting central
terminal points (PI. V. fig. 7), occur abundantly in the thin circular membrane which
surrounds each of the parietal gaps. As a rule, the two rays are equally long, but forms
also occur with two long unequal rays, and others which by the very great shortening of
one ray have become monacts.

Both in the outer and inner trabecular framework rosettes occur, scattered quite
irregularly in the form of oxyhexasters, whose short principal rays bear usually three
(PL V. fig. 2), seldom four, straight, long, more or less stout terminal rays. The rosettes
represented by Wyville Thomson on PI. V. figs. 3, 8, I have only discovered in those
greatly damaged fragments which were much mixed with the spicules of other
Hexactinellida, and which were collected ofi" the coast of Brazil (Station 124). These
latter rosettes probably do not belong to Euplectella suberea, but appear to me to have
originated in other Hexactinellida, and to have become accidentally embedded in these
specimens. At least in sections of Euplectella suberea (collected to the west of Gibraltar),
I have never been able to find such rosettes in situ.

The dermal skeleton consists of a layer of dagger-like delicate h}q)odermal hexacts
with rough conically pointed extremities, which have their tangential rays so apposed to
one another under the bounding membrane that a rectangular meshwork is formed, while
the proximal ray penetrates deeply into the parenchyma, and the distal ray, extending
to the very tip of every minute tubercle of the skin, bears the somewhat freely projecting
floricome (PI. VI. fig. 3).

It is noteworthy that on the top of those flat archings which project outwards
between every four parietal pores, the much prolonged distal ray of a specially large
hexact usually projects at right angles to the surface, and instead of carrying a floricome

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

is suiTounded by a bundle of diacts, which are provided with prougs ou their pointed
extremities (PI. V. fig. 6), and have four conical swellings on the axial intersection.

The floricomes of the outer skin exhibit a form quite similar to that in Euplectella
aspeiyillum, but it is important to note that they are almost twice as large. I usually
found seven terminal rays on each principal.

The gastral skeleton exhibits on the inner side of the quadrate ledge-like network
which projects into the large lumen of the tube the same slender pentacts which occur
in similar positions in Euplectella aspergillum ; on the other hand, on the inner side
of the diverticula-like chambers which underlie the external gentle archings, slender
dagger- like hexacts occur, wdth a prolonged distal ray, and a proximal ray projecting
radially into the inner lumen (PI. VI. fig. 3), raising the gastral membrane into pointed
tubercles, and bearing on its end a floricome just like the similar dermal hexacts. It is
interesting, however, that this gastral floricome does not agree either in size or form with
the above-mentioned dermal floricome. The number of the terminal rays belonging to
each principal amounts to ten or twelve in the former, but only to seven in the latter.
The cup formed by these terminal rays is relatively broader (the diameter of the whole
floricome is only half as large as in the case of the dermal floricome) (PL V. fig. 4).

Fragments of Euplectella suherea, Wyville Thomson, were collected by the Challenger
in three different places : —

(1) In 600 fathoms, lat. 36° 25' N., long. 8° 12' W.; Station IV.; west of Gibraltar.

(2) In 1090 fathoms, lat. 35° 47' N., long. 8° 23' W.; Station V.; w^est of Gibraltar.

(3) In 1600 fathoms, lat. 10° 11' S., long. 35° 22' W. ; Station 124; between
Pernambuco and Bahia.

The figure (PL V. fig. 1) is reconstructed from somewhat defective, and yet in some
respects adequately preserved, specimens found to the west of Gibraltar (Stations IV., V.).

In a short review of the results of the French " Travailleur " expedition, Milne-
Edwards^ has noted (1881) that "two fine specimens of Euplectella suherea were taken
off the Berlingues at 3307 metres." And Filhol" adds, " Les Euplectella suberea sont
largement repandues dans I'Atlantique nord. Pendant la croisiere du Talisman nous les
avons draguees a diverses reprises par des fonds variants entre 900 et 2300 metres. En cer-
tains points elles etaient d'une extreme abondance etdevaient couvrir d'assez vastes capaces."

3. Euplectella cucumer, R. Owen.

Since I had the opportunity of seeing the original specimen of Euplectella cucumer,
Owen, which is preserved in the British Museum, and of examining it so far as was
possible without injuring the beautiful specimen, I shall here take the opportunity of
briefly expressing my opinion in regard to this interesting form.

' Comptes rendus, xciii. pp. 871-931 ; Ann. and Mag. Xat. Hist., ser. 5, vol. ix. p. 46.
^ La vie au fond des mers, p. 282, pi. iii.

REPORT ON THE HEXACTINELLIDA. 77

In opposition to the opinion which has been expressed by several authors, e.g.,
Marshall/ to the eflfect that Euplectella cucumer is not specifically distinct from Euplec-
tella aspergillum, I must point to the following essential difi"erences. Whilst Exiplectella
aspergillum has the form of a bent tube, which becomes wider towards the upper
extremity, Euplectella cucumer consists of a perfectly straight, bellied tube, devoid
both of the externally projecting ridges and of the cuff on the upper margin. The
regularity of the arrangement of the round parietal apertures, which is frequently inter-
fered with in Euplectella aspergillura, is everywhere manifest in Euplectella cucumer.
The fusion of the skeletal spicules which appears in all parts of the tube of Euplectella
aspergillum soon after the attainment of its full size, has taken place in this obviously
old specimen of Euplectella cucumer only to a very limited degree in a few regions.

It thus appears to me that (on the whole) Euplectella cucumer is more nearly related
to Euplectella suherea than to Euplectella aspei'gillum. From Euplectella suherea,
however, it may be distinguished first by the compact sieve-plate, consisting of strong,
firmly united spicules, and secondly, by a character which Owen noted, and which Carter
subsequently announced as a distinctive peculiarity, — the possession of strong hexacts,
1 to 2 cm. in length, regularly distributed in such a way that one always occurs in
the middle space between every four adjoining parietal gaps, with its strong distal ray
projecting freely from the summit of the raised arching of the outer wall.

From such considerations I believe I am justified in maintaining that Euplectella
cv^umer, Owen, is an independent species.

4. Euplectella jovis, Oscar Schmidt (PI. VI. figs. 4, 5).

According to Oscar Schmidt's account Euplectella jovis — found in the Gulf of Mexico,
in 423 fathoms in the neighbourhood of St. Lucia, and in 416 fathoms in the vicinity of
Granada — agrees in many points with Euplectella suherea, Wyville Thomson. The
straight tube, which extends to a length of 50 cm., is provided with a basal tuft and a
terminal sieve-plate ; but the external ridges are absent. A projecting wreath of spicules
surrounds the sieve-plate. The round parietal gaps are arranged in spiral rows, though
not quite regularly (PI. VI. fig. 4). On the inferior extremity 0. Schmidt found a trans-
versely placed inferior sieve-plate which divided the inferior conical portion of the tube,
which was filled with mud, from the principal vital portion.

As the spicules never become fused the tube remains soft and flexible.

The intersections of the longitudinal and transverse strands of fibres forming the
c^uadrate lattice-like network are here also principally supported by the mutually apposed
tangential rays of the strong pentacts, whose distal radial ray projects freely outwards, for
a distance of 1 or 2 cm. beyond the surface. An intersection of the fibrous lattice- work, pro-

1 Zeitschr.f. iviss. ZooL, Suppl.-Bd. xxv. pp. 148, 209.

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

vided with such a pentact and distal ray, always alternates both in the longitudinal and in
the transverse direction, as in Euplectella suherea, with an intersection without a pentact.

Both the tangential rays and the basal portion of the distal ray are surrounded by
numerous greatly extended comitalia which are for the most part triacts. Over the
meshes of the lattice-work which are not perforated by parietal gaps, there extend spiral
strands of thin, elongated parenchymalia, which cross each other obliquely, and which are
also for the most part triacts.

In the circular membrane which surrounds every parietal gap there occur near the free
inner margin numerous short blunt monacts of a sceptre-like form, whose five reduced
rays project at the one end as short rounded stumps or hemispherical knobs (PL VI. fig. 5).
The occurrence of this form has induced 0. Schmidt to dedicate the species to Jupiter.
Besides the fully formed sceptre spicules numerous diacts also occur with projecting
cruciate axial knots, which show very clearly how the sceptres have arisen.

More characteristic than those sceptres, which also occur singly in Euplectella
suherea, are the clasps or fibulae which are scattered with comparative uniformity
throughout the whole parenchyma of Euplectella jovis. These exhibit a straight, thick,
middle portion with a central node and two curved hook-like terminal rays which run out
to fine points (PI. VI. fig. 5). The latter are as a rule clearly marked ofi" from the middle
part, but their curves usually lie in difi"erent planes, turned towards opposite sides, so that
when seen in a certain direction they seem to present the figure of an S. I regard these
clasps, which I also met with in Holascus fihulatus, as derived from oxyhexasters with
bent terminal rays. I do not derive them from simple hexacts, because on each side a
terminal ray springs from the corresponding principal ray. Important also in this respect
is the generally very clearly developed median node, on which four lateral tubercles may
sometimes be seen. Here and there triacts or tetracts may also be seen whose rays have
the same form as in these diact fibulae. This interpretation also harmonises ^uith the,
fact that in Euplectella jovis the oxyhexasters which are present in all other species of
Euplectella do not occur, but are, as might be expected, converted into the clasps.

On the outer surface of the distal ray of the sword-hke hexact-hypodermalia, large
floricomes occur with a small number of terminal claws exactly as in Euidectella suherea.

5. Euplectella owenii, Herklots and Marshall (PL VI. figs. 1, 2).

1868. Herklots and Marshall, Arch. Neerl. des sci. nat., iii. jx 435.

1875. Marshall, Zeitschr. f. wiss. Zool., Bd. xxv., SuppL, p. 142.

1876. Marshall, Zeitschr. f. wiss. Zool., Bd. xxvii., p. 113.

Of that elegant Japanese Euplectella, which was first described by Herklots and Mar-
shall and named Euplectella oiceni, no specimens were brought home by the Challenger
Expedition. There occurs, however, a relatively well-preserved specimen of this species
among those Japanese Hexaetinellida which were handed over to me by Dr. Doderlein of

REPORT ON THE HEXACTINELLIDA. 79

Strassburg. Including the basal tuft this form lias a leugth of 20 cm. The portion
which projects freely from the mud is 12 cm. in length, and its cross section is oval, the
long axis below measuring 35 mm., but above, just below tlie dome-like arched sieve-
plate, only 25 mm.; the short axis beneath measures 30 mm., and above 22 mm. A
compai'ison of these figures wdth those recorded by Marshall for his specimen, shows that
we have here to deal with a I'elatively small, and probably young specimen. In form
it agrees throughout with Marshall's sketch and figure,^ exhilnting a straight tube
whose diameter is greatest somewhat beneath the inferior third part, and diminishing
upwards at first very gradually, finally somewhat more rapidly, becoming narrowest
just beneath the terminal sieve-plate. Both cuff and external ledges are entirely
absent. The inferior extremity is continued with a more gradual narrowing into the
long, soft, hair-like basal tuft.

The arrangement of the round parietal gaps, which are almost 1 mm. in diameter, in
regular, transverse and longitudinal rows is very manifest. Between these parietal gaps,
which lie in pit-like depressions, a rectangular lattice-work of transverse and longitu-
dinal elevations projects both in the outer and inner surfaces (PI. VI. fig. 1). A
system of rectangularly crossed longitudinal and transverse bands of fibres, of which the
latter project most internally, while the former cross them transversely on their outer
side, serves for the support of this lattice-work, which forms somewhat narrower meshes
at the upper than at the lower end. Transverse fibres also occur, and these become
interwoven with the system of longitudinal fibres, sometimes occurring above, and some-
times beneath the latter. The longitudinal and the transverse bands of fibres do not
always form, as in Euplectella aspergillum, simple and compact bundles, but frequently
consist of two separate, but parallel bundles which lie close together, or at some distance
from one another. Llarshall describes these double bands of fibres as somewhat constant
and characteristic of the species. In the specimens before me the division only occurs
here and there, and is most marked in the longitudinal bundles (PI. VI. fig. 1 ). The
circular fibrous bands exhibit this peculiarity only in the upper region of the sponge.
I believe, however, that this peculiar condition of the bands of fibres is of essential
importance for the characteristic architecture of the sponge, and especially for the
nature and method of its growth. Since the spicules do not here become firmly bound
to each other, the elements of each individual band of fibres may readily become
laterally separated. If this occurs in every two parallel bundles by a simple division
both of the longitudinal and circular bands, perfectly simdar new rows of quadrate
meshes become formed, which are ari-anged strictly in longitudinal and transverse
directions. A few particularly strong fibres always lie in the axis of each of these
longitudinal and transverse bundles of fibres. These are the much prolonged rays of
those strong tetracts whose axial cross corresponds exactly to the intersections of the

1 Loc. cit., pL xii. fig. D.

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

buudles. The comitiilia, which for the most part compose the fibrous bundles, are, on the
other hand, very much prolonged triacts with short transverse rays ; more rarely they
are diacts or tetracts. Inferiorly and in the outer part of the longitudinal bands of
fibres numerous diacts always occur with abundant barbs and with inferior anchors on
whose knob-like thickened extremities an indefinite number of anchor teeth form a pro-
jecting wreath. I have not found pentact anchors such as occur in the basal tuft of
Euplectella aspergillum (Pi. III. figs. 23). Between the bundles of longitudinal and
transverse fibres, which intersect at right angles, oblique or spiral filjrous bands alone
occur. These are, on the whole,* somewhat more loosely compacted than those formed
from the tetract principalia just described, and consist almost exclusively of very long
thin triacts with short transverse rays, and more rarely in part also of diacts and
tetracts. Sometimes some sj^iral bands also extend over the outer side of the longi-
tudinal and the inner side of the transverse beams, or pass over some longitudinal
bands externally, then along their inner side between the longitudinal and transverse
beams for a certain distance, again passing outwards, and so on. Since each of the
quadrate principal meshes, which are from 3 to 5 mm. in breadth, encloses a circular
parietal gap, the obliquely crossed bands of fibres cannot lie along the diagonal line of
the meshes, but pass close to the parietal foramina, and assist in forming their margins.
In this way a woven mesh work is formed, which — as Marshall has pointed out — somewhat
resembles that of our common cane-bottomed chairs. The parenchymalia which lie between
the long spicules of the principal bands of fibres are for the most part triacts, pentacts
and diacts with slightly bent rays, also small normal hexacts and oxyhexasters less
regularly arranged. In general they resemble the corresponding structures of Euplectella
aspergillum so completely that I may refer to the description and figures given for that
species. It is only in the circular memlirane, which surrounds the parietal foramina, that
the parenchymal spicules — as has also been noted by Marshall — present a somewhat
difi'erent character, since in this locality and in the neighbourhood of the free margin
substantial spindle-like diacts with four or two median tubercles occur in great numbers.
These Marshall" has happily designated compass spicules. Passing from the outer
portion of the iris-like circular membrane towards the interior, all transitions from
the circular hexacts to those peculiarly deformed diacts may be observed. While
externally greatly prolonged curved diacts occur beside the isolated strong hexacts and
the numerous pentacts, further towards the interior it may be seen that tetracts and
triacts of the same strength predominate, while just at the innermost margin compass
spicules are almost exclusively found.

There may also be observed a system of slender dagger-shaped hypodermalia with

1 The long tetracts which are represented by Marshall in his diagraniatic figure {loc. cit., pi. xv. fig. 54, cc), as
principalia of the oblique spicular bands, I have sought for iu vain in this young specimen.

2 Loc. cit., pi. XV. fig. 57.

REPORT ON THE HEXACTINELLIDA. 81

greatly prolonged proximal rays, and distal rays which raise the dermal membrane into
pointed elevations, and bear on their extremities somewhat freely projecting floricomes.

Bundles of very fine raphide-like spicules, which lie close upon the outer ray of the
hypodermalia, appear to me to occur here in greater abundance than in Euplectella
aspergillum.

The gastralia are slender penta.cts with prolonged distal rays, while then- atrophied
sixth ray often appears as a rounded, more or less projecting tubercle.

The composition of the much arched sieve-plate does not differ essentially from that
of Euplectella aspergillum. In the basal root-tuft I found anchor-like structures
similar to those of the latter species.

Among the Japanese Hexactinellida of Dr. Doderlein there occurs — in addition to
the specimen of Euplectella oiveni upon which the foregoing description is founded —
a completely macerated and much injured, though coherent tubular skeleton, which may
also with probability be referred to Eup)lectella oweni. This forms a tube of 32 cm. in
length, somewhat compressed on one side. At the (3 to 5 cm.) wide extremity, the
spicules seem loose, but they are fused below into a firm lattice-work. Since not only
the position and arrangement of the bands of fibres, but also the structure of all the
larger and smaller spicules which I was able to isolate by tapping, fuUy agree with those
of Euplectella oiveni, I do not doubt that this specimen is simply a very large and old
fragment of Eup)lectella oweni, in which the usually unfused spicules have at a later stage
become soldered together.

6. Euplectella crassistellata, n. sp. (PL XIII. figs. 5-7).

In the middle of the Pacific (Station 274, lat. 7° 25' S., long. 152° 15' W., depth
2750 fathoms, bottom Radiolarian ooze) there were found some plate-like fragments
from 4 to 6 cm. square, and about 1 mm. thick, along with a narrow tuft of siliceous
spicules 6 cm. long, to which a small part of the plate-like mass was still firmly attached.
Round or oval apertures from 1 to 1"5 mm. in diameter occur here and there on the plate.

I must regard these fragments as a new species of Euplectella, although many
characters of this genus are not definitely indicated on account of the insuflicient
preservation of the fragments.

The principal spicules of the plate are represented by diacts which run out to a
point at both ends, and are usually provided with a node-like thickening in the middle.
These lie scattered without any recognisable regularity of arrangement, somewhat near
the inner surface, and parallel to it.

Besides these and some long thin diact comitalia which are applied closely to the thick
principalia, only a few extended spicules are to be found in the parenchyma. Rosettes,
however, and a peculiar form of oxyhexaster occur in surprising abundance. These oxy-

(ZOOL. CHALL. EXP. — PART LIU. — 188G.) Gfgg 11

ei4

..^(M^^U— »'- l.vuloutthocxtvomitics

.1 .1, hit, . M. Mil. hg. (.).

\\\» ^^((.)^ (I, ho other KupkH-telhdan

. ^,^v^^^v^ v4vv^ miv\sll. W »v thickno^ of the wall

'/, /^^

vv>

y^v^^^'-"' "^--

\ :»Uv\vvf»t exactly agret?
m^ «uxlv 0*008 mm.

\

. ;■ ffi»

M aat a

REPORT ON THE HEXACTINELLIDA.

The most impo
whose greatly proL
another in a naannc
described. The un]
There is not, hower
lattice- work. An ii
tudinal and transvci

Delicate and mi
closely apposed to i
fibres contain numc,
ing on their infer i'
(PL XIV. fig. 5).
marked externally 1
numerous spicules ^
scattered rosettes o<
work. Among the
long bundles of v(
occasionally seen pei

t parts of the quadrate framework consists of strong pentacts,
■d tangential rays are placed in apposition to, and across one
milar to that occurring in the species of Euplectellidse already
ed ray extends in a radial direction to near the external skin,
a special pentact at each of the intersections of the beams of the
section without a pentact in fact usually alternates, in the longi-
direction, with one which is so provided (PL XIV. fig. 2).
extended comitalia, consisting of diacts, triacts, and tetracts, are
strong pentact rays. Besides these the longitudinal strands of
; long diacts equipped with upwardly directed barbs, and exhibit-
xtremities a knob-like thickened anchor-head with three teeth
intersection of the axial canal lies above the latter, and is usually
bur barbs arranged in a cruciform manner. In addition to the
a smaller number of rays, delicate regular hexacts and sparsely
iinder the parenchymalia which lie outside the quadrate lattice-
ittes I found, near the external skin, some lophiohexasters with
fine, straight, perfectly parallel terminal rays, and I have also
tly unconnected oxyhexasters, but whether these belonged to this
sponge I was forced leave doubtful, though since then I have been led to regard this
as very improbable.

The absence of ii oxyhexasters, as are so abundantly present in other Euplectellid
species, is very stril , and may perhaps be explained by the fact that the specimen in
question is very yo , as it seems to me not impossible that certain forms of spicules
only originate at ; mewhat late period. Such characteristic spicules as the thick
pentacts of Etq:)lec aspergillum, or the compass spicules of Enj)lectella oiveni, are at
any rate entirely a t from the cii'cular membrane surrounding the parietal apertures.
In the dermal s ton the delicate dagger-like hypodermal hexacts with a prolonged
proximal ray, whi' re so characteristic of the family of Euplectellidse, exhibit the

id arrangement, and in most cases bear a floricome on the pro-
vIV. fig. 2). The floricome in no way difi"ers from that described
lum. On some of these hypodermal hexacts a bundle of freely
cts may be observed in place of the floricome (PL XIV. fig. 2).
;uished by four prongs which project cross-wise in the middle, by
outer and inner ends, and by fine lateral teeth or prickles. These
lateral prickles arc >wever, sometimes absent (PL XIV. figs. 3, 4). The gastralia are
simple pentacts wit rolonged radial distal rays.

As characteristic -eculiarities of this young sponge the following features may be
noted : — (1) the bunies of pointed diacts which project freely from the lateral wall and
are provided with mdian nodes ; (2) the three-toothed anchors of the longitudinal

familiar connection
jecting distal ray (
in Exiplectella aspi
projecting peculiar
These diacts are di
the sharp points on

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

fibrous bands, and (3) tlio lophioliexasters which lie near the outer surface, but yet
underneath the dagger-like hypogastralia. This form perhaps belongs to the genus
Holascus.

Genus 2. Regadrella, 0. Schmidt.
1880. 0. Sclimidt, Die Spongien des Meerbusens von Mexiko, p. 6L

Regadrella i^hcenix, 0. Schmidt (PI. XIII. figs. 1-4).

Although Regadrella phoenix, which has been described by Oscar Schmidt, is in
general characters closely allied to the genus Eiq^lecteUa, remarkable differences occur
which may well suffice for the establishment of the genus. We have here, as in the
Euplectella, to deal with a tube whose walls are much perforated by round apertures,
and whose transversely truncated superior extremity is covered by a watch-glass-shaped
arched sieve-plate, and bordered by a cuff-like wreath of spicules. The basal tuft is
entirely absent, and the skeletal framework consists not of longitudinal and transverse,
but of oblicjue, irregularly interwoven strands of fibres, while the rosettes scattered in the
parenchyma are essentially distinct from those of the various species of Euplectella.
The inferior extremity of the tube has become converted into a compact and substantial
cup by extensive fusion of the spicules. The cup is fixed by a knobby base on the
stony substratum, while towards the upper end it passes quite gradually into a
progressively looser spicular framework. After the death of the sponge the part of the
skeleton which is not united by siliceous matter becomes separated from the rest of the
body, but the basal part persists, and so admits of the occurrence observed by 0. Schmidt,
that several generations encapsule one within the other, the younger forms settfing within
the remnants of their predecessors.

Without entering upon a detailed description of all the individual forms of spicules, I
win confine myself to noting the differences between some Regadrella spicules and the
corresponding spicules of the genus Euplectella.

The spicules which project freely from the undulating curved margins of the terminal
sieve-plate are hexacts, whose prolonged free distal ray is equipj^ed with scaly or prong-
like protuberances (PL XIII. fig. 2).

AU the rosettes which are abundantly scattered in the parenchyma are distinguished
from the corresponding rosettes of Euplectella by the fact that their slightly bent
terminal rays, three or four of which S23ring from every short principal ray, do not run
out to simple points, but become divided at their narrowed extremities into four
transversely directed and cruciately disposed, hook-like, backwardly bent, fine prickles.
These forms should thus be designated not oxyhexasters, but rather discohexasters.
With regard to fig. 3 on PI. XIII., which represents a rosette of this kind from Regadrella

REPOET ON THE HEXACTINELLIDA. 85

phoenix, I may here note that the representation of the fine transversely arranged
prickles on the extremities of the terminal rays has not Ijeen successful.

The floricomes which are attached to the tips of the dagger-shaped hypodermalia
differ from those of Ewplectella aspergillum in their somewhat greater size and stronger
terminal rays, as well as in the longer claws of the terminal plates ; they thus more
nearly resemble those of Euplectella suherea (PL XIII. fig. 4).

Since no specimens of Regadrella i^hcenix occur among the HexactineUida collected
by the Challenger expedition, I have figured (PL XIII. fig. 1), from a photograph, a
specimen which was kindly given to me for examination by Professor Oscar Schmidt.

The localities for this species, as at present known, are, according to Oscar Schmidt,
the Barbados, in 221 and 288 fathoms of water, and Santa Cruz, in 248 fathoms, in the
Gulf of Mexico.

Subfamily 2. Holascin^ (PI. XIV. figs. 6-13; Pis. XV.-XIX.).

Euplectellidse in which the lateral wall is solid, that is, not perforated by parietal
gaps. The principal spicules are not fused together, and form with their longi-
tudinal and transverse rays a quadrate network. The hypodermalia are dagger-shaped,
and have a somewhat swollen distal ray beset with prongs, and frequently extended by
apposed pointed diacts, but in other cases probably bearing a floricome. The
hypogastralia are either simple pentacts without a proximal ray, or hexacts in which
the inward projecting (proximal) ray probably can bear a floricome.

Genus 1. Holascus, u. gen. (PI. XIV. figs. 6-13; Pis. XV.-XVIL).

Tubular in form, with a compact wall, the outer surface of which shows no pit-like
depressions, while a lattice-work of longitudinal and transverse ledges projects internally.
The upper transversely truncated extremity is provided with a thickened margin which
is destitute of a spicular wreath, and is closed by a compact sieve-plate, while the
inferior extremity runs out into a basal tuft. The network of strands which serves
for the support of the body-wall consists of greatly prolonged, longitudinal and trans-
verse rays of compact hexacts, pentacts or tetracts, which lie close to one another,
forming a tolerably firm framework, and which are also surrounded by a coating of thin
comitalia.

In the parenchyma, in addition to various other spicules, oxyhexasters occur, or instead
of these in other cases fibulas.

The sword-shaped h}T3odcrmalia bear no floricomes, but are externally extended by
apposed pointed and narrow diacts.

86 THE VOYAGE OF H.M.S. CHALLENGEE.

1. Ilolascus steUatus, n. sp. (PI. XIV. figs. 6-13; PI. XV. figs. 6-23).

Station 325, east of Buenos Ayres ; lat. 36° 44' S., long. 46° 16' W.; depth,
2650 fotlioms ; bottom, blue mud.

At the above locality, the specimen figured in PI. XV. fig. 6 was dredged, as also
the greatly injured fragment of a second specimen represented in PI. XV. fig. 14.
The two s^Decimens difi'er, moreover, not only in their size, but also in the individual
forms of the spicules, so that notwithstanding the general agreement, we have perhajjs
to deal with two difi'erent species.

In the former specimen the almost cylindrical tube, which is slightly widened in the
middle (PL XV. fig. 6) is 8 cm. long and 13 mm. broad. Upon the upper transversely
truncated margin, a terminal circular pad projects somewhat outwards, and this supports
transversely the stretched retiform sieve-plate which has been figured by Wy ville Thomson
(PI. XV. fig. 6), but which is no longer preserved in the object before me. From their
inferior extremity, which is only slightly narrowed, there extends, somewhat to the side, a
badly-preserved tuft of fibres about 2 cm. in length. The outer surface of the wall,
which is from 1'5 to 2 mm. in thickness, appears to the unaided eye to be roughened
by small tubercles, while, with the help of a lens, numerous small pointed elevations
showing a uniform distribution and a regular arrangement may be observed in transverse
and longitudinal rows. On the inner side pit-like depressions about 1 mm. in breadth
likewise occur.

The principalia, which constitute the groundwork of the quadrate lattice-like frame-
work of the tubular wall, are hexacts and pentacts with a distal radial ray. The long
longitudinal and transverse rays are so applied to each other that the transv^erse rays
come to lie altogether on the inner side of the longitudinally dii-ected rays. According
to the figure given by Wyville Thomson (PI. XV. fig. 7) it might be thought that the
lattice-like network lies close to the inner side of the wall of the tube, and that it
consists of nothing but pentacts. Such, however, is not the case. It extends, on the
other hand, between the inner and middle third of the thickness of the wall, and consists
for the most part of hexacts, between which pentacts with distal radial rays only here and
there appear. Numerous thin comitalia, with a variable number of rays, accompany the
transverse rays of the principalia. Near the margin of the body, outside the longitudinal
strands, long diacts also occur, with pointed upper ends, and with anchor-hooks on their
inferior extremities. These diacts are, as a rule, quite smooth above, but bear towards
the lower end barbs, which become gradually longer towards the end where the ray
passes into a conical pointed thickening, from which usually four, seldom more, anchor-
teeth project obliquely outwards and upwards (PI. XV. fig. 13). Projecting freely down-
wards for a greater or less distance they form the basal root-tuft.

Among the parenchymalia we must also mention the long, thin, terminally pointed

REPORT ON THE HEXACTINELLIDA. 87

diacts, wliicli are either scattered at random, or aggregated in bundles. Several forms of
rosettes also occur, and in greatest abundance, simple oxybexasters with short principal rays,
and with two to four straight or hook-like, moderately diverging, terminal rays on the
extremity of each of the principal rays (PI. XIV. fig. 13 ; PI. XV. figs. 9, 11, 12). The
number of the terminal rays may vary from one to three on each of the main rays (PI.
XIV. fig. 8; PI. XV. fig. 11). These remarkable skeletal elements from whose globular
centre four, three, or two rays spring, which are greatly bent in one plane, or even
spirally twisted, I regard as arrested derivatives of oxyhcxasters (PI. XIV. figs. 10-12;
PI. XV. fig. 10). In the outer part of the parenchyma, graphiohexasters with close
bundles of almost parallel fine terminal rays appear (PL XV. fig. 19).

Whether the elegant discohexasters represented in PL XIV. fig. 9, which are provided
with an S-like terminal ray, and with very small terminal discs, really belong to this
species, or are not rather intruded bodies, I have lately begun to doubt.

These and similar discohexasters also aj^pear in that fragment of Holascus which is
represented in PL XV. fig. 14, and the individual spicules are figured after WyviUe
Thomson in figs. 15 to 23 of the same plate.

The dermal skeleton consists of sword-shaped hypodermaUa with greatly prolonged
proximal rays which run out to sharp points at the extremities, with a thick scaly or
toothed distal ray, and with four transverse rays intersecting at right angles, obliquely
pointed at the extremity, and of median length. Upon these, as on the proximal ray,
small pointed elevations may often be observed.

The tangential arms of these hypodermalia always lie somewhat beneath the dermal
membrane, which is raised up in a conical elevation by the distal ray. Close to the outer
portion of the proximal ray, and over the whole distal ray, — even extending beyond the
outer extremity of the latter, — thin diacts are disposed which run out to points at both
ends. These may serve as defensive weapons in place of the floricomes which are here
absent (PL XIV. fig. 6).

The gastral skeleton consists of hexact sword-shaped hypogastralia, in general
resembling the hypodermalia, but somewhat more delicate. Thin diacts are here and
there apposed to the spicular rays, and are even more delicate than those of the outer
skin.

2. Holascus fibulatus, n. sp. (PL XV. figs. 1-5; PL XVI.).

A species of Holascus, markedly characterised by the possession of numerous double-
hooked fibulae, is included among the sponges of the Challenger expedition. Three
specimens were obtained, one (represented in PL XV. fig. 1) in a trawling to the south
Australia (Station 160, lat. 42° 42' S., long. 134° 10' E.), at a depth of 2600 fathoms,
from a red clay bottom ; while the other two were got in a trawling which was made in

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

a locality lialfway between Kerguelcn Island and the Cape of Good Hope — the one
(PI. XVI. fig. 9) (Station 146, lat. 46° 46' S., long. 45° 31' E.) from a depth of 1375
fathoms and a bottom of Globigerina ooze; the other (PI. XVI. fig. 1) (Station 147,
lat. 46° 16' S., long. 48° 27' E.) from a depth of 1600 fathoms, and a bottom of Diatom
ooze. Although none of these specimens is wholly uninjm-ed, the combination of all
the three affords a clear conception of the form and structure of the species. The foi-m
can be most clearly recognised from what is really the most macerated specimen, but
which is preserved in its entire length (PL XVI. fig. 9). This consists of a conical tube,
narrowed downwards and running out inferiorly into a basal tuft ; while the upper
transversely truncated extremity is bounded by a narrow marginal ridge and closed by a
sieve-plate which extends within the latter, and is arched slightly outwards in its
central portion. The specimen obtained at Station 160 (PI. XV. fig. 1) represents only
the much injured lower end with the basal tuft, while of the decidedly larger specimen
from' Station 147 (PL XVI. fig. 1) only the upper portion with the sieve-plate and the
relatively well-preserved soft parts persist.

The whole outer surface of the sponge exhibits the fine small points which occurred on
Holasciis stellatus. Here also on the inner side of the tube there extends a system of
intersecting longitudinal and transverse ledges which form quadrate meshes with central
pits (PL XVI. fig. 1). The framework of the terminal sieve-plate, though not quite
regularly constructed, exhibits approximately radial and circular strands of beams, from
the intersections and nodes of which small prickles project outwards (PL XVI. figs. 9, 10).

The principal framework of the wall of the tube lies towards the inner surface, and
consists of strong smooth tetracts. The somewhat long longitudinal rays cross the shorter
transversals externally. Numerous comitalia with a variable number of long narrow rays
are attached both to the longitudinal and transverse rays of the principalia (PL XV.
fig. 2 ; PL XVI. fig. 2).

On a transverse section of the somewhat thick wall of the tube (of the fragments
figured in PL XVI. fig. 1) a system of rough hexacts is observed, \\dth rays disposed in
radial, longitudinal, and transverse directions, and apposed to one another to form a
framework of beams enclosing cubical meshes (PL XVI. fig. 2). The outermost and
innermost of these hexacts correspond in position and direction with the hexact
hypodermalia and pentact hypogastralia, to the long parenchymal ray of which they are
symmetrically joined. The other ii-regularly scattered parenchjnnalia consist of isolated
graphiohexasters with bundles of long, very delicate, parallel terminal rays, and of those
characteristic fibulae, which I am inclined to derive from greatly reduced oxyhexasters
with bent terminal rays. One can frequently observe at some distance from the central
nodes and upon each of the two rays a boundary line, at which the straight central,
thicker segment passes into the thinner terminal portion (PL XV. fig. 3c, d). I regard this
straight, inner, thicker portion as corresponding to the principal ray, the bent outer

REPORT ON THE HEXACTINELLIDA. 89

portion, on the other hand, as equivalent to the terminal. The fact that the fibulae are
provided with bent terminal rays, instead of the oxyhexasters which are present so
abundantly in other species of Holascus, is also in agreement with my theory. It has
lately been shown that neither the oxyhexasters, nor the prickly small discohexacts,
represented in the diagrammatic section {PI. XVI. fig. 2), belong to the species, but have
been accidentally intruded.

The dermal skeleton consists of rough sword-shaped hexact hypodermalia with
greatly prolonged proximal rays, and toothed, slightly thickened distals, on which several
(frequently four) narrow pointed diacts are disposed, and with their outer extremities
extended for a greater or less distance beyond the points of the distal ray (PI. XVI.
fig. 2).

The gastral membrane is supported by the transverse rays of the rough pentact-
hypogastralia, whose long distal ray penetrates radially into the parenchyma.

The spicules of the basal tuft of fibres consist for the most part of very long diacts,
which are smooth on the outer pointed extremities, but are beset further down with barbs,
and bear at their inferior extremities a conical pointed knob, from the side of which three
or four strong anchor-teeth project outwards and upwards. The intersection of the axial
canal lies, as a rule, at some distance from the inferior extremities, and corresponds
usually with the four lateral prongs, which are cruciately arranged (PL XVI. fig. 11).
In some long spicules which run out to points on both ends, I have also found the axial
cross of the central canal in the middle, and the hook-like teeth so directed on the two
sides that their points were turned away from the centre (PI. XVI. fig. 12).

The skeleton of the tolerably compact terminal sieve-plates chiefly consists of strong
hexacts, which form quadrate meshes by the apposition of the transverse rays. Their
distal, freely projecting ray is spindle-shaped, thickened, and sparsely covered with small
conical teeth, while the short proximal is a simple cone, and entirely tubercular.
(PI. XVI. fig. 8). The four cruciate, long, transverse rays are smooth, and gradually run
out to a point. To these transverse rays, but more especially to the outwardly projecting
distal, diacts with pointed extremities are closely affixed (PL XVI. fig. 8).

Holascus ■polejaevii, n. sp. (PL XVII. figs. 1-5).

The single, probably young, specimen figured on PL XVII. fig. 1, represents a new
species of Holascus, which I dedicate tothe meritorious investigator of the Calcarea
and Keratosa, Dr. Polt^jaeff. This form was trawled to the south of Australia
(Station 157, lat. 53° 55' S., long. 108° 35' E.), from a depth of 1950 fathoms, and a
bottom of Diatom ooze. The specimen has been injured at the upper end, so that both
the terminal sieve-plate and the marginal ridge are wanting.

The sack-shaped, somewhat thin- walled body has a length of 10 mm., and a

^ZOOL. CHALL. EXP. — PAUT LIII. — 1886.) Ggg 12

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

breadth of 4 mm. From the narrowed lower end there arises a basal tuft, 4 "5 mm. in
length.

The spicules, which principally serve for the formation of the quadrate lattice-like
network of the parietal skeleton, are simple hexacts with smooth, frequently somewhat
bent rays (PI. XVII. fig. 2), to which some comitalia are here and there apposed. Of
loose parenchymalia there are present, in addition to some small regular hexacts, simple
thin diacts, and three different kinds of rosettes. Among the latter the well-known
oxyhexaster form, with three long, diverging, terminal rays on each of the moderately short
principal rays, is most frequent (PI. XVII. fig. 3) ; less frequently an oxyhexaster occurs,
from each of whose strong principal rays, which are expanded outwardly in petaloid
fashion, a bundle of eight or more strong, straight, moderately diverging, terminal rays
proceeds (PL XVII. fig. 4). Graphiohexasters only occur scattered through the outer
region of the parenchyma, and are provided with long bundles of fine, slightly diverging,
terminal rays (PI. XVII. fig. 2).

The sword-shaped hypodermalia have a long, often somewhat bent, proximal ray, a
slightly thickened, scaly, pronged distal ray, and four simple, smooth, pointed, transverse
tangential rays. Close to the proximal and distal rays, narrow diacts occur here and there
which run to a point on both ends, and exhibit central thickened knots. As in the other
species of Holascus, these diacts extend for a considerable distance beyond the extremity
of the hypodermalia (PI. XVII. fig. 5).

The hypogastralia are simple pentacts whose long distal ray may be applied to the
corresponding proximal ray of a principal hexact or of one of the hypodermalia (PL XVII.
fig. 2).

Of the basalia only the upper part is preserved, and this does not differ essentially
from that of the basalia of Holascus Jihulatus.

4. Holascus ridleyi, n. sp. (PL XVII. figs. 6-8).

In the neighbourhood of the Philippines (Station 211, lat. 8° 0' N., long. 121° 42' E.),
the dredge brought up from a blue mud ground, and a depth of 2225 fathoms, a portion
(about 5 cm. long, and 2 cm. broad) of the lateral wall of a Holascus, the siliceous spicules
of which though agreeing indeed in some points with those of Holascus iwlejaevii, are so
markedly difi"erent in others that it must be described as a distinct species. I shall name
this after Mr. Stuart 0. Ridley, the meritorious investigator of the Challenger Monactellida.

The principal skeletal framework consists of substantial, long, smooth-rayed tetracts,
forming the quadrate meshes of the side wall, and, closely ajiposed to the latter, delicate
comitalia with a variable number of rays (PL XVII. fig. 6).

The intersections of the longitudinal and of the transverse bands of fibres, which lie
more towards the interior, do not all contain the axial cross of a principal tetract. In

REPORT ON THE HEXACTINELLIDA. 91

fact, intersections with and without such a cross alternate quite regularly both in the
longitudinal and transverse direction.

The parenchyma is richly provided with oxyhexasters, each of the principal rays of
which bears four or three straight diverging terminals (PI. XVII. fig. 8). Here and there
I also found an oxyhexaster form with much curved terminal rays (PI. XVII. fig. 7),
though it remains doubtful whether this type really belongs to the sponge. On the other
hand, more frequently, and undoubtedly belonging to the sponge, characteristic hexact
spicules occur which, like those of Holascus polejaevii represented in PI. XVII. fig. 4,
bear on each of the strong principal rays (which are widened and petaloid towards the
exterior) a bundle of eight or more markedly diverging terminals. The terminal expansion
of the principal rays is even larger, and more sharply separated from the inner portion
than in the above-mentioned species.

The hypodermalia resemble those of Holascus polejaevii, and are, like the latter,
extended outwards by the longer apposed, thin, pointed diacts.

The hypogastraha on the other hand are not pentacts but slender hexacts, in which
the somewhat swollen proximal ray, projecting freely inwards into the gastral cavity, is
beset with small prongs.

Genus 2. Malacosaccus, n. gen. (Pis. XVIII., XIX.).

These sponges are saccular or tubular forms, with a flabby readily pliable wall,
whose outer surface appears uniformly even, while on the inner surface numerous,
larger or smaller, round openings of the efi"erent passages occur in irregular distribution.
The inferior extremity has not been observed, but the superior exhibits a wide opening
surrounded by a narrow smooth margin.

The principalia are represented by hexacts with long, thin, flexible rays, which are
disposed in radial, longitudinal, and transverse directions, becoming apposed to one
another, and partly interwoven to form a cubical lattice-work. Oxyhexasters and disco-
hexasters also occur in the parenchyma.

On the projecting rays of the sword-shaped hexact hypodermalia and hypogastralia,
floricomes occur. The epidermal floricomes are larger and stronger than the epigastral.

1. Malacosaccus vastus, n. sp. (PI. XVIIL).

In the south of the Indian Ocean, at a locality about halfway between the Cape of
Good Hope and Kerguelen (Station 146, lat. 46" 46' S., long. 45° 31' E.), there was
trawled, from a depth of 1375 fathoms, and from a bottom of Globigerina ooze, a flabby
plate which could be folded like a woollen cloth. Some fragments belonging to the

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

same form were also obtained. Three sides of tliis irregularly quadrangular plate, which
measured about 1 foot square, exhibited much torn margins, but on the fourth the natural
border was preserved. On the very probable supposition that this border, which is
about 40 cm. in length, represents the upper margin of the sponge, and that no great
portion is wanting, the whole form would be that of a sack or cup, whose upper
aperture would be about 13 cm. in diameter. At the much destroyed margin, opposite
to the intact border, and, therefore, probably the lower, the sponge is as much as 1 cm.
in thickness, and decreases gradually from this point to the upper somewhat sharp
margin. In order to convey an idea of the character of the outer and inner surfaces I
have figured a portion from the middle in its folded position (PI. XVIII. fig. 1). While
the outer surface appears uniformly fiat, and merely exhibits numerous subdermal
cavities of varied size, the inner surface bears numerous round excurrent orifices be-
longing to the efi"erent passages, and varying from 3 to 12 mm. in diameter. These are
irregularly distributed, at intervals of 5 to 15 mm., and show at the bottom that they
are ramified. Between these wide excurrent orifices a comparatively flat surface is seen,
only perforated here and there by small apertures.

The principalia forming the supporting framework of the entire sponge are hexacts
with long smooth and flexible rays,^ which are accompanied by long, thin, and very flexible
comitalia. The parenchyma also contains numerous hexacts, of median size, with tubercled
rays, further smaU hexacts with smooth rays, and finally rosettes of two kinds. Among
the latter the oxyhexasters, which are present in great abundance, first attract attention.
They bear on each of their short principal rays, which are somewhat exj)anded
outwardly, four long narrow terminals (PI. XVIII. fig. 10). The second form of rosette,
which belongs to the discohexaster type, is smaller, and bears, on the gently convex
transverse terminal disc of each principal ray, a bundle of about thirty delicate
terminals which become broader towards the outer extremities, and which are beset
laterally with numerous small barbs, and provided terminally with a four-pronged
transverse disc (PL XVIII. figs. 3, 5).

The hypodermalia of the dermal skeleton have a slightly spindle-like thickened
distal ray with scaly teeth, a long smooth proximal which runs to a point, and four
simple, smooth pointed, transverse rays (PI. XVIII. figs. 2, 8). Whether the large and
strongly-developed floricomes which occur very abundantly scattered on the outer
surface in many regions (PI. XVIII. fig. 4), are regularly attached to the distal rays of
the hypodermalia, as in the Euplectellinse, I cannot definitely assert, since I have not
found them in this position. It is, however, quite possible that, on account of the great
softness of the whole plate, they have been pushed out of their original position.

The hypogastralia are Likewise hexacts with a prolonged parenchymal ray, but the
freely projecting proximal is not thickened into a spindle, nor scaly and j)ronged, like
1 In PI. XVIIL fig. 2, the rays of these hexacts are represented much too short and too straight.

REPORT ON THE HEXACTINELLIDA. 93

the projecting distal of the hypodermalia. All the hypogastralia are somewhat uniformly
rough. It has not been determined with certainty whether the numerous floricomes which
are represented on PI. XVIII. figs. 6, 7, 9, and which appear abundantly on certain
portions of the inner surface, really adhere to the projecting proximals of the hypogastralia.
These inner floricomes difler from the outer, moreover, not only in their smaller size, but
also in the fact that the slightly bent terminals, which occur to the number of six or more
on every principal ray, bear small barbs in addition to the many pronged terminal plate
(PI. XVIII. figs. 6, 7).

2. Malacosaccus unguiculatus, n. sp. (PL XIX.).

To the south of Sierra Leone (Station 348, lat. 3° 10' N., long. 14° 51' W.) there was
collected from a depth of 2450 fathoms, and on grey mud, a saccular sponge, 3 cm. in
length and 1"5 cm. in breadth, with walls measuring from 1 to 2 mm. in thickness. The
inferior extremity was torn ofi", and the superior margin only preserved in an injured
condition. The outer surface is slightly rough, but on the whole uniformly even, while
the inner surface, on the other hand, is perforated by numerous canalicular orifices of
varied breadth (PI. XIX. fig. 1).

The principalia of the parenchymal skeleton are represented by slender hexacts with
long, narrow-, smooth and flexible rays, which are disposed radially, longitudinally and
transversely, and apply themselves by corresponding rays to one another, or to the
prolonged parenchymals of the hypodermalia and hypogastralia, so as to form a loose
network which permits of the bending and folding of the whole sponge-wall. The
parenchyma also includes numerous oxyhexasters with slender, straight, or slightly bent
terminal rays, of which three occur in most cases on each of the comparatively short
principals (PI. XIX. fig. 7). Delicate discohexasters also occur, with thin terminal rays
which are somewhat thickened outwardly, and which bear upon their outer extremities
a whorl of four or more backwardly bent, thin, transverse spines (PL XIX. fig. 3).

The hexact hypodermalia have a thickened distal ray beset with scaly teeth, and a
greatly prolonged proximal, which, like the moderately long transverse rays, is seldom
quite smooth, but is, as a rule, more or less richly beset with small pronged elevations
(PL XIX. figs. 2, 4).

The hypogastralia are very similar to, but are distinctly more slender than the
hypodermalia (PL XIX. fig. 2).

Floricomes are not found on the outer surface but appear here and there on the inner.
They bear three or four terminals on every principal ray. The thickened outer extremity
of every terminal ray runs out into an overhanging plate with two or three strong
terminal claws (PL XIX. figs. 5, 6).

Although I did not find these floricomes upon the inwardly projecting proximal ray of

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

the hypogastralia, but here and there even beneath the gastral membrane, I am still
inclined to believe that here too they have been inserted on the tips of the hypogastral
spicules.

Subfamily 3. T^geein^, F. E. Schulze (Pis. VII.-XL).

The wall of the saccular or tubular body is perforated by apertures of various
sizes, irregular in shape and arrangement. The lattice-like trabeculse of the skeleton
form, for the most part an u-regular network of partially cemented principal spicules.
The outer end of the distal ray of each dagger-shaped hypodermal hexaster bears
a floricome.

Genus 1. Tiegeria, u. gen.

Tsegeria pulchra, n. sp. (PI. VII.; PL VIII.; PI. XL figs. 1-3).

In the neighbourhood of the Fiji Island, Kandavu (Station 174c, Lit. 19° 7' 50" S.,
long. 178° 19' 35" E.), the trawl brought up, from a depth of 610 fathoms, on a bottom
of coral mud, an elegant Euplectellid, which is figured on PL VIL, after the restoration
made by Wyville Thomson from a somewhat damaged specimen.

The thin-walled saccular body, which expands somewhat above the middle to a
maximum diameter of 6'5 cm., exhibits a circular section, and has a length of 20 cm.
Near the lower blind sack-like end there is a compact, tangled, somewhat lateral, basal
prolongation which grows on the firm substratum. The upper extremity bears a circular
opening, 3 cm. in diameter, which is surrounded by a somewhat firm margin, and
overarched by a beautiful corona of long, curved, siliceous spicules which bend towards
the centre. The lateral wall of the body is only from 2 to 3 mm. in thickness,
and is penetrated by numerous irregularly scattered, approximately circular, tolerably
large parietal apertures, varying from 3 to 4 mm. in diameter. These gaps are disposed
at intervals of from 1 to 2 cm. in the median portion of the sponge, but are, above and
below, somewhat more widely apart. Between these larger orifices, smaller round pores
here and there occur, varying from 1 to 2 mm. in diameter. The numerous, light,
roundish spots, however, which may be observed in the spirit specimen, and also in the
figure on PL VIL, occurring on the external surface between the above noted apertures,
are neither holes nor pits from the outside, but represent pit-like hoUowings on the inner
surface which do indeed in many cases become, at a later period, artificially opened and
converted into canals which pass completely through the wall.

The larger beams of the supporting skeletal framework, together with the delicate
comitalia which surround them, are for the most part fused, by cementing matter

EEPORT ON THE HEXACTINELLIDA. 95

and synapticula (PL VIII. fig. G), into a continuous network, which seems to be firmest
on the inferior portion of the sack, becoming more and more loose towards the upper
end. AYhile the principal strands of fibres are irregularly disposed, for the most part
obliquely, on the inferior and middle thirds of the body, the upper third exhibits a more
regular arrangement of the longitudinal beams and circular transverse ridges.

The latter lie, as in all Euplectellidaj, on the inner side of the former. The strong
principalia of the lattice-work are in the regular upper portion almost exclusively
tetracts, in the larger and uu)i\' irregular inferior portion, on the other hand, triacts and
diacts predominate. The comitalia consist chiefly of triacts and diacts.

Closely disposed spicules, with from six to two rays, also occur in the parenchyma,
viz., somewhat numerous, small, smooth or spinose, regular hexacts, with rays running
out to fine points (PL XL fig. 2); more rarely small, smooth discoh exacts (0"17 mm. in
diameter), with from four to six hooks on the arched terminal discs, and finally, disco-
hexasters irregularly scattered in somewhat large numbers (PL XL fig. 1; PL VIII. fig. 4).
]\Iost of these discohexasters exhibit the same characters and dimensions as the discohexacts,
which have just been mentioned. The very short principal rays usually run out into two,
less frequently into three, long terminal rays, which are, at their origin, united in a bow-
like manner, and which, narrowing in the middle, again increase in strength towards the
terminal hemispherical disc, which bears six strong hooks (PL VIII. fig. 4). Hexasters
with somewhat longer and narrower terminal rays occur scattered at intervals, and on
their transversely truncated extremities, four to six small hooks project transversely.
In the neighbourhood of the outer surface graphiohexasters occur with long tufts of
parallel, straight, very delicate, terminal rays (PL VIII. fig. 5).

The sword-like hexact hypodermalia are delicate, and run out to fine points. To their
distal ray a floricome is always attached, which does not differ essentially from the well-
known floricome of Euplectella asj^ergiUum (PL VIII. fig. 3).

The gastral skeleton consists of somewhat stronger pentacts, in which both the
prolonged distal and the four tangentials are either transversely rounded ofi", or terminate
in a knob-like swelling (PL VIII. fig. 7).

The thickened margin which surrounds the upper terminal opening bears two circularly
arranged rows of hexacts with strongly developed distal rays. In the spicules of the
outer circle the straight distals, which measure from 2 to 3 mm. in length, are directed
obliquely outwards and upwards, while the slightly lient, strong, distal rays of the inner
circle, which are 1^ cm. in length, bend together in dome-like fashion over the opening,
so that only a central space, about 8 mm. in breadth, remains (PL VII.).

The folded layer of simple sacciform chambers surrounds clefts or short canals, which
either open freely by wide internal openings, or are arched over by a continuation of
the gastral skin (PL XL fig. 1). The inner trabecular framework occurs in a thin layer
on the inner side of the eff'erent passages which lead from the wide openings of the

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

chambers, while the outer trabecular framework either entirely fills the space between
the chambers and the outer skin, or is perforated by the subdermal lacunae and afferent
canals.

The name of the genus Twgeria is formed from the family name of my wife, " Taeger."

Genus 2. Walteria, n. gen.

Walteria jlemmingii, n. sp. (PI. IX.; PL X.; PL XL figs. 4-6).

The fragment of a hitherto unknown Hexactinellid, which is very faithfully
represented in its natural size on PL IX., was collected in the Pacific to the north of the
Kermadec Islands (Station I^Oa, lat. 29° 45' S., long. 178° 11' W.), from a depth
of 630 fathoms, on volcanic mud. The sponge apjiears to be broken across the middle,
and to have had about double the length of the fragment preserved, which is 15 cm.
long by 8 cm. broad, becoming narrowed towards its extremity into a conical funnel-
shaped tube only 6 mm. in width. Since the terminal portion is also broken off at this
place it is impossible to determine whether w^e have before us the inferior extremity by
which the sponge was fixed or the free upper end.

The thin wall of the tube consists of a framework wath strands of varied strength,
seldom more than 1 mm. in thickness, and surrounding polygonal meshes of very diverse
size and irregular form, which may attain the size of 1 cm. It is only towards the
narrowed funnel-like end that the strands arrange themselves in a system of somewhat
converging longitudinal and transverse bands of fibres, which come to lie closer and
closer to one another, forming quadrangular meshes, \\'hich are only crossed by small
secondary beams. Here and there in the expanded main portion of the sponge longitudinal
and transverse beams may also be recognised. Most of the beams in the lattice-like
network are not smooth, but are beset with small knobs which project externally,
and which occur sometimes isolated, sometimes in grouped arrangement. Even on
examination with the naked eye, but more obviously by employing a lens or the
microscope, the network strands may be seen to be supported by a firmly united sUiceous
framework, which is covered by a distinct and strongly-developed soft skin. The latter
becomes elevated into numerous small short tubes which project obliquely or at right
angles to the surface, and are each tenanted by the hydranth of a commensal hydroid
poly|5e. While the hydrophyton extends into the deeper layer of the rind, the shortly-
stalked or sessile hydranths project transversely outwards (PL XI. fig. 4). There can
be no doubt that the hydranth, by its simple presence and especially by the continual
stimulus exercised on the surrounding sponge substance by its expansion and
contraction, has caused the formation of the projecting tubes. I have not observed
anything of the nature of a perisarc. The hydroids seem to be entirely naked, and lie

REPORT ON THE HEXACTINELLIDA. 97

in simple lacunte, which arc, for the most part, formed by a rupture of the inner tra-
becular framework. Wherever a hyclranth arises, the chamber layer, the outer trabecular
framework, and the outer wall, become perforated by a canal opening to the exterior.

The chamber layer is but slightly folded, a circumstance which is in harmony with
the slight thickness of the whole body- wall.

The principal lattice-like framework of the skeleton consists of long compact spicules
with a varied number of rays, but especially triacts and diacts, which, with the more
delicate comitalia that surround them, are, for the most part, disposed in bundles, and
firmly united by means of numerous synapticula (PI, X. figs. 3, G). Tlie circular beams
lie as before on the inner side of the longitudinal, while the oblique strands are
irregularly interwoven throughout the entire lattice-work. In the narrowed funnel-
shaped terminal portion of the sponge the spicules of the framework are more delicate,
shorter, and more amalgamated.

The looser parenchymalia are in part extended spicules among which diacts seem to
predominate, which are provided with conical or rounded rough extremities, and in part
small delicate hexacts with fine transverse prickles, and lastly rosettes of two kinds. The
one type includes discohexasters with few, usually three, long terminal rays on each of
the six short principals. Where the principal ray divides an irregular tubercular
thickening occurs, and from this the somewhat distant, narrow, diverging terminals which
separate from one another project outwardly, while the extreme ends bear small discs with
from four to six transversely disposed, inwardly bent, thin hooks or claws (PI. X. fig. 5).

The other rosettes are, it is true, likewise discohexasters, Init they may be distinguished
from those just described in difierent respects.

The entire appearance is essentially diff'erent since, besides the stellate, the spherical
form also prevails (PL X. fig. 1 ; PI. XL fig. 4), on account of the great number of equally
long terminal rays which bear hemispherical transverse terminal discs. Each of the six
short principal rays passes at first into a discoid expansion, which bears on its arched
outer surface numerous (about thirty) terminals, which increase somewhat in strength
towards the exterior, and are disposed in a radiating and divergent fashion. The hemi-
spherical terminal discs of these terminal rays have a sharp-pronged margin which extends
inwards in a somewhat campanulate manner. In some cases the number of the terminal
rays on these rosettes is less, each principal ray bearing only about seven terminals. It is
noteworthy that the rosettes, which have on the whole a spherical appearance, and are pro-
vided with hemispherical terminal discs, always occur only in the neighbourhood of the outer
skin (PI. XI. fig. 4), whereas the stellate forms which are provided with transversely disposed
terminal tubercles are, on the other hand, scattered throughout the whole parenchyma.

Whether the bundles of very fine rhaphide-like spicules, which I found here and there in
the parenchyma in the neighbourhood of the outer surface, are to be regarded as the broken
off terminal rays of graphiohexasters or as independent groups of spicules, I cannot

(ZOOL. CHALL. EXP. — PAItT UII. — ISSG.) Ggg 13

98 THE VOYAGE OF H.M.S. CHALLENGEE.

with certainty determine, since I was never able to find a completely intact graphio-
hexaster. I am, however, inclined to the former opinion, since I have always found a
number of these rhaphide bundles together, and in the neighbourhood usually a small six-
rayed cross, with a discoid expansion of the cylindrical principal rays, from which numerous
small points projected like the broken ends of fine terminal rays. The whole exactly
resembled the middle portion of a graphiohexaster whose terminal rays had been broken off.

The sword-like hexacts of the dermal skeleton are distinguished by the strength of
the four tangential and of the distal rays, all of which terminate in rounded off extremities,
while the j^rolonged proximal gradually decreases in diameter towards the extremity,
and finally terminates in a conical point. The distal ray corresponding to the hilt of the
sword bears, sometimes in the middle or towards the outer extremity, a club-like
thickening (PI. XI. fig. 6).

On the distal ray of most of the hypodermalia a fioricome occurs, which is distinguished
from those of Euplectella aspergillum by the greater number (fifteen) of the S-like
curved terminal rays on each principal, and by the wider cup-like form of each whorl of
terminals (PI. X. fig. 2).

A special gastral skeleton is always absent where the large beams of the lattice-work
directly form the inner wall. Where this is not the case, it consists of simple hypogastral
pentaets, with four equally long intersecting rays, lying on the parenchymal side of the
gastral membrane, while the prolonged unpaired fifth ray penetrates widely into the
parenchyma (PI. XL fig. 2).

The extremities of all the five rays are simply rounded or slightly knobbed ajid
thickened, only the prolonged distal is sometimes narrowed towards the extremity,
and may even terminate in a point.

I have named this memorable sponge in honour of my esteemed friend and colleague
Professor Walter Flemming of Kiel.

Since the single and much damaged specimen of Walteria flcmmingii is inhabited
by many comparatively large commensal hydroid polypes, which have doubtless pro-
duced the peculiar tubules running obliquely or at right angles to the peripheral strands
(PI. X. figs. 4, 6 ; PL XL fig. 4), the question arises whether we have here to do with a
normally formed individual, or not rather with one essentially modified. In specimens
without these invading polypes, the wide lumen of the parietal gaps may be in
great part filled up by the soft portion of the body-wall, which is here restricted almost
exclusively to a cortical layer for the supporting beams.

I know of at least one case in another group of sponges, namely, that of a
Myxilla, found abundantly near Trieste, which is normally a compact bulbous body, but
which when invaded by Stephoscyphus mirabilis becomes bush-like, resembling a tuft
of the common heath Calluna vulgaris.

REPORT ON THE HEXACTINELLIDA. 99

The following seven genera are not yet sufficiently known to enable their systematic
position to be determined, but they all appear to belong to the family Eu2)lectellida), and
hence are treated of here.

Genus 1. Ilabrodictyum, Wyville Thomson.
This genus contains only one species, Ilabrodictyum speciosum.

Habrodictyum speciosum (Quoy and Gaimard).

1833. Quoy and Gaimard, Voyage de TAstrolabe, Zool., vol. iv. p. 302.

1836. Deshayes and Milne-Edwards in Lamarck, Animaux sans Vertfebres, vol. ii. p. 586.

1849. Owen, Trans. Zool. Soc. Lond., vol. iii. p. 205.

1865. Bowerbank, British Spongiadse, vol. i. p. 174.

1 866. Gray, Ann. and Mag. Nat. Hist, ser. 3, vol. xviii. p. 487.

1867. Gray, Proc. Zool. Soc. Lond., 1867, p. 530, 531.

1868. "Wyville Thomson, Ann. and Mag. Nat. Hist., ser. 4, vol. i. p. 114.

1869. Bowerbank, Proc. Zool. Soc. Lond., 1869, p. 323.

1872. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. ix. p. 442.

1873. Carter, Op. cit., ser. 4, vol. xii. p. 349.
1873. Gray, Op. cit., ser. 4, vol. xiii. p. 284.

Quoy and Gaimard ^ gave a description and figure of a form designated Alcyoncellum
speciosimi which differed essentially from any sponge then known. The genus Alcyon-
cellum had been some years before, 1830, defined by Blainville'^ in the following manner : —
" Corps fixe, mou, subgelatineux, soUdifie par des spicules tricuspides, phytoide, a
branches peu nombreuses, cylindriques, fistulaires, terminees par un orifice arrondi, a
parois epaisses, composees de granules reguliers, polygones alveoliformes, perces d'un pore
k I'exterieure et a I'int^rieure."

The generic definition is followed by this note ^ : — " Espece L'alcyoncelle specieux,
A. speciosum, Quoy et Gaimard, Zoologie, Astrolabe, msc." As the following " obser-
vation " adds : — " Ce genre a ete etabli par Mm. Quoy et Gaimard pour un corps
organise, rapport^ dans leur dernier voyage et qu'ils ont bien voulu soumettre h notre
observation. Quoique sa forme rappele un peu celles des cellaires, il est cependant
evident, que c'est aupres des alcyons et des eponges, qu'il doit etre place. Mais ensuite,
pour determiner, si c'est un alcyon proprement dit ou un Spongiaire, il faudrait savoir.
si chaque grain celluliforme contient un polype ; toutefois, comme cela nous paralt peu
probable, nous nous sommes determine h en faire un faux alcyon ou un spongiaire."

The description of Alcyoncellum speciosum, Quoy and Gaimard, is, on the other
hand, just as published in 1833 by Quoy and Gaimard,* for the genus {Alcyoncellum) : —

1 Voyage de la Corvette rAstrolabe, Zool., 1883, vol. iv. p. 302-303, and Atlas, vol. ii., Zoophytes, pL xxvi. fig. 3.

2 Diet. d. Sci. Nat., vol. Lx. p. 492. 3 Blainville, loc. cit, p. 492.
* Voyage de I'Astrolabe, Zool., p. 302.

100 ' THE VOYAGE OF H.M.S. CHALLENGER.

" Spongiaire, lamelleux, dont la cliarpcnte est formee de filets tres delies, accoles les uns
anx autrcs et cntrecroises de maniere h former des mailles nombreuses, arrondies, assez
reguliferes et semblables h. celles d'une dentelle," and for the species {Alcyoncellum
spcciosum) : — " Alcyoncellum cylindricum, cavum extremitate rotundum, album, reticulis
lapidicis elegantissime contextum." " Cette singuliere production donnant lieu au
genre ci-dessus represente un cylindre creux de sept a huit pouces detendue, en forme,
de phallus, arrondi et un peu dilate h, une extremity, ouvert d I'autre, k parois minces,
formees de filets tres delies, lachement accoles les uns aux autres, entrecroises dans tous
les sens de maniere a former des nombreuses mailles arrondies, presque regulieres
comme celles de la dentelle ou bien des sieges tisses en rotang. Ce qui fait, que tout la
masse est a jour. En voyant I'dlegante lilancheur et la r^gularite d'un tel tissu, on a de
la peine a se persuader qu'il est le produit d'une reunion d'animaux. On aime mieux en
voir un seul au fond de la mer travailler a se faire ce logement pour un but quelconque,
en tirant de sa propre substance, comme le font certaines chenilles, la matiere, qui se
petrifie aussitot qu'elle est en contact avec I'eau. Ce zoophyte habite, nous a-t-on dit, de
grandes profondeurs d'oii il a ete a,mene jDar une sonde. Les eclats qu'on remarque a
une de ses extremites indiquent qu'il doit etre fixe. Nous le devous a M. Merkus
gouverneur des Moluques, qui s'est plu a favoriser avec la plus grande obligeance nos
recherches d'histoire naturelle pendant le temps que nous avons passe dans les iles qu'il
administre."

In Blainville's work,^ besides a repetition of the generic diagnosis of Alcyoncellum
above quoted, which is also contained in the Diet. d. Sci. Nat.,^ these words also
occur : — " Espece I'Alcyoncelle gelatineux, A. gelatinosum, Quoy et Gaimard," and then
there follows the same "observation" which has already been quoted above (p. 99)
from the Diet. d. Sci. Nat. But the figures of Alcyoncellum gelatinosum which are
given by Blainville '^ undoubtedly represent a calcareous sponge belonging to the family
of Sy cones (Haeckel).

In the second volume of Lamarck's Histoire naturelle des animaux sans vertebres,
which appeared in 1836, Milne-Edwards stated the characters of the genus Alcyoncellum
in the following words : — " Sj^ongiaire lamelleux, dont la charpente est formde de filets
tres delies, accoles les uns aux autres et entrecroises de maniere a former des mailles
nombreuses arrondies, assez regulieres, et semblables a celles d'une dentelle." In this he
evidently refers exclusively to that siliceous skeleton which Quoy and Gaimard had
figured and described as Alcyoncellum speciosum, without even mentioning the older
species Alcyoncellum gelatinosum, Blainville.

While Max Schultze, Bowerbauk and Gray, as already mentioned, regarded Alcyon-
cellum speciosum, Quoy and Gaimard, as identical with Eup>lectella aspergillum, Owen,

1 Manuel d'ActiuoloKie, 1834. ' Vol. Ix. pp. 492, 529.

^ Manuel d'Actinologie, Atlas, pi. xcii. fig. 5.

REPORT ON THE HEXACTINELLIDA. 101

and while Bowerbank included in his genus Alcyoncellum, in addition to the well-known
species Alcyoncellum aspergilliim, the two species designated Alcyoncellum corhicula
and Alcyoncellum rohustum, Gray separated the genus Alcyoncellum from Euplectella,
and divided it into two genera, Corhitella and Heterotella.

Corhitella, which included the single species (figured) Corhitella speciosa, was
characterised thus : — " The tube clavate, rather irregular, rounded at the end, formed of
slender fascicules of open elongate filiform spicules, placed in longitudinal transverse and
oblique directions, forming an irregular network ; " the genus Heterotella, on the other
hand, with the single species Heterotella corhicula, is defined as follows : — " The tube
short, rather irregular, conical, truncated, irregularly netted. Skeleton formed of thick
bundles of very numerous slender spicules, placed in all directions, and forming an
irregular network, similar to the network of the lid of Euplectella."

Soon after this, however, AVyville Thomson, in his renowned contribution On the
Vitreous Sponges,^ again united the two genera of Gray, Corhitella and Heterotella, into
one, on account of the great similarity of their form, structure, and siliceous spicules, and
named the genus Hahrodictyon. His characters for this new genus are as follows :' —
" Sponge body subcylindrical, tubular, attached by a slightly contracted base. The
walls of the tube composed of a perfectly irregular network of bundles of siliceous needles
loosely and irregularly arranged in sheaves crossing one another at low angles, and
connected by a small quantity of soft mucilaginous sarcode. The spicules of the skeleton
all essentially of the hexradiate form, free and separate from one another, or rarely
connected in groups of two or three. The spicules of the sarcode numerous ' floricomo-
hexradiate stellate,' and various simple and branched modifications of the hexradiate
type." The two forms Hahrodictyon speciosuni and Hahrodictyon corhicida were
examined afresh in respect at least to the skeletons which had already been studied by
Bowerbank and Gray, and which are preserved in the Museum of the Jardin des
Plantes in Paris. One of these which bears the museum label "Alcyoncellum
coriicrJa, Valeneienne, donne par M. Saches 1857," and which was u?im.edi Heterotella
corhicula by Gray, exhibits a beaker-like inferiorly anchored tube, 10 cm. in
height and 5 cm. broad in the superior part, while its terminal opening is closed by a
transverse sieve-plate. The lateral wall of the tube consists of a very irregular
network in which the beams form irregular, roundish meshes, and are composed of loosely
united rod -like spicules varying in length up to 15 mm. No definite arrangement of
the beams of the network in transverse and longitudinal bands is exhibited, nor is there
any indication of externally projecting ridges.

While the long rod-like spicules, which are somewhat thickened and rough at both
ends, merely exhibit in the middle four cruciately disposed lateral tubercles, as indications
of the six-rayed type, numerous well-developed six-rayed sj^icules also occur, which are,

1 Ann. and Mag. Nat. Hist, ser. 4, vol. i. p. 114, 1868. * Loc. dt, p. 126.

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

in p;n't, quite regularly formed with six long, smooth, conical rays, and in part more
irregularly with bent and knotted rays. Wy ville Thomson describes as peculiar " flesh
spicules " the small regular hexradiate forms which cover the skeletal framework and are
abundantly scattered throughout the soft parts, and also the elegant structures which
were termed " floricomo-hexradiate stellate" by Bowerbank, in which each of the six
principal rays is continued into a small, outwardly curved, richly pronged, terminal plate,
and into a petaloid system of small branches which are bent in an S-shaped manner, and
widened at their extremities. The network of fibres forming the terminal sieve-plate,
which extends all round the sharply truncated lip-like upper walls of the tube, consists of
the very same elements as the wall of the tulie, but appears more closely woven, and
exhibits large six-rayed spicules between the rod-like forms, which are in general some-
what short. Besides this completely preserved skeleton, which has been figured by
Wyville Thomson from a photograph, the JMuseum of the Jardin des Plantes in Paris is
said to contain a second injured specimen and a fragment of a third.

As to the other form, designated by Gray Corhitella speciosa — Habrodictyum
speciosum, Quoy and Gaimard — Wyville Thomson was able to study only the single
specimen contained in the Museum of the Jardin des Plantes, and labelled ' Alcyon-
cellum corhicula Val. tir^ par 80 brasses de profondeur dans la rade de St. Denis de
Bourbon par M. Leschenault 1819." This is probably identical with the specimen
first described and figured as Alcyoncellum speciosxim by Quoy and Gaimard in their
great work. It exhibits a lattice-like tube 19 cm. long, which gradually widens from a
base which, when fully grown, is 32 mm. broad, to a free upper extremity which is
60 mm. in breadth. The lateral wall of the tube does not, as in HeteroteUa corhicula,
end in a sharply truncated upper margin, nor become closed by a flat transverse sieve-
plate, but, without changing in texture, bends inwards, and so forms a curved arch
which serves to close the tube above. While the rod and spindle-like spicules which
compose the lattice-like framework, and also the larger six-rayed sj^icules, resemble the
corresponding elements of the other species except in a few differences in size, and while
both the above-mentioned small, stellate six-rayed forms and "floricomo-hexradiate
stellate spicules " recur in similar form and size scattered through the soft body, there is
further a very abundant occurrence of a flesh spicule, which has not been certainly observed
in Habrodictyum corbicida, namely, a small hexradiate form with its branches divided
longitudinally in a forked or trifid manner. It was the great abundance of this form,
which Bowerbank tenned a " bifurcate rectangulated hexradiate spicule," which mainly
induced Wyville Thomson to regard Habrodictyum speciosum, Quoy and Gaimard, in every

' The circumstance already recorded by Gray (Ann. and Mag. Nat. Hist., ser. 4, vol. i. p. 173) is striking, and one
by no means easy to reconcile with the above supposition, namely, that in the Voyage de 1' Astrolabe Quoy and Gaimard
note that their specimen was presented to them by Mr. Merkus, governor of the Moluccas. In order to reconcile the
two statements it must be accepted that the elegant form which Mr. ilerkus presented did not come from the Moluccas,
but from the Isle of Bourbon.

REPOET ON THE H EX A CTLN'ELLIDA. 103

Other respect closely related to Hahrodictyum corhicula, as a quite distinct and separate
genus from the latter, and further, the fact that in Hahrodictyum speciosum no terminal
sieve-plute is present, led Wyville Thomson to object to the opinion which had been
expressed by Boxverbank,' to the effect that in Euphctella and allied sponges " the
openings of the Ud and those of the tube stand to one another in the relation of oscula
and pores." He maintains that on the contrary " each of the large openings of the wall
is occupied by an exhalent orifice and that inhalation takes place as usual by minute
pores in the interstices between the spicules of the skeleton."

Carter says in his treatise On flexactinellidse, ■ " Alcyonellum speciosum and
Alcyonellum corhicula appear to me to belong to one and the same species," which
opinion I assent to.

The genus Hahrodictyum probably also includes Eudictyiim elegans, Marshall, which
Marshall has briefly described from a hoUow, canal-like form (presers'ed in the Museum of
the Amsterdam Zoological Garden), agreeing with Hahrodictyum speciosum, Quoy and
Gaimard, in the form and structure of its irregular, lattice-like skeleton, though
exhibiting in the interior a looser spicular work (of interstitial connective tissue —
Flockengewebe). The latter consists, according to Marshall, chiefly of long (up to 1 cm.)
rod-like spicules, between which slender, sword-shaped hexradiate forms occur, besides
more delicate six-rayed and five-rayed spicules, with irregulai'ly developed prickles,
and further, small crowded forms, with six, five, four or three rays, which, like the two-
rayed so-called compass spicules, probably form wreaths round the dermal ostia.
Particularly characteristic are the six-rayed spicides, 0'2 to 0"3 mm. in axial length,
which bear on the extremity of each ray a disc with seven prongs. Marshall's bristle-
like spicules also occur, as well as the well-known " floricomo-hexradiate rosettes," which
do not differ essentially fr-om those of Euplectdla.

In the memoir which appeared iu 1876,^ Marshall characterised the family of the
Euplectellidse as follows : — " Pollakid Hexactiuellidse of tubular form, monozoic, osculum
closed by a sieve-plate; in the_ parietal tissue longitudinal, circular and spiral bundles.
Dermal skeleton chiefly formed of hexradiate spicules, between which lie the smaller dermal
pores. Large crateriform elevations of the parietal tissue also occur. There are also a con-
siderable number of dermal ostia which lead directly into the body ca^-ity, and which may
be closed by compass spicules. The rosettes belong to the ' floricomo-hexradiate ' type."

In the genus Euphctella, Marshall notes two species, namely, Euplectella aspergillum,
Owen, and Euplectdla oweni, Marshall and Herklots, and in the genus Hahrodictyum,
Wyville Thomson, only the single species Hahrodictyum speciosum, Quoy and Gaimard,
leaving it imdecided whether the sponge {Eudictyum elegans) he had described, from the
Amsterdam Museum, is an independent form, or belongs to the Hahrodictyum speciosum.

I British Spongiadfe, vol. i. pp. 176, 177. ^ Ann. and Mag. Nat. Hist., ser. 4, vol. rLL p 368.

^ Ueber die Yerwandtschaftsverhaltnisse der Hexactinelliden, Zeitschr. f. iciss. ZooL, Bd. .\xvii. p. 128.

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

Although I ha\'e no specimens of this genus Ilahrodictyum for special examination, it
follows from Wyville Thomson's above quoted (p. 101) generic diagnosis and figures ' that
the two species distinguished by Wyville Thomson, Habrodictyum corbicida, Valenciennes,
and Habrodictyum speciosum, Quoy and Gaimard, are very closely related to Regadrella
phasnix, Oscar Schmidt. Whether the agreement goes so far that the two species must
be included in a common group I cannot decide, though I regard this as by no means
improbable. Wyville Thomson was certainly right in uniting Gray's two genera
Heterotella and Corbitella, and in entirely dropping the generic name Alcyoncellum ,
which really referred to a calcareous sponge.

Habrodictyum agrees with Regadrella in this, that the skeletal framework of the
tube-wall is formed of an irregular network of fibrous bundles, which in the upper part
are only united by means of the soft parts of the body, but which towards the somewhat
diminished inferior extremity become gradually more and more firmly welded together
by siliceous matter, and finally pass into a knotted base which grows directly upon the
solid substratum. The irregular distribution of the parietal apertures is characteristic of
both genera. The spicules described and in part figured by Wyville Thomson do not,
on the whole, differ much from the spicules of Regadrella. Only the rosettes, which are
scattered throughout the parenchyma, and which m Regadrella are provided with a trans-
verse terminal cross belonging to the terminal rays, exhibit in Habrodictyum speciosum
true oxyhexasters with proportionately long principal rays and shorter pointed terminals.

While in Habrodictyum corbiculct the wall of the tube is affirmed to be sharply
separated from the transversely disposed terminal sieve-plate by a lip-like margin, such
a separation is entirely wanting in Habrodictyum speciosum, since the lattice-like net-
work of the tube-wall passes directly and without change into the gently arched terminal
plate. Whether Habrodictyum corbicida contains the parenchymal oxyhexasters which
are so abundantly present in Habrodictyum speciosum, has not been determined with
certainty.

Genus 2. Eudictyum, Marshall.
This somewhat doubtful genus contains only a single species, Eudictyum elegans.

Eudictyum elegans, Marshall.

In his investigations into the Hexactinellida ^ Marshall has described, under the
name Eudictyum elegans, a specimen in the Museum of the Amsterdam Zoological
Garden, which is perhaps identical -^-ith the above-mentioned Habrodictyum speciosum
of Wyville Thomson, and at any rate, very closely allied to it. According to Marshall
the tissue of the wall of the hollow club-shaped sponge shows longitudinal and trans-

• Ann. and Mag. Nat. Hist, ser. 4, vol. i. pi. iv. ^ Zeitschr.f. wi^s. Zool, Bd. xxv. p. 211, 1875.

REPORT ON THE HEXACTINELLIDA. 105

verse bundles, wliicli run irregularly ami usually form anastomoses ; the dermal ostia
are usually irregularly distributed between them. At the uppci' end is a strongly
injured sieve-like plate wliicli is not so distinct from the tissue of the wall as in
Euplectella. This trabecular network seemed to Marshall to consist of unsoldered
spicules. The loose mass of spicules lies on the inner side of the lattice, and consists of
uniaxial spicules 1 to 10 mm., long, slender daggers, weak five or six-rayed forms with
irregularly developed rays, very small spicules with several rays (three to six), spicules
with compressed rays and similarly formed diacts which are like a compass needle.
Hexacts with axes, from 0'2 to 0'3 mm. long, which bear at the end of each ray an
umbel with seven rays, are especially characteristic. Finally, there were found five-
rayed spicules and floricomes which could not be distingaiished from those of
Euplectella.

Genus 3. Dictyocalyx, n. gen.
This genus contains only the one species, Dictyocalyx gracilis.

Dictyocalyx gracilis, n. sp. (PL XII. figs. 1-7).

The framework of siliceous beams, which is shown in its natural size, from, a
photograph, in PL XII. fig. 1, was trawled in the South Pacific (lat. 22° 21' S., long.
150° 17' W.) from a depth of 2385 fathoms, and a red clay bottom (Station 281).
From a compact conical basis, which has been attached to some solid body by a basal
surface of 5 mm. in breadth, and which is narrowed upwards to a diameter of 3 mm.,
there arises an irregular retiform framework of beams, resulting in a cup-like form, 25 mm.
in length, and about 18 mm. in width above. One of the sides appears to have opened
inferiorly, and to have been again closed above.

The beams of this framework, where they spring from the massive base, measure
from 2 to 3 mm. in thickness, but become thinner upwards by gradual ramification.
They consist of greatly jirolonged spicules, which vary in the number of their rays, but
which are for the most parts diacts, cemented externally in a cpiite irregular fashion. In
the meshy conical basal portion numerous hexacts occur, soldered between the larger
beams.

After a more careful examination of the entire specimen I detected in various places,
but especially at the points of union of several intersecting beams, small patches of an
adherent soft substance which partly covered the beams. Although such insignificant
remnants of the soft body were no longer available for sections, it was still possible to
detach them in small fragments from the lattice-like framework, and to detect in them a
number of isolated spicules, which throw at least some light on the true character and
systematic position of the sponge.

(ZOOL. CHALL. EXP. — PAHT LIII. — 1886.) Ggg 14

106 THE VOYAGE OF H.M.S. CHALLENGEPv.

In the fir.st place, sleiulor sword-shaped hcxacts occur in great numljers. In these
both the prolonged ray and the four rectangularly intersecting transversals are smooth,
and terminate in a truncated point, \Yhile the sixth ray opposite the former is thickened
in its middle portion, and beset with outwardly-directed prongs. I am inclined to
believe that these sword-shaped hexacts represent the hypodermalia of the dermal
skeleton, and that they bore on their distal ray those iioricomes which were found here
and there, though indeed very much scattered, and which are figured in a fragment in
PI. XII. tig. 7. This lioricome form is distinguished by the small number (three to five)
of the terminal claws, and by a peculiar median outward bending of each of the six
terminal rays borne by each principal.

Four different forms of rosette also occur. In the first place, discohexasters in which
each principal bears four terminal rays, slightly bent in an S-like manner ; these
terminal rays increase in thickness towards the exterior, and bear terminally a large hemi-
spherical terminal plate, with small marginal teeth (PI. XII. fig. 2). In the second place,
discohexasters with four straight terminal rays, each of which bears on its unthickened
extremity a deep campanulate terminal disc whose margins are divided into parallel
pointed teeth (PL XIII. fig. 4); thirdly, discohexasters half the size of the above, in
which each of the principal rays bears a bundle of from twenty to thirty thin terminals
varying in length, and terminating in small four-pronged transverse discs (PI. XII.
fig. 6); and fourthly, graphiohexasters with a bundle of very fine somewhat diverging
rhaphides, which are borne on the discoid terminal expansion of each principal ray
(PI. XII. fig. 5). As somewhat loose parenchymalia, the scattered simple hexacts and
diacts with central intersection nodes are to be noted.

The reference of this sponge to the Euplectellidse, and to a position somewhere in the
neighbourhood of Tiegeria, I base upon the character of the main skeletal framework,
and on the great abundance of sword-shaped fioricome- bearing hexacts, which doubtless
belong to the dermal skeleton.

Perhaps the variety A of Oscar Schmidt's Rhahdodictyum delicattvm^ is identical
with my Dictyocalyx gracilis, as is suggested by the figure,' and the short but very
apposite description of the framework which Oscar Schmidt has given in the words :
" an airy wall arises from a thick pillar-like base." I would also have accepted
0. Schmidt's designation for this Hexactinellid, had the name selected by 0. Schmidt
appeared to me to be more suitable for his variety B, to which his description
(especially of a peculiar rosette) chiefly refers. This variety B is likewise included in
the Challenger material, and will be immediately descril)ed.

' Spongien des Meerbusens von Mexico, p. 46. - Loc. cit., pi. vii. fig 3r(.

REPORT ON THE HEXACTmELLIDA. 107

Genus 4. Rhahdodictyum, 0. Schmidt.
This genus contains only one species.

Rhahdodictyum delicatum, 0. Schmidt (PI. XX.).

Of the two Hexactinellids (dredged in the neighbourhood of Bequia from a depth
of 1591 fathoms) which Oscar Schmidt has distinguished as varieties of one species —
Rhahdodictyum delicatum — it appears to me, as already mentioned, that the variety A is
identical with my Dictyocalyx gracilis. For the variety B I retain the original name.
0. Schmidt describes this latter form as a " simple or l:)ranched tube with a perforated
wall, formed of interwoven cords. The cords consist of amalgamated or loosely fused
hexradiate spicules. The rays, which are arranged in layers in the longitudinal direction
of the cords, are for the most part strikingly prolonged, so that the tissue has the appear-
ance of being composed of irregularly intersecting rods. The free hexradiate spicules are
slender, smooth when quite young, but subsequently covered with spines. The rays are
very pliable. A beautiful rosette form occasionally occurs, in which each of the six rays
is provided with eight intersecting umbels."

Oscar Schmidt's figure,^ and still more his description, have convinced me that these
skeletons dredged by the Challenger in the neighbourhood of the Bermuda Islands
(Station 56), from a depth of 1075 fathoms, and on coral mud, belong to Rhahdo-
dictyum delicatum, 0. Schmidt. They exhibit slender, almost tubular cups, with a length
of 6 cm., and much broken at the upper ends (PI. XX. fig. l). The basal portion, which
is from 6 to 10 mm. in breadth, is attached by means of a terminal expansion to some
solid body, and becomes gradually widened upwards to twice this diameter or more. The
wall of the tube, which in the larger specimens measures as much as 4 mm. in thickness,
is radially perforated by numerous round holes. These are from 2 to 4 mm. in width,
and become somewhat broader towards the exterior, at the same time increasing in
diameter towards the upper end of the sponge. The arrangement of these parietal
apertures is tolerably irregular in the inferior portion, but towards the superior extremity
acquires more and more the character of two somewhat steep, intersecting sjjiral rows.
The siliceous framework which forms the supporting wall is usually from 1 to 3 mm.
broad between these foramina, and consists of the greatly prolonged hexact-rays firmly
united by soldering and synapticula (PI. XX. figs. 2, 3. 4).

Since the three specimens which are at my command are all incomplete and wholly
macerated there is no trace of the soft tissue nor of the looser spicules of the parench}'Taa,
skin, or gastral membrane, so that no conception of the structure of the entire sponge can
be obtained suflicient to enable us to determine whether we have here to deal with a

' hoc. cit., pi. vii. fig. 3, B.

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

Euplectellid or not. Jt is not even certain whether the parietal apertures observed in the
skeleton are really open during life, or whether they are covered by the soft parts.

Genus 5. Rhahdoi'^ectella, 0. Schmidt.
Only one species has yet been referred to this genus.

Rhabdopectella tintinnus, 0. Schmidt (PI. XII. figs. 8-12).

Among the Hexactinellida from the Bay of Mexico which have been studied by
Oscar Schmidt, a form occurs represented by several fragments and by one whole young
specimen, in which the cup-shape manifest in the young form at least, the method in
which the spicules are united by fusion and synapticula in the basal part, but
more loosely in the upper portion, and further the form of the loose spicules present in
the soft parts, justify its reference to the Euplectellidae, though we do not know either
the form or structure of the entire adult sponge. The inferior portion of the mature
Rhabdopectella tintinnus, as figured by 0. Schmidt,^ presents so firm and stalk-like a
mass that in spite of the muddy character of the ground in question, I am far from
accepting the opinion of Oscar Schmidt, that a root-tuft must have been present, for the
expanded basal plate of the young specimen by no means excludes the supposition that
the sponge was fixed to some hard body.

The wide-meshed lattice-like framework of the cup-shaped bod}', composed as it is of
greatly prolonged hexacts and numerous diacts, may well be compared with that of Rega-
drella and other firmly sessile EuplectelUdse. Of the looser spicules Oscar Schmidt has
carefully described and figured several noteworthy forms. Among these the floricomes, which
agree throughout with Euplectella floricomes, deserve special attention ; further, there are
discohexasters of different kinds — first, those with small transversely disposed stellate plates
on the thin extremities of the long secondary rays, of which two or three are attached to
every principal ray ; secondly, those with hemispherical, marginally toothed terminal umbels
borne by the thin, but externally conically thickened terminal rays, of which four occur
on each principal, and in which the S-shaped curvature (PI. XII. fig. 8) produces a mutu;il
entanglement; thirdly, a form resembling the latter but with long, parallel, marginal
prickles which run back from the hemispherical terminal disc, close to the axis of the
terminal ray ; and finally, a very small, on the whole spherical form in which each of the
long, thick, principal rays bears five strong terminals. These five terminals are arranged
in a regular manner, so that a somewhat shorter straight ray forms the direct extension
of the principal, while the four other longer and slightly bent rays are disposed in a
cross, and run obliquely outwards. Every terminal ray is provided with a hemispherical

^ Loc. cit., pi. viii. fig. 9.

REPORT ON THE HEXACTINELLIDA. 109

terminal umbel, from the lateral margin of which about sixteen pointed teeth project. It
is remarkable that these hemispherical or globular terminal umbels are not of equal size,
the central having but about half the diameter of the four surrounding it (PI. XII.
fig. 11).

In a fragment of this species which Professor Oscar Schmidt had kindly placed
at my disposal, I found a few other rosette forms, namely, first, oxyhexasters with two
diverging thin terminal rays on a very short principal (PI. XII. fig. 9), just like those so
frequent in diff"erent species of Euplectelln, and secondly, small discohexacts in which
upon the outer extremities of every simple c3dindrical principal ray, a hemispherical
hollow terminal disc with marginal prongs was so adjusted that the cavity seemed to
be directed not inwards but outwards.

The stars which Oscar Schmidt observed in Rhahdopectdla tintinnus, and which he
compared to the Tethya-stavs, I have likewise seen, and have figured them in PI. XII.
fig. 12. I regard them as oxyhexasters in which the strong terminal rays have at their
base grown together laterally on the very shoi't principal rays.

Finally the rough bow-shaped spicules (PI. XII. fig. 10) deserve to be mentioned.
The great resemblance between these forms and those of Euplectella jovis has already
been noted by Oscar Schmidt.

Genus 6. Hertwigia, 0. Schmidt.
This genus includes only the single species mentioned below.

Hertwigia falcif era, 0. Schmidt.

From a knotted and branched basis, according to Oscar Schmidt's representation,
there arises an irregular labyrinth of tulx'S with thin leaf-like walls, supported l>y
a lattice-like framework of obliquely crossed rods and fibres. Owing to the defective
preservation of the upper much damaged portion it was not possible to frame any
satisfactory conception of the whole form. Among the looser siliceous spicules, Oscar
Schmidt mentions — (l) hexradiate spicules and derivative five- and three-rayed forms,
which are for the most part rough towards the point, while others exhibit a fir-tree-like
ray with at most very short prickles ; (2) rosettes with four intersecting umbel teeth ;

(3) rosettes with larger teeth on the backwardly bent margin of the terminal umbel ;

(4) the " specific Euplectella rosettes," that is to say, floricomes ; and (5) two kinds of
sickle-rosettes, one of which carries on the hemispherical terminal disc of each principal
ray several whorls of sickle-like terminal rays, while in the other four large sickle-like
teeth are inserted on every principal ray. Kods with numerous oblique lateral prickles
at one end are sc, ttered here and there. Oscar Schmidt also mentions delicate

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

siliceous webs whose threads are frequently provided with small terminal hooks or
umbels. I would venture to suggest that we have here to deal witli the retiforni
siliceous skeleton of certain Eadiolarians in which terminal hooks or pronged terminal
discs frequently occur quite similar to the hexasters of Hexactinellida.

This form was found in the neighbourhood of Dominica, on muddy ground, and at a
depth of 6 1 1 fathoms.

In spite of the characteristic floricome described by Oscar Schmidt, it must still
remain doubtful whether this form really belongs to the Euplectellidse.

Genus 7. Hyalostyhis, n. gen.
This genus contains only one species, Hyalostyhis dives.

Hyalostylus dives, n. sp. (PI. LXX).

This Hexactinellid (PI. LXX. fig. 1) was dredged in the Mid South Pacific Ocean,
lat. 39° 41' S., long. 131° 23' W. (Station 289), from a red clay ground at a depth of
2550 fathoms. The soft loose body measures 5 '5 cm. in length, while from the lower
end there extends a much damaged awd-shaped stalk, 11 cm. long, and 2 to 1 mm.
thick. The whole sponge has been apparently so much compressed and distorted in the
dredge, that from its present state it is ditticult to decide certainly as to the original
form of the living organism, though a reconstruction is, to a certain extent, still
possible. In general form the body of the sponge resembles a laterally mucli-corapressed
cone, of which one of the narrow sides is uniformly rounded and the other much
folded. The broader upper end is not transversely, but obliquely truncated, as the
rounded, somewhat S-curved narrow side is rather longer than the other, which exhibits
a simple convex contour. The upper surface is contracted in funnel-like fashion, and
exhibits an irregularly folded infundibular opening. The lower end of the body is not
drawn to a point, but is slightly rounded. The stalk springs from the base of the
rounded, unfolded narrow side, and becomes gradually slender throughout its length.

As the sponge w^as given to me for investigation in an undoubtedly much altered form
and condition, several questions as to its anatomical structure, and these of essential
importance in determining its systematic position, have unfortunately to be left un-
answered. Thus it must remain doubtful whether the cup-form of the body represents
the original shape, whether the smooth external surface is really dermal, whether the
folded portion of one side is merely a part of the external surface, whether the styliform
stalk represents the intact form or only a fraction of the original, and so on.

Microscopic examination soon revealed that the tissue was not sufficiently well
preserved to admit of tlie recognition of the arrangement and structure of the chambers,

REPORT ON THE HEXACTINELLIDA. Ill

&c. My description is therefore necessarily restricted to the siliceous spicules, which in
delicacy and variety of form are not surpassed bj- those of any other sponge.

The spicules of the parenchyma are long, filiform, slightly curved, smooth diacts ; the
ends are always knobbed and beset with fine spines. The middle point is generally marked
by a swollen ring, or by yb?<?' cruciate, or less frequently, two opposite bosses (PI. LXX. figs.
6,7). Between these long diacts which are isolated and disposed in strands, separate, pecu-
liarly curved oxydiacts occur, and in some situations {e.g., just below the skin) they are
even more abundant. They measure about 0'3 mm. in length, and are covered not all round,
lint on one side, with minute tubercles and warts. Except on this side they are smooth.
The curvature of these drawn-oat s]iiudles may be generally described as like that of a
corkscrew, but both the pointed ends are slightly bent outwards (PI. LXX. figs. 5, 8). I
was not able to detect the axial canal, probably on account of the roughness of part of the
surface. Besides the above, the parenchyma contains four difierent kinds of rosettes.
One extremely rare form exhibits at the end of the comparatively short, simple, principal
ray, six somewhat markedly diverging, long, straight terminals, arranged in a corona.
The basal half of these terminal rays is very thin and delicate, while the outer, terminal
half consists of a thicker, cylindrical, terminally rounded portion, like a Typha-s^\k.&
(PI. LXX. fig. 2). Another somewhat common rosette bears on the short, simple,
principal ray a bundle of long S-shaped terminals, with toothed terminal plates which
project outwards (PI. LXX. fig. 3). The filamentous terminal rays, which are slightly
thickened at their outer ends, are arranged like petals, but of unequal length — the
median being" longer than the external.

A third form of rosette is characterised by the inverted bell-shaped umbels formed by
the long terminal rays, six of which diverge from the end of each of the short simple
principals. These terminal umbels exhibit on the margin of a small, flatly convex,
transverse disc, at the end of the terminal ray, about eight fine umbel rays, which extend
parallel to one another towards the centre of the rosette, and form along with the
terminal umbel an inverted bell-shaped structure (PL LXX. fig. 11). The fourth and most
frequent form of rosette bears at the end of each of the short, simple, principal rays a
bundle of long, thin, radiately disposed terminals, which gradually increase in diameter
away from the base, and terminate in small, slightly convex, transverse discs with
toothed margins (PI. LXX. fig. 12). The number of terminal rays on this, often
apparently spherical, discohexaster varies greatly. Each principal ray may bear from
twenty to forty distinct terminals.

The dermal skeleton consists of smooth hexacts, mostly of median size. Their strong,
freely projecting distal ray is club-shaped, i.e., gradually increases terminally into a swollen
knob, and occasionally exhibits several tubercles on its outer end (PI. LXX. figs. 4, 6).
The four tangential rays, which are always disposed at right angles, are also thickened
terminally, and the conically pointed external end is roughened. The proximal ray is

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

usually without any, or with only a very slight terminal swelling, and indeed, as a rule, ends
in a usually conical, roughened point. The length of the distal and proximal radial rays
varies greatly. The tangentials measure, on an average, 0"2 to 0'3 mm. in length, and
the distals are not unfrequently of the same size, but the latter may in some situations,
e.g., near the lower end of the body, attain more than double the length of the former.
The proximal ray exhibits a similar variability, measuring sometimes only O'l mm. in
length, and in other cases 0"4 to 0'5 or more, while in thickness it always falls consider-
ably below that of the distal.

Here and there, on the external surfece, I found simple cruciform tetracts, with
smooth knobbed rays, pointed at the extremity (PI. LXX. fig. 10). The fact that,
for the most part, only a few dermalia were to be found on the external surface, is
probably to be traced to the injuries which the sponge seems to have suffered by
attrition. I found these dermal hexacts most abundantly and in best preservation at
the pointed basal extremity of the body, just above the origin of the stalk. They were
also abundant, though somewhat displaced, in the deep folds of the already mentioned
plicated side of the body, which, if flattened out, would form part of the external
surface.

The stalk is chiefly composed of very long needles of various thickness, closely
arranged, parallel to the longitudinal axis, and, especially in the inferior portion,
frequently bound together by transverse synapticula (PI. LXX. fig. 9). On the external
surface, between the longitudinally disposed diacts, I found forms the same as those
above described as parenchymalia, as well as similar hexacts with club-shaped distal
rays, and simple cruciform tetracts, somewhat bent towards the surface.

I am not in a position to report so definitely as to the gastral skeletal elements, since
I was not able to study them in their natural disposition. In the interior of the body,
in the numerous septa, hexacts and pentacts occur in abundance, with somewhat long,
smooth rays, which end in slightly club-shaped thickenings ; and it seems to me
probable that these function as gastralia, though I was unfortunately unable definitely
to determine their disposition.

It is clear that this curious form, obtained from the depths of the South Pacific, at a
great distance from land, must occupy a peculiar position in the system of the Sponges.
On the one hand, the character of the hexact dermalia with their diverging distal rays
would suggest an alliance with the family of the Euplectellidse, in which, further,
such remarkable twisted oxydiacts alone occur (in Rhahdoplectella tintinnus, PI. XII.
fig. 11); while, on the other hand, all the microscopic characters of the body, and
the sharply defined, long, narrow stalk indicate a closer resemblance to such forms as
Crateromorpha, Aulochonc, and especially Caulocaly.v (PL LXIX.), to which we shall
have again to refer in detail.

REPORT ON THE HEXACTINELLIDA, 113

Family II. Asconematid^.

Lyssacina of sack-, tube-, beaker-, or mushroom-like form, fixed either directly or by
means of a round stalk. The body-wall, which is not perforated, forms a thin soft plate
or a compact mass, which represents either the thick wall of a beaker, or the arched
plate of a mushroom-like body. In the latter case the gastral surface has become the
convex outer side.

Both dermal and gastral surfaces are densely and uniformly beset with hexact and
pentact piauli, in which the freely projecting fir-tree-like ray is prominently developed,
while the parenchymal ray is usually small or entirely atrophied. The four transverse
rays, which are cruciately disposed, lie embedded in the bounding skin, and the whole
spicules are accordingly designated autodermalia and autogastralia. Below these pentact
hypodermalia and hypogastralia occur.

The rosettes which lie scattered between the diact or hexact principalia are for the
most part discohexasters.

Subfamily 1. Asconematin^.
Sessile sack-, cup-, or tube-like Asconematidse, with a thin, flabby, pliable wall.

Genus 1. Asconema, Saville Kent.

1870. S. Kent, Monthly jMicr. Journ., Nov., p. 241 (Asiwiema setuhalense).

1871. Gwyn Jeffrey.s, Proc. Roy. Inst., N. 54, p. 258.

1872. Gray, Ann. and Mag. Nat. Hist., vol. ix. p. 442.

1873. Thomson, Depths of the Sea, p. 429.

1874. Carter, Ann. and Mag. Nat. Hist., vol. xii. p. 349.
187,4. Gray, Ann. and Mag. Nat. Hist., vol. xiii. p. 284.

1875. Marshall, Zeitschr. f. wiss. Zool., Bd. xxv., Suppl., p. 142.

1876. Marshall, Zeitschr. f. wiss. Zool., Bd. xxvii. p. 113.
1880. Norman, Ann. and Mag. Nat. Hist., voL vi. p. 430.

1880. 0. Schmidt, Spongien des Meerbusens von Mexico, ii. p. G5.

1881. Milne-Edwards, Comptes rendus, xciii. p. 876.
1885. Filhol, La vie au fond des mers, p. 288.

History. — Among the marine Sponges preserved in the Lisbon Mu.seum of Natural
History, Saville Kent detected in 1870,^ on the occasion of the " Noma " expedition, some
large, but only partially preserved cup- or sack-shaped specimens of " felt-like consistence,
composed of an interlacement of long filiform siliceous fibres or spicules, and, interspersed
among these, hexradiate spicula of various sizes and minute multiradiate ones with
capitate extremities." For these specimens he erected a special genus, Asconema, and
characterised the single representative species, Asconema setuhalense, in the following

1 Monthly Micr. Journ., p. 245, 1870.
(zool. ch.\ll. EXP. —part liii. — 1886.) Ggg 15

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

words : — " Sponge body, vase- or sack-like, of large dimensions, expanding superiorly,
often upwards of two feet in height; composed of interlacing fasciculi of long filiform
fibres or spicula. Individual filiform spicula smooth, finely canaliculate, varying in
diameter from ^^jij^th to 5^xjth of an inch, occasionally possessing a central or excentral
inflation, as at plate Ixiv. fig. 4. Hexradiate spicula of two types, the one large, with
smooth alternate radii, the other of smaller but more varying size, with obtuse extremities
and entirely erectly spinous surfaces. Scattered through this sponge there also occa-
sionally occur simple alternate spicula, clavate and erectly spined at either extremity (see
plate Ixiv. fig. 5). Average diameter of the minute multiradiate spicule ^-^th of an inch."
From the fact that the base was absent in all the specimens, Saville Kent was led to
conclude that the lower portion of the cavity, where the sponge is fixed by its base to
the sea-bottom, was filled with mud, and that it must thus have been constantly torn
oft' during the process of fishing up the sj)onge.

Fig. 3.—Asconcma sctuhaUn.sc, Sav. Kent, one-eiglith the natural size, from Wyville Tliomson's Depths of the Sea, p. 429.

A giant specimen of this species, with a diameter of 3 feet, was dredged by Gwyn
Jeff"reys' ofi" Cape St. Vincent on the coast of Portugal, on rocky ground and from a
depth of 374 fathoms.

In his Classification of Sponges,"^ Gray forms Kent's genus Asconema into a special
family, the Asconematidas, with characters essentially similar to those noted by Kent in
regard to the genus and the single known species. A large and comparatively well-

■ Proc. Roy. Inst., No. 54, p. 258, 1871. ^^ Ann. and May. Nat. Hist., ser. 4, vol. i.\. p. 458, 1872.

REPORT ON THE HEXACTINELLIDA. 115

preserved specimen of Asconema setiibalense, which was dredged "in 374 fathoms on
rocky ground off Cape St. Vincent, during Mr. Gwyn Jeffrey's cruise in 1870," is described
by Wyville Thomson,^ and depicted in an excellent woodcut. He calls it " a complete
vase of very elegant form, nearly ninety centimetres in diameter at the top, and about
sixty in height. The sponge came up folded together, and had much the appearance of
a piece of coarse greyish-coloured blanket. Its minute structure is however very
beautiful. It consists, like Holteuia, of two netted layers, an outer and an inner, formed
by the symmetrical interlacing of the four cross branches of five-rayed spicules; and as
in Holtenia and Rossella, the sarcode is full of extremely minute five and six-rayed
spicules, which, however, have a thoroughly distinct character of their own, with here
and there a very beautiful rosette-like spicule, another singular modification of the
scxradiate type characteristic of the group. Between the two netted surfaces the sponge
surfoce is formed of loose curving meshes of loosely aggregated bundles of long single
fibres, sparsely mixed with spicules of other forms. This sponge seems to live fixed
to a stone. There are no anchoring spicules, and the bottom of the vase, which in our
two specimens is a good deal contracted and has a square shape something like an
old Irish " mether," has apparently been torn from some attachment."

Carter, in his paper on the Hexactinellida,^ has noted certain agreements between
Asconema and CnUeromorpha. He says — " In Asconema the small sexradiate spicules
with short-spined arms (of which the vertical one on one side is often deficient) and
formed together in groups, recall to mind the same kind of spicules (which form a
rectangularly reticular network) on the surface of C rater omoiyha meyeri and Rossella
velata. Indeed, so far as this goes, Asconema might be considered a sessile vase-like
representative on the coast of Portugal of the cup-like head of Crateromorpha found
about the Philippines only."

Some siliceous elements figured by Kent as skeletal spicules of Asconema setubalense,^
which do not exhibit the hexradiate type, were declared by Carter not to belong to this
Hexactinellid but to have been most probably intruded from a PackastreUa abyssi.

With regard to the affinities of this form among the Hexactinellida, Marshall * conjec-
tures that Asconema, which he refers as an aberrant form to his Pleionacidge, is closely
related to Lanuginella, Schmidt, that the latter, in fact, is probably a young Asconema.

Norman reported* that, during the voyage of the " Travailleur," "a little bunch of
the strong coarse spicula of the great Asconema setubalense was dredged in the Bay of
Biscay, in about 600 fathoms."

As a second species belonging to the genus Asconema, Kent, Oscar Schmidt
described Miis Asconema Icentii. This was found in from 300 to 1500 fathoms, in the

I The Depths of the Sea, 1873. = Ann. and Mag. Nat. Hist, ser. 4, vol. xii. pp. 369, 370, 1873.

s Loc. cit., figs. 10, 11. * Zeitschr.f. miss. ZooL, Bd. xxvii. p. 121.

» Ann. and Mag. Nat. Hist., ser. 5, vol. vi. p. 436. " Spongien des Meerbusens vou Mexico, p. 65, 1880.

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

neighbourhood of Grenada, Martinique, Guadeloupe, and Bequia. He speaks of this
new species as follows : — " It occurs in two varieties ; the one resembles a shallow or but
moderately deep cup, which is provided inferiorly with a rounded or short, often
somewhat irregularly twisted, pointed stalk ; the other is sack-like, with an irregular
margin, with irregular pocket-like divisions and excavations, which are separated by
somewhat thin lappet-like partitions, just like a beggar's sack. The latter forms are a
foot in length and seem to be sunk in the mud up to the upper margin." In a netted
covering-layer, which is particularly well develojDcd in the beaker-like variety, 0.
Schmidt found fir-tree-like spicules. " On the margin of the beaker-like variety the
covering-layer usually projects as a plaited fringe, and extends from the outer towards
the inner surface. Above this there projects a dense row of imperfectly developed
hexradiate spicules, and needles with an axial cross. The whole body is unusually rich
in double ' quirls,' ^ which vary greatly in dimensions and in individual form."

Oscar Schmidt notes especially that although Asconema kentii agrees completely in
the form, consistence, and especially in the felt-like condition of its walls with Asconema
setubalense, Kent, the agreement by no means extends to the spicules.

The dried specimen of this species was kindly lent to me by Oscar Schmidt, and is
figured in PI. XXX. fig. 9. That it does not belong to the genus Asconema, but to
Hyalonema, will be pointed out in the detailed description of the characteristics of the
species Hyalonema hentii, which will be given below.

In a short report on the voyages of the French ships " Travailleur " and " Talisman,"
Filhol ^ gives a woodcut and the following notices of Asconema setubalense : —
' L' Asconema setubalense n'avait ete trouve jusqu'au voyage du Talisman que siu- les
cotes du Portugal. Lors de la campagne de ce dernier bateau, nous I'avons recueilli sur
les cotes du Maroc, au voisinage du cap Bojador, par 410 metres. Les deux exemplaires
que nous en avons obtenus vivaient fixes par leur base sur des roches ou sur des coraux
(Lophohelia) dont le chalut contenait de nombreux debris." And on p. 289 : — " Les
Asconema ont ete trouvees avec les Aphrocallistes,"

The genus contains only one species.

Asconema setubalense, Saville Kent (PI. XXI).

A thin-walled cup in which the wide superior opening is surrounded by a simple,
smooth, thin margin, while the narrowed inferior extremity seems to be firmly fixed.
Among the sponges of the expedition of H.M.S. "Triton," which were entrusted to me
for review, I observed the fragment figured in PL XXI. fig. 1, and some small
specimens of a Hexactinellid, which, after comparison with Kent's original in the

' Quirl = a short round stick provided with a toothed disc on one extremity.
2 La vie au fond des mers, p. 285, fig. xcii., 1885.

REPORT ON THE HEXACTINELLIDA. 117

British Museum, I regard as identical with the sponge named Asconema setuhalense by
Saville Kent. The specimens in question were trawled at Station 4 of the "Triton"
expedition, from a depth of 327 to 430 fathoms, and were preserved in absolute alcohol.
By means of this fitly selected medium the soft parts were especially well preserved.

The priucipalia of the parenchyma are strong, smooth diacts, 1 cm. or more in
length, pointed at both ends, and usually exhibiting in the middle a small knot-like
swelling. They lie laarallel to the bounding surface at diff"erent levels, and are distri-
buted with comparative irregularity. Every principal ray is surrounded by a layer of
thin diact comitalia which extend over its surface in gently undulating curvature.

The parenchyma also contains a rich abundance of thin rod-like diacts of small size,
separately or in bundles, and disposed in different directions. Of these many exhibit a
central swelling (PI. XXI. figs. 7-10), and either terminate in sharp points (PI. XXI.
figs. 9, 10), or are provided with rough knob-like swellings on one or at both extremities
(PI. XXI. figs. 7, 8).

Between all these rod-like spicules four different kinds of rosettes occur in ii-regular
distribution: — namely (1) simple oxyhexasters with long terminal rays, two or three in
number, and running out in a diverging manner from the principal (PL XXI. fig. 12);
(2) smaller oxyhexasters in which each of the broad, moderately short principal rays
bears a bundle of three to twelve fine terminals (PI. XXI. fig. G) ; (3) quite small disco-
hexasters in which each of the broad principals bears a strongly-developed, diverging
brush-like bundle of numerous fine terminal rays with terminal knobs (PI. XXI. fig. 3) ;
and (4) discohexasters twice the size of the above, with short principal rays, each of
which exhibits six long, thickly barbed, cylindrical terminals. These are disposed either
in a tuft or round a funnel-shaped space, and terminate in an arched terminal disc with
several prongs (PL XXI. fig. 11).

The dermal skeleton is supported by tolerably large, simple, smooth hypodermal
pentacts. In the dermal membrane itself lie the four cruciate transverse rays of small
pentacts, from each of which a distal radial, as long as the transverse, projects outw^aixls,
while the proximal radial ray is reduced to a smaU rounded peg. The five developed
rays of these autodermalia are thickly beset with small outwardly directed prongs, while
the proximal stump appears to be smooth (PL XXI. fig. 4).

The gastral skeleton is very similar. Here also large, smooth, pentact hypodermalia
(not introduced into the diagrammatic figure 2) are present. The gastral membrane
contains the cruciate transverse rays of pentacts, in which the proximal ray projects
inwards into the gastral cavity, while the radial is rudimentary. The five developed
rays, like those of the autodermalia, are beset wdth prongs. Between the pentact auto-
gastralia, hexacts also occur here and there. In these the distal ray enters into the
parenchyma, and has the same length and development as the proximal I'ay opposite to
it (PL XXI. figs. 2, 5).

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

Genus 2. Aulascus, n. gen.
This genus only contains one species.

Aulasciis johnstoni, n. sp. (PI. XXII. figs. 1-3).

The loose and readily pliable tube-like fragments which are figured in PI. XXII. fig. 1 ,
have a waU from 2 to 4 mm. in thickness, and were dredged in the southern part of the
Indian Ocean, about halfway between the Cape of Good Hope and the Kerguelen
Islands, at Station 145a (lat. 46° 41' S., long. 38° 10' E.), from a depth of 310 fathoms,
and a bottom of volcanic sand. AVhile the inferior portion of both tubes is entirely
destroyed, a simple, sharp, smooth margin occurs on the somewhat widened upper portion
of one of them. Both the outer and the inner surfaces are covered by a continuous thin
skin, through which irregularly scattered cavities of variable widths (up to 2 mm. in
diameter) may be seen.

In the parenchyma the principalia consist of moderately large hexacts with rough
ends which run out gradually to a point, and of numerous straight or slightly bent diacts,
which occur partly in bundles and partly isolated, and are provided with a central
thickening and with rough ends running out gradually into points. While the hexacts
are usually disposed at right angles to the bounding surfaces, the diacts are for the most
part parallel to these or arranged obliquely in different directions.

Between the principalia numerous irregularly scattered rosettes appear. Among
these discohexasters predominate, and each of their principal rays is divided into two or
three diverging straight terminals. The latter decrease in size towards the exterior, are
densely beset with numerous short hooks, and terminate in a small, slightly arched,
transverse disc which runs out into four or more prongs. In some cases the outer
extremities of the terminal rays are so delicate, and the terminal transverse disc so
small, that the latter can only be seen with high magnifying powers, and the whole rosette
resembles an oxyhexaster.

The plumicomes (0'05 mm. in diameter), indicated in PL XXII. fig. 2, and figured in
fig. 6 from another sponge, occur here and there. The extremity of each of the moderatelj'^
long, round, principal rays bears a tolerably thick, round, transverse disc, on the convex
outer side of which the fine terminal rays, which are bent in an S-shaped manner, arise in
concentric circles and together form a cup.

The dermal skeleton is supported by moderately large pentact hypodermalia with
rough ends, which gradually run out to fine points. The abortive distal ray is some-
times indicated by a round tubercle. The dermal membrane itself contains the four
cruciate transverse rays of the hexact autodermal pimdi, in which the freely projecting
distal ray is beset with scaly prongs, like a fir cone, in its larger outer principal portion,
while the base remains naked (PI. XXII. fig. 3). The pole-like pointed proximal is

REPORT ON THE HEXACTINELLIDA. 119

provided, like the similarly formed four transverse rays, with fine outwardly directed
prongs, and is smooth only on its inmost portion.

The gastral skeleton is quite similar. Here too pentacts of median size occur as
hypogastralia. These have rough pointed extremities and a knob-like rudiment of the
sixth ray, while the four cruciate transverse rays of the hexact pinuli, in which the
freely projecting proximal ray is covered with scaly prongs like a fir cone, lie in the
gastral membrane itself. The distal which is as long as the transverse rays, is like the
latter beset, except in its inmost portion, with small outwardly directed prominences.
They diS'er from the autodermalia only in this, that the freely projecting fir-cone-like ray
is here somewhat more slender, although in both it runs out to a point.

Subfamily 2. Sympagellin^,

Ovoid, thick-walled, usually (perhaps always) stalked goblets, with smooth, thin upper
margin. Between the principal b exacts small discohexasters and long diacts.

Genus 1. Sympagella, 0. Schmidt.

1870. 0. Schmidt, Grundziige einer Spongienfaima des atlant. Gebietes, p. 15.

1872. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. ix. p. 457.

1873. Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. xi. p. 283.
1873. Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. xiL p. 360.

1875. Marshall, Zeitschr. f. wiss. Zool., SiippL, p. 142, 1875.

1876. Marshall, Zeirschr. f. wiss. Zool., Bd. xxvii. p. 127.
1881. Milne-Edwards, Comptes rendus, xciii. p. 931.

History. — Under the name Sympagella nux, 0. Schmidt described and figured a
Hexactinellid obtained off Florida from a depth of 98 to 123 fathoms; it presented the
form either of a single individual with an ellipsoidal body, about 1 cm. high, with a
terminal osculum and a simple stalk, or of a branched stem with several terminal
individuals of similar nature.

In the membrane covering the external surface of the body and lining the gastral cavity
0. Schmidt found pentact pinuli with a rudimentary sixth ray, and in the latter situation
also "nodular hexacts." In the parenchyma were numerous small hexaets with three
Itarbs situated on the er^d of each ray. In the stalk were long tubercula united by
transverse hour-glass-shaped connectives. It was probably on account of these ladder-
like structures that Gray' in 1872 placed Sympagella along with Farrea in his family
Farreadse.

In 1873 Carter^ discovered in Sympagella nux "rosettes with rays multitudinous, of
unequal length, without heads, flexed outwards and arranged en fleur-de-lis ; pappiform."

' Notes on the Classification of the Sponges, Ann. and Mag. Nat. Hist., ser. 4, vol. ix. p. 457.
'■^ Ann. and Mag. Nat. Hist., ser. 4, vol. xii. p. 361.

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

SympagcUa mix also appears in Milne-Edwards' list, published in 1881, of the
specimens collected by the " Travailleur " on the coasts of Spain and Portugal.
This genus contains only one species.

Sympagella nux, 0. Schmidt (PI. XXII. figs. 4-9).

The much injured specimen of Sympagella nux, 0. Schmidt, which is figured in
PI. XXII. fig. 4, was obtained on coralhne mud in the neighbourhood of the island of
St. lago, one of the Cape Verde Islands, from a depth varying from 100 to 128 fathoms.
The body is branched like the horns of a stag, and from its round principal stem, which is
1*5 mm. in thickness, two somewhat bent more delicate side branches arise on the same
side. The upper extremity of the principal stem, which is broken ofi' beneath, forms an
oval body, which is 1 cm. long and 6 mm. broad. The superior transversely truncated
extremity of the latter bears the orifice (2 mm. in width) of a simple gastral cavity. The
latter is slightly narrowed towards the base of the body, and ends blindly without being
continued into the stalk. While the outer surface of the body-wall (which is 2 mm. in
thickness) is surrounded by a continuous skin, through which small cavities can be seen
only here and there, larger efi"erent passages appear on the inner surface, and open
directly into the gastral cavity. It is indeed possible that the gastral membrane may
have been torn or otherwise injured where it extended over the inner openings of the
efferent canal system. The upper lateral branch exhibits the lower fragment of a torn-
ofi" body.

The principalia of the parenchjnnal skeleton consist, as in Aulascus, of simple, regular,
moderately strong hexacts, with somewhat rough pointed extremities, and of numerous
diacts, varying in strength, length, and direction. In these the centre swelling sometimes
exhibits four cruciately disposed knobs, sometimes an annular elevation, or sometimes
only the merest trace. The diacts are straight or slightly bent, and run out to similar
points at both ends. A slight roughness frequently occurs in the neighbourhood of the
terminal points.

With regard to rosettes, discohexasters are particularly abundant in the parenchyma,
and are provided with short principal rays, and long diverging terminals uniformly thin,
or slightly thickened towards the outer end. The latter are smooth or very slightly
roughened, and terminate in a watch-glass-shaped arched disc, provided with six to eight
marginal prongs. Each principal ray bears three, or less frequently four, terminal rays.
These are strong at the base, diverge in an arch-like curve, and then continue in a
straight or slightly bent course.

Besides these discohexasters, the plumicomes which are found in Aulascus johnstoni
also occur, but I have found them only in scattered distribution. Finally, the parenchyma
contains here and there peculiar structures, which Oscar Schmidt called " roller stars "

REPOET ON THE HEXACTINELLIDA. . 121

(Walzensterne). Short strong prickles, usually three at each extremity, diverge from
both ends, and frequently also from the middle of a somewhat thick, straight, or slightly
bent roller-like body. I regard these roller stars as reduced oxyhexasters, in which only
one axis, with the two princijial rays and the associated terminals, has attained full
development, while the other four rays arc either entirely atrophied, or are reduced to
short transverse prickles (PI. XXII. fig. 7, a, h, c).

The dermal skeleton consists of simple, moderately strong, pentact hypodermalia and
pentact autodermalia in the form of pinuli. In the latter the fir-tree-like distal ray, which
measures 0"1 mm. in length, is naked at the base, broadest in the middle, and runs out
to a point superiorly, while the four transverse rays lying in the dermal membrane bear
on their larger outer portion small points and irregularities. A small, rounded, smooth
peg appears in the place of the atrophied proximal ray (PI. XXII. fig. 8).

The simple pentact hypogastralia completely resemble the hypodermalia, nor do the
inwardly projecting gastralia differ much from the corresponding (inwardly projecting)
dermalia. These are very long, outwardly bent, slender hexacts, in which the narrow
pointed proximal ray which projects freely into the gastral cavity is beset only with
short prongs, which run obliquely outwards and attain a length of 0'5 mm. or more,
while the similarly armed pointed distal, which projects into the parenchyma, measures
only 0'17 mm. in length, and the similarly formed four cruciate transverse rays, are only
about O-I mm. long (PI. XXII. figs. 5, 9).

The stalk exhibits the same general structure as the body, except that the hexact
principalia fall into the background as compared with the extraordinarily prolonged
diacts, which are all arranged parallel to the long axis, and which, as Oscar Schmidt
accurately describes,^ are firmly united laterally by numerous synapticula. The
size of the hjqpodermalia decreases downwards, and they are, moreover, pronged and
rough.

Discohexasters, plumicomes, and "roller stars" may be found in the sto.lk in almost
as great abundance as in the body.

Genus 2. Polyrhabdus, n. gen.
The genus contains only the one species, Poli/rhahdus oviformis.

Polyrhabdus oviformis, n. sp. (PI. XXIII. figs. 1-8).

In the Antarctic Ocean (lat. 62° 2G' S., long. 95° 44' E.) there was trawled from a depth
of 1975 fathoms, and a bottom of Diatom ooze, the considerably damaged and triturated
egg-shaped specimen represented in PI. XXIII. fig. 1. It measures 2 cm. in length and

■ Loc. cit., pi. i. figs. 10-12.

(ZOOL. CHALL. ESP. — PART LIU. — 1880.) Ggg 16

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

12 mm. in its greatest diameter. The narrower upper end bears the circular opening of
the simple gastrie cavity, which is G mm. in diameter, and 10 mm. in depth. The body-
wall, which is iuferiorly from 3 to 4 mm. in thickness, becomes very gradually attenuated
upwards to the narrow smooth border. It is impossible to tell whether this oval body
was fixed upon a stalk, since the inferior extremity is greatly triturated.

Most of the somewhat thickly crowded parenchymal spicules are straight diacts of
variable length and of moderate thickness. They usually exhibit on either end a
slender olive-like thickening, which is beset with small points and prongs. Such
irregularities also occur where the thickening is absent, on the simple rounded ends
(PI. XXIII. fig. 7). In the middle the rods exhibit either four cruciately disposed
tubercles (PI. XXIII. fig. 6) or a circular wreath. Less frequently the central swelling
is insignificant or entirely absent, and the central portion appears often to be marked
only by the intersection of the axial canals. Between the very irregularly disposed
diacts, which lie, however, for the most part parallel to the oiiter surface, there occur
moderately large, usually radially disposed hexacts, whose rays run out to points for a
greater or less distance, and are frequently beset with pointed tubercles, either aU over
or at the extremities (PL XXIII. fig. 5).

Besides these principal spicules numerous rosettes occur in the form of discohexasters
with six, eight, or more similar thin diverging terminals on eveiy principal ray. The
watch-glass-like bent terminal disc of every terminal ray runs out into six or more
pointed marginal prongs (PL XXIII. fig. 3).

The dermal skeleton includes strong pentact hypodermalia ; I have not, however,
seen either these or the autodermalia in situ. The latter are hexact piuuli in which the
thicker oval distal ray, which resembles a fir cone, is beset all round with scaly prongs,
while the proximal and the four transverse rays are smooth up to the pointed roughened
extremities (PL XXIII. fig. 4).

The hypogastralia resemble the hypodermalia ; the autogastralia diifer from the
autodermalia in this, that their freely projecting ray is narrower and longer than in the
latter (PL XXIIL fig. 2).

Genus 3. Balanites, n. gen.
Containing only one species. Balanites pipetta.

Balanites pipetta, n. sp. (PL XXIII. figs. 9-14).

The single specimen, which resembles a clay pipe or cigar holder, and which is
figured in PL XXIII. fig. 9, was trawled in the Antarctic at Station 157 (lat. 53° 55' S.,
long. 108° 35' E.), from a depth of 1950 fathoms, and a Diatom ooze bottom. The oval
body, which measures 5 cm. in length and 1'5 cm. in thickness, is continued by a

KEPORT ON THE HEXACTINELLIDA. 123

gradually narrowed inferior portion into the obliquely inserted, round, hollow stalk,
which is 3 mm. in diameter, and has been torn off at a distance of 3 cm. from the body.
The upper transversely truncated extremity of the latter exhibits the circular aperture
(9 mm. in width) of a simple gastral cavity, which is 2 mm. in depth, and ends beneath
in a blind sack-like manner. The tolerably compact body-wall is about 5 mm. in thick-
ness beneath, and becomes gradually attenuated upwards to the narrow smooth border
of the opening. I was not able to observe a direct continuation of the gastral cavity
into the cavity of the tubular stalk, though such a connection may be present.

Among the principalia of the parenchymal skeleton the strong regular hexacts with
simple smooth rays occupy the first place. Diacts of variable length, however, with
pointed or rounded extremities, occur scattered or in bundles and disposed in different
directions. They are for the most part destitute of any central swelling, but the
intersection of the axial canals is frequently recognisable.

Between these large parenchjrmal spicules numerous small regular hexacts occur, with
strong rays which decrease in diameter towards the extremities, which are beset with
small thorns and bear at the very tip a small arched terminal disc with six or more
bent marginal ^jrongs. Sometimes, too, one or other ray may be found to be divided
almost to the middle into two or three diverging branches. These exceptional forms
lead to a rosette which occurs here and there, and is figured in PI. XXIII. fig. 12.
Here four to six strong diverging terminals, half the length of the principals, stand on
each extremity of the proportionally long, smooth or slightly tuberculated, principal rays.
These straight or slightly bent thin terminal rays are provided with lateral thorns, and
bear on each of their extremities a transverse disc provided with several marginal prongs.
Another rosette which is figured in the same plate (figs. 10, 11) occurs far more abundantly
than the form mentioned. This consists of long smooth principal rays which are
terminated by a small discoid expansion. The flat outer surface of the disc bears a
tuft of short delicate terminal rays provided with pronged terminal discs. The outermost
circle of terminals, are bent in an S-like manner.

The essential agreement between the dermal and gastral skeleton is here particularly
manifest. The transition from the one to the other occurs quite imperceptibly on the
rounded margin of the large oscular opening (PL XXIII. fig. 14).

Both hypodermalia and hypogastralia are strong and tolerably large pentacts, in
which each of the strong radial principals is beset on its middle portion with isolated
spines, while the four transverse rays are smooth throughout.

The autodermalia are hexact pinuli with broad, scaly, pronged, oval, fir-cone-like distal
rays, while their moderately long proximals, like the somewhat longer transverse rays,
remain quite smooth (PI. XXIII. fig. 13) up to the conically narrowed, rough or pronged,
terminal portion. The autogastralia difi'er from the autodermalia only in their smaller
breadth, and in the more slender, smooth basal portion of the freely projecting scaly ray.

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

Subfamily 3. Caulophacin.e.
Fungiform, with a long cylindrical tubular stalk.

Genus 1. Caulophacus, n. gen.

Funo-iform, with a long tube-like stalk. The gastral surface of the body has become
convexly arched by eversion. The dermal surface has either remained convex or has
become concave by a turning over of the marginal portion.

Caulophacus latus, u. sp. (PI. XXIV.).

About halfway between the Cape of Good Hope and the Kerguelen Islands (Station
147, lat. 46° 16' S., long. 48° 27' E.), from a depth of 1600 fathoms and Diatom
ooze ground, a Hexactinellid was trawled, which exhibited in its external form a marked
resemblance to a flat mushroom. The body is a circular disc, 15 '5 cm. in breadth, with a
thickened median portion continued downwards l)y means of a conical neck into the
rounded hollow stalk. The latter is 5 mm. in thickness, and is bent obliquely to the
side in its upper portion, which is alone preserved. While a shallow depression occurs in
the middle of the upper surface, the gradually narrowed marginal portion, which finally
terminate!? in a narrow smooth border, is bent slightly downwards (PI. XXIV. fig. 1).
On making a section at right angles to the surface (PL XXIV. fig. 1), the afferent and
efferent canals are clearly recognisable, and they may even be detected through the
uninjured dermal and gastral membranes. These canals are disposed at right angles to
the bounding surface. The cavity of the tubular stalk, which is 2 mm. in width, is
continued upwards into the efferent system of lacunae in the median part of the disc
(PL XXIV. fig. 1).

Since the upper usually slightly convex surface of the body, which exhibits only in
the middle a shallow depression, corresponds to the gastral surface, it is only natural that
all the chambers of the richly folded chamber layer should have their wide excurrent
openings turned towards this convex surface, and their blind sack-like extremities on the
other hand directed towards the concave dermal region. Here, too, as in all other cases,
the water enters through the dermal membrane, and passes through the sieve-like net-
work of the chamber walls in flowing from the outside inwards.

The principal parenchymalia are moderately strong, simple, regular hexacts, with
smooth rays, each of which runs out to a sharp point, and smooth diacts varying in breadth
and length, and provided on both sides with a rough rounded terminal portion. Between
these there is an abundant though scattered occurrence of small prickly hexacts
(0*25 mm. in diameter) with arched and marginally pronged transverse terminal discs

KEPORT ON THE HEXACTINELLIDA. 125

(PI. XXIV. fig. 7) exactly like those which we have already noted iu Balanites. Simple
oxyhexasters also occur with short principal rays, and long, frequently somewhat rough,
terminals. The division of one or several principal rays often does not take place, so that,
as in PI. XXIV. fig. 8, some rays become divided into two or three terminal rays, while
others which remain undivided are straight and run out to a sharp point.

Rosettes are represented only by the form figured in PL XXIV. figs. 4 and 6. Each
of the six smooth and moderately long principals passes into a narrow conical tuft of
eight or more almost similar, straight, fine terminal rays, each of which is somewhat rough
towards the extremity and is provided with a small, convex, marginally pronged, terminal
plate. Here and there, abnormally, one or other of the secondary rays does not occur on
the extremity of the principal ray but laterally, and is inserted more or less close to the
axial node as if it had moved down to this position.

The pentact hypodernialia have a strong radial ray which is wholly or partially beset
with spines, while the four transverse rays which are spread out beneath the dermal
membrane but rarely exhibit small spines, and usually appear somewhat roughened on
their bluntly pointed extremities.

The autodermalia are hcxact piuuli in which each of the freely projecting (usually
0'05 mm. long) distal rays resembles a crowded oval fir cone with a short smooth stalk
(PI. XXIV^. fig. 10), or is more rarely narrower and longer (PI. XXIV. fig. 2).

The four slightly conical transverse rays lying in the dermal membrane, and the
somewhat shorter proximal ray, arc onl}^ beset with small prongs on the outer portion.

While the hypogastralia do not difi"er essentially from the hypodernialia, the auto-
gastralia which project over the whole convex upper side of the body of the sponge difiier
essentially from the autodermalia, inasmuch as the freely projecting, and at the same
time distally directed ray, is very long (1 mm. or more), narrow, drawn out to a fine
point, and also beset with relatively short prongs (PL XXIV. fig. 5). On the other
hand, the ray directed towards the parenchyma is greatly reduced, or has entirely
disappeared, so that the majority of autogastralia no longer represent hexacts but
pentacts, in which the short, conical, ti'ansverse rays are beset with small prongs either
on the outer extremities alone or all over.

The difference between the short crowded liexact autodermalia and the long pointed
pentact autogastralia is very clearly indicated on the outer margin of the disc where
both border on one another, yet it may also be readily recognised on the other side by
the shorter projecting distal ray and the short proximal ray of the outermost auto-
gastralia. Both are really but difterent members of the same series.

It is noteworthy that in those autodermalia with a greatly prolonged distal ray, which
occur here and there between the short and crowded forms, the inner proximal ray is
usually absent (PL XXIV. fig. 3).

In the parench}Tna of the stalk the principal hexacts are in the minority, and the

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

long diacts with rough rounded extremities are accordingly all the more abundant.
These are here, moreover, not so uniformly disposed in a longitudinal direction as in
thinner stalks, but occur with some irregularity, though for the most part longitudinally.
These principalia are extensively united into a firm framework by means of synapticula.
It is to be regretted that the dermal skeleton of the stalk has been rubbed oif, and also
that the inner lining of the cavity of the stalk is no longer intact.

Caulophacus elegans, n. sp. (PL XXV.; PI. XXVI. figs. 1-3).

A second species of Caulophacus was trawled to the east of Japan (Station 241, lat.
35° 41' N., long. 157° 42' E.) from 2300 fathoms, on red clay ground. Three specimens
were collected, two of them small forms with short imperfect stalks. One of these, as
figured on PL XXVI. fig. 1, show^s a marked resemblance to Caulophacus latus. The
straight, round, hollow stalk, 1"5 mm. in diameter, is continued by a trumpet-like expan-
sion into the slightly curved superior extremity with a convexo-concave circular disc, with
somewhat overhanging margins. In the specimen figured the diameter of the disc
measures almost 2 cm., and the greatest thickness 5 mm. Another much injured
specimen is somewhat smaller. A tolerably distinct form is figured in PL XXV. fig. 1,
from a somewhat larger third specimen. The hollow stalk is somewhat bent at the end,
and though on the whole straight, is slightly curved here and there in an undulating
manner. It measures 3 mm. in diameter, and bears terminally a biconvex lens-shaped
disc, 4 cm. in breadth and r5 cm. in thickness, with a somewhat downwardly bent
margin. In spite of some superficial ii'regularities on the upper side and several
shallow folds below, both surfaces appear on the whole to be uniformly arched, though the
upper is the more convex. The stalk, though broken off" below, attains a length of 9 cm.
I am inclined to beheve that the piece of stalk which was brought up along with the
former, and which exhibits a similar structure and a length of 15 cm. (PL XXV. fig. 2),
really belonged to the kame specimen, and formed the inferior extremity of the stalk,
which would accordingly measure 24 cm. in length.

As in Caulophacus latus, the principalia consist of strong, regular, smooth hexacts,
and long diacts with rounded, more or less rough, extremities. These have an isolated
or a bundle-like distribution. Each of the rays of the hexacts exhibits a bulging swelling
in the middle, and a rounded terminal point. The diacts usually exhibit a central
axial canal intersection, but as a rule there are no knob-like protuberances or ring-like
swellings to represent the remnants of abortive rays. Between the principalia the same
small regular pronged hexacts, with marginally pronged, arched, terminal plates, occur in
great numbers, resembling those which we have akeady become acquainted with in
the other species, and in Balanites pipetta (PL XXIV. fig. 7). Simple division of a ray,
as shown in PI. XXV. fig. 7, may occasionally be observed. On the other hand.

REPORT ON THE HEXACTINELLIDA. 127

regularly developed discohexasters are of by uo means rare occurrence. These are
provided with numerous (eight or more) prong-bearing, long, terminal rays, which are
placed upon the terminal expansion of the short smooth principal ray, and do not differ
essentially in other respects from the rays of the discohexacts (PL XXV. fig. 6). It is
remarkable that these many-rayed discohexasters lie almost exclusively under the
surface of the disc, and that scattered between them similar rosettes occur, with a few
(often only four or three, or even two) pronged terminal rays on every principal. One
can detect a tolerably continuous series of transitions between the simple pronged
hexacts and those many -rayed discohexasters.

Sometimes I also observed rosettes with numerous long, smooth, pointed, terminal
rays, i.e., oxyhexasters. The terminal rays are inserted on a terminal expansion of the
short principal rays like those of the discohexasters.

The slight differences in the numerical proportions of the discohexacts and disco-
hexasters, in the thickness of the terminal rays and in the dii'ection of the same, which
are discoverable between the smaller fungiform specimens and the larger forms with lens-
shaped bodies, I regard as insufficient for the differentiation of particular species.

The dermal skeleton resembles very much that of Caulophactcs latus. Here, too, the
proximal ray of the pentact hypodermalia is usually beset with prongs (PL XXV. fig. 8),
and the autodermalia are hexact pinuli with broad, short, fir-cone-like outer rays,
while their slightly conical transverse rays and the similarly formed proximal ray only
exhibit a sUght irregularity of surface on the terminal portion (PL XXV. fig. 5).

The gastral skeleton lying on the convex upper side diff'ers from the corresponding
skeleton of Caulophacus latus only in the fact that the autogastralia are not pentact
but hexact pinuli, in which both the developed proximal, and the four transverse rays, are
smooth and only slightly rough on the extremity, while the freely projecting, scaly, pronged
distal ray is not so narrow and pointed as in the latter, but becomes outwardly broader
and more swollen, so as to terminate in a knob-like rounded extremity (PL XXV. fig. 4).

The parenchyma of the tube-like stalk contains, just as in Caulophacus latus, some-
what rough, rod-like diacts, which are rounded at both extremities. They are greatly
increased at the expense of the hexacts, which have disappeared, and are almost all
approximately parallel to the long axis of the stalk. It is only in the inferior portion of
the stalk that the diacts ai'e firmly united by means of synapticula.

The pentact-hypodermalia of the dermal skeleton are roughened only on the extremity
of the proximal, and of the four transverse rays.

The autodermalia of the stalk are pentact pinuli in which the proximal ray is
atrophied to the size of a small tubercle, while the freely projecting distal, on the
contrary, has become a squamous pronged ray, 075 mm. or more in length. This
increases in breadth outwardly, and, like the autogastralia of the discoid upper surface
terminates in a knob-like thickening.

128 THE VOYAGE OF H.M.S. CHALLENGEE.

Genus 2. Trachycaulus, u. gen.
Only one species is referred to this genus.

Trachycauhis guiiittii, n. sp. (PI. XXVI. figs. 4-9).

In the middle of the southern part of the Pacific (Station 289, lat. 39° 41' S., long.
131° 23' W.) there was trawled from a depth of 2550 fathoms and a red clay ground,
the round stalk — 12 cm. in length and 2 to 3 mm. in diameter — which is figured in
PI. XXVI. fig. 4. At the lower thinner end this stalk is slightly bent ; it is otherwise
straight, and is at the ujiper end broken across transversely. The surface of this hollow
and tolerably firm stalk is covered with hairs. Although nothing remains of the sponge
body belonging to the stalk, I cannot doubt from the structure of the fairly well-preserved
dermal skeleton that we have to deal with a form closely related to the genus Ccmlo-
phaciis. Nevertheless the presence of a very peculiar rosette, and the absence of a form
characteristic of the two known species of Caulophacus, even in the stalk, namely, of
the frequent dispersalia, pronged discohexacts and discohexasters, prevent one from
referring this specimen to that genus.

Greatly prolonged rod-like smooth diacts with rounded rough extremities form a firm
lattice or ladder-like framework for the parenchyma, being bound together in parallel
longitudinal rows by means of numerous transverse synapticula. Slender regular
hexacts with smooth pointed rays occur in very various positions in the parenchyma,
partly as integral parts of the lattice-work, partly only in external connection with it, or
else quite freely on its surface.

Especially remarkable, however, on account of their size ( = 0'4 mm. in diameter) and
singular form are the rosettes with short principal rays, upon each of which four long
sickle-like terminal rays are placed. These forms occur scattered in the subdermal spaces
between the hexacts which have just been mentioned (PI. XXVI. figs. 7, 8). They
recall the sickle rosettes of Oscar Schmidt. The inner portion of each terminal ray is
very thin ; outwardly the ray becomes graduall}^ thicker, and bears on its extremity a
more or less sharply hooked curve finally terminating in a backward bent point.
Great variations occur in the form and size of this hook-shaped terminal portion which is
sometimes quite irregularly bent, or else divided into two or three diverging points. •
Occasionally the four hooks which bend towards one another may fuse together.

There is a striking agreement between these forms and the " sickle rosettes " found
by Oscar Schmidt in his Hcrtwigia falcifera, and figured by him in his Spongien des
Meerbusens von Mexico, Taf vi. fig. 8.

The dermal skeleton is composed of greatly prolonged hexact autodermal pinuli, in
which the four transverse rays and the almost equally long proximal ray are simple.

REPORT ON THE HEXACTINELLIDA. 129

smooth, or at most somewhat rough towards the pointed ends, while tlie delicate, freely
projecting, fir-tree-like distal, which is two or three times longer, is beset with oblique
outwardly directed prongs, and runs gradually to a point towards the outer extremity
(PL XXVI. fig. 9).

The specific name I have given is in honour of my friend and former colleague in
Graz, Professor Gurlitt.

Family III. R os s ell id .^i (Pis. LIII.-LXIX.; PL GIL).

Goblet- or beaker-shaped, with walls of varying thickness. Some rest either directly,
or by means of a longer or shorter cylindrical stalk, upon a solid basis, others are rooted
in mud by means of a basal mass of spicules. The external surface of the body is in
some smooth and naked, in others armed with prominent pleuralia of varying length.
The simple wide gastral cavity opens by a simple, round, more or less broad, oscular
aperture, the margin of which is either naked or armed with a border or circlet of spicules.
A special characteristic of the family is to be found in the fact that the distal ray of the
dermalia is ahoays absent. The dermalia occur as pentacts, tetracts, diacts, or even
monacts. The gastralia have usually no freely projecting proximal ray, but in some cases
they occur as fully developed hexacts.

Genus 1. Lanugindla, 0. Schmidt (PL LIII. figs. 3-5).

1869. 0. Schmidt, Mittheil. dos naturw. Vereines fiir Steiermark, jjp. 89, 2G1.

1870. O. Schmidt, GrundziigB einer Spongieufauna des atlant. Gebietes, p. 13.
1870. Sav. Kent, Montlily Micr. Journ., voL iv. p. 247.

On a specimen of AphrocalUstes from St. lago, one of the Cape Verde Islands, Oscar
Schmidt found in 1870 some small spherical or ellij^soidal sponges of very elastic consist-
ence, exhil)iting a central cavity and a wide superior osculum. The outer surface, apart
from spicules projecting here and there, was smooth— just as if varnished. The interior
contained, according to 0. Schmidt, prominent, smooth or finely spiuose hexacts and simple
oxy diacts with intersecting axial canals. In the outer layers, between the projecting
needles, hexasters occurred in which each of the short principal rays was soon divided
into four or five tuberculated branches, with transverse terminal plates. In the external
dermal layer, and at the margin of the osculum, numerous sexradiate spicules occur,
besides quadriradiate forms filling up the interspaces, and in part provided with sharply
cornered knotted extremities. The smoothness of the outer surface seemed to be due
to these four-rayed sjaicules.

Oscar Schmidt named the newly discovered Hexactinellid Lanuginella, because it

(ZOOL. CHALL. EXP. PART LIII. — 1880.) Ogg 17

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

felt iu the dried condition like a delicate woollen cocoon, and he added the specific name
pupa on account of the shape already noted.

Only one species is known, Lanuginella pupa, 0. Schmidt.

Lanuginella pupa, 0. Schmidt (PL LIII. figs. 3-5).

In the vessels which contained the spirit specimens collected at Station 192, off Little
Kj Island, I found, along with the large Pheronema gigantewn and several specimens of
Polylophus philippinensis, a number of oval and spherical structures 2 to 8 cm. in
diameter, which turned out on closer inspection to be small sponges. Some of them
were seen to be representatives of the Lanuginella pupa described by Oscar Schmidt,'
while the others were young forms of Poli/lophus j)ltilipi'>inensis which will be again
referred to in the description of that form.

Lanuginella pupa, which occasionally grows on other Hexactinellida, has the form
either of a completely closed smooth sphere, 2 to 3 mm. in diameter, or of a larger oval
body with a somewhat flattened basal pole, and a round oscular opening about 1 mm. in
breadth at the narrow upper end {PI. LIII. fig. 3a, b, 4, 5).

On a longitudinal section of the larger ovoid specimen, one observes the central
longitudinal gastral ca-vdty, which is rounded off at the lower end, opening superiorly of
course in the already mentioned osculum. Into this gastral cavity, which is Lined by
a subgastral trabecular network, the sack-like chambers open, either directly, or by
means of canal-like efferent passages, and in this case the whole chamber layer is deeply
folded. Between the smooth external network or dermal membrane and the folded
chamber layer, there is an external or subdermal trabecular space, including a subdermal
trabecular network and the subdermal sj)aces or afferent canals which penetrate tlie latter
(PL LIII. fig. 5).

The parenchyma of the sponge contains, as 0. Schmidt noted, medium sized oxyhexacts
with long straight or slightly curved rays usually disposed radiaUy and tangentially.
Besides these principal forms,, somewhat long, straight, or slightly curved oxydiacts occur,
with a central axial cross, over which cruciate tubercles often project outwards as indications
of the undeveloped rays. The disposition of these oxydiacts, which are often somewhat
roughened terminally, is for the most part perfectly or approximately tangential, though,
in some cases, it may be more or less divergent (PL LIII. fig. 5).

Between these large supporting spicules, we have to note the ii-regular scattered
occurrence of small regadar oxyhexacts with thin rays, and of numerous small disco-
hexasters, in which each of the short, simple, cylindrical princijmls bears three, four, or
five long diverging terminal rays with four to six-toothed, somewhat recurved, transverse
discs at their extremities (PL LIV. fig. 3; PL LIII. fig. 5). Besides these, the

1 Grundziige einer Spongienfauna ties atlant. Gebietes, p. 13.

REPORT ON THE HEXACTINELLIDA. 131

parenchyma includes small rosettes in varying abundance and irregular distribution. One
of these forms admits of the designation plumicome. Each of the six simple cylindrical
principal rays bears a small plano-convex transverse disc, from which very fine S-shaped
terminals arise in concentric rows in perianth-like fashion, very much as in the plumicome
of Poli/Iophus philipinncnsis represented in PI. LIV. figs. 4, G. Finally, there are very
minute isolated rosettes, in which the short simple principal rays bear terminal transverse
discs, having a thick brush-like fringe of somewhat long and very thin, radially disposed,
knobbed terminals, like those in the rosette of Rossella antarctica, figured in PL LV.
fig. 6.

I must further note that I was not able to find these two extremely delicate and
inconspicuous forms of spicules in all the specimens which I examined, but only in a few.

The hj'podermalia are simple smooth oxypentacts in which the unpaired proximal
ray is always straight and somewhat long, and penetrates radially, like a nail, far
into the parenchyma. The four tangential rays, intersecting at right angles, extend
close beneath the dermal membrane, corresponding to the curvature of the general
sponge surface in exhibiting a more or less marked inward curvature (PI. LIII.
fig. 5). In the quadrate meshes, which are formed by the apposition of these tangential
rays of adjacent hypodermal pentacts, smaller pentacts of a similar type occur, with their
tangential rays disposed parallel or diagonally to the tangentials of the larger forms.
The dermal membrane itself contains exclusively numerous cruciate autodermal tetracts
in which the rays are more or less roughened, somewhat narrowed towards the extremity,
and ending finally in a conical point or in a slightly blunted fashion.

At the lower somewhat truncate basis of the egg-shaped sponge, the hypodermal
oxypentacts are modified into long anchors by the thickening and more marked curvature
of the tangentials, and by the decided thickening and elongation of the proximal
tangential ray. These anchors are gradually more and more protruded from the sponge
body, and may serve for the attachment of the sponge to its soft substratum. At the
margin of the oscular aperture, I have sometimes observed long, pointed, rod-like needles,
projecting radially to a more or less marked extent. These may be fitly regarded as
marginalia.

In the evidently much younger and completely closed spherical specimen, the skin
exhibits only hypodermal oxypentacts, and simple, strongly developed, cruciate auto-
dermal tetracts. Internally, between the simple oxyhexacts, moderately long oxydiacts
occur, with central nodes of intersection. Numerous discohexasters are also present
(PI. LIII. fig. 4).

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

Genus 2. PohjJophus, n. gen.

1872. Gray, Ann. ami Mag. Nat. Hist., ser. 4, vol. x. p. 137.

1873. Gray, Op. cit., ser. 4, vol. xi. p. 234.
1873. Carter, Op. cit., ser. 4, vol. xii. p. 361.
1875. Carler, Op. cit, ser. 4, vol. xv. p. 118.

1877. Marshall and Meyer, Mittheil. Zool. Mus. Dresden, vol. ii. p. 261.

History. — Among the Hexactinellida which Dr. A. B. Meyer brought with him from
Zebu (Philippines), Gray found in 1872 a comparatively large specimen which bore
peculiarly disposed projecting tufts of needles on the posterior half, while the anterior
extremity exhibited the wide aperture of a spacious gastral cavity. He compared this
sponge with Tetilla 'polyura, 0. Schmidt, while Carter, to whom it was handed over, at
once detected a close relationship between this form and his Rossella antarctica. Gray
therefore named it Rossella philippinensis. In 1873 Gray received some young buds of
the same sponge, which were also obtained from Dr. A. B. Meyer. These were briefly
described under the title Psetalia glohidosa, Gray. In 1875, however, Carter published
a detailed description with good figures of the form and structure both of the older
specimen and the young buds [Psetalia glohidosa, Gray). Both the general structure
and the form of the spicules exhibited the close resemblance which those specimens bore
to Carter's Rossella antarctica. Finally, Marshall and Meyer subjected these results to a
close scrutiny in 1877, and described in detail the form and structure both of the entire
sponge and of the various spicules.

The characteristic anchors found in the root and tuft-spicules were regarded
by Marshall and Meyer not as homologous, but as analogous to the anchors in the
tuft-spicules of Euplectella, Semperella, and Hyalonema, since the double cross was
not found in these latter genera in the anchor head itself but further up in the shaft, so
that the anchor teeth could not be looked upon as modified rays of hexradiate spicules.
The nine distinct forms of spicules observed were found by Marshall and Meyer to be
distributed as follows : —

A Appendicular spicules :— I C. Spicules of the parietes : —

1. Eoot-spicules of anclior-like fomi. 5. Gastral and facial tive-rayeJ spicules.

2. Uniaxial spicules of the spicular wreath. I 6. Six-rayed spicules.

7. Uniaxial spicules.

8. Rosettes.

B. Spicules of the dermal skeleton : —

3. Five-rayed spicules.

4. Four-rayed spicules with tubercles. D. Gastral spicules : —

9. Small six-rayed spicules.

A young specimen, 5 mm. in diameter, in which the anchor-tufts were tolerably
well developed, already showed the same spicular forms in a similar arrangement. A
specimen preserved with its soft parts in spirit was found to be fiUed ■ftitli small green
and grey granules ; some cell -nuclei were seen, and here and there a fragment of

REPORT ON THE HEXACTINELLIDA. 133

siliceous matter or a drop of fat. On the outside of the sponge, in fine tangential
sections, the small dermal pores could be detected in the quadrate meshes of the dermal
skeleton.

This genus contains only a single species.

Polylophus philippinensis, (Gray) (PL LIII. figs. 1, 2 ; PI. LIV.).

Some very young spherical or egg-shaped specimens of Polylophus [Rossella, Gray)
philqypinensis, Gray, were procured, along \vith Lanugmella pupa, 0. Schmidt, near Little
Ki Island (Station 192) from a depth of 129 fathoms. But besides these, numerous
adult forms of the same species were collected near the Philippine Island, Zebu (Station
209, lat. 10° 14' N., long. 12,3° 54' E.), from a depth of 95 fathoms and a blue mud
bottom. Some of these specimens were as large as a man's fist. They agree exactly
with the careful description and figure given by Carter^ and by Marshall. All the
specimens exhibited a short, thick-walled, cup-form, with a wide round upper opening to
the simple sack-like gastral cavity. Into the latter the efferent canals open with more or
less wide round apertures, while the outer surface, covered with a fine lattice network,
is elevated into numerous mammilla-like papillae. From the summit of each of these
radially disposed papiUse, which are especially thick and long on the lower surface of the
body, a thin bundle or tuft of long hair-like siliceous spicules projects. On the upper
and lateral surfaces these tufts of needles stand out radially, stifi" and straight ; on the
median and basal papillae, however, the siliceous hairs are much longer, and are apposed
to one another to form a long, broad, loose root-tuft, by which the sponge is anchored in
the mud (PI. LIV. fig. 1). Very frequently, on the larger specimens, some or most of
the papillary elevations are thickened terminally in a club-shaped fashion, and more or
less markedly constricted at the base, so as to form pear-shaped appendages represented
by a whole series of stages varpng greatly in size and difierentiation. Each protuberance
begins as a small conical, subsequently pear-shaped, structure 2 to 3 mm. in length, and
attains the size of a hazel nut. Fully developed forms exhibit lateral radial papillae,
and a basal radial tuft, while at the upper pole a circular aperture communicates with
the central cavity. The whole form of a young Pohjlophus is thus assumed, and it
remains connected to the mother organism merely by a few spicules. These give way of
themselves, and a small form is set free which imdeniably resembles the mother sponge
in all essential points (PI. LIV. fig. l). It is noteworthy that almost all the larger
specimens exhibit that tendency to form buds which has been repeatedly noted by
earlier observers. Between the papillae the external skin-covering appears to be as
smooth as the concave interior surface of the gastral cavity. On the simple, somewhat
sharp-edged oscular margin, there is no trace of a cufl'-like marginal fringe of spicules.

• Ann. and Mag. Nat. Hist, ser. 4, vol. xv., pi. x. fig. 1.

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

The principal supporting spicules of the loose parenchyma consist of large or
medium sized smooth oxyhexacts, with straight rays, and of straight or slightly curved
oxydiacts, which are roughened towards their conically pointed extremities, and exhibit
at their middle point four cruciate transverse tubercles, or an annular swelling, or
at least the intersection of axial canals. These oxydiacts either lie isolated, or are
for the most part disposed parallel to the surface. The larger and stouter forms
are usually accompanied and ensheathed by a number of long thin oxydiacts or comitalia.
As to smaller spicules, the parenchyma includes numerous regular oxyhexacts with more
or less roughened rays, and also a large number of oxyhexaster forms which are very
characteristic of this genus. In these the principal rays are comparatively long, simply
cylindrical, or narrowed in a slightly conical fashion, and are always somewhat roughened
terminally, where they divide into two or three short, diverging, conical terminals
(PI. LIV. fig. 5). When only two terminal rays are developed at the end of the
principal, the planes of forking of the two rays in the same axis of the spicule are at
right angles to one another. Finally, the parenchyma contains peculiar rosette forms,
occurring in regular, but not very abundant distribution. These belong to the ph;mi-
come type, but are not so minute as those of Lanuginella puj^ct. They attain to about
the same diameter as the above-described oxyhexasters, namely, about O'l mm. The
short, cylindrical princi]:)al rays bear terminally a plano-convex transverse disc, from the
outer convex surface of which several concentric whorls of S-shaped terminal I'ays arise.
These go to form a perianth-like form with several whorls, enclosing a central funnel-
shaped space (PI. LIV. figs. 4, 6). The thi'nnest portion of these delicate S-shaped
terminals lies near their origin, while the outer end becomes thickened in a somewhat
club-shaped fashion, exhibiting, however, a slight attenuation at the extremity.

The dermal skeleton consists especially of hypodermal oxypentacts of variable size,
in which the long proximal ray is always straight and radially disposed, while the
four tangentials which go to form a quadrate lattice-work are usually curved gently
inwards. Like the proximal ray, the tangentials frequently appear to be roughened
near their conical extremities. Between these substantial pentact hypodermalia, the
dermal membrane includes small cruciate autodermal tetracts in great abundance and
in rectangular distribution. In these, the rough cylindrical rays, which are rounded off'
or even truncated at their ends, are curved slightly inwards, so that the whole spicule
appears to be uniformly arched with the convexity outwards (PI. LIV. figs. 2, 7). Some-
times, at the point of intersection of the four arms, an inward projecting boss or
tubercle persists as an indication of the undeveloped fifth (proximal) ray. In a few of
these dei-malia, the fifth ray is actually developed, and exactly resembles the
tangentials.

The gastral skeleton differs strikingly from the dermal. The hypogastral pen-
tacts, as we should naturally expect, are wholly absent. Instead of them, there are

REPORT ON THE HEXACTINELLIDA. 135

strongly developed, smooth oxydiacts, with slight cui-vature or twisting, and on these
numerous oxyhexacts, of variable but not verj^ considerable dimensions. In the latter
the four tangential rays are tangentiall)^ disposed in the gastral membrane, while the
internal, usually longer radial projects freely into the gastral cavity, and the external
radial into the subgastral trabecular space. All the rays of the smaller gastral oxy-
hexacts run out gradually to a point, and are of approximately equal length, while in the
larger gastral oxyhexacts the internal free ray is decidedly longer than the others, so
that a dagger-form results. All the efferent canals are clothed internally with small
delicate oxyhexacts, in which the inner radial ray projects to a greater or less distance
into the canalicular space (PI. LIV. fig. 2).

The long radially projecting spicules, which arise in a tuft from the terminal pole of
each of the numerous lateral and basal papillse, are smooth oxypentacts. Their interior
end runs gradually to a point, while the outer terminates in a four-toothed anchor, in
which the four cruciately disposed, strongly developed teeth do indeed at first arise at
right angles to the shaft, but become more or less markedly curved inwards (PI. LIV.
%9).

The basalia forming the root-tuft only differ from these lateral radially projecting
pleuralia in their greater strength and length, and also in the fact that many of them,
which have arisen from the curvature of the lower pleuralia, extend no longer quite
radially, but are more or less markedly curved downwards (PI. LIV. fig. 1).

A good general survey of the structure of this species can be obtained by making-
sections of the small, pear-shaped buds. A combined diagram representing this is given
in PL LIII. fig. 2. It will be seen that the sack-shaped chambers, taken as a whole,
make up a much folded layer, which forms the boundary between the water-passages
leading from the external skin and those leading into the common gastral cavity. The
circular oscular aperture of the gastral cavity appears at a later stage at the broad distal
pole of the pear-shaped bud. The bud at the same time pushes itself outwards on the
bundle of pleuralia belonging to the papiUa. It forms for itself papillae w^th pleural and
basal tufts of spicules. Finally it becomes detached and rooted in the mud.

As in the above-described genus, Lanuginella, the basalia (and pleuralia also) are
nothing but long drawn-out and protruded spicules of the hypodermal pentact series.
This may be presumed by comparing the young basalia or pleuralia with the adjacent
hypodermalia, but the fact is distinctly demonsti'ated by studying sections of young,
spherical or oval specimens, 2 to 5 mm. in diameter (PL LIII. figs, la, h) (probably
developed from ova), which were found in the same glasses as the Hexactinellids from
Station 192, Little Ki Island. At the lower, often somewhat truncate pole, opposite the
oscular aperture, between the ordinary hypodermal pentacts, other forms occur in which
the four tangential rays do indeed lie in the dermal membrane, but are bent more
markedly inwards than the others, and are beginning to resemble the anchor-teeth of the

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

basulia, while others of similar structure are bent somewhat outwards above the level of
the skin, and others, again, especially in somewhat larger sponges, arc drawn out into
true anchors, which exactly resemble the basalia of the adult specimens.

Genus 3. Rossella, Carter (PL LV., and woodcut, fig. 4).

1872. Carter, Ann. and Mag. Nat. Hist., vol. ix. p. 409 {Ros^scUa antardira).

1872. Bowerbank, Op. cit., vol. x. pp. 58-61.

1872. Gray, Op. cit., vol. x. p. 134 {Rossella philippinensk).

1873. Carter, Op. cit, vol. xi. p. 275.

1873. Gray, Op. cit., vol. xi. p. 234 (Pectalia (jlohidosa).

1873. Wj'ville Thomson, Depths of the Sea (Rossella vclata).

1873. Carter, Ann. and Mag. Nat. Hist., vol. xii. p. 361.

1874. Gray, Ojx cit., vol. xiii. p. 284.

1875. Carter, Op. cit., vol. xv. p. 113.

1875. Willemoes Suhm, Zeitschr. f. wiss. Zool., vol. xxv. p. 25.

1876. Marshall, Op. cit., vol. xxvii. p. 113.

1877. Zittel, Abhandl. d. IL CI. k. haier. Akad. d. ■^^^ss., vol. xiiL
1877. Marshall and Meyer, Mittheil. Zool. Mus. Dresden, vol. ii. 261.

History. — Among some Antarctic sponges which were dredged by Captain Sir James
Ross during his voyage of discovery and research in the Southern and Antarctic region
in 1839-43 (lat. 74i° S., long. (?), and lat. 77^° S., long. 175' W., from a depth of
300 fathoms), Carter found, in 1872, two remarkable and hitherto unobserved forms
of siliceous elements — the one a five-rayed spicule in which the longer ray, measuring
about 1 cm. in length, ran out to a fine point at one extremity, but passed at the other
into four approximately cruciate, straight or slightly bent transverse arms, disposed at
right angles; the second, a four-toothed anchor with a long (at least 4 cm.) shaft and
cruciately disposed, somewhat strong, recurved teeth, which ran out into simple points.
In some anchors the straight shaft was continued to the other side of the anchor teeth, so
that with the projecting point the entire structures became hexradiate.

While all parts of the anchors were smooth, the rays of the five-rayed spicules, which
are hkewise round, were thickly beset with very fine microspines ; on the four pointed
transverse arms, however, numerous larger slightly bent tubercles also occurred with
their free points turned away from the intersection of the arms. Although the sponge
bodies to which these spicules belonged were not discovered, it seemed certain that the
two kinds of spicules belonged to one and the same sponge, which Carter named Rossella
antarctica. This procedure certainly seemed at first somewhat rash, and Bowerbank did
not fail to declare his disapproval.^ Further discoveries, however, soon served to justify
Carter's opinion.

In the same year, Gray found a second species of the same genus among the sponges

' Ann. and Mag. Nat. Hist., ser. 4, vol. x. p. 58.

REPORT ON THE HEXACTINELLIDA. 137

which A. B. Meyer had sent to the British Museum from Zebu, one of the Philippine
Ishands, and named it Rossella philippinensis} It was a longish oval specimen as large
as a walnut ; its transversely truncated upper end bore the wide round orifice of a deep
central cavity, while, from the inferior half of the otherwise smooth body, a cylindrical
bundle of long siliceous spicules ran out radially at different distances from one another,
and then bent downwards into a tuft. That this sponge belongs to the genus
Rossella was confirmed by Carter in a communication addressed to Gray ; the four-armed
spicules of the skin with their somewhat backwardly bent branches were to him sufficient
proof of the fact. Yet, as Carter remarks, this form may be readily distinguished from
Rossella antarctica by the fact that the arms of the outer spicules are smooth, and not
beset, as in the latter, with delicate microspines.

Carter has also directed attention to the great similarity between the spicules of
Rossella phiUpinnensis and those found in the genus Crateromorplia, Gray.

Wy ville Thomson," gave an account of a third species of Rossella, which was dredged
in 651 fathoms, to the west of the opening of the Strait of Gibraltar. The oval body of
tliis remarkably elegant sponge, described as Rossella velata, bears superiorly (as in
Pheronema) a single large round osculum, but instead of forming a cup uniformly lined
with a netted membrane, the oscular cavity divides at the bottom into a number of
branching passages as in Pheronema annse described by Leidy.

" A delicate outer veil about a centimetre from the surface of the sponge is formed by
the interlacing of the four secondary rays of large five-rayed spicules which send their
long shafts from that point vertically into the sponge body. The surface of the sponge
is formed of a network of large five-radiated spicules arranged very much as in
Pheronema."

In a paper on Sarcohexactinellidan Sponges,^ Carter has noted that in Rossella velata
and Rossella philippinensis "the minute equi-armed hexradiate spicules pass from the
equi-armed hexacts with bifurcated and pointed extremities to the same with capitate
extremities, and lastly into an unclescribed form where the ends of the arms are terminated
by a small conical tuberculated inflation presenting a short straight spine on the apex,
which spine is surrounded by almost innumerable linear filaments rising each from one of
the tubercles, attaining various heights and bending outward like the expanded petals of
a tubular flower, forming one of the most exquisite objects in nature. It might be named
' pappiform' flexed and simple in contradistinction to another kind in which the filaments
are straight and capitate."

The generic diagnosis of Rossella was given by Carter in his Review of the
Hexactinellida * in the following words: — "Rosette few- or many-rayed: rays few of
equal length straight and pointed or spinocapitate ; or multitudinous, of unequal length,

' Ann. and Mag. Nat. Hist., ser. 4, vol. x. pp.- 137, 138. " The Depths of the Sea, p. 418, 1873.

3 Ann. and Mag. Nat. Hist., ser. 4, vol. xi. p. 279. * Ann. and Mag. Nat. Hist., ser. 4, vol. xii. p. 3G1.

(ZOOL. CHALL. EXP. PART LIII. 1886.) Ggg 18

138 THE VOYAGE OF H.M.S. CHALLENGEE.

withuut lieacls, flexed outwards and arranged en Jieur-de-lis ; pappilbrm, or sometimes
many-rayed with rays straight and capitate. Anchoring spicules all smooth, stout and
terminating respectively in heads oifouv equally stout recurved spines or hooks."

In 1875 Carter happened to discover among the treasures of the British Museum "a
glass jar containing two small specimens of the veritable RosseUa antarctica dredged up
by Sir J. Eoss in 300 fathoms 74^° S. lat.," longitude not given.

"The general form of this sessile or fixed sponge was sack-like (compressed) with the
upper end truncated and open and the lower one conical and closed. External surface
uniformly cribellate and monticular, covered by a thin layer of spicular lattice-work and
surrounded by three forms of projecting spicules : — viz., (1) stout linear smooth nearly
straight fusiform acerate spicules, finely pointed at each end, constituting an erect beard
round the aperture; (2) anchoring spicules which increase in number, size, and length
towards the lower or conical end ; (3) crucially headed or veil spicules projecting chiefly
from the monticules over every part of the external surface but the aperture, consisting
of a shaft whose pointed or inner end is fixed in the sarcode of the body, and whose free
or outer one is terminated by four long arms spread out horizontally so as to intercross
with those of its neighbours, and thus form a general veil-like covering separated from
the body by the length of the shafts between the body and their heads respectively ;
shaft smooth or only microtuberculate over the imbedded end, arms more or less flexuous,
fine-pointed, parting from the head of the shaft at difi"erent angles, covered almost
throughout with minute spicules closely approximated, amongst which here and there is-
a much larger sjjine curved and inclined outwards or from the head of the shaft."

Among the seven other forms of spicules which Carter has described from the body
of the sponge the following are especially noteworthy : —

" (1) Very minute sexradiate rosettes with numerous straight capitate rays, and (2)
sexradiate rosettes with thick sparsely spined arms, whose inflated ends supjsort four or
more indistinctly capitate rays ; rays raicrospined, thick at first, then becoming finely
attenuated and terminating in a hardly perceptible capitate inflation ; rays at first straight
and parallel like the prongs of a dinner-fork, becoming more or less divergent towards-
their extremities."

In his systematic review published in the year 1875,^ Carter placed the genns Bossella
along with Cratcromorpha in a grouji called the Eosettifera within his family of the
Sarcohexactinellidse.

The characteristics which Marshall " in 1876 assigned to the genus RosseUa ran thus:
— " Monozoic, root-tufts springing from papilla-like hillocks on the j^arietes, dermal
skeleton composed of five-rayed spicules. In RosseUa antarctica a peristomal spicular
wreath (whether present in the other species is doubtful)."

C/iaracter of the Genus. — Thick-walled, ovoid or cask-shaped goblets, with a superior,

' Ann. and Mag. Nat. Hist., ser. 4, vol. xvi. p. 199. ^ Zeitschr.f. wiss. Zool., Bil. xvii. p. 127.

REPORT ON THE HEXACTINELLIDA. 139

circular, smooth-margined opening to the deep saccular gastral cavity. A group of
diact and pentact pleuralia, whose tangential rays form a kind of veil, project radially
from each of a number of small, regularly disposed, rounded elevations of the
surface. In the neighbourhood of the simple unarmed oscular wall a number of strong
isolated diacts project upwards. Between the dermal layer and the sieve-like gastral
layer which stands above it the soft parts form a deeply folded plate with alternating
iuhalent and exhalent radial funnel-shaped canals.

The parenchyma contains oxyhexasters with very short main rays and various disco-
hexasters. The spicules of the dermal membrane are almost exclusively pentacts.

1. Rossella antarctica, Carter (PI. LV.).

Of the two species of Rossella hitherto known, viz., Rossella antarctica, Carter,
<and Rossella velata, Wyville Thomson, the former is represented in the Challenger
collection by numerous specimens, which vary considerably in size, and were collected at
four different stations.

Several specimens, from 2 to 4 cm. in height, growing on small fragments of bivalve
shells and similar objects, were dredged to the south-east of Prince Edward Island
(Station 145, lat. 46° 43' 0" S., long. 38° 4' 30" E.), from a depth of 140 fathoms, and
volcanic sand ground. Numerous specimens of very various dimensions, up to
30 cm. in length and 15 cm. in breadth, were dredged to the south of the Kerguelen
Islands (Station 150, lat. 52° 4' S., long. 71° 22' E.), from a depth of 150 fathoms, and
a coarse gravel ground. Other forms of almost equal dimensions Avere obtained in
Christmas Harbour, Kerguelen, at various depths, from volcanic mud ground. All
these more or less well-preserved spirit specimens were attached at their base to stones,
either directly or by means of small processes. Finally, several specimens, attaining a
length of 20 cm., were trawled to the east of Buenos Ayres (Station 320, lat. 37° 17' S.,
long. 53° 52' W.), from a depth of GOO fathoms, and a green sand ground. These also
were attached, either directly or by means of small processes and prolongations, to stones
or other solid bodies.

The general form of this sponge may be described as barrel- or keg-like, or else
as resembling an elongated pear. The lower, sometimes somewhat narrowed, solid
end is either attached over its entire breadth to some large solid body, or is
fastened by short lateral processes of irregular form to various smaller objects. I have
never found the peculiar loose root-tuft which Carter represents in his diagrammatic
figure ; ' but I have noticed such a structure on one of the two original specimens
preserved in the British Museum, and it is quite possible that this modification, as

' A nn. and Mag. Nat. Hist., ser. 4, vol. xv. pi. x. fig. 4.

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

present on one of the specimens examined by Carter, is conditioned by the sojaewhat
looser nature of the substratum.

On the middle of the upper end there is a sharp cornered, round, oscular aperture,
leading directly into the equally wide, deep, gastral cavity, which has a cylindrical form,
and is rounded off in the blind lower portion. The cavity occupies about a third of the
entire diameter of the body (PL LV. fig. 1). The whole external surface of the body is
beset with numerous regularly distributed, but somewhat undefined, papiUa-like eleva-
tions, which attain a height of 2 to 4 mm., and occur, according to the size of the
sponge, at variable distances of 2 to 10 mm. or more. From each of these papillse a
tuft of radially directed pleuralia projects. Of these spicules some run out to a point,
while others, at a distance of 3 to 8 mm., give off transversely disposed tangentials,
which go to form a veil over the surface of the sponge. Near the oscular opening the
papillfe become less conspicuous and more crowded, but cease altogether close to the
oscular margin. Instead of them, a large number of isolated pointed radial spicules
occur, which form a sort of marginal fringe, not round the sharp edge of the osculum,
however, but at a slight distance further out. Nor do they form a sharply-defined, cuff-
like fringe. I observed similar prostalia round the oscular margin of the original
specimens described by Carter, but besides these some more transversely directed needles,
forming a sort of flat collar, which I did not find developed on the Challenger ajsecimens.
In the uniformly even, but fine velvet-like sieve-network of the gastral membrane,
numerous round apertures could be seen. These measure about 2 mm. in diameter, and
occur in uniform distribution at intervals of 2 to 3 mm. over the whole gastral surface
of the sponge. They represent the main efferent canals, which alternate with the
afferent passages, and are uniformly disposed at right angles to the lateral wall. In cross
sections the parenchyma is seen as a deeply folded layer (PL LV. figs. 1,7).

The principal supporting spicules of the loose parenchyma consist of longer or short,
straight or gently curved diacts, which vary in strength, and are either isolated,
or are irregularly disposed in strands throughout the body. A strongly-developed beam
is usually ensheathed by a layer of thin comitalia. Most of these diacts are smooth except
at tlie extremities, whicli are rough or tubercled, and either end in conical points or are some-
what rounded off and occasionally club-shaped. The central portion usually exhibits an
annular thickening, or four cruciately disposed tubercles. Even when the middle is cpiite
smooth and without j)rojection, a trace of the axial cross is often apparent. Between
the strands of diacts, there is here and there a tolerably abundant occurrence of rough or
even spinose regular hexacts of less than medium size. Numerous minute, rougli, regular
hexacts also occur. There is a very abundant occurrence of small rough oxyhexasters
with a varying number of rays divided into terminals. Frequently only one ray is
thus split, but usually three or four, and rarely all the six. The forking of the j^rincipal
ray always takes place at a slight distance from the frequently thickened node of inter-

REPORT ON THE HEXACTINELLIDA. 141

section, so that the whole spicule comes sometimes to look like a many-rayed star. The
principal i"ay is usually divided only into two, but not unfrequently three diverging
terminals are thus produced, and less frequently more tliau three. Besides these
oxyhexasters, there is a very abundant occurrence of various discohexaster forms, which
are in part somewhat smaller than the oxyhexasters just described, but in part larger.
These also exhibit short simple principal rays, and comparatively long divergent
terminals, with small four- to six-toothed, transverse and somewhat recurved terminal
plates. The number of terminal rays varies very considerably in the smaller
forms ; there are usually three to six or even more on each principal ray (PL LV. figs.
6, 15).

In the larger discohexasters, which are also rather divergent in form, the simple
cylindrical principals are somewhat longer, and divide into from three to six S-shaped thin
terminals, which are grouped together in a slender perianth-like bundle (PL LV. fig. 14).
Carter has already noted this peculiar form of rosette as characteristic of the species.
Of quite isolated and exceptional occurrence in the parenchyma is the small simple
discohexact, figured in PL LV. fig. 8. It may possibly not belong to this sponge, but
have originated as an intrusion from without.

The dermal skeleton includes, in the first place, medium-sized hypodermal
oxypentacts, in which the rays are for the most part smooth, and only roughened at the
conically pointed ends. The four cruciate tangential rays are rarely disposed exactly in
the dermal membrane, but are usually somewhat below it, or pushed out beyond,
though in both cases parallel. In some cases the four tangential rays do not intersect at
right angles, but are all pushed together to one side, as we shall afterwards have to note
in regard to the pleural prostalia. In much closer connection with the dermal membrane
are the small, finely spinose, autodermal pentacts. In these the four cylindrical tangential
rays, which are tolerably straight, and at most gently incurved, with rounded or somewhat
conically pointed ends, form a rectangular meshwork within the dermal membrane. The
proximal radial ray, which has approximately equal length and similar characters,
projects into the subdermal space (PL LV. figs. 2, 3). The undeveloped sixth distal
radial ray is almost always represented by a rounded tubercle or knob. In rare cases,
especially near the base, I found, between the pentacts, isolated cruciate tetracts of
similar form and equal size (PL LV. fig. 4).

The gastral skeleton, which lines the inner surface of the gastral cavity, essentially
resembles the dermal, Ijut dificrs in this, that the finely spinous gastralia, whose
tangential rays extend within the gastral membrane, are not pentacts, but weU-developed
hexacts (PL LV. fig. 5) in which the internal radial ray projects into the gastral ca^dty,
and gives the internal surface that peculiar velvety appearance which we noted above.

The radial tufts of pleural prostalia, which project from the papilke over the whole
external surface of the body, consist of four to eight spicules which project for 1 to 2 cm.

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

The strong, smooth, principal radial ray, which is always pointed towards the interior
of the sponge-body, is either simply pointed at the outer extremity, so that simple
oxydiact forms result, or gives origin to four exactly tangential rays (6 to 10 mm. in
length), which do not, however, intersect at right angles, but form with one another
three acute angles of 40' to 50°, so that the two outermost rays together form an angle
of 100° to 150°, rarely of 180° (PI. LV. figs. 9, 13). As from three to five of these
pen tacts are grouped together in a tuft, and so disposed that the points of the tangential
rays meet at an angle, a most beautiful veil is formed, which spreads over the sponge at
a distance of 5 to 12 mm. from the surface. Besides these peculiar pentacts, each tuft of
parietal prostalia includes several of the above-mentioned radial cliacts, which project
freely for 3 to 6 mm. beyond the surface of the pentact-veil. Near the oscular
margin the pentact prostalia disappear, and the oxydiacts become more prominent. The
latter are sometimes so numerous and closely grouped that they form a thick WTeath of
externally directed and pointed spicules (PI. LV. figs. 1, 7). The long, smooth, radial
ray of the pleural pentacts, is, as regards its surface, in marked contrast to the four
tangentials which spring from its outer extremity, and run out in a straight or slightly
curved course to a gradually narrowed point. As Carter has carefully described, the
whole surface of these tangentials is so thickly and uniformly beset with fine pointed
tubercles that it presents a thoroughly rough appearance. Between tliese small
tubercles, but at greater intervals, strongly developed spines project obliquely outwards
like the prickles of a rose. They occur with tolerable imiformity, but without
recognisable law, and gradually decrease in height towards the narrowed end (PI. LV.
figs. 9, 13).

The spicules of Rossella antarctica undergo peculiar modification at the base, where
the sponge is either attached directly to some solid body, or fixed by means of processes
to various smaller objects. The parenchyma contains, in the first place, very abundant
small discohexasters, like those represented in PI. LV. fig. 6, but with more numerous .
and shorter terminals rising from comparatively broad basal discs.

We have also to note that here all the longer spicules, and especially the long diacts,
exhibit terminally a club-shaped thickening with large tubercles, instead of the small
points or roughnesses which occur elsewhere. Finally, in the region where the sponge
comes into contact with the substratum, manifold modifications occur in the form of
outgrowths, fusions, and not unfrequently irregular reticulations, similar to those
represented in PI. LXIV. fig. 3, from Rhabdocalyptus mollis. Such modifications only
occur where some irritation is exerted on the sponge through contact with foreign
bodies.

Whether the spicule figured in PL LV. fig. 12, which 1 found isolated in the basal
portion of a Rossella antarctica (Station 320), really belongs to the sponge or is a foreign
intrusion, I cannot determine.

REPORT ON THE HEXACTINELLIDA.
2. Rossella velata, Wp'ille Thomson.

143

A beautiful species of Rossella was dredged by Dr. Gwyii Jeflreys in the Athuitic,
off the mouth of the Strait of Gibraltar, from a depth of 651 fothoms, and is figured

Fig. i.— Rossella velata, Wwille Tliomson.

and shortly described by Wyville Thomson.' Of this form the Challenger Expedition
procured no specimen, but I obtained for investigation a specimen dredged by Wyville

1 Depths of the Sea, pp. 418, 419.

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

Thomson on the " Porcupine " expedition, off the Strait of Gibraltar, and very
probably the same form as was figured in the Depths of the Sea, p. 419. To this
dried specimen the following description refers.

In contrast to the cylindrical or barrel-like shape of most of the specimens of
Rossella antarctica, Carter, the form of Rossella velata, Wyville Thomson, is
perfectly ovoid, 6 '5 cm. long, by 4 "5 broad. The superior pole exhibits a circular,
sharp-edged aperture, 1"6 cm. in diameter, — the opening of the equally broad, cylin-
drical, gastral cavity (4 cm. in depth) into which the efferent canals open. The
roundish apertures of the latter are especially wide in the blind basal portion of the
cavity, and become gradually narrower towards the upper end. The external surface
of the body is not uniformly smooth, nor beset merely with minute, sharply-defined
papilla?, as in Rossella antarctica, but is rather to be described as hillocky, with
numerous gently convex protuberances, from 5 to 6 mm. in breadth, and not sharply
defined from the reticulate, connected, intervening depressions. They are distributed
over the whole surface, with some regularity, at intervals of about 10 mm. between the
summits. The network of beams, which extends radially in relation to these projecting
eminences, is doubtless covered during life by the fine rectangular lattice-work of the
smaller dermalia. The whole external surface of the sponge is covered by the greatly-
developed system of most beautiful prostalia, which project radially, in small groups,
from the apices of the hillocky elevations. These tufts of pleuralia consist of
simple pointed diacts and pentacts, in which the four tangential rays arise at
right angles to the radial ray, at a distance of about 10 mm. from the surface of
the body. They intersect, not at acute angles, but in a perfect cruciform fashion;
they are also directed mutually at right angles. At the upper end of the sponge,
near the oscular margin, only the radially disposed, long oxydiacts persist, forming a
marginal fringe, which attains the conspicuous length of 3 to 4 cm. At the lower pole
of the sponge-body, on a flat surface measuring 6 to 7 square cm., thick tufts of
spicules arise from the apices of hillocky elevations. The tufts bear twenty or more
basalia, 5 to 8 cm. in leno-th, and taken tosrether form a loose basal tuft. On most of
these long basalia one can recognise, even with the naked eye, at the outer extremity,
a small four-rayed anchor.

The parenchyma contains, as in Rossella antarctica, medium-sized oxyhexacts and
oxydiacts, which are frequently roughened towards the pointed extremities. The
middle portions of the oxydiacts are either smooth, or provided with an annular swelling
or with four cruciate projections — traces of the undeveloped rays. Small oxyhexasters
with short principal rays, each bearing two long divergent terminals, are very
abundant. Between these there is a somewhat abundant, but locally variable occur-
rence of discohexasters similar to the above, but with toothed, somewhat incurved,
transverse discs at the ends of the terminal rays, and also of discohexasters in which the

REPORT ON THE HEXACTINELLIDA. 145

short, somewhat terminally cxpauded principal rays, bear several (three, four, or more)
long terminals with toothed transverse terminal discs. The rosettes with slender
groups of S-shaped terminals, which occur abundantly in Rossella antarctica (PI. LV.
fig. 14), are absent in this species, and their place is taken by small plumicomes, in
which the short principal rays bear broad, plano-convex, transverse discs, from the
convex external surface of which several whorls of fine S-shaped terminals arise in
perianth-like fashion, as in the plumicomes of Pohjlophus 'philippinensis (PI. LIV.
figs. 4, 6).

The dermal skeleton contains, besides simple medium-sized hypodermal oxypcntacts,
strong rough or toothed pentacts with rounded or truncated, straight, cylindrical rays,
such as occur in Rossella antarctica. In these forms there is a very constant occurrence
of a knob-like or hemispherical projection at the distal side of the point of intersection —
the evident rudiment of a sixth distal radial ray.

The gastral skeleton does not differ essentially from that of Rossella antarctica.

The long basalia, which are always smooth inside the body, but sometimes have
lateral barbs outside the body, bear on their extremities four markedly recurved, simply
pointed anchor-teeth, in which the central canal can be detected.

Genus 4. Acanthascus, n. gen. (Pis. LVL, LVII.).

Thick-walled goblet-shaped forms firmly attached at their base. The deep simple
gastral cavity opens superiorly in a round, smooth-margined osculum. From the external
surface of the body isolated, or more rarely grouped oxydiact pleuralia project radially.
The parenchyma contains oxyhexasters with short principal rays, and various disco-
hexasters. The dermalia are small roughened tetracts and pentacts.

1. Acanthascus grossxdaria, u. sp. (PI. LVL).

In the vicinity of Possession Island (Station 148, lat. 46° 47' S., long. 51° 37' E.),
from a depth of 210 fathoms, on a hard ground of gravel and shells, an egg-shaped
sponge was dredged, which in size and form resembled a large goosel^erry. The lower
pole had been broken ofi". At the superior pole there is a circular, smooth- margined
oscular aperture, leading into a saccular gastral cavity 2 '5 cm. in depth. Through the
smooth internal skin of the cavity the efi"erent canals can be seen, 1 to 2 mm. in width ;
the larger afferent canals were, in the same way, visible through the external skin. From
the outer surface, which is on the whole uniformly arched and smooth, numerous simple
smooth and pointed needles, irregularly distributed, project obliquely outwards and
upwards, and attain a length of 5 to 10 mm. Near the oscular opening these prostalia are

(ZOOL, CHALL. EXP. — PART LIII. 1886.) Ggg 19

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

especially long aud strougly developed, aud liave an approximately perpeudicular disposi-
tion (PI. LVI. figs. 1, 2).

The principal spicules of the parenchyma are large and medium-sized oxydiacts of
varying length, which are tubercled towards the pointed ends, and are in the middle
either smooth or provided with an annular swelling, or with four projecting bosses or
ray-rudiments. They occur either in isolated distribution or in strands which vary
in direction. They are, for the most part, parallel to the outer and inner bounding
surface, or disposed radially to the same. Some especially large and thick diacts lie
longitudinally near the gastral surface. I have hardly ever observed large hexacts,
though here and there isolated medium-sized regular oxyhexacts occur, with sUghtly
spinous rays. Small hexact forms also occur with smooth rays. On the other hand, all
through the parenchyma, in tolerably uniform though scattered distribution, oxyhexasters
occur with short principal rays, and with a varying number of long divergent smooth
terminals, two of which are usually borne at the end of each principal ray (PI. LVI.
fig. 7). Besides these, we have to note the less abundant occurrence of small disco-
hexasters with four or more long, tolerably straight terminals, each provided with a small
four-toothed, terminal, transverse disc (PI. LVI. fig. 9). Not unfrequently, also, some-
what larger discohexasters occur, in which each of the short simple principaLs bears from
three to six or more slightly S-shaped terminals. These are united in a usually somewhat
slender, perianth-like bundle, and are terminally roughened on their outer extremities,
which always bear a four-toothed disc (PL LVI. figs. 3, 4). In quite isolated distribution
small discohexasters occur, with moderatelj^ short principal rays terminally expanded
into a disc, which bears on its outer convex surface a large number of delicate diverging
terminals, forming a brush-like tuft, and bearing on their extremities minute four-toothed
discs (PL LVI. figs. 10, 11).

In the dermal skeleton, medium-sized, smooth hypodermal oxypentacts occur, in
which the proximal ray is radially disposed, while the four long tangentials, intersecting
at right angles, follow the superficial curvature of the sponge in being slightly curved
inwards.

In the large rectangular meshes of these pentact hypodermalia, numerous autodermal
pentacts and tetracts occur in the dermal membrane. They agree exactly with the small
rough dermalia in the genus RosseUa. It has to be noted, however, that the number of
tetracts exhibiting a simple right-angled intersection of rays (PI. LVI. fig. 5) is in
proportion to that of the otherwise very similar pentacts (PI. LVI. fig. 6) decidedly
greater in this Acanthascus than in RosseUa (PI. LVI. fig. 2). These rough dermal
pentacts when compared wdth those of the genus RosseUa exhibit a slight divergence
in this, that there is no boss- or knob-like rudiment of the sixth distal ray. The tetracts
are also without any such rudimentary ray (PI. LVI. figs. 5, 6).

On the inner side of the body-wall, the special gastral skeleton is represented by a

REPORT ON THE HEXACTINELLIDA. 147

sheath of finely spinous oxyhexacts (PL LVI. fig. 12), in wliich the radial ray, directed
towards the centre, projects into the lumen of the gastral space, while the opposite radial
ray intrudes into the subgastral space, and the tangential rays, intersecting at right
angles, lie in the gastral membrane (PL LVI. fig. 2).

The medium-sized oxydiacts, which project freely from 3 to 5 mm. outwards and
upwards beyond the external surface, are unfortunately for the most part broken. They
occur somewhat irregularly and are not ver}^ numerous, being probably to a large extent
lost. I have not discovered any pcntact prostalia similar to those which occurred in
Rossella, and must therefore conclude that they are absent.

2. Acanthascus duhius, u. sp. (PL LVIL figs. 8-13).

To the south of Puerta Bueno, in Patagonia (Station 310, lat. 51° 27' 30" S., long.
74° 3' W.), from a depth of 400 fathoms and a blue- mud ground, a Hexactinellid was
trawled, of which unfortunately only the basal part, which had grown on the ramifications
of a colony of Lopliolielia, was preserved, and that imperfectly. What remains is about
the size of a man's fist, and represents the basal portion of a cup-shaped sponge, which
in its entirety must have been about four times as long. The remnants of the external
surface still preserved seem to be approximately smooth, but it is impossible to decide
as to the presence or absence of elevations with freely projecting spicules. On the
interior gastral surface of the wall (2 to 3 cm. in thickness), there are large round
apertures, 3 to 8 mm. and more in diameter, — the openings of the efferent canals into the
gastral cavity. The latter must have been at least 5 cm. wide.

The tissues are a good deal macerated, so that the skeleton is somewhat loose, and would
have altogether fallen to pieces were it not that the larger spicules of the parenchyma
are thoroughly bound together by means of synapticula and siliceous cementing masses.
The somewhat thick parenchymal skeleton consists of numerous oxydiacts, either isolated
or in strands, and often bound together as above noted. They vary in length, but
attain no considerable thickness. Especially in the lower portions of the sponge which
had grown on the Lophohelia, the parallel, straight, rod-like spicules are very thoroughly
bound together by numerous transverse synapticula (PL LVIL fig. 12). The terminal
portions of these spicules are usually tuberculated or rough, and are either gradually
narrowed or conicaUy pointed, or else rounded ofi", and sometimes even somewhat clubbed.
The regular oxyhexacts which occm- here and there between the numerous long
diacts, are under medium size, either spinous or smooth, and in rare instances exhibit
thickened or tubercled central nodes (PL LVIL fig. 13). Here and there small smooth
oxyhexacts occur, with slender, straight rays, which do not exceed in size the numerous
rosettes about to be described. Among these there is a special abundance of oxyhexasters

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

witli very short principals, and comparatively strong and long divergent terminals, whicli
vary in number from two to four, though they are generally three (PI. LVII. fig. 8).
By the extreme shortening of the princijjals, these terminal rays are often drawn so near
to the centra] node, which is usually thickened in such cases, that the whole spicule
appears as a many-rayed star round the central node. Were there not countless
transitional forms of hexasters wdth six distinct principals, leading up to the many-rayed
stellate form, there might be real doubt as to the nature of the latter. As it is, however,
the stars can be without difficulty traced back to the triaxial type. I observed the
isolated occurrence of discohexasters with short principal rays, and with a varying
number of long terminals. As figured on PI. LVII. figs. 9, 10, 11, their resemblance to
similar structures in Acanthasciis grossularia is evident.

As to the dermal skeleton, I have observed some smooth hypodermal oxypentacts of
medium size, and numerous smaller dermal pentacts, with straight, rough rays which are
conically pointed at their extremities, or else rounded ofi" and even truncated. In the
latter form there is an almost constant occurrence of a small boss or elevation repre-
senting the undeveloped sixth distal ray. On the other hand, I have never observed any
cruciate tetracts. On the inner gastral surface numerous hexacts occur, with straight,
rough rays, similar to the above-described dermal pentacts.

The marked resemblance in form, structure, and skeleton between the basal portion of
this form, and the corresponding part of Acanthascus grossulaiia, seems to justify the
assumption that the upper portions must also have been like one another. We would
therefore expect the presence of oblique projecting pleural oxydiacts and a simple oscular

margin.

3. Acanthascus cactus, n. sp. (PI. LVII. figs. 1-7).

Among the dried specimens of Japanese sponges which Dr. Doderlein brought with
him from Enosima, there was a pear-shaped, somewhat laterally compressed, thick-
walled form, figured in PI. LVII. fig. 1. This Hexactinellid resembles a sack or beaker
in form, is 9 cm. long by 5 cm. broad, and 4 cm. thick. By its narrower end directly,
and also by means of several l)asal projections, 3 to 4 mm. in diameter, the sponge is
attached to a firm substratum. The superior aperture of the simple gastral cavity,
which is about 3 cm. in width, is surrounded by a somewhat sharp-edged smooth margin.
On the lower end also, between the basal attaching protrusions, there is an irregularly
contoured aperture about 8 mm. in diameter, but possibly the result of subsequent
damage.

The external surface of the body is beset with conical elevations varying in height
up to 8 mm. They are largest in the middle of the sponge, and decrease in height
towards the oscular margin. They occur at intervals of 1 to 2 cm., and bear on their

REPORT ON THE HEXACTINELLIDA. 149

apex a tuft of divergent pointed spicules, which project freely for 2 cm. or more. The
general form of the body, and the occurrence of these spherical elevations at approxi-
mately equal intervals on the outer surface, and especially the tufts of spicules pro-
jecting from their apices, produce a close resemblance between this sponge and
certain prickly cactuses. I have for this reason given this form the specific designation
cactus.

While the whole external surface of the body is covered with a hue meshed dermal
network, the simple, smooth internal surface is covered with a firm gastral mem-
brane, which is abundantly perforated by numerous round apertures, about 1 mm. in
breadth.

The parenchymal skeleton exhibits numerous thickly distributed oxydiacts in
variable disposition. These are, for the most part, surrounded by oxydiact comitalia
of but slight development. Near the conically pointed ends these diacts are usually
somewhat rough, but are otherwise smooth. In the middle they are usually quite
smooth, though less frequently exhibiting a median thickening, or else four cruciate
or two opposite elevations. Besides the slender comital oxydiacts, numerous isolated
spicules of the same sort occur.

I have not observed in the parenchyma any simple hexacts of medium size, but
only small, slender, isolated oxyhcxacts, resembling in size the rosettes about to be
described.

There is an abundant occurrence of the familiar oxyhexaster forms with short principal
rays, and long, smooth, diverging terminals, which are here somewhat strongly developed
(PI. LVII. fig. 3). Between these, spicules occur of similar size and structure with
six undivided rays, each of which exhibits, at the same distance from the common
node of intersection as the point where the principals divide in the oxyhexasters, a
sharp flexure, succeeded by a small curve, and finally by a straight portion continued
on to the point. Spicules thus modified may be regarded as reduced derivatives of the
above described oxyhexasters.

There is besides a not unfrequent occurrence of irregular rosettes, which seem at first
sight to be quite distinct from the Hexactinellid type of spicule, since they almost
always exhibit eight principal rays. These somewhat thick, but not exactly cylindrical
rays, which spring from a thickened central node, arc often beset with knot-like protrusions.
Each bears a bundle of four to eight thin straight terminals of equal length, which
diverge slightly from one another, without being sharply marked ofi' from the principals.
Each terminal ray bears on its extremity a minute disc (PI. LYII. figs. 4, 5 ; c/l
PI. LXV. fig. 3). I am of opinion that some of these apparent principals have arisen
by the splitting of real or primary principal rays. This is suggested, for instance, by
the fact that at their base they are not cylindrical, but somewhat convergent and
flattened. From the base of one ray a smooth link may be seen passing to the base of

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

an adjacent ray. We have, besides, to note the (sometimes very regular) occurrence of
a protuljerance between the bases of the rays, which is occasionally drawn out into a
long radially projecting spine. Or it may be that besides the six principals divided
into terminals other independent rays are present, running out into simple points
(PI. LVII. fig. 4).

Finally, the parenchyTna includes a tolerable abundance of very minute disco-
hexasters, which are in some regions disposed with especial thickness below the external
and under the gastral skin. In these small discohexasters the short simple principals
bear broad, plano-convex, transverse terminal discs, from the convex external surface of
which numerous delicate divergent terminals arise, which are of equal length, and bear
minute transverse discs like those in the rosettes of other sponges as figured in PI. LVII.
fig. 11, PI. LXI. fig. 7, PI. LXV. fig. 4.

The greater part of the dermal skeleton consists of autodermal tetracts, in which the
rays, intersecting at right angles, and inclined slightly inwards, are cylimlrically
thickened, or even somewhat club-shaped at their rounded extremities, and are beset
throughout their entire length with fine spines (PI. LVII. fig. 6). In many places, and
especially in the hillocky elevations, from which the radial spines project, besides these
autodermal tetracts, pentacts also occur in which the four tangential rays are altogether
similar to those of the tetracts, while a fifth proximal ray, springing from the node of
intersection, projects in a radial direction inwards (PI. LVII. fig. 7). Between these
tetracts and pentacts isolated smooth tetracts occur, double the size of the others, and
with their rays intersecting at right angles.

The gastral skeleton consists of small gastral pentacts, which correspond exactly to
those of the dermal skeleton. Between these there is a sparse occurrence of tetracts also
like the dermal forms (PL LVII. fig. 2).

Genus 5. Batlujdorus, n. gen. (Pis. LVIIL, LIX.).

Saccular or bladder-shaped forms, with thin loose walls, smooth or spinous external
surface, and thin round oscular margin, which is (always Vj provided with a cufi-like
fringe of projecting spicules. The parenchyma contains, besides large diacts and hexacts
of various kinds, oxyhexasters and in some species discohexasters. The dermal skeleton
includes, besides the familiar smooth hypodermal oxypentacts, autodermal oxytetracts,
and in some species similar diacts or even monaets. In the gastral skeleton, on the other
hand, only roughened oxyhexacts occur.

KEPORT ON THE HEXACTINELLIDA. 151

1. Bathydorus fimbriatus, n. sp. (PL LVIIL).

In the North Pacific, at Stations 241 and 248, and from considerable depths, the
trawl brought up fragments of a somewhat large, bladder-shaped sponge, which was
found to be a Hexactinellid of the very simplest structure.

At Station 248, lat. 37° 41' N., long. 177° 4' W., from a depth of 2900 fathoms
and a red clay ground, the somewhat injured, though coherent upper portion of a bladder-
shaped sponge was trawled. The loose smooth wall is only 1 to 1*5 mm. in thickness,
and becomes gradually thinner towards the sharp oscular margin, which is surrounded
by a fringe of long diact marginalia. The diameter of the bag, which is almost cylin-
drical in its lower region, measured from 4 '5 to 5 cm. The lumen is gradually narrowed
towards the upper end, measuring only 2'5 cm. in diameter at the superior terminal
opening (PI. LVIIL fig. l). The almost uniformly smooth external surface exhibits
minute, regularly distributed, round holes, of about \ mm. in diameter. Somewhat
larger round openings may be detected on the internal surface of the gastral membrane,
which is also almost uniformly smooth.

An inspection of flattened portions of the wall, or better still, an examination of cross
sections at right angles through the wall, reveals distinctly the thimble-like form
of the ciliated chambers, which are disposed in a very simple and regularly folded layer
between the two reticulate limiting membranes, and supported by the external subdermal
and internal subgastral trabecular framework. From the larger, lacunar, subdermal
spaces, duct-like diverticula pass between the externally arched folds of the chamber
layer. The lumen of each of these cupola-like folds is free from the trabecular scaffold-
ing, and opens into the lacunar subgastral spaces; the latter do not, however, open
directly into the general gastral cavity, but only into the loose meshwork of the gastral
membrane which stretches smoothly over them (PI. LVIIL fig. 2).

The larger parenchymalia consist of slender diacts which vary in length, and do not
exceed 0'08 mm. in thickness. They are rough and rounded off at their ends, and are
frequently somewhat club-shaped in consequence of thickening. A central swelling or forma-
tion of tubercles, which take the form of four cruciate, or two opposite elevations, is present
in many cases and absent in others. Between these long supporting spicules, which are
almost always disposed quite, or approximately parallel to the surface, numerous oxy-
hexasters occur, _,with a variable number of long terminal rays, which are either straight
or gently curved (PL LVIIL figs. 4, 7). This curvature of the long terminal rays
frequently assumes an S-shaped form, and results in the formation of minute, delicate,
three or four-rayed perianth-like whorls at the end of the principal rays (PL LVIIL fig. 7).

The dermal skeleton is supported by medium-sized, smooth, hypodermal ox}q)entacts,
in which the four tangential rays, intersecting at right angles, lie close below the dermal
membrane, and are opposed to the corresponding tangentials of adjacent pentacts to

152 THE VOYAGE OF H.M.S. CHALLENGEE.

form a quadrate lattice-work. They are usually curved slightly inwards. Opposite the
straight proximal radial ray which penetrates almost through the thickness of the body-
wall, a small distal protuberance is occasionally developed, though usually absent. These
pentacts, especially in the neighbourhood of the oscular margin, sometimes protrude
beyond the surface of the sponge, and form with their tangential rays, as in RosseUa, a
sort of veil over the slcin ; but this diflferentiation occurs only locally and exceptionally.

The dermal membrane itself includes a large number of small, simple, roughened
oxytetracts, in which the arms, intersecting at right angles, seem to be usually some-
what bent inwards (PI. LVIII. figs. 2, 5).

The gastral skeleton seems wholly destitute of hypogastralia, and consists merely of
tbe very numerous gastral oxyhexacts, in which pointed rays are thickly beset Avith fine
spines. The ray which projects radially inwards into the gastral cavity is usually cj^uite
as long as the five others (PI. LVIII. figs. 2, 6). The marginalia, which project in cufi"-
like form round the superior oscular margin, are smooth oxydiacts, 8 to 10 mm. in
length, without definite central swelling or formation of knots, and with simple pointed
ends.

At Station 241 (lat. 35° 41' N., long. 157° 42' E.), from a depth of 2300 fathoms and
a red clay ground, the trawl brought up some flat fragments, 1 mm. in thickness, which
corresponded in their whole appearance to the above described Bathydorus. They
exhibited the same structure as the above, with the single exception of the parenchymal
oxyhexasters, in which the long terminal rays were never so thickly disposed in perianth
fashion, as was frequently the case in the above-described form. Here they all diverge
more or less, as was indeed occasionally the case above. The spinous gastral hexacts
exhibit somewhat longer radial rays than do those in the form already described. I do
not think, however, that these slight difi"erences justify the erection of a distinct species,
but believe rather that we have here to deal merely with a local variety, and therefore
refer these fragments without hesitation to Bathydorus Jimhriatus.

2. Bathydorus stellatus, n. sp. (PI. LIX. figs. 1-5).

In Messier Channel, ofi" Patagonia, right in front of Port Grappler (Station 307,
lat. 49° 24' 30" S., long. 74° 23' 30" W.), from a depth of 140 fathoms and a blue mud
ground, the trawl brought up a small inconspicuous sponge form, about 15 mm. in length.
The thin- walled saccular body was divided superiorly into two terminal tubes, 6 mm.
in breadth, with terminal apertures. On the otherwise smooth external surface, some
thin pointed spicules project obliquely outwards and upwards for 5 to 6 mm. At
the terminal oscular aperture no distinct crown of marginalia is to be seen, but this is
perhaps the result of slight injury.

The parenchjTual spicules are long, narrow, straight or gently curved diacts, with or

REPORT ON THE HEXACTINELLIDA. 153

without central nodes. The somewhat roughened or tul^ercled terminal portions are
rounded ott', or not unfrequently slightly club-shaped at their extremities. Between
these a large number of oxyhexasters occur with long, slim, but roughened terminal rays,
which vary considerably in number. The principal rays are much sliortened, and this
goes so far in some hexasters, that the spicule has quite the appearance of a simple, many-
rayed star (PI. LIX. fig. 2). A slight irregularity frequently occurs at the base of these
rays, in the occurrence of a slight union or something of that nature, which also points
to the reduction above referred to. The presence of these star-shaped pareuchjTnal
spicules led me to give this form of Bathydorus the specific title stellatus. Finally, in
the parenchyma, I have observed the isolated occurrence of minute delicate discohexasters,
in wdiich each of the six short principal rays bears, on a terminal expansion, three to six
long diverging terminal rays with a toothed terminal portion (PI. LIX. fig. 4).

The dermal skeleton contains medium-sized, smooth, hypodcrmal oxypentacts, with
rough conically pointed ends, and more or less roughened autodermal tetracts, with
rounded extremities, disposed flatly in the dermal membrane (PL LIX. fig. 3). Between
these, pentact forms occur with uniform proximal radial rays, and sometimes diacts also
w^th central nodes (PL LIX. fig. 5).

The gastral skeleton, which is not well preserved, consists of rough or spinous oxy-
hexacts. The pleural prostalia are simple or strongly developed oxydiacts, about 1 cm.
in length.

3. Bathydorus spinosus, n. sp. (PL LIX. figs. 6-9).

Near Penguin Island (Station 147, lat. 46° 16' S., long. 48° 27' E.), from a depth of
1600 fathoms and a Diatom ooze ground, the trawl brought up a thin-walled sack, 6 cm.
in length and 3 cm. in greatest diameter, but reduced to 2 cm. in width towards the
superior, irregularly contoured terminal aperture. The surface of the sack exhibited
irregularly scattered pointed spicules, projecting for about 6 mm., while the oscular
margin bears a corona of marginalia. It is possible that this sponge form does not
represent a distinct species, but belongs to the above described Bathydorus stellatus,
which it resembles in general structure, and especially in the nature of the siliceous
spicules. There is, however, a certain difference about to be noted, which seems to
warrant at least a provisional separation of the two forms. I have, therefore, regarded
the distinction as really specific.

While the long diacts of the parenchymal skeleton are not in any way essentially
diff"erent from the long diacts of Bathydorus stellatus, I find that the abundant
oxyhexasters differ from those of the above species in this, that the principal rays are not
usually so markedly shortened, even to disappearance, as they are in the above ; the

(ZOOL. CHALL. EXP. — PART LHI. — 1886.) Ggg 20

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

terminal rays are, fiirtlier, for the most part somewhat wavy in their curvature (Ph LIX.
figs. 8, 9). I did not discover any small discohexasters.

The dermal and gastral skeletal elements correspond almost perfectly to those of
Bathydorus stellatus, though it may be noted that the gastral oxyhexacts are here very
varied, sometimes rough, sometimes spinous, and rarely quite smooth, as represented in
PI. LIX. fio-. 7. In general, in spite of the larger size of the specimen, the rays of its
hexact gastralia are less strongly developed than in the much smaller specimens of
Bathydorus stellatua.

The pleural and marginal prostalia are, like those of Bathydorus stellatus, simple
smooth oxydiacts, 8 to 10 mm. iu length.

4. Bathydoms haculifer, n. sp. (PI. LIX. figs. 10-18).

In the middle of the South Pacific (Station 286, lat. 33° 29' S., long. 133° 22' W.), from
a depth of 2335 fathoms and a red clay ground, the trawl brought up a small fragment
of a sponge which apparently belongs to the genus Bathydorus. The specimen, as
figured in PI. LIX. fig. 10, is a tolerably smooth, approximately semicircular plate,
2 mm. in thickness. From the smooth surface some isolated simple oxydiacts project
obliquely for 10 to 20 mm., while the other surface appears uniformly rough.

The spicules of the parenchyma are long, narrow, smooth diacts, with rough, conically
pointed, rounded, or more rarely slightly club-shaped ends, with or without central nodes,
exactly as in the other species of Bathydorus. Medium-sized simple hexacts occur here
and there. Between these there is a very abundant occurrence of oxyhexasters with
short principal rays and long straight terminals (PI. LIX. figs. 12, 13). A few disco-
hexasters also occur with short principal rays, each of which bears usually three diverging
terminals, bent convexly outwards at their base, and carrying on their extremities
minute, transverse, four-toothed, somewhat convex discs (PI. LIX. fig. 18).

The dermal skeleton contains medium-sized hypodermal pentacts, with rough,
rounded, or somewhat conical extremities, and small autodermal diacts or monacts,
which are rough all over, are rounded ofi" at their ends, and exhibit at the central point
a definite swelling of the axial canal, or the trace of undeveloped rays, usually in the
form of two opposite, or more rarely four cruciate tubercles (PI. LIX. figs. 14, 15, 16, 17).
The hypodermal pentacts form, by the apposition of their opposite tangentials, a quadrate
lattice-work. The autodermal diacts and monacts occur in irregular disposition in the
dermal membrane (PI. LIX. fig. 11). In the diacts the two rays belong as a rule to the
same axis, but forms occasionally occur, as figured in PL LIX. fig. 16, where the two
rays form a right angle. In the monacts, which occur in tolerable abundance, the end
which represents the centre of the original six-rayed form, and which therefore contains
the node of intersection of the axial canals, exhibits a club-shaped thickening, and it may

REPORT ON THE HEXACTINELLIDA. 155

be that three, or less frequently five, boss-like projections remain, which correspond in
their position to the undeveloped rays (PI. LIX. fig. 14).

The gastral skeleton is unfortunately not preserved, and this fact may make it
somewhat doubtful whether I am correct in regarding this form as a thin-walled
Bathydoriis, and not as a representative of a genus Rhahdocalyptus, now to be
described.

Genus 6. RJiahdocalyjJtus, n. gen. (Pis. LXIV., LXV.).

The body has the form of a moderately thick-walled cup or sack, with smooth
external surface. It is attached by a narrowed base or short stalk to a solid body.
The wide round oscular opening has a gradually sharpened smooth margin. The
interior surface exhibits, between the close round excurrent apertures of the numerous
efiierent canal or lacunar passages, an irregular ridged network.

The principal pareuchj'mal spicules are more or less long diacts. Besides disco- and
oxyhexasters of various sorts, eight-rayed rosettes occur, with several disc-bearing
terminal rays borne on the end of each medium-sized principal.

The dermal membrane contains either only rough diacts, or also rough pentacts,
tetracts, and monacts. The gastralia are rough oxyhexacts.

1. Rhahdocalyptus mollis, n. sp. (PI. LXIV.).

Among the dried Japanese Hexactinellids which Dr. Doderlein collected near Enoshima,
the form figured in PL LXIV. fig. 1 (one-third natural size) is conspicuous because of
its large dimensions. It is a moderately thin- walled, laterally compressed cup, 35 cm.
in height and 20 cm. in breadth at the superior aperture. The diameter of the round
compact stalk is about 6 cm. The large cup bears on its side a smaller, more elongated
form, 15 cm. in length, and 6 cm. in width at its orifice; while beneath the latter there
is a csecal protrusion or boss.

The wall, which measures in its lower portion 4 to G mm. in thickness, becomes
gradually thinner towards the upper end, and terminates in an undulating or crisped,
smooth, slender, oscular margin, without a fringe of spicules. At the lower portion of the
cup, just above the solid, somewhat tuberculated base, there is a round hole, 5 mm. in
diameter, which establishes a communication between the gastral cavity and the water
outside. The stalk is attached to the firm substratum by a slight basal expansion,
and includes here and there some intruded material, especially soil debris. The lower
surface of the stalk, where it is fixed to the substratum, exhibits the familiar thin but
firm reticulated plate, which is developed in all Hexactinellids at their point of attachment
to foreign bodies. The outer surface of the body is not quite uniformly curved, but exhibits

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

gentle elevations and depressions, which may be, however, the results of desiccation.
The inner surface of the wall of the cup exhibits here and tliere inconspicuous, ridge-like
elevations with a thin edge.

The extremely delicate dermal membrane, which forms a fine lattice-work is pre-
served onl}- in a few sheltered portions. The parenchymal skeleton is represented
by a feltwork of spicules, which are disposed in strands 2 to 4 mm. broad, and
enclosing roundish spaces of various sizes. The gastral skin, which also forms an extremely
delicate and fine network, exhibits a feltwork of spiculav sti'ands with meshes somewhat
narrower than those in the dermal skeleton.

The subsequent influence of damp has caused this specimen to fall in, so tliat the two
halves of the wall of the cup have been united. I was therefore unable to discover any-
thing definite as to the nature of the inner surface or of the oscular margin. Dr. Doderlein
told me, however, that this specimen, even when still well preserved, " exhibited a marked
lateral compression. The inner wall or cup exhibited ridged elevations. The free upper
margin had no marked plaiting, and ended in a slender smooth edge without a fringe of
spicules. At the base of the cup the larger apertures of the efferent canals were apparent."

A second smaller sj)ecimen of this species, measuring 13 cm. in height and 8 in
breadth, agrees closely with the above. It also exhibits a laterally compressed cup-like
or saccular shape, with a thin oscular margin, but is attached to the firm substratum at
several places, over a comparatively broad expansion. Two roundish apertures, 4 mm.
in diameter, occur near the lower end.

The principal portion of the parenchymal skeleton consists of long slender diacts,
occurring either isolated or disposed in strands. They exhibit a central nodal thickening,
and the rough ends are sometimes conically pointed, sometimes simply rounded, and
occasionally thickened in a club-shaped fashion. Between these there is a scattered
occurrence of the familiar oxyhexasters, in which the principal rays are very short, and
frequently almost aborted, while long terminal rays, present in variable number, are
somewhat curved at their base, but otherwise quite straight on to the pointed outer end.
Oxyhexacts occasionally occur in which the rays are twisted at a point corresponding to
that at which the terminal rays arise from the principals in the oxyhexasters. They
are doubtless degenerate oxyhexasters, which are again almost reduced to simple oxy-
hexacts, retaining only a trace of their metamorphosis in the twisting at the base.

In certain regions, especially in the subdermal trabecular spaces, there is an abundant
occurrence of large rosettes with terminal rays bearing terminal discs. These spicules
but rarely exhibit the ty^^ical number of principal rays, but as a rule eight are present.
When only six principals are present, as in the form represented in PL LXIV. fig. 7,
they intersect as usual at right angles in a somewhat thickened node, and are rather
thick and cylindrical. At the slightly expanded outer end they divide into three to six
straight or somewhat S -shaped terminals, which diverge slightly in a tuft, and attain a

REPOET ON THE HEXACTINELLIDA. 157

leugth double that of the principals. Each terminal l)ears at its extremity a small,
convex, transverse disc, with recurved marginal teeth (PI. LXIV. fig. 7). In these
discohexasters the central node is occasionally much thickened, and provided in the angle
between each two principal rays with radial, tubercle-like rounded processes, which may
also be drawn out into simple spines. In other cases, one or two of the principals are
especially thick, and split up externally into several terminals ; and this modification
appears to me to indicate the way in which the numerous rosettes have arisen, which
bear eight principal rays arising at approximately equal angles from the central node.
The variations in the rays of these rosettes are so numerous that it is impossilile to
attempt to describe all the modifications. I shall only note that not unfrequently the
central node becomes swollen into a conspicuous sphere, from the surface of which, besides
several broad principals, numerous terminals also arise, evidently by the basal splitting
of the principals. The splitting of a particular portion may thus increase till the whole
principal is divided.

Under the skin these peculiar discohexasters occur in abundance, but in other regions,
and especially in the subgastral trabecular space, peculiarly modified oxyhexasters,
oxyhexacts, and remarkable diacts, derived by reduction from the latter, occur abundantly.
In numerous oxyhexasters and oxyhexacts, the curved basal portion of the otherwise
quite straight, gradually pointed rays, bears a coating of fine spines or barbs, directed
obliquely inwards. These barbs are very numerous, and sometimes so long that those
of adjacent rays almost unite (PI. LXIV. figs. 8, 9). On these spinous oxyhexacts the
rays are sometimes curved, and this not unfrequently takes the form of a spiral twnsting
of the two rays on the same axis (PI. LXIV. fig. 10). If it happen, as is by no means
unfrequent in these spinous spicules, that the number of rays is reduced, forms result
like those represented in PL LXIV. fig. 11, in which a spherical, spinous, central body
bears at its two opposite poles two spinous rays, which are twdsted half round in a
spiral and then continued in a straight course to end in a simple point, or to be divided
into several pointed terminals. But the multiplicity of structure in these apparently
reduced forms is so extremely great, that I will not begin to give a detailed account of
the multitudinous modifications.

The dermal skeleton is supported by medium-sized hypodermal oxypentacts, in which
the long smooth rays sometimes exhibit a simple curvature, but are, as a rule, quite
straioht. In the dermal membrane itself, numerous autodermal diacts occur, in which
the rough rays, lying in one axis, end conically or are somewhat rounded oft'. The centre
usually exhibits two or four projecting tubercles (PL LXIV. fig. 5), but these are in
other cases entirely absent (PL LXIV. fig. 6). There is an isolated occurrence of well-
developed tetracts with rays crossed at right angles, and even pentacts with a ray
penetrating the parenchyma and resembling that of the diacts. Monacts occur less
frequently than in Bathijdorus hacuUfer, but still in tolerable abundance. They may

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

be with, or without lateral tubercles at the expanded end. Finally, I should mention
that here and there small discohexasters occur with numerous terminal rays on the
expanded ends of the principal rays, like the similar structures in RhabdocalyjJtus
roeperi, represented in PI. LXV. fig. 4.

The gastral skeleton consists of strong, rough oxyhexacts, similar to those in the
genus BatJiijdorus.

2. Rliabdocalyptxis rceperi, n. sp. (PI. LXV.).

To the south of Puerta Buono, in Patagonia (Station 310, lat. 51'' 27' 30" S., long.
74° 3' W.), from a depth of 400 fathoms and a blue mud ground, the trawl brought up a
cup-shaped Hexactinellid, 16 cm. in length and 9'5 cm. in maximum breadth. The solid
inferior portion of the sponge is contracted into a cylindrical stalk, 2 '5 cm. in thickness.
This stalk is bent to the side, as if to be attached to some solid body, but the rest has
been unfortunately torn away. The somewhat bulging wall of the cup has a thickness
of 8 mm. at the lower part, but decreases gradually towards the upper margin, and ends
in a simple smooth margin, bent slightly outwards. The sharp, uniform, circular edge
exhibits no projecting marginalia. Through the thin even lattice- work of the dermal
membrane are seen the irregularly scattered, elongated, angular or spindle-shaped pits,
from which roundish canals extend towards the centre. On the inner surface round
sharply contoured depressions of various sizes occur, into the bottom of which the more
or less wide excurrent canals of the efferent system open (PL LXV. fig. 1).

Besides this large specimen a small, fiat, wide cup- or basin-like form was obtained at
the same locality. It measured 2 cm. in maximum breadth and 1 cm in height, but was
also torn away from its connection, and exhibited only a torn base, 1 "5 cm. in breadth.
From this the side wall of the body, measuring about 2 mm. in thickness, arises, pro-
jecting obliquely upwards and outwards, and continued on to a smooth, circular, sharp-
edged mai'gin, measuring 2 cm. in diameter (PL LXV. fig. 2). In perpendicular cross
sections through the wall of the cup one can readily detect the alternating afferent and
efferent canals, and the folds of the chamber layer which extends between them.

The main portion of the parenchjrmal skeleton consists of long slender diacts with
rough pointed or slightly club-shaped ends, with or without central nodes, like those
which have been ah-eady so often described in this family (PI. LXV. fig. 11). There is
an isolated occurrence of small, weakly-developed oxyhexacts ^dth fine straight rays
(PL LXV. fig. 6), and a greater abundance of oxyhexasters wth short, sometimes very
short, principal rays, bearing long, diverging, straight or wavy, fine terminals (PL LXV.
fig. 5). The number of terminals on each principal ray varies from two to three, while
less frequently only one is present.

Not unfrequently the principal rays are so short that they can hardly be detected.

REPORT ON THE HEXACTINELLIDA. 159

being hidden by a spherical central thickening, from which the long terminals directly
arise, resulting in a many-rayed form.

In certain regions, and especially below the external skin, in the subderinal space,
there is a tolerably abundant occurrence of those eight-rayed discohexasters, which were
already described in Rhahdocaly2)tus mollis. Here also there are eight medium-sized
principals not exactl}^ round, but somewhat angular. These form approximately equal
angles with the central node, which is beset with roundish tubercles. Terminally the
principals tlivide into three to four fine, long, diverging terminals, which bear on their
extremities small, toothed, transverse discs.

We have also to note the quite isolated occurrence of very small delicate disco-
hexasters (PL LXV. fig. 4), in which the expanded ends of the principal rays bear
numerous fine terminals, with transverse discs on their extremities.

The dermal skeleton includes medium-sized hypodermal oxypentacts, in which the
long, smooth, occasionally somewhat curved rays, are roughened towards the pointed
extremities. The dermal membrane itself contains a large number of pentacts, tetracts,
diacts, and occasionally monacts, in which the rough straight rays, beset with small
pointed spines, are rounded off at their outer ends, or provided with blunt points
(PL LXV. figs. 7, 8, 9). At the middle of the diacts and at one end of the monacts,
four or five ci'uciately disposed hemispherical tubercles (or sometimes two in one axis)
project laterally as rudiments of the undeveloped rays (PL LXV. figs. 7, 8).

The gastral skeleton, and that of the skin which lines the wide efi"erent canals,
consists exclusively of comparatively large oxyhexasters in the usual disposition. The
weakly developed straight rays, which are of approximately equal length, have an
irregularly tubercled or roughened surface (PL LXV. fig. 10).

There are no marginalia in this species.

Genus 7. C rater omorpha. Gray (Pis. LXL-LXIII., LXVIL, LXVIIL fig. 2).

1872. Carter, Ann. and Mag. Nat. Hist., vol. x. p. 110 {Crateromorpha mei/eri).

1872. Gray, Ann. and Mag. Nat. Hist., vol. x. p. 13G.

1873. Carter, Ann. and Mag. Nat. Hist., vol. xi. p. 275.

1873. Carter, Ann. and Mag. Nat. Hist., vol. xi. p. 349.

1874. Graj', Ann. and Mag. Nat. Hist., vol. xiii. p. 284.

1875. Carter, Ann. and Mag. Nat. Hist., vol. xvi. p. 1.

1875. Marshall, Zeitscbr. f. wiss. ZooL, Bd. xxv., Suppl., p. 112.

1876. Marshall, Zeitschr. f. wiss. ZooL, Bd. xxvii. p. 113.

1877. Zittel, Mitth. d. Baler. Akad., xiii. p. 1.

History. — Among the sponges which Dr. A. B. Meyer sent home from Zebu (one of the
Philippine Islands), to the British Museum, there is an oval cup- or bell-like form with
a round but comparatively slender stalk. This form was distinguished by Gray as

I GO THE VOYAGE OF H.M.S. CHALLENGER.

Crateromorpha nieyeri, and was first described by Carter.' The approximately round
stalk, which was as long and as thick as one's finger, and connected by a trumpet-like
expansion with the " cup- or goblet-shaped body," was found by Carter to be perforated
1iy a lai-go number (twelve) of longitudinal canals " which open into the vents of the
bottom of the cu}) wliere the stem joins the latter." At the lower end the stalk was
expanded and adapted for fixing the sponge to the sandy bottom of the sea. The
margin at the brim of the cup was extremely thin, but the wall became thicker towards
the base. The principal forms of spicules which Carter found in the goblet-shaped
body were " (1) straight, fusiform spicules terminally swollen and spinose, also more or
less swollen and smooth in the centre, where the central canal has a hexradiate cross,
opposite to the ends of which two or four tubercles may occur. These spicules form
bundles crossing one another at right angles, or approximately so, and thus support the
dermal network. (2) Nail-like or cruciate sj^icules, in which the arms are smooth, straight
and pointed, the shaft a little longer than the rest, also occur somewhat sparsely. (3)
Minute smooth hexradiate spicules, in which each arm of the cross, just after leaving the
centre, divides into two long divergent spines, are present in myriads. (4) In the pore-
bearing area cruciate spicules occur, in which the arms, arising at right angles from the
centre, are more or less expanded at the ends and spinose throughout. (5) The stem
contains, in addition to the rod-like spicules forming a woolly mass, large, smooth, thick
spicules, four-twelfths of an inch in length, acerate and swollen at both ends. These
occur distributed longitudinally over the surface of the stem."

Gray"^ repeats Carter's description in somewhat different language, and establishes
on the strength of this specimen the family Crateromorphidse with the following
characteristic : — " Sponge cup-shaped, attached by an elongated pedicel formed of
numerous short spicules. Body of sponge covered externally with hexradiate spicules,
the outer ray of which is aborted, placed in longitudinal and transverse lines, making a
square mesh ; hollow with large oscules which diminish in size as they reach the margin
of the cup. Stem formed of numerous cylindrical tubes situated in a spiculous felt,
ending in a Inuich of filaments sunk in the mud."

In 1876, Marshall" gave the following diagnosis of the genus: — " Polyzoic, with
pseudogaster (?) of beaker form, outer side of the beaker-like cup perforated by cylindrical
.spaces (the stomachic cavities 1), others of the same kind running longitudinally to spaces
in the base of the cup. In the internal skeleton the hexradiate forms of spicules dis-
appear, and uniaxial forms occur. Body cavities covered by peculiar small hexradiate
spicules with bifurcated rays (?). Dermal skeleton of five-rayed spicules. No proper
root-tuft. Kosettes with numerous secondary knobbed rays."

Crateromorpha meyeri is regarded by him as a pollakid HexactinelHd which has

' Ann. and Mag. Nat. Hist., ser. 4, vol. x. p. 112. - Ann. and Mag. Nat. Hist., ser. 4, vol. x. pp. 1.36-137.

' Zeitschr. f. wiss. Zool., Bd. xxvii. p. 126.

REPORT ON THE HEXACTINELLIDA. 101

degenerated in several points, as may be seen in tlie preponderance of uniaxial spicules,
and also in the absence of a special root-tuft.

Character of the Genus. — The somewhat thick-walled cup- or bowl-shaped body is
firmly attached to the substratum by means of a more or less long and compact cylindrical
stalk. The sharpened circular oscular margin usually bears a membranous, perpendicular,
annular fringe, though occasionally naked. The external surface appears smooth and
sometimes meshed. On the internal surface the numerous efferent canals open, through
roundish apertures of variable size, directly into the large gastral cavity. The parenchyma
includes oxyhexasters and discohexasters of variable form and size. The dermal
membrane contains small rough tetracts and pentacts, and sometimes also diacts, which
are blunted terminally or thickened in club-shaped fashion. The gastral membrane
contains roughened pentacts.

1. Crateromorpha meyeri, Gray (PI. LXL).

Near Zebu, Philippine Islands {Station 209), from a depth of 95 fathoms and a
blue mud ground, several well-preserved specimens of Crateromorpha meyeri were
obtained.

The sponges are 10 to 12 cm. in height, and of a beautiful wine-glass-like form. An
irregular expansion (basal plate) attaches the sponge to the firm substratum, and the
cylindrical stalk, about 5 cm. in length and about as thick as one's little finger, is con-
tinued through a trumpet-like enlargement into a widely open bulging cup, about 5 cm.
in breadth. The wall of this cup is inferiorly about 2 cm. in thickness, but becomes
narrower upwards and finally ends in a delicate, thin, ti\insparent, straight or slightly
outwardly curved lamellar fringe, with a sharp edge, which measures 4 mm. in height,
and about 0"5 mm. in thickness. The external dermal surface of the whole sponge is
smooth, but through the fine rectangular lattice-work of the dermal membrane the
irregular roundish cavities of the afferent canals may be seen (PI. LXL fig. 1). On the
much firmer internal surface there are numerous roundish openings of varying width.
These belong to the eS"erent canals. The smaller apertures in the neighbourhood of the
lattice-work are still covered by the gastral membrane, but the larger open freely.
Inferiorly the openings of the efferent system of canals become larger, so that finally,
at the foot of the gastral cavity, only more or less narrow septa remain between them.
(PI. LXL fig. 2). The larger septa extend radially from the centre towards the lateral
wall, and produce a radially symmetrical division into four, which recalls a similar con-
dition in species of Hyalonema. In longitudinal sections the relation of the afferent
and efferent canals is very distinctly seen, and between them the much folded layer
of ciliated chambers (PI. LXL figs. 2, 3).

In PI. LXL fig. 1 an attempt is made to reproduce the peculiar grey-yellow colour,

(ZOOL. CHALL, EXP. — PART LHI. — 188G.) Ggg 21

1G2 THE VOYAGE OF H.M.S. CHALLENGER.

exliibited by these forms of Crateromorpha, and by the great majority of the spirit-pre-
served Hexactinellida here described. It is noteworthy that the marginal fringe which
surrounds the superior circular opening (about 40 mm. in width) is not yellow, but
usually white. The stalk also appears somewhat lighter in colour than the dark grey-
yellow body-wall (PI. LXI. fig. 1).

The main mass of spicules in the parenchyma of Crater omm'pha meyeri consists,
as in Batliydorus and Bhabdocalyptns, of long thin diacts, which are roughened termin-
ally, and exhibit either rounded off or bluntly pointed extremities. Some spicules bear
a distinct central thickening, or four cruciate, or less frequently two opposite hemispherical
tubercles, while the others exhibit no trace of such structures. The diacts vary greatly
in length and thickness, and are sometimes thickened terminally in a slightly club-shaped
fashion. Less frequently they are gradually narrowed and finally terminate in points,
while the middle portion is so thickened that the whole spicule acquires a spindle-shaped
form. Between these diacts of various length, medium- and small-sized oxyhexacts occur
in irregular distribution. In these the rays are frequently of considerable length, and
are uniformly narrowed towards the extremity, finally terminating in a point.

There is an abundant occurrence of oxyhexasters with very short, sometimes
almost aborted, principal rays, and long slender terminals, which are usually bent
slightly outwards at the base, but are otherwise straight up to their pointed extremities.
Each short principal bears two to four of these terminals. Less frec[uently, and chiefly
on the subgastral trabecular s^^ace, small discohexasters occur, in which the moderately
short principal rays are terminally expanded into a transverse disc. The convex surface
of this disc bears a brush-like tuft of fine terminal rays, with toothed transverse discs on
their extremities (PL LXI. fig. 7).

The dermal skeleton includes medium-sized hypodermal oxypentacts, with moderately
long and strongly developed straight rays, which are roughened near the extremity, but
are elsewhere smooth. The four tangential rays, which are cruciately disposed round
a central node, are at first slightly expanded and decrease gradually in thickness
towards the pointed ends. Besides these hypodermalia which do not always form a
quadrate lattice-work with their tangential rays, a large number of small dermal pen tacts
occur, with rough straight rays, of approximately equal length, and frequently somewhat
club-shaped at their extremities. The four tangentials intersecting at right angles are
often inclined slightly inwards. From the point of intersection a knob-like tubercle
sometimes projects outwards, representing the undeveloped external radial ray (PL LXI.
fig. 10) ; it is, however, generally absent. Between these dermal pentacts, perfectly
similar tetracts not unfrequently occur, in which the four rays, intersecting at right
angles, are usually curved shglitly inwards.

In the gastral skeleton, as in Rhabdocalyptus, there are no large hypogastralia. The
autogastralia are all small rough pentacts, in which the straight rays are not club-

REPORT ON THE HEXACTINELLIDA. 163

shaped terminally, but simply rouuded, or else with a Ijlunt point (PI. LXI. fig. 9). The
node of intersection is frequently thickened, and usually exhibits a slightly projecting
tubercle, representing the undeveloped sixth ray. While this sixth ray is undeveloped
in the gastraLia proper, it appears distinctly on the spicules which serve to support the
lining of the efferent canals, and exactly resembles the other five rays. The canalaria
are therefore oxyhexacts, and the sixth ray usually projects for a greater or less distance
into the lumen of the efferent canals. While they form a continuous layer on the internal
surface of the wider canals, they occur more sparsely in the narrower branches, and may
finally be altogether absent (PI. LXI. fig. 3). There are no special marginalia, since the
lamellar oscular fringe described above ends in a simple smooth edge.

The stalk demands special notice. Though its structure essentially resembles that of
the cup-shaped body, there are several peculiarities worth noting. It is penetrated by
several anastomosing longitudinal canals which open above into the gastral cavity of
the cup, and receive laterally the short efferent canals of the wall.

The long diacts of the parenchyma have almost exclusively a longitudinal disposition,
and are more or less abundantly and firmly united by transverse synapticula, so that the
whole stalk, being penetrated and supported by a continuous and connected framework,
is very compact and firm. Only near the external surface, and near the upper end where
the stalk joins the cup, are the parenchymal spicules united to a less extent, or not at
all. Between the diacts there is a comparatively abundant occurrence of small thick
hexacts, which may be isolated, but generally have one or two opposite rays laterally
united by a siliceous mass to the long diacts (PI. LXI. figs. 5, 6).

The degree of amalgamation and connection by means of synapticula increases from
above downwards, so that the basal enlargement is traversed by a dense and stony siliceous
framework.

The oxyhexasters which occur so abundantly in the parenchyma of the cup-shaped
body are less frequent in the stalk, and are only found in those portions where the union
of the spicules by amalgamation or by synapticula has not occurred to a very marked
extent. I have observed that small discohexasters only occur isolated.

The skeleton of the smooth external skin-layer exhibits medium-sized hy|)odermal
pentacts and small rough dermal pentacts or tetracts, with rounded or slightly club-
shaped ray- tips, like those which occurred in the dermal skeleton of the cup-shaped body.

On the inner surface of the large longitudinal efferent canals of the stalk, similar
roughened pentacts also occur, with rounded or slightly club-shaped ray-tips. These
correspond to the pentact gastralia of the gastral cavity in the superior cup-shaped
portion.

This form, which Bowerbank received in 1877 from Dr. A. B. Meyer, and described
as an anomalous Hyalonema,^ is not a Hyalonema, but certainly Crateromorpha

1 Proc. Zool. Soc. Land., p. 461, 1877.

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

meyeri. This was further proved by the examination of a hxrge sj^ecimen of the same
form, preserved in the Zoological Museum at Dresden, which I was allowed to study
through the kindness of Hofrath A. B. Meyer.

I also found a di-ied specimen of Crater omorpha meyeri among the sponges which
Dr. Doderlein collected at Enosima in Japan.

2. Crateromorpha thierfelderi, n. sp. (PI. LXII. figs. 1-4).

The stalked cup-shaped form, represented in its natural size in PI. LXII. fig. 1 , was
trawled near the Little Ki Island (Station 192), from a depth of 129 to 140 fathoms, and
a blue mud bottom. It measures 10 cm. in total length, and 4 "5 cm. in maximum
breadth. The bulb-like or thick spindle-shaped body is 5 cm. long, and exhibits a
simple bulging gastral cavity, about 2 '5 cm. in width, opening superiorly, by a circular
osculum 15 mm. broad, and continued inferiorly into the lumen of the stalk. The stalk
is from 6 to 8 n^n. in breadth. The lower end of the latter and a large piece of the
upper waU of the body have unfortunately been torn away. In its bulging portion the
wall is about 8 mm. in thickness, but it decreases rapidly towards the upper end, ending
finally in a short, thin, smooth, projecting fringe. Inferiorly, on the other hand, the
thickness of the body- wall decreases to about 3 mm. as it joins the stalk.

WhUe the external surface of the sponge appears smooth, the internal surface exhibits
numerous round apertures, belonging to the efferent canals. These pores vary in width
up to 5 mm.

The parenchyma contains besides strongly developed slightly curved diacts (PI. LXII.
fig. 4) and medium-sized hexacts, also numerous oxyhexasters, with terminal rays bent
gently outwards (PL LXII. fig. 2). Besides these, there are small comparatively regular
discohexasters 0'05 mm. in diameter, with five to eight equal-sized terminals on each
principal ray (PI. LXII. fig. 3).

The dermal membrane contains rough tetracts, and less frequently similar pentacts,
with a more or less strongly developed internal ray. In the gastral membrane only
roughened pentacts occur.

I have named this form in honour of my friend Professor Albert Thierfelder of
Rostock. WhUe it presents a certain resemblance to Crater omorpha meyeri, it is also
very closely allied to Crateromorpha murrayi about to be described, and occupies in fact
an intermediate position between these two species.

3. CrateroTuorpha murrayi, n. sp. (PI. LXIIL).

In the vicinity of the Little Ki Island (Station 192), from a depth of 129 to 140
fathoms and a blue mud ground, the trawl brought up, along with numerous other

REPORT ON THE HEXACTINELLIDA. 165

Hexactinellids, the beautiful stalked form represented in PI. LXIII. fig. 1. The cup
measured 12 cm. in height, and as much in maximum breadth— at the circular superior
opening of the somewhat depressed and yet bell-shaped body. This form I have named
Crateromorpha murrayi in honour of my much esteemed friend John Murray. The
tubular stalk, which is broken oif inferiorly, has a parietal thickness of 3 mm., and a
diameter of 2'5 cm. Just beyond the trumpet-shaped expansion, where the stalk joins
the body, tlie latter exhibits a rounded boss-like protrusion (1*5 cm. in length), directed
outwards and upwards. The external surface of the body and of the stalk is smooth, and
covered by a fine quadrate lattice-work, through which numerous roundish incurrent
apertures are visible. The internal surface of the wide gastral cavity exhibits numerous
round excurrent apertures of the efferent system of canals, which are small in the
neighbourhood of the superior margin, but become gradually larger towards the base of
the cup, aud occur at last so close together, that a network of more or less slender septa
protrudes between them into the gastral cavity.

The thickness of the body-wall is on an average between 2 to 3 cm. , and decreases
oradually upwards to the smooth sharp-edged margin, on which no distinct cuff-like
fringe was discoverable.

Of the larger spicules of the parenchymal skeleton the most abundant are those
slender diact forms which are beset at both ends with small pointed spines. The ends are
thickened in club-shaped fashion, bluntly pointed, or less frequently simply rounded off. In
the middle these spicules are in some cases smooth, while in others they exhibit the familiar
annular swelling, or else four cruciate or two opposite hemispherical nodes. Beside these
we have to note the occurrence— characteristic of the species — of thick diacts of medium
length (2'5 to 4 cm.), which are curved in an Indian bow fashion, or else gently twisted
in the middle. They are here especially well developed, attaining a thickness of 0'15 mm.,
and gradually decrease in diameter towards the extremities, where they end in blunt
points (PI. LXIII. fig. 4). The terminal portion may be smooth, as represented in
PI. LXIII. fig. 4, or laterally beset with numerous minute pointed spines.

There is in the parenchyma a remarkably sparse occurrence of large or medium-sized
hexacts, an important item in distinguishing this species from the closely related Cratero-
morpha thierfelderi.

Between the large parenchymalia there is an abundant occurrence of oxyhexasters
with short principal rays, and two to four long divergent terminals, which have either
a perfectly straight course, or are somewhat curved terminally (PI. LXIII. fig. 5). Less
frequently, and especially in the neighbourhood of the gastral surface, another form of
rosette is represented by small discohexasters, in which the somewhat short principal rays
bear on their terminal transverse discs numerous fine terminals, which vary in length,
and bear transversely on their extremities small convex toothed discs (PI. LXIII.
fig. 6).

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

The dermal skeleton is composed of medium-sized, strongly developed hypodermal
pentacts, in which the smooth rays, gradually narrowed towards the extremity, are either
simply conical or somewhat roughened and rounded at the tips. Besides these, the
dermal membrane includes numerous teti'acts and pentacts, and more rarely diacts, with
central nodes. These dermalia correspond exactly to those of Cratcromorpha meyeri.
Their straight and rough rays intersect at right angles, and the ends are thickened in
somewhat club-shaped fashion.

The gastralia, which form a connected layer lining the gastral cavity, resemble the
dermalia, and consist, for the most part, of pentacts with small tubercles or bosses in the
place of the sixth ray. Hexacts also occur, in which the projecting radial ray extends
for a greater or less distance into the gastral cavity. The canalaria lining the inner
surface of the wider efferent canals resemble those just described. The skeleton of the
stalk does not essentially differ from that of the bell-shaped body.

4. Crateromorpha tumida, n. sp. (PI. LXVIL; PL LXVIII. fig. 2).

Near the Banda Islands (Station 194a, lat. 4° 31' S., long. 129° 57' 20" E.),
from a depth of 360 fathoms and a volcanic mud ground, the trawl brought up two
specimens of a Crateromorpha, in which the very soft body proper has been badly
preserved, and so much injured by subsequent friction during transport, that only a
cloudy formless mass remains attached to the firm stalk. Fortunately a figure
(PI. LXVIL fig. 1) of one of the specimens had been previously made under Sir Wyville
Thomson's direction, which shows at least the form of the body before ft was so much
damaged. The specimen measures 17 cm. in length. The cylindrical stalk is firmly
attached to a compact basal plate, and from its thinnest portion, just above the basal
plate (where it measures about 5 mm. in breadth), gradually increases in diameter on to
the base of the body, where it attains a breadth of 20 mm. Where the smooth and firm
stalk passes into the body, there is a trumpet-shaped expansion with projecting radial
ridges and depressions between. The body proper measures 13 cm. in breadth by 8 cm.
in height, it exhibits irregular lateral bulgings, and has on the whole the form of a broad
cup with very thick walls and with sharp superior margin. The irregular undulating
edge of the superior aperture is approximately circular and measures almost 10 cm. in
diameter.

The irregularly ridged and furrowed external surface of the body is covered by the
delicate lattice-work of the dermal membrane, through which one can see the numerous
round lacunte (1 to 4 mm. wide) of the afferent canal system. The somewhat firmer
internal surface of the wide, almost hemispherical gastral cavity, exhibits in its upper
marginal region comparatively small apertures of the efterent canal system, which
beeome gradually larger, however, towards the lower surface. It is noteworthy, as a

REPORT ON THE HEXACTINELLIDA. 167

distinction from Aulochone, that from the lower portion of the gastral cavity no canal
leads downwards into the stalk, but that the floor of the gastral cavity has the same
parenchymatous structure as the lateral wall, and exhibits the same round canalicular
apertures as in the latter. The stalk is not a tube with a simple canalicular lumen as in
Aulochone, but is penetrated by an irregular lacunar and canalicular system, which opens
superiorly into the efferent canal system of the lower portion of the body.

The greater part of the parenchymal skeleton consists of long, straight or slightly
curved, smooth diacts, with or without central nodes, and with rough club-shaped ends
(PI. LXVII. figs. 2, 3). They occur either in isolated distribution or in strands. Between
these a large number of oxyhexasters occur, in which the comparatively short and strong
principal rays bear from two to four diverging, long, strongly developed terminals. The
extremities of these terminal rays are never bent in hook-like fashion as in Crateromorpha
■murrayi and Crateromorpha thierfelderi. Sometimes one or other of the principal rays
remains undivided, and not unfrequently both the rays on one axis, so that the latter
extends in a simple straight course to a pointed extremity (PI. LXVII. fig. 5). Between
these very numerous oxyhexasters, many-rayed discohexasters occur in very much less
abundance, but of about the same size as the former. Each of the short principal
rays bears about eight straight, substantial, cylindrical terminals, which are of equal
length, and extend, with uniform divergence, radially outwards. On their extremities
they bear convex, marginally toothed, transverse discs (PI. LXVII. fig. 6 ; Pl. LXVIII.
fig. 2).

The hypodermalia of the dermal skeleton are moderately large pentacts, with straight,
cylindrical, generally smooth, but terminally roughened and somewhat club-shaped rays.
The tangential rays are disposed at right angles to one another and to the proximal radial
ray.

The dermal membrane itself contains autodermalia in the form of cruciate tetracts,
with rough cylindrical rays, which are usually simply rounded off at their extremities, or
less frequently swollen in club-shaped fashion. Diacts also occur, with rays crossed
at right angles, and with or without four or two central protuberances. Sometimes,
through the abortion of one of the tetract rays, a triact form results (PI. LXVII.
fig. 4).

The gastral and canalicular skeleton is destitute of hypogastralia and hy]3ocanalaria.
It consists of rough pentacts, with rays crossed at right angles, and rounded ofi' or some-
what swollen at their extremities.

The spicules of the stalk do not difter from the corresponding forms in the body
proper (PI. LXVIII. fig. 2). In the lowermost portion of the stalk, however, there is,
by means of synapticula, a more or less extensive amalgamation and connection of
the long diacts, which are longitudinally disposed in approximately or perfectly parallel
courses (PI. LXVII. fig. 8).

168 THE VOYAGE OF H.M.S. CHALLENGEE.

The second specimeu, which is not figured, has about the same size as the above, and
resembled it exactly, though much looser in texture. Not only is the greater part of the
body wanting, but the lower end of the conical, expanded, basal portion is quite gone.
The region where the stalk passes into the body is somewhat broader than in the former
specimen.

Genus 8. Aulochone, n. gen. (PI. LXVI. ; PI. LXVIII. figs. 1, 3-7).

A cylindrical cup is borne on a long tubular stalk. The oscular margin is folded
outwards and backwards in such a way that a portion of the gastral surface has
become part of the outer wall of the cylindrical body, while the other portion of the gastral
wall surrounds the funnel-shajDcd gastral cavity, and is continued downwards into the
membranous lining of the tubular sheath.

In this way the dermal membrane limits the body only on the inferior somewhat
hollowed surface, where the cup is attached to the top of the stalk.

This concave, somewhat hollowed, inferior surface of the body exhibits a fine-meshed,
quadratic dermal lattice-work, while the cylindrical external surface of the body and the
internal funnel-shaped gastral surface exhibit the numerous large and small roundish
openings of the efferent water canals.

The parenchyma contains, between the large diacts, numerous discohexasters with
terminal rays of variable length. In the dermal and gastral membranes small rough
pentacts have a preponderant or exclusive occurrence.

1. Aulochone cylindrica, n. sp. (PI. LXVI. ; PI. LXVIII. fig. 1).

A sponge form, strikingly different in external appearance from the Crater omorpha
species just discussed, but nevertheless closely allied, as the character of the siliceous
spicules plainly shows, was trawled to the north-easf of the Kermadec Islands (Station
171, lat. 28° 33' S., long. 177° 50' W.), from a depth of 600 fathoms and a hard ground
bottom. The sponge consisted, in the first place, of a broken stalk, about 4 mm. in
diameter, and 2 '5 cm. in present length, but probably mucli longer when intact. On
this the body proper is seated, resembling the vertebral centrum of a bony fish in the
infundibular concavity of its terminal surface, and in the cylindrical appearance of the
sides (PI. LXVI. fig. 4). The upper terminal surface is deeply concave in funnel-like
fashion, and this funnel-like space is continued directly into the cylindrical tubular
cavity formed by the upward prolongation of the stalk. The opposite inferior surface,
on the other hand, is hollowed out only to the extent of forming a circular furrow,
while the central portion is occupied by the trumpet-shaped connecting portion between
stalk and body (PL LXVI. figs. 1, 2, 4). Nor is this broad annular furrow on the

REPORT ON THE HEXACTINELLIDA. 169

lower surface quite uniformly hollowed out, like the funnel-like superior surface, but is
somewhat interrupted by four cruciately disposed radial folds which project inwards
(PI. LXVI. figs. 2, 3).

Towards the lower surface with its groove-like concavity, the cylindrical surface of
tlie sides is sharply defined off by a smooth margin with a sharp edge, which projects
downwards at a sharp angle. From the u^iper terminal funnel the cylindrical lateral
surface is much more irregularly separated, exhibiting manifold inward folds and less of
a sharp margin (PL LXVI. figs. 1, 4).

The whole lower surface, from the trumpet-shaped attachment of the stalk to the
sharp-edged bounding margin, is covered by a very smoothly expanded dermal network
with fine square meshes. Through this, large round lacunae 10 cm. and more in
width can be seen (Pi. LXVI. figs. 2, 3). The character of the cylindrical lateral surface,
and of the upper funnel-shaped concavity continued into the central canal of the
stalk, differs from that above described in the less uniform development of the bounding
membrane, and in its penetration by numerous small and large irregularly disposed
round apertures, the latter measuring up to 6 mm. in diameter. These gaps represent
the external apertures of numerous efferent canals (PI. LXVI. figs. 1, 4).

From vertical sections it may be seen that it is only from the inferior grooved surface
that the more or less wide afferent canals penetrate upwards into the loose parenchyma
of the body, while the corresponding efferent passages all open on the lateral or on the
superior funnel-shaped surface (PI. LXVI. fig. 4). From this it follows that not the superior
funnel-like surface alone, but the external lateral cylindrical surface also, corresponds to
the gastral surface of other Hexactinellids. The true dermal surface, on the other hand,
is represented wholly by the groove-like concavity on the lower surface of the body, and
by its direct continuation, the outer surface of the stalk. We have thus to do with an
evcrsion of the originally inner surface of the cup and with a partial overgrowth of the
wall, so that a portion of the originally internal gastral surface, that is to say, the upper
portion, has become the external lateral surface, while the whole originally lateral external
wall has been confined by the outward and downward curvature of the sharp oscular
margin to the inferior concave surface of the present sponge body. This theory of the
present shape is in complete harmony with the arrangement of the chamber layer and
with the distribution of the siliceous spicules. In a vertical section of the downwardly
bent, overhanging, sharp edge of the lateral margin, that is to say, of the original everted
oscular rim (PI. LXVIII. fig. 1), it maybe seen that the folded chamber layer is disposed
in such a way that all the chambers have their convex external surface turned to the
delicately latticed skin which covers the grooved inferior surface of the sponge body.
The efferent apertures of the chambers, on the other hand, are all directed towards the
cylindrical lateral surface, which therefore corresponds to the gastral surface of other
sponges, while the inferior surface is really the dermal.

(ZOOL. CHALL. EXP. — PART LIII. — 1887.) Ggg 22

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

The principal spicules of the parenchymal skeleton are slender, smooth diacts, of
varied length. They exhibit rough ends thickened in club-like fashion. Less frequently
simply rounded or pointed diacts occur. Many of these long diacts exhibit no swellings
or lateral processes at their middle point, while others bear four cruciate, or two opposite
tubercles — the rudiments of undeveloped rays.

Between these long jiarenchymal spicules a large number of irregularly scattered
oxyhexastere occur. They exhibit strongly developed, short principal rays, and two to
four long, straight, divergent terminals on each principal (PI. LXVI. fig. 6). It ought
to be noted that oxyhexasters very frequently occur in which one principal ray is con-
spicuously longer than the others, so that the whole sj^icule deviates considerably from
the rosette form. Other irregularities occur, such as the displacement of one of the
terminal rays from the common whorl to the side of the associated principal ; spinous
ramifications or irregular twistings occur here and there as abnormal, or perhaps as
pathological modifications (PI. LXVI. figs. 11, 12). In isolated cases, I observed disco-
hexasters with moderately short princijDal rays, bearing on their expanded ends numerous
long, thin terminals, radially disposed, and each terminating in a transverse disc with
clasp-like teeth (PI. LXVI. fig. 7). In the subdermal spaces of the connecting portion
between the stalk and the body, large discohexasters of floricome-like pattern occur, as
is also the case in Aulochone lilium (PI. LXVIII. fig. 5). The 2>erianth-like groups of
terminal rays are in this species, however, somewhat thinner at their basal and median
portions.

The dermal skeleton consists of strongly developed rough pentacts, of variable size,
the rays having rounded or club-shaped ends. Between these pentacts, tetracts of similar
structure occasionally occur. The gastral skeleton on the upper funnel-like surface, and
on the cylindrical sides, consists of rough pentacts similar to the above (PL LXVI.
fig. 10).

On the much curved regions, both dermalia and gastralia exhibit a backward curvature
of the four tangential rays. On the superficial side a small tubercle is occasionally seen
where the sixth ray has not been developed.

In the skin which lines the large canalicular cavities of the body, I have never found
any special canalaria.

The skeletal elements of the stalk do not essentially diff'er from those of the body
generally. The longitudinal and transverse, i.e., circular, bundles of spicules usually
alternate in their arrangement. Between the above, the oxyhexasters and small disco-
hexasters already referred to occur, while the large floricome-like discohexasters are found
more abundantly in the subdermal spaces. The dermalia are strongly developed, and
the inwardly projecting radial proximal ray is often strikingly shortened. The gastralia
are frequently less rough in their median portion, while the club-shaped thickened ends
always exhibit numerous minute spines.

REPORT ON THE HEXACTINELLIDA. 171

2. Aulochone lilium. n. sp. (PI. LXVIII. figs. 3-7).

Near the Meangis Islands, north-east of Celebes (Station 214, lat. 4° 33' N., long.
127° 6'E.), from a depth of 500 fathoms and a blue mud ground, the dredge brought up
a comparatively well preserved long-stalked type of Aulochone, which measured 23 cm. in
total length. The rounded pillar-like portion of the body measured 7 cm. in transverse
diameter and 4 cm. in height, while the tubular stalk measured 6 mm. in thickness and
19 cm. in height. The stalk forms a funnel-shaped expansion with laterally projecting
radial folds before passing into the body proper, and the lumen of this stalk opens
into the funnel-shaped gastral cavity within the loose body. The body, which has been
somewhat compressed both in capture and preservation, exhibits on its lower side a
number of radially projecting folds, which radiate out from four cruciately disposed
principal ridges. Between these radial septa there are pouch-like depressions, which are
closed externally by means of a sharp-edged margin which projects perpendicularly
downwards, and is directly continuous with terminal expansions of the radial folds
(PI. LXVIII. fig. 3). The pouch-like depressions of the lower surface thus appear to be
rounded ofi" externally.

The funnel-shaped excavation on the superior surface is bordered by pad-like folds of
the body -wall, and is not distinctly limited superiorly or externally. In fact, the gastral
wall is directly continued by a rounded superior marginal pad into the somewhat longi-
tudinally folded outer surface of the body. The latter represents, indeed, the everted
gastral surface, and is limited towards the inferior dermal surface only at the inferior
external sharp-edged margin of the body generally (PI. LXVIII. fig. 3). It is thus
evident that the whole architecture of this sponge agrees with that of Aulochone
cylindrica, and this is ecjually true of the internal structure, and especially in regard
to the disposition of the folded chamber layer. The convexity of all the chambers
within the body is directed towards the cavities, which are pushed in from the lower
surface, and enveloped in the dermal membrane; in the stalk, on the other hand, the
convexity is towards the external dermal surface. The openings of the chambers
within the body are directed towards the external and superior gastral wall, while in the
stalk they are turned towards the central lumen of the tube.

The principal spicules of the parenchymal skeleton are long, thin, or slightly thickened
diacts, with rough, club-shaped, or simply rounded ends. There is rarely any annular
thickening, or development of four cruciate, or two opposite bosses. The rays are for the
most part uniformly cylindrical, or slightly thickened in a spindle-like manner, and are
perfectly smooth except at their roughened ends. They are either quite straight or gently
curved. The abundant parenchymal oxyhexasters correspond exactly in form and size
to the oxyhexasters of Aulochone cylindrica; they differ, however, markedly in this, that
the long terminals are completely covered with small spines, being in fact very rough

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

(PI. LXVIII. fig. G). Another divergent rosette form is represented by a floricome-like
discohexaster, which occurs in considerable abundance, especially in the subdermal
trabecular space and on the transverse terminal expansions. This form exhibits six
comparatively short and simple principal rays, each equipped with a whorled tuft of long,
fine, S-shaped terminals. These bear on their extremities marginally toothed plates, which
overhano- towards the exterior. The number of terminals on each principal ray varies.
Sometimes there are but six in a single whorl, sometimes eight, ten or more, and they are
not always of equal length (PI. LXVIII. fig. 5).

The parenchymal spicules of the stalk differ from those of the body in this, that the
long diacts are stronger, and in part, at least, in the inferior portion of the stalk, directly
united by synapticula. Besides oxyhexasters and discohexasters, peculiar hexacts occur
with short but thick rays, which meet in the centre in a sort of trumpet-shaped basal expan-
sion, forming a thick node of intersection. The outer rounded extremity is always beset
with small spines, which sometimes extend over the entii-e length. The rays are either of
equal length, as in PI. LXVIII. fig. 7, or to some extent unequal, in that the rays on the
same axis are equal to one another, but different from those on the other axes. Less
frequently pentacts or tetracts occur of similar form.

The dermal and gastral skeletons are similarly composed of pentacts of variable size, in
which the rays are thickly beset with small spines, either throughout as in the dermalia, or
over the greater part of their length, that is with the exception of the central portion. At
their extremities the rays are either simply rounded off, or thickened in club-shaped fashion.
The dermalia have their tangential rays usually inclined somewhat inwards, and exhibit
no distal tubercle at the node of intersection. The gastraba, on the other hand, have
their tangentials disposed at right angles to the frequently reduced radial ray, and very
often exhibit a tubercle or rounded knob, which projects into the gastral cavity, in the
position of the sixth ray.

Genus 9. Caulocalyx, n. gen.
This genus contains the single species Caulocalyx tener.

Caulocalyx tener, n. sp. (PI. LXIX.).

A fragment of a remarkable sponge form, imperfectly preserved, but indubitably
referable to the family Rosellida3, was trawled in the Mid South Atlantic, to the west of
Tristan da Cunha (Station 333, lat. 35° 36' S., long. 21° 12' W.), from a depth of 2025
fathoms and a Globigerina ooze ground. It consisted for the most part of a fibrous stalk,
. frayed out inferiorly, and measuring 4 cm. in length, by 2 to 4 mm. in breadth. Towards
the upper end radial lateral folds gradually increase in height, so as to form a cup-shaped

REPORT ON THE HEXACTINELLIDA. 173

body measuring 4 to 5 cm. in transverse and longitudinal diameter. This loose body-
was unfortunately very much injured. While the upper funnel-shaped terminal surface,
limiting the gastral cavity, is so abundantly penetrated by roundish apertures of variable
size that only a network of more or less broad septa and free smooth margins remain,
the external, much torn surface is studded with numerous thin and pointed spicules
which project freely for 2 to 4 cm. beyond the general surface of the body (PI. LXIX.
fig. 1).

The large spicules of the body-parenchyma consist for the most part of long, rather
thin, smooth diacts, which are rough at their ends, and terminate without any marked
swelling in a simple conical point. At their centre some exhibit a distinct annular
swellino- containing an intersection of the axial canal. Most of the spicules, however,
remain uniform and cylindrical throughout.

Between these long and slightly bent or looped diacts, a varying number of diverse
rosettes occur. The majority are rather large discohexasters, in which the moderately
short and simple principal rays usually bear four long divergent terminals. The latter
are very thin at their base, but increase gradually in thickness towards the exterior, and
finally terminate in a relatively broad, convex, transverse disc, which usually bears a
fairly large number of marginal teeth (PI. LXIX. figs. 3, 4, 7, 8). In the more strongly
developed forms, the thickened external ends of the terminal rays exhibit small irregularly
scattered tubercles (PL LXIX. fig. 7).

Between very weakly developed rosettes of this type, with thin terminal rays
(PL LXIX. fig. 3), and rather strongly developed forms, twice or three times as broad
(PL LXIX. fig. 4), many transition types occur.

In addition to the above, we have to note the rather frequent occurrence, especially
in the subgastral trabecular spaces, of small plumicomes, in which the moderately short,
simple principal rays each bear on their ends a broad convex transverse plate. From
the convex external side of the latter, a large number of S-shaped, pointed, terminal rays
arise, disposed in perianth-like fashion, and more strongly developed in the outer than
in the central portion (PL LXIX. fig. 6).

The dermal skeleton is formed of oxjrpentacts, in which the tangential rays, inter-
secting at right angles, and often somewhat bent, are laterally beset with scattered, slightly
curved spines. The radially directed proximal ray, which is usually much longer, and
frequently curved with faint undulations, is, on the other hand, either quite smooth, or
merely ecpiipped with minute tubercles (PL LXIX. fig. 2). An inconspicuous protuber-
ance may be occasionally observed on the distal side of the node of intersection. The
gastralia are somewhat rough oxyhexacts, in which the four tangential rays are straight
or slightly curved, and conically pointed at their ends. The moderately long, freely
projecting ray is either conically pointed as above, or exhibits a slightly clubbed
thickening in front of the terminal point, while the usually longer radial ray opposite

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

the lattei- is either conically pointed as above (PL LXIX. fig. 5), or becomes gradually
narrower towards the end.

In the fibrous rudiment of a stalk, the majority of the long diacts, which vaiy ex-
ceedingly in thickness, have their ends thickened in club-like fashion, or thickly beset
with spines, while on the surface of many which are otherwise smooth small spines occur
on the middle portion. Towards the lower end of several of the long diacts, which
are disposed in longitudinal bundles parallel to one another, one finds spines, bosses, or
protuberances of some kind projecting laterally. These seem to represent the rudi-
mentary traces of the synapticula which are subsequently seen in their developed
form.

The radially disposed, freely projecting prostalia, which are found on the side and
probably were also on the superior margin of the cup-shaped body, are oxydiacts measuring
2 to 4 cm. in length. Their freely projecting surface is, for the most part, somewhat
thickly studded with numerous minute, quite irregularly scattered, conical bosses or
spines.

Genus 10. Aidocalyx, n. gen.
This genus contains only one species, Aidocalyx irregularis.

Aulocalyx irregularis, n. sp. (PL LX.).

OS" Marion Island, south-east of the Cape of Good Hope (Station 145a, lat.
46° 41' S., long. 38° 10' E.), from a depth of 310 fathoms and a volcanic sand bottom,
several much injured and partially macerated specimens of a Hexactinellid were dredged.
The lattice framework still hung together, and the general form was that of a broadly ex-
panded cup with complex, much folded or diverticulated wall (PL LX. fig. 1), similar to
that observed in Periphragella elisae, Marshall. Although the fragments obtained were
only from 3 to 4 cm. in height, it seems probable that the uninjured cup was at least
twice as high. Of the soft tissue only some very small fragments remained.

The skeletal framework, which is moderately thick at the base and as hard as stone,
becomes gradually looser and more delicate towards the upper end, and finally so sparse
and thin, that, as in Regadrella, Dictyocalyx, Rhahdocalyptus, Hertioigia, &c., one is
inclined to believe that in the upper, most recently added portion no firm fusion of the
skeletal elements has occurred.

As to the general structure of the soft tissue, it must be noted that between the two
sieve -like, perforated, bounding lamellae of the body-wall (the dermal and gastral mem-
brane), the membrana reticularis forms in simple folds a connected series of thimble-like
diverticula comj)osing the chamber layer. The complication of this folding increases

EEPORT ON THE HEXACTINELLIDA. 175

with the thickness of the wall, and thus differs very greatly in the different regions of
the body. The subdermal trabecular framework, extending between the dermal membrane
and the chamber layer, exhibits only here and there irregular gaps, while the similar
subgastral framework following the concave curvature of the chamber layer exhibits
wide lacunae, which extend into the gastral membrane. A thin layer of the trabecular
framework also extends on the inner surface of each concave diverticulum (PI. LX. fig. 3).
The connected lattice framework of the skeleton consists of medium-sized hexacts,
which are united in an irregular fashion either by fusion or by synapticula (PI. LX.

fig-3).

As can be readily seen in the older portions of the framework, with the help of the
much enlarged axial canals, the six rays stand, as a rule, at right angles to one
another, but frequently bend from their original direction, and become, as they approach
the rays of other hexacts, united with the latter by lateral mooring or by synapticula.

From the complete absence of a regular orientation and mode of connection, it seems
to me preferable to refer Aulocalyx not to the Dictyonina but to the Lyssacina. In the
closer regions of the skeletal framework, between the medium-sized hexacts, numerous
small forms occur with short, thick rays, rounded off at the ends, and with a much expanded
central portion. These are irregularly scattered, and are united with the former, so that
a very close, spongy skeletal meshwork often results.

As to the isolated spicules which occur between the beams of the skeletal framework,
either scattered quite irregularly in the parenchyma, or limited to certain regions of
the same, I ought first to mention the fine, small, regular oxyhexacts, which occur
irregularly, but in considerable abundance, through the whole parenchyma (PL LX. fig. 3).
They always exhibit a slight roughening of all their rays. Besides these, numerous
in*egularly distributed, small, regular discohexasters occur, in which, from each of the
moderately large, simple, smooth main rays, a tuft of S-shaped thin terminals project in
perianth-like arrangement. Each of the terminal rays exhibits a somewhat thickened
end, bearing a convex terminal disc with marginal teeth (PI. LX. figs. 4, 6).

The length of the main rays and the number of terminals on each are the same in
one and the same discohexaster, but vary in the different rosettes, of which two very
different specimens are figured in PI. LX. figs. 4, 6.

Another quite unique and characteristic form of rosette occurs, only however in the
•subdermal trabecular space, but there in comparative abundance. It is distinguished
by its conspicuous size and by its peculiar structure, as represented in PI. LX. figs. 3, 5.

These perfectly regular rosettes have a diameter of 0'4 mm. From each of the
relatively short, smooth main rays, six terminals arise, arranged in a whorl, and disposed
at equal angles in a funnel-like manner. These long terminals are either straight or some-
what convexly curved at their narrow base, and gradually increase in thickness in club-like
fashion toward the simply convex, or more rarely somewhat knobbed external end. They

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

are beset with numerous, somewhat recurved, thin, pointed barhs, wliich surround the
ray in irregular disposition (Pi. LX. fig. 5).

The dermal skeleton is formed of medium-sized oxypentacts, with comparatively long
straight rays, which are, throughout their whole length, somewhat thickly beset witli
small pointed spines. The very ends of the four tangential rays are blunt and conical,
while the proximal ray usually runs more gradually to a point. In the position of the
undeveloped (sixth) distal radial ray, a rounded tubercle sometimes occurs. The equally
large gastralia exhibit the same structure and similar disposition (PI. LX. fig. 3).

Several fragments of a connected skeletal framework, belonging to the same species,
were also trawled between the Marion and Crozets Islands (Station 147, lat. 4G° 16' S.,
long. 48° 27' E.), at a depth of 1600 fathoms, and a Diatom ooze ground. The com-
pletely macerated skeleton contained no trace of soft tissue or of isolated spicules.

Genus 11. Euryplegma, n, gen.
This genus contains only the one species Euryplegma auriculare.

JEuryplegma auriculare, n. sp. (PI. GIL).

Off Eaoul or Sunday Islands, to the north-east of New Zealand (Station 170a,
lat. 29° 45' S., long. 178° 11' W.), from a depth of 630 fathoms, and a volcanic mud ground,
two specimens of a plate-shaped Hexactinellid were trawled. They were, however, partly
macerated and only fragmentary. The best preserved but broken specimen has the form
of a semi-involute ear-shaped plate (PL GIL fig. 1 ). It measures 1 7 cm. in height, and
6 to 7 cm. in breadth. The macerated skeletal fragment of the other specimen forms a
semitubular stalk or basal portion 2 cm. in diameter, and this is continued on into a
closed plate 3 to 5 mm. in thickness, which forms the lower portion of a shallow funnel
(PL GIL fig. 2). The upper portion is unfortunately broken off.

The convex outer surface of the involute ear-shaped specimen (3 to 5 mm. in thick-
ness) seems comparatively smooth, but is penetrated by numerous long oval holes which
lead into the runnels and ducts of the afferent canal system (PL CII. fig. 1). Near the
upper end several of these cavities are covered by a membranous continuation of the
thin sieve-like perforated dermal membrane (PL GIL fig. 3). The concave internal
surface is very rough and uneven in comparison with the outer surface. In the upj^er
portion there is a distinct system of longitudinal ridges with interjacent furrows about
2 mm. in breadth. Somewhat further down they are covered by transverse, arc-like,
broad flat zones, so that transverse ridges 4 to 5 mm. in breadth cross at right angles
over the deeper longitudinal ridges, and project into the interior. Still further down
transverse arcs exhibit knots or cushion-like elevations about 3 mm. in diameter, and

REPOllT ON THE HEXACTINELLIDA. 17"

between these there are pit-like depressions of irregular form (PL CII. fig. 1). The
summit of each of these projecting knots generally bears the apertui'e (0'5 to 1 mm. in
breadth) of an efferent canal. Similar apertures occur in varying abundance between the
nodes. The firm basal portion of the specimen is quite destitute of soft parts for a
distance of 2 to 3 cm., and must have been quite dead for some time (PI. CII. fig. 1).

These projecting ridges above referred to are readily brought into relation with the
macerated skeletal structure exhibited by the lower portion of the other specimen
(PI. CII. fig. 2). This fragment shows very distinctly that on the outer convex surface
there is a reticulate framework, enclosing long, oval or angular meshes, and composed of
beams which lie parallel to the smooth arched surface. Internally there lies a system of
approximately parallel, or rather somewhat fan-shaped divergent longitudinal ridges which
measure 1 to 2 mm. in breadth, and now and again divide. Between these there are
longitudinally disposed canals of equal breadth, which are covered internally by a layer
of beams for the most part transverse. From this latter layer numerous round
perforate swellings and internally open tubes project into the interior.

The round excurrent apertures of the efferent canal system, which measure about
2 to 3 mm. in width, are disposed in longitudinal and transverse rows intersecting at
right angles (PI. CII. fig. 1).

From the macerated skeleton, or better still, from longitudinal and transverse sections
of the whole body, it is clear that the latter consists of a much folded plate 2 to 2"5 mm.
in thickness, in which the bounding surfaces are formed by the sieve-like dermal and
gastral membrane, while between these the ranch folded chamber layer is supported by
means of the familiar trabecular framework (PL CII. fig. 3). Many of the entrance
apertures of the aflereut canals are closed by a second external membrane split
off from the dermal. In its peripheral portion this membrane is supported by the sub-
dermal trabecular framework, but is subsequently quite freely expanded (PL CII. fig. 3).

The dictyonal framework, which penetrates throughout the entire sponge body,
consists of moderately large simple hexacts which are soldered together, or occasionally
connected by synapticula in an irregular fashion.

Only in a few regions, and especially in the neighbourhood of the dermal surface, have
the dictyonal hexacts the typical form with straight rays disposed at right angles to
one another, on which account I had formerly referred this genus to the Dictyonina.'
But they are usually curved and elongated in very various ways (PL CII. fig. 3),
so that an intersection and fusion of the rays of adjacent hexacts may occur at
very various angles, and the dictyonal framework in fact resembles rather the skeletal
framework of several Lyssaciua than the ordinary type of Dictyonina. The whole
skeletal beam is uniformly thick and cylindrical, with numerous irregularly scattered,
minute pointed spines (PL CII. fig. 3).

1 tiber den Bau und das System der Hexactinelliden, Ablmndl. d. 7;. Preuss. Akad. Berlin, 1886, p. SO.

(ZOOL. CHALL. EXP. — PART LUI. — 1887.) Ggg 23

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

The parencbj-malia lying between the beams of the dictyonal framework consist of
small, simple, and regular oxyhexacts, with delicate, straight, and rough rays,
which run out with a slightly conical terminal pointing, and of discohexasters which
are much smaller than the oxyhexacts, but vary considerably in size and form. Some
of them bear on the short simple principals, three to five, rarely more, S-shaped,
whorled, somewhat markedly divergent, moderately thin, terminal rays with convex
marginally toothed transverse discs (PL CII. fig. 5). In other, usually somewhat larger
discohexasters of a similar type, the short simple principal rays bear six to ten long
terminals, which are similarly S-shaped and disposed in perianth-like fashion, only more
closely packed together and with less broad terminal discs (PI. CII. fig. 4) than in the
preceding form. Both forms are connected, however, by numerous intermediate
types.

The dermalia are oxypentacts in which the straight or slightly incurved tangential
rays are about 0'2 mm. in length, while the straight radial proximal is twice or three times
as long. All the five rays are cylindrical, and are smooth except on the conically
pointed terminal portion. A distal tubercle sometimes appears as the rudiment of the
undeveloped sixth ray. The gastralia have quite the same form and an approximately
equal size. They are largest on the gastral surface of the body, and become smaller and
more weakly developed the nearer they are to the blind ends of the efierent canal system.

Tribe II. AMPHIDISCOPHOEA, F. E. Schulze (Pis. XXVII.-LIL).

The bounding surface-membranes always contain amphidiscs. Hexasters are wholly
absent from the parenchyma. There is always a tuft of basal fibres, by means of which
the cup- or club-shaped sponge body is rooted in the mud. The ciliated chambers are
not exactly thimble-shaped or sharply defined from one another, but appear rather as
somewhat irregular diverticula of the membrana reticularis.

Family Hyalonematid^, Gray (Pis. XXVII.-LIL).
Both dermal and gastral membranes bear numerous pentact pinuli.

Subfamily 1. Hyalonematin^, F. E. Schulze (Pis. XXVII.-L.).

The usually cup-shaped body exhibits a more or less sharp-contoured, circular
oscular region, and is only exceptionally split in an ear-shaped fashion.

REPORT ON THE HEXACTINELLIDA. 17 9

Genus 1. Hijalonema, Gray (Pis. XXVII.-XLL).

1832. Gray, Synopsis of the Contents of the British Museum, p. 79 {Hyalonema sieholdii).
1835. Gray, Proo. Zool. Soc. Lond., vol. iii. p. 63; abstracted in Oken's Isis, 1837,

p. 128.
1850. Milne- Edwards and Jules Haimes, Monograph of the British Fossil Corals, p. 81.
1850. Gray, Ann. and Mag. Nat. Hist., ser. 2, vol. vi. p. 306.
1857. Gray, Proc. Zool. Soc. Lond., vol. xxv. p. 278.
1859. Gray, Ann. and Mag. Nat. Hist., ser. 3, vol. iv. p. 439.

1859. Brandt, Symbolae ad jwlypos hyalochietides spectantes {Spongia spinicrvx and Sjwmjia

ocfannjra).

1860. (Jray, Ann. and Mag. Nat. Hist., ser. 3, vol. v. p. 229.

1860. M. Schultze, Die Hyalonemen.

1861. Ehrenberg, Monatsber. d. k. preuss. Akad. d. Wiss. Berlin, p. 450.

1861. von Marten.s, Monatsber. d. k. preuss. Akad. d. Wiss. Berlin, p. 480.

1862. Bowerbank, Phil. Trans., vol. clii. p. 747.

1862. Siiss, Verhandl. d. k. k. zool.-bot. Gesellsch. Wien, xii. p. 85 {Hyalonenia
parallduiit).

1864. Barboza du Bocage, Proc. Zool. Soc. Lond., p. 265 {Hyalonema lusitanicwin).

1865. Barboza du Bocage, Proc. Zool. Soc. Lond., j'- 662.

1866. Gray, Ann. and Mag. Nat. Hist., ser. 3, vol. xviii. p. 287.
1866. Bowerbank, Ann. and Mag. Nat. Hist., ser. 3, vol. xviii. p. 397.
1866. Siiss, Ann. and Mag. Nat. Hist., ser. 3, vol. xviii. p. 404.

1866. Gray, Ann. and Mag. Nat. Hist., ser. 3, vol. xviii. p. 485.

1867. Bowerbank, Proc. Zool. Soc. Lond., p. 18.

1867. Gray, Proc. Zool. Soc. Lend., p. 350 {Carteria japonica).

1867. Max Schultze, Ann. and Mag. Nat. Hist., ser. 3, vol. xix. p. 153; Archiv f. mikrosk.

Anat., Bd. iii. p. 206.
1867. Ehrenberg, Monatsber. d. k. preuss. Akad. d. Wiss. Berlin, pp. 298, 843.
1867. Barboza du Bocage, Ann. and Mag. Nat. Hist, ser. 3, vol. xx. p. 123.

1867. Semper, Archiv f. Naturgesch., Jahrg. xxxix. p. 84 (Hyalonema (Semperella)

schultzei).

1868. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. i. p. 161.

1868. Gray, Ann. and Mag Nat. Hist., ser. 4, vol. ii. pp. 36, 263, 264, 320, 373, 388, 390,

423.
1868. Bowerbank, Proc. Zool. Soc. Lond., p. 118.
1868. Haeckel, Jenaisch. Zeitschr. f. Med. u. Naturwiss, Bd. iv. p. 64.
1868. Loven, Ofversigt k. Vetensk.-Akad. Fcirhandl., xxv. p. 105.
1868. Loven, Archiv f. Naturgesch., Jahrg. xxxiv. p. 82.
1868. Lov^n, Ann. and Mag. Nat. Hist., ser. 4, vol. iii. p. 423.
1868. Wright and Wyville Thomson, Ann. and Mag. Nat. Hist., ser. 4, vol. ii. p. 320.

1868. Semper, Verhandl. d. phys.-med. Gesellsch. Wiirzburg, vol. i. p. 29.

1869. Wyville Thomson, Phil. Trans., vol. chx. p. 70 (Hyalonema loveni).
1869. Bowerbank, Ann. and Mag. Nat. Hist., ser. 4, vol. iii. p. 172.

1869. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. iii. p. 192.

1870. Percival Wright, Quart. Journ. Micr. Sci., vol. x. p. 73.

1870. liarboza du Bocage, Journ. Sci. math. Acad. Lisboa, ix. p. 69.

1871. Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. viii. p. 330.

1872. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. ix. p. 324.

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

1872. Sars, On some remarkable forms of animal life from the great deeps off the Norwegian

Coast, vol. i. {Hi/<donema longissimum and Hi/aloncma parvum).

1873. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. xii. pp. 76, 347.

1873. Wyville Thomson, Depths of the Sea, p. 420.

1874. Gray, Ann. and JIag. Nat. Hist., ser. 4, vol xii. p. 284.
1874. E. P. Wright, Proc. Eoy. Irish Acad., vol. i. p. 549.

1874. Hudlow, Trans. Asiatic Soc. Japan, vol. i.

187r). Higgin, Ann. and Mag. Nat. Hist., ser. 4, vol. xv. p. 377 {Hyalonema cehuense).

1875. Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. xvi. p. 1.
1875. Iviistermann, Archiv f. mikrosk. Anat., Bd. xL p. 282.
1875. Bowerbank, Proc Zool. Soc. Lond., p. 607.

1875. Marshall, Zeitschr. f. wiss. Zool, Suppl.-Bd. xxv. p. 142 {Hyalonema tlwmsonh and

Hyalonema affine).

1876. Marshall, Zeitschr. f. wiss. Zool., Bd. xxvii. p. 113.

1876. Willemoes Suhm, Zeitschr. f. wiss. Zool., Bd. xxvii. p. xcvii.

1877. Wyville Thomson, The Atlantic, p. 273 {Hyalonema toxeres).
1877. Sollas, Ann. and Mag. Nat. Hist., ser. 4, vol. xx. p. 285.

1877. Bowerbank, Proc. Zool. Soc. Lond., p. 461 {Hyalonema anomalum).

1877. Young, Ann. and Mag. Nat. Hist, ser. 4, vol. xx. p. 425 {Hyalonema smithii).

1878. Carter, Ann. and Mag. Nat. Hist., ser. 5, vol. ii. p. 157.

1878. Maggs, Second Annual Report Devonsh. and Cornwall Nat. Hist. Soc, p. 21.

1879. Agassiz, Bull. Mus. Comp. Zool, vol. v. p. 289.

1880. U. Schmidt, Die Spongien des Meerbiiseus von Mexico, Heft iL

1880. Norman, Ann. and Mag. Nat. Hist, ser. 5, vol. vL p. 436.

1881. Milne-Edwards, Comptes rendus, t, xciii. pp. 876, 931.

1885. Armauer Hansen, Spongiadie of the Norwegian North Atlantic Expedition {Hyalonema

arcticuni).

1886. Filhol, La vie au fond des mers, p. 278.

History. — The first important scientific account of a Hyalonema is given by Gray^
{loc. cit., p. 79). After his description of the Pennatulidse, tlie following remarks
are found : — " Here also is placed the extraordinary glass-rope {Hyalonema., Gray), the
axis of which is formed of numerous transparent siliceous fibres, slightly twisted together
so as to look like a rope of spun glass ; the fibres appear to be somewhat similar to the
calcareous spicula of the Alcyonia. These corals are found with their tapering base
inserted in a sponge on the coast of Japan. No animal hitherto discovered, except the
inhabitant of this extraordinary substance, secretes pure silex."

Gray gave a more accurate description of this sponge in 1835,^ regarding it as a
Gorgonia-like coral with the following generic diagnosis : — " Hyalonema. Corallium
simplex, subcylindricum, ad basin cortice coriaceo tuberculato tectum ; tuberculis spar.sis,
depressis, polypiferis. Axis e spiculis numerosis, elongatis, filiformibus, subcoutortis,
siliceis constans. Polypus ignotus. Hyalonema sieboldii. Hab. apud Japoniam. Dr.
Siebold."

Against the view held by Gray that the siliceous threads are to be regarded as

' Synopsis of the Contents of the British Museum, ed. 27, 1832, London, 12 mo.
^ Proc. Zool. Soc. Lond., pt. iii. pp. 63-65.

REPORT ON THE HEXACTINELLIDA. 181

the skeleton of the adherent polype, Valenciennes expressed the belief that we have here
to deal with spougoid spicules, and that the polypes occurring upon these belong to the
group of the Zoantharia, and are to be looked upon as parasites.^

Gray* at once defended his views, and pointed out that it was improbable that the
polypes should always be found associated with the sponge, or that the association
should be so intimate as it was, and noted further that the axis skeleton of the well-
known Gorgonise presents a concentric lamination similar to each of the long sUiceous
threads.

In a further communication by Gray on the same subject,' the family of the
" Hyalonemidse " is regarded as a family of corals, with the single genus Hyalonermi, with
characters which agree, essentially, though not verbally, with the above diagnosis of the
senus given in 1835.

Gray reported the presence of ring-shaped or spiral projections on the outer surface
of the concentrically laminated long spicules, and communicated the result of the
chemical analysis of these by W. Prout and T. J. Pearsall, both of whom gave their
composition as hydrated silicic acid.

It is curious that Gray here designates as Hyalonema mirabilis the only known species,
which he had described in his first communications as Hyalonema sieholdii, and that he
quotes the Synopsis of the British Museum, 1830, p. 118, as relevant literature. This
quotation, which has passed into se%'eral later papers by other observers, is proved to be
false, inasmuch as in the library of the British Museum no edition of the Synopsis of
the year 1830 is to befouud, while the edition immediately preceding, published in 1827,
contains no communication on Hyalonema, either at the page cited or an}^"here else.
Since I was unable, for want of time, to continue the search further during my stay in
London, I requested Mr. Stuart 0. Ridley to do so, and the result of his accurate examin-
ation— for which I here express my best thanks — has shown that in none of the earlier
editions of the Synopsis of the British Museum, down to edition xxvii. of the year 1832 — •
in which the words already quoted are to be found on page 79 — does any communication
on Hyalonema occur. This notice, however, is repeated in the following editions — xxvii.
to xl. — ^without essential change, and further observations are first added in edition xli. of
the year 1840. No alternative is therefore left but to believe that Gray made a false
quotation from memory, or had perhaps in error taken the note on his museum label or
such like for the scientific name which he himself had adopted. This is important, inas-
much as it is not the specific name Hyalonema mirahilis used by Gray in the year 1857,
but that already given by Gray himself to the same species in 1835, Hyalonema
sieholdii, which has the priority.

I Milne-Edwards and Jules Haime, Monograph of the British Fossil Corals, j). 81.
- Ann. and Mag. Nat. Hkt., ser. 2, vol. vi. p. 306.
^ Proc. Zool. Soc. Land., 1857, pt. xxv. p. 278.

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

On the systematic position of Ilyalonema, Gray in 1859^ asserted that within the
group of polypes with feathered tentacles or Alcyonaria, and close to the " SabulicolaB "
which are rooted in the sand, and to the " Rupicolse " or Gorgonidse attached to
the rock, a third division might be formed, which on account of its rooting on sponges
might be named " Spongicolse," with the single genus Hyalonema.

The opinion of Gray as to the polype nature of the remarkable glassy and filamentous
tuft, and which became fixed to a sponge, was shared by Brandt,'^ who in 1859 also
described the glassy spicular bundle as the axial skeleton of the cortical layer of the
polype. Brandt thought that two difierent genera of polypes could be distinguished,
viz., Hyalonema with plain discoid polypes, and Hyalochiete with cylindrical terminally
funnel-shaped solitary polypes, both of which he united in the single family
Hyalochastides. In the genus Hyalonema he described two species, namely, Hyalonema
sieboldii. Gray, and Hyalonema affine, Brandt ; of the genus Hyalochaeta only the single
species, Hyaloch^ta possieti.

The loose fleecy mass in which the conical extremity of the glassy filamentous tuft is
usually embedded was referred to by Brandt as a " parasitic sponge " which had destroyed
the soft parts of the polype stock at its point of attachment, he also distinguished,
though he did not accurately describe, two species of this sponge as " Spongia spinicrux "
and " Spongia octancyra." The attachment of a group of Hyalonema to a Pholas-
bored stone by means of their conical projecting tufted extremity, which occurred in the
case of a specimen from Japan, Brandt regarded as artificial.

Of this work by Brandt a critical review by Gray^ appeared in 1860. Gray asserted
that the two genera distinguished by Brandt with their three species were varieties of
one and the same species. Ehrenberg* also expressed an opinion during the same year
as to the nature of this remarkable organism ; he regarded the whole as a Japanese
artifice, compounded of long sponge spicules and a polype.

A thorough investigation of several dried Hyalonemata from Japan was first under-
taken by Max Schultze. As a result of this he briefly expressed his views in regard to
the relation of the polype and sponge," to the efl"ect that the siliceous filiform cords and
the sponge body form an inseparable whole, and that accordingly the Hyalonema are to
be regarded not as polypes but as sponges. In the same year he published the results of
his accurate analysis of the formation and structure of the sponge, as well as of the
adherent polype crust, in a special monograph, illustrated by beautiful plates, entitled
Die Hyalonemen. The opinion expressed in the preliminary communication, that the
crust surrounding the spicular tuft also belonged to the sponge body, he soon retracted.
He believed that the large cylindrical sponge body became attached by means

1 Avn. and May. Nat. Hist., ser. 3, vol. iv. p. 439. ' Symbolai ad polypos hyalochaetides spectantes.

3 .47171. and Mag. Nat. Hist., ser. 3, vol. v. p. 229. * Monatsber. d. k. prevss. Akad. d. IViss. Berlin, p. 173.

5 Comptes rendus, April 23, 1860, vol. 1. p. 792.

REPORT ON THE HEXACTINELLIDA. 183

of its broad truncated basal plate to a solid substratum, and that the long spicular
tuft, whose inferior compact portion is surrounded by the crust of a parasitic p olyi)e,
Palytlioa fatua, projected from the narrowed extremity. Passing over his accurate
description of the long siliceous filaments, as well as of the elegant, very manifold, small,
siliceous spicules of the sponge body proper, we need only here refer to the important
observation made by Max Schultze, to the effect that both in the long tuft-spicules, and
in the many-rayed or rod-shaped spicules of the body itself, and even in the axial part
of the remarkable " Amphidiscs," a fine central canal extends, which is usually intersected
at the middle by one or two transverse canals which cross it at right angles.

While Ehrenberg,' in opposition to this opinion of Max Schultze, still maintained that
we had to deal with an artificial Japanese production, von Martens examined the debated
organism in Japan, and in the same publication (p. 480) essentially confirmed Max
Schultze's opinion.

A theory of the nature of Hyalonema, similar to that expressed by Max Schultze in
his first communication in the Comptes rendus, was expressed some years later by
Bowerbank in one of his papers on the Anatomy and Physiology of Sponges.^
Bowerbank united Hycdonema with Halichondria, Isodictya and Spo7igilla, in the
suborder of his Silicea. with a " spiculo-reticulate skeleton," — the skeleton being " con-
tinuously reticulate in structure but not fibrous." The genus Hyalonema, Gray, was
here characterised by Bowerbank in the following manner : — " Skeleton an indefinite
network of siliceous spicula, composed of separated elongated fasciculi reposing on
continuous membranes, having the middle of the sponge perforated vertically by an
extended spiral fasciculus of single, elongated, aad very large spicula, forming the axial
skeleton of a columnar cloacal system."

Siiss in 1862^ called attention to a fossil from the Carboniferous limestone of
Yorkshire, which had been already described by M'Coy as Serpula parallela, which
exhibited a bundle of from fifteen to thirty or more round, smooth, parallel rods of the
thickness of a knitting needle, and each provided with a central canal. This he named
Hyalonema parallelwn.

In 1864 Barboza du Bocage made a communication^ on a new species of the genus
Hyalonema, which was discovered off the coast of Portugal at great depths. He named it
Hyalonema lusitanicum. Bocage regarded the form and peculiarly regular arrangement
of the polypes, which partly surrounded the siliceous spicular tuft and were provided
with forty tentacles, as especially characteristic of his new species. His diagnosis runs
thus : — " Hyalonema polypario elongato fibris setaceis, hyalinis, spiraliter tortis, corio
polypigero ab apice usque ad | longitudinis totse involutis polypis dilatatis, ellipticis valde
aggregatis, parum elevatis, per series longitudinales ac spirales regulariter digestis."

1 Monalsher. d. h. preuss. Akad. d. Wiss. Berlin, 1861, p. 450. ^ pj^n Tram., 1862, p. 1113.

3 Verhandl. d. h k. zool.-bot. Gesellsch. Wien, 1862, p. 85. * Proc. Zool. Soc. Lond., 1864, p. 266.

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

On the sponge body Bocage made no observations ; be simply notes that the
covering of polypes extends without interruption to the pointed extremity of the fibrous
cord, where they become somewhat diminished in size. At a later period he ^ supplied
some further data as to the locality of his Hyalonema lusitanicum — near the mouth of
the Tagus — and also confirmed his previous report from the examination of some new
material.

The account given by Barboza du Bocage seemed to Gray to be well calculated to
support the opinion which he still firmly maintained, that the long fibrous bundle
was the skeletal axis of the polype covering. He utilised it in a renewed defence of his
views,^ admitting, however, that he had been so far mistaken in his examination of the
individual polypes, since these possessed not eight feathered, but twenty simple tentacles.
Soon after this Bowerbank elaborated ' his theory in regard to the nature of
Hyalonema. He now maintained that not only the long tuft of glass-spicules and
the cylindrical or conical basal mass, but also the uneven irregular rind surrounding
the fibrous tuft, were all parts of the sponge.

A memoir by Max Schultze * is noteworthy as containing the proposal to unite the
two genera Euplectella, Owen, and Hyalonema, Gray, into a special group of siliceous
sponges with the title " Lophospongise."

An accurate description of the sponge named by him Hyalonema mirabile was now
given by Bowerbank.' The cylindrical elevations on the rind of the siliceous tuft, which
had been described by most observers as polypes, were recognised as the oscula of
the sponge. The form described by Barboza du Bocage Bowerbank * looked upon as not
specifically distinct from his Hyalonema mirabile.

In his comprehensive sponge system published in 1867,' Gray designated the large
cylindrical body, which he regarded as the whole sponge, in which the long siliceous
filaments were embedded, by a new generic name " Carteria," assigned to this genus a
place in his family of the " Esperiadse," in the order of the Acanthospongia, and gave to
it the following diagnosis : — •" Sponge massive, irregularly reticulated, shallow, formed of
abundant agglutinated filiform needle-like spicules, with four- and six-rayed stellate,
cruciform and birotulate spicules." As the only known genus he mentioned Carteria
japonica { = Hyalonema mirabile, Bowerbank, Spongia octancyra, Brandt, and Spongia
spinicrux, Brandt).

In two communications published in 1867,- Ehrenberg maintained his opinion
that the entire body was artificially compounded by the Japanese, but communicated
in the second paper a letter from Barboza du Bocage, in which the latter declared

' Proc. Zool. Soc. Land., p. 662, 1865. 2 ji^n. and Mag. Nat. Hist., ser. 3, vol. xviii. p. 287, 1866.

3 Ann. and Mag. Nat. Hist., ser. 3, vol. xviii. p. 397.

* Archivf. raikrosk Anat., 1867, p. 206; Ann. and Mag. Nat. Hist., ser. 3, vol. xix. p. 153.

5 Proc. Zool. Soc. Land., p. 18, 1867. " Proc. Zool. Soc. Loud., p. 902, 1867.

' Proc. Zool. Soc. Lend., p. 492. ^ Monatsber. d. k. jimiss. Akad. d. PFiss. Berlin, pp. 298, 843.

REPORT ON THE HEXACTINELLIDA. 185

that he had been convinced of the correctness of Max Schultze's opinion by finding
two large specimens of Ilyalonema with sponge bodies but without the polype crust.

The sponges which Semper described in 1867^ and in 1868," under the designa-
tion of Hyalonema schultzei, I shall discuss when considering the genus Semperella.

In 186S Gray made several communications on Hyalmiema ; ^ in the first of
these he added to his Carteria jaiwnica a second species, Carteria lusitanica
{ = Hyalonema lusitanicum, Bocage). In the second he combated Max Schultze's con-
ception of Hyalonema, and in the third he expressed the opinion that the sponge
described a short time previously by Loven as Hyalonema boreale * in no way belonged
to the genus Hyalonema, Gray (or Carteria, Gray), but was from the form of its
simple spindle-shaped spicules to be referred to the Halichondridse, in close relation to
Halichondria Jicus, Johnston {Ficidina, Gray). In his later memoir published in
English,'* Loven described a North Sea sponge, with a club-shaped body and a slender
stalk fixed in the sand by means of root-like processes, and bearing a certain external
resemblance to Hyalonema sieholdii. Loven conjectured that the Japanese and Portugese
HyalonenKita were, like his Hyalonema horeale, rooted in mud by means of their long
tuft-like spreading spicules, and that accordingly the broad cylindrical sponge body of
these forms did not represent the base, but was fixed as an upward projecting mass upon
the root-tuft.

In the meantime Barboza du Bocage had also discovered, off the coast of Portugal, some
sponges which resembled the Hyalonema horeale. In his report on these forms," which
he had been at first inclined to regard as young specimens of Hyalonema lusitanicum,
he expresses his conviction that these, along with the Hyalonema horeale of Loven,
should be placed in a special new genus " Lovenia."

Gray now announced^ that the Japanese collectors, who sometimes found the
diverging extremity of the glassy spicular tuft covered with mud, affirmed that this free
tufted extremity was embedded in the mud or sand of the sea-bottom. He was
himself inclined to regard this view as correct, and figures * a Hyalonema placed
in this position. It is noteworthy that Gray now returned to the name Hyalonema
sieholdii, which he first applied in 1835 to the supposed polypes, and thus abandoned
the designation Hyalonema mirabile, which had been used for a longer period.
For the Hyalonema lusitanicum described by Barboza du Bocage, he adheres to a
generic name which had already been used in 1867, viz., Hyalothrix, and he names
the specimen in question Hyalothrix lusitanica. Both in the case of Hyalonema
sieholdii and of Hyalothrix lusitanica, he admits varieties with and without the adhering
sponges.

' Archivf. Natiirgesch., Bd. xxxix. p. 84. 2 Verhandl. d. phys.-med. Gesellsch. TViirxhurg, 1868.

3 Ann. and Mag. Nat. Hist., ser. 4, vol. i. p. 161, &c. ■" Ofversigt k. Vetensk.-Akad. Forhandl, 1868, p. 105.

» Ann. and Mag. Nut. Hist., ser. 4, vol. ii. p. 81, 1868. ' Ann. and Mag. Nat. Hist., ser. 4, vol. ii. p. 36, 1868.

7 Ann. and Mag. Nat. Hist., ser. 4, vol. ii. p. 271, 1868. ^ Loc. cit., p. 275.

(ZOOL. CHALL. EXP. — PART LIII. 1887.) Ggg 2-1

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

Percival Wright,' who dredged Ilyalonema lusitanicum from a depth of 480 fathoms
south-west of Setubal, and had observed it in the fresh condition, found it emljedded
in mud up to that part of the siliceous spicular tuft which was not covered by Palythoa,
while the upper extremity of the tuft bore a sponge body with several oscular openings.
In the polypes forming the crust of the stalk movements of the tentacles were observed.

Hyalonemas were dredged in a similar situation in the Gulf Stream by Wyville
Thomson from a depth of 550 fathoms.

In his memoir on Holtenia carijenteri,- Wyville Thomson refers to a Hyalonema
loveni as a new species, without however describing or characterising it.

Percival Wright^ now described in an elaborate paper the Hyalonema found off the
coast of Portugal by Barboza du Socage, Wyville Tliomson and himself, and named it
Hyalonema mirabile. On a specimen with an oval cup-shaped body, about 8 inches in
length and 4 inches in breadth, " a number of irregular large openings (oscula) " could
be seen in an upper pitting, and these " are covered over with a delicate open sarcode
network, the edges of the meshes of which are thickly lined by the spicules called
'spiculate cruciform spicules' by Bowerbank." A figure of this, given on his pi. iii.,
shows the sieve-net on the upper truncated surface of the sponge body, extending
evenly over the oscular opening and over the layer of the " spiculate cruciform spicules "
in the net beams.

The two sponges described by Sars in 1872, in his work "On some remarkable
forms of animal life from the great deeps ofi" the Norwegian coast," and named as
Hyalonema longissimum and Hyalonema parvum, manifestly do not belong to the
Hexactinellida any more than his Hyalonema horeale.

In the work entitled Depths of the Sea, Wyville Thomson reported in 1873 on
some Hyalonemata which he had dredged to the north of the Butt of Lewis from a depth
of 450 to 500 fathoms. In these the root-tuft alone measured 40 cm. or more. He
identified these forms as Hyalonema lusitanicum, Barboza du Bocage, and figured {loc.
cit, p. 421) a slender specimen about 34 cm. in length. It is noteworthy that this figure
does not include the sieve-net which was observed l^y Percival Wright in his Portuguese
specimen (described under the title Hyalonema mirabile), covering the oscular openings
of the upper truncated surface. Instead of this a central conical boss projects about
2 cm. above the upper bounding surface, which is provided with large openings.

Higgin next described, in 1875,* a Hyalonema from Zebu (one of the Philipijine
Islands) as Hyalonema cehuense. This form resembles, indeed, in many respects the
Japanese Hyalonema sieboldii, but according to Higgin's description and figures differs
from that species chiefly in the following points : — the form of the body resembles that
of a hewer's mallet ; the polype crust of the stalk is absent ; the " large stout acerate

1 Arm. and Mag. Nat. Hist., ser. 4, vol. ii. p. 323. ^ Phil. Trans., p. 701, 1869.

3 Quart. Joum. Micr. Sci, p. 73, 1870. * Ann. and Mag. Nat. Hist., ser. 4, vol. xv. p. 377.

REPORT ON THE HEXACTINELLIDA. 187

spicules " occur not only in the interior of the body but also in the superficial dermal
layer ; the rays of the small hexradiate spicules in the interior are not herd but straight,
and are, up to the pointed extremity, beset with lateral prongs.^

Higgin succeeded in discovering in his Hyalonema cebvjense the true position of the
dermal spicules, and Klistermann^ observed in the same year, in Hyalonema sieboldii,
the presence of four cruciately disposed anchor-teeth on the extremity of the long tuft
spicules. Kiistermann also observed, close to the extremity of a tuft-spicule, that the fine
axial canal was crossed by two short transverse canaliculi disposed at right angles to one
another.

An elaborate description of the minute structural relations of Hyalonema sieboldii
has been given by MarshalP in his admirable Researches on tlie Hexactinellidse. He
calls attention to the peculiar sieve-like perforated plate which covers the upper truncated
extremity of the sponge-body. In the large hollow cavity of this species he also
succeeded in discovering round embryos of the size of a millet grain or a pea, and
exhibiting a central cavity with a small round excurrent opening. The wall consisted for
the most part of smooth hexradiate spicules. On the outer surface spindle-shaped
spicules occurred ; and long pointed rod-like forms penetrated the wall here and there in
a radial direction, and projected for a considerable distance beyond the surface. One of
the specimens of the Leyden collection, described by Max Schultze, appeared to
Marshall, on account of its diff"erent dimensions, and especially on account of the peculiar
condition of the dermal skeleton, which was said to exhibit anastomosing bands of long
elastic uniaxial spicules with large cross spicules at the anastomoses, to justify the
institution of a new species, which he named Hyalonema affine. Important deviations
from the Hyalonema sieboldii type were exhibited by a small (only 7 cm. in length)
Hyalonema which Wyville Thomson collected to the north of the Shetland Islands, from
a depth of 550 fathoms, and which has been named Hyalonema thomsoni. Instead of
the sieve-net, which extends across the upper truncated extremity of Hyalonema
sieboldii, a central slender cone, 1 cm. in length, projects freely, and from its expanded
base four cruciately disposed narrow ridges extend, forming the ujjper border of four
septa which traverse the cavity of the sponge in a radial direction. Another peculiarity
of the Hyalonema thomsoni consists, according to Marshall, in the thumb-like prongs
which are spirally disposed on many of the tuft-spicules, and which at theii- union
with the ridge-like projecting basal portion of the spicule exhibit a canaliculated
appearance.

AVyville Thomson* then briefly noted and figured some particularly striking forms of
Hyalonema from the rich collection of the Challenger Expedition. One species, obtained
in the vicinity of St. Thomas, from a depth of 390 fathoms, and named by Wyville

1 Loc. cit., pi. xxi. figs. 4, 5. » Archivf. mikrosk. Anat., Bd. xi. p. 268.

3 Zeitschr.f. wiss. ZooL, 1875, Suppl.-Bd. xxv. p. 142. * The Atlantic, 1877, vol. i. p. 273.

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

Thomson IlyaJonema toxeres, is distinguished by its remarkably strong, bow-shaped,
spindle-like spicules.

In the same year, 1877, M. and J. Young^ reported on fossil remains from the
carboniferous strata of Cunningham Baidland, near Dairy in Ayrshire, which included not
only the liyalonema parallelum, M'Coy, already described by Siiss, but a second species,
Hyalonema smithii, Young and Young. Besides the long, straight, smooth, knitting-
needle-like spicules which Suss refers to in Hyalonema parallelum, the Youngs found
also short five-rayed and six-rayed spicules.

Among the sponges which A. B. Meyer brought home from the Philippines and
New Guinea, and handed over to Bowerbank for description, there was a cup-shaped
specimen which Bowerbank^ briefly described in 1877, and in spite of the entirely
distinct form of the spicules identified as a Hyalonema from which the basal tuft had been
torn off. This he named Hyalonema anomalum,. Dr. A. B. Meyer had the kindness
to hand over to me at my request a portion of this same specimen belonging to the
Dresden Zoological Museum, and I have been able to convince myself that we have here
to deal not with a Hyalonema, but with a badly-preserved specimen of Crateromorpha
meyeri ; and with this the description given by Bowerbank himself agrees.

In the abundant sponge material which was collected during the deep-sea expedition
by Agassiz and Pourtales in the Gulf of Mexico, and entrusted to Oscar Schmidt for
examination, this renowned spongiologist found only two specimens of Hyalonema}
The larger specimen consisted only of a torn off s^jicular tuft partly covered by a
crust of Palythoa. The other, which was distinctly smaller, was however completely
preserved, and bore a sieve-net over its upper extremity. The latter was identified by
0. Schmidt as Hyalonema sieholdii.

In the Bay of Biscay, during the French expedition of the " Travailleur," a specimen
of Hyalonema lusitanicum was dredged, accoi'ding to Norman's* account, from a depth
of 600 fathoms.

In the report by G. Armauer Hansen on the Sponges of the Norske Nordhavs
Expedition, a badly-preserved Hexactinellid is described and figured (1885). It was
obtained from a depth of 1081 fathoms, lat. 63° 17' N., long. 1° 27' W., on Biloculina
clay. Armauer fianscn had for examination {loc. cit., 'p. 19) " several specimens with
round hollow stems of somewhat variable thickness, measuring up to 3 cm. in length, and
surrounded at the one extremity by a very loose almost cotton-like substance." Since
I observe a discohexact among the siliceous spicules figured {loc. cit., pi. v. fig. 10),
and since " the stem is composed exclusively of spicules which are truncato-spinose at
both extremities," I can in no way agree wath Armauer Hansen's opinion that this

' Ann. and Mag. Nat. Hist, ser. 4, vol. rs. p. 425. ^ Proc. Zool. Soc. Land., p. 461.

2 O. Schmidt, Spongien des Meerbusens von Mexico, 1879 and 1880, p. 64.
* Ann. and Mag. Nat. Hist., ser. 5, vol. vi. p. 436.

REPORT ON THE HEXACTINELLIDA. 189

Hexactinellid, which he names Hyalonema arcticum, belongs to the genus Hyalonema.
No Hyalonema possesses a " round hollow stem," with rough, rod-like spicules, rounded
off at both ends ; and in no Hyalonema are there rough discohexacts, such as are
figured {loc. cit., fig. 10a). On the other hand, I conjecture from the description and
from the figures of Armauer Hansen that we have to deal with a species of my genus
Caidophacus, and certainly at least with an Asconematid.

Character' of the Genus. — The body is generally in the form of a thick-walled cup,
from the shallow cavity of which a central conical boss projects. From the upper
sharply defined border, which is distinguished by a fringe of marginal spicules, an
independent sieve-net-like perforated membrane extends in many species over the
whole oscular opening. From the centre of the lower, usually somewhat narrowed end
of the body, a tuft of long and strongly developed basal spicules projects downwards.
The tuft consists wholly or for the most part of four-toothed anchors, which serve to
moor the sponge in the loose mud. The bundle of spicules is continued superiorly into
the cone which rises in the gastral cavity. In most species, and perhaps even in all, the
upper portion of this basal tuft is surrounded by an encrustation of Palythoa, which
begins just below the end of the body proper, extending in difierent individuals to a
very varied distance, but never growing on the lower somewhat bushy divergent portion
of the tuft.

Both dermal and gastral skeleton are composed (l) of strong pentacts, with the
unpaired ray sunk more or less deeply into the parenchyma, (2) of pinuli which have
their tangential basal cross inserted in the .skin, and (.3) of amphidiscs which are
radially disposed with the median axial cross portion embedded in the skin, and with
one end projecting freely outwards, while the other extends into the parenchyma, or into
the subdermal or subgastral spaces as the case may be. Amphidiscs more rarely occur
tangentially in the skin, but are then represented only by minute forms. In some species
the gastral skeleton is continued without much change into the eff"erent ducts, i.e., into the
canalicular skeleton. The parenchyma contains large and small oxyhexacts, cruciate
tetracts, triacts, straight diacts, and occasionally isolated monacts.

The chambers are not so distinctly marked ofi" as glove-finger or thimble-shaped
sacs as in most of the other Hexactinellids. They appear rather as more irregular
and less sharply defined diverticula of the membrana reticularis.

I shall divide the genus Hyalonema, Gray, into the two subgenera Hyalonema, sens,
strict., and Stylocalyx — a distinction based on the presence or absence of an oscular
sieve-plate.

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

Subgenus 1. Htjalonema, sens, strict. (Pis. XXVII. -XXIX.; PI. XXX. figs. 9-17;

PI. XXXIII. ).

The superior aperture of the gastral cavity is covered by a sieve-plate, which extends
from the side (with its annular, cufi-like fringe of fine marginalia) either right across
the mouth of the cup, or sunk down into the latter, towards the inner surface, in funnel-
shaped fashion.

1. Hyalonema sieholdii, Gray (PL XXVII. figs. 1-13).

Among the numerous species of Hyalonema obtained by the Challenger Expedition,
the familiar Japanese form, Hyalonema sieholdii, Gray, one of the best and first known
Hexactinellids. was unfortunately not included. A collection of Japanese Hexactinellids,
however, entrusted to me by Dr. Doderlein, contained a tolerably well-preserved dried
specimen of this sponge, which is figured after a photograph (one-third natural size, not
one-half as stated on the plate) on PL XXVII. fig. 1. I was able further to compare
this with several di'ied and preserved specimens in the Royal Museum, Berlin, which
were brought from Japan by Dr. Hilgendorf.

The total length of the specimens examined varies from 50 to 80 cm. The body
proper measures 6 to 15 cm. in thickness, and occupies 10 to 15 cm. of the above total
length, the remainder representing the length of the spirally twisted, basal tuft. The
upper portion of the latter where it joins the body, and where, for a distance of 10 to 20
cm., it is beset by Palythoa fatua, M. Schultze, only attains a diameter of 5 to 10 mm.
Gradually, however, the tuft becomes looser, and breaks inferiorly into a bushy bundle.
Since the form and external appearance of the whole sponge have been already described
and figured in detail by Max Schultze and others, I shall simply restrict myself to
emphasizing the important points, and devote more attention to certain structural
relations which have been hitherto less intimately investigated. The general form of the
body is cylindrical, transversely truncated at the upper end, and inferiorly narrowed in a
somewhat conical manner, rounded off, and finally ending in a small inconspicuous
marginal pad. The superior external margin bears a fine fringe, composed of slender
needles projecting in wreath-like arrangement (marginalia). The upper truncated surface
is formed by a sieve-plate, which is stretched right across, or protrudes with a slight
convexity. The component lattice-work of the sieve-plate is not, however, uniformly
developed, but is usually divided into four distinct portions by four zones of unperforated
membrane, which unite in the centre to form an irregular cross. The dividing zones
measure from 3 to 10 mm. in breadth, and the four irregularly triangular or roundish
perforated sieve portions vary in width from 2 to 4 cm. Each of the four perforated
sieve-regions corresponds to an equal sized, deep portion of the gastral cavity, while the

REPORT ON THE HEXACTINELLIDA. 191

imperforate limiting zones lie just above the gastral septa, and arc indeed partially-
united with them. This union of the septa of the gastral cavity with the sieve-plate
usually occurs only in the centre and at the outer ends, so that one can see
the septa, in their middle portion, ending with a free convex margin just below
the plate. In some cases, however, each septum unites along its whole extent
with the corresponding zone of the sieve-plate. Commensal Anthozoa occur in
extremely variable number all over the lateral surface, and even on the cruciate
imperforate areas of the sieve-plate. They protrude from circular apertures, which are
2 to .3 mm. in width and possess a firm peripheral layer. In some specimens they^
cover the whole lateral surface so thickly that the distance between them is not more
than from 3 to 10 mm., while they also occur aljundantly (PI. XXVII. fig. 2) on the
dividing zones of the sieve-plate. In other specimens they occur singly only here and
there, though they never seem to be wholly absent.

Apart from these Anthozoa tubes, the surface of the s^ionge is comparatively smooth.
The 23ores of the dermal membrane covering the surface are for the most part micro-
scopically small. The extreme inferior end of the body, adjacent to the Palythoa
encrustating the basal tuft, consists of an inconspicuous but compact circular cushion,
varying greatly in breadth in the difi"erent specimens. In dried forms, the fine reticulate
pattern which is distinctly seen over the whole lateral surface of the body is in this
region no longer recognisable.

When the terminal sieve-plate is removed, the gastral cavity is revealed, and is
seen to be di\'ided into four wide spaces by the four broad, cruciately disposed, radial
septa, with a central conical elevation. The free upper margins of these septa, if not
fused with the sieve-plate, are rounded off and convex, and extend from the body
margin on the outside to the conical elevation in the centre. The cavities lying
between the septa become gradually narrower downwards, and receive from the
sides and from below the wide lacunar efferent canals of the body-wall. In a longi-
tudinal section near the central axis, a continuation of the basal tuft is seen as a some-
what markedly twisted strand of spicules, prolonged up the central column and gradually
narrowing towards the conical elevation, as was indeed long ago observed and figured by
Max Schultze. Since the efferent canals are very wide and often lacunar, and since the
subdermal spaces with the afferent canals penetrating inwards from them have a similar
character, it evidently follows that the layer betw^een the two systems of canals cannot
be by any means broad.

The skeletal elements of Hijalonema sieholdii which always remain completely
isolated were thoroughly studied in 1860 by Max Schultze, and so excellently described
in his monograph, that I must here simply content myself wdth referring to that admirable
memoir, and with a brief review of the different forms of spicules and of the manner of
their disposition.

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

The parencliymal spicules supporting the soft body arc represented by a large number
of simple spindle-shaped diacts, varying in length up to 3 mm. Both their ends are
simply pointed or rounded off, and not unfrequently provided with small spines. While
most of these spicules appear to be smooth centrally, some exhibit there an annular
swelling, and others two opposite, or four cruciate tubercle-like elevations, into which two
or four cross branches of the axial canal are seen to penetrate. Even in the diacts
which are smooth centrally sometimes similar cross branches from the axial canal can be
detected.'

The spindle-shaped spicules are, for the most part, not perfectly straight but slightly
bent, lying in strands or somewhat irregularly scattered. Slender diacts are also
occasionally to be found beset towards both ends wdth inwardly directed hooks
(PL XXVIl. fig. 3). Between the spindle-spicules there is a somewhat sparse
occurrence of proper oxyhexacts and derivative spicules, the latter with five to
three rays, or even with two opposed at right angles. Larger smooth hexacts very rarely
occur. Somewhat more abundantly, but yet rarely, slender hexacts are found with
distally directed teeth, as represented in PL XXVII. fig. 13. Similar hexacts with
curved rays (PI. XXVII. fig. 10), as figured by Max Schultze {loc. cit., Taf. iv. fig. 4), I
have only very rarely seen — so rarely, indeed, that I doubt whether they have not
found their way in from some other species of Hyalonema, and are not really foreign
to Hyalonema sieboldii. I am also doubtful whether the peculiar amphidiscs,
which were found so abundantly in the limiting membranes, are also proper to the
parenchyma.

The dermal skeleton is mainly composed of strongly developed pentact hypogastralia,
which form by their mutually apposed tangential rays a comparatively wide-meshed
rectangular lattice-work, while the strands of the finer network of the skin are supported
by tangentially disposed diacts. The narrowed ends of the somewhat blunt rays are
frequently to some extent covered with tubercles, or are at least rough. They usually
exhibit the same character on the same pentact, but in diff'erent spicules vary so far at
least, that some are pointed and others quite blunt, some relatively smoother and
others more or less markedly beset with terminal protuberances. The distal (sixth)
ray has so completely disappeared that only the merest hint persists in the form of a
slight prominence.

On the hypodermalia and on the dermal strands of diacts extended between
them there are seen countless autodermal pinuli, which are here exclusively
pentacts. The four basal rays intersecting at right angles lie wholly in the dermal
membrane ; the somewhat long distal ray, which is drawn out into a long fine point, is
always at right angles to the surface of the skin, and thus projects freely into the water
perpendicular to the body-surface. There is no proximal sixth ray, or its presence is

' M. Schultze, Die Hyalonemeii, Taf. iii. and iv.

REPORT ON THE HEXACTINELLIDA. 193

only suggested. The comparatively strong and straight basal rays, which are sparsely
covered with small externally directed teeth, are only about 0"05 mm. in lengtli. At
first gradually narrowing, they end in a short conical point. The distal ray, which has
an average length of 0'3 mm., is in some regions decidedly shorter, while it may, on the
other hand, attain double the above length. It is as a rule perfectly straight, strongly
developed at the base, and very gradually diminishing in diameter up to the fine narrow
point. It is covered with strong, distally directed spines or teeth, which stand out
somewhat markedly towards the base, and become shorter and more closely apposed
towards the extremity (PI. XXVII. fig. 12).

In some positions on the lateral surface of the body the pinuli are specially long and
somewhat bent (PI. XXVII. fig. 11). The dermal skeleton of the superior terminal
sieve-plate does not differ in essential characters from that of the lateral wall. Instead
of the large pentact hypodermalia, however, strands of diacts of various sizes pre-
dominate.

The marginalia forming the marginal fringe of the whole sieve-plate are straight or
but slightly curved oxydiacts, 1 to 1'5 mm. in length. Four distinct, cruciately disposed,
broad and rounded protuberances from the point of intersection represent the tangential
rays. The freely projecting distal portion bears externally directed spines, and
narrows very gradually to a fine point. The proximal portion, which is inserted in
the parenchyma, is either smooth or but sj^arsely beset with small pointed elevations,
which stand out transversely or are somewhat turned towards the apex, i.e.,
internally. The proximal portion is always thicker and less pointed than the distal
external portion, and its length is to that of the latter generally in the proportion
of 1 : 2.

Though it cannot be doubted that these marginalia belong to the rank of the
dermalia, and are most nearly related to the autodermal pinuli, no distinct transitional
forms are to be observed ; they can indeed be distiuguished with equal sharpness from
the pinuli of the sieve -plate and from those of the sides of the body.

Special attention must be directed to those remarkable elements of the dermal
skeleton which were designated birotulate spicules by Bowerbank, and amphidiscs by
Max Schultze and others. They consist of a straight main stem, both ends of which
bear a similarly composed campanulate umbel formed of a varying number of radiate
processes, varying in length, and either of a leaf-like form, or narrow like the ribs of
an umbrella. This form of spicule is t}^ical and characteristic of the whole family of
the Hyalonematidse. Manifold variations occur in shape and size, and these are partially
characteristic of the different genera and species, and thus useful for diagnostic purposes.
That we have here really to deal with diacts is evident, in spite of the absence of any
intersection of the well- developed axial canal by one or two small transverse canals, from
the fact that amphidiscs are not unfrequently found, in which four, cruciately disposed,

(ZOOL. CHALL. EXP. — PART Lm. — 1887.) Ggg 25

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

conical or rovindcd elevations project at the proper position. Here and there structures
may be observed in wbich, instead of such rudiments of the four abortive rays, the
latter, or at least two of them are j^resent as well equijjped with many-ribbed umbels as
are the two rays usually alone developed. Carter has previously reported the occurrence
of this abnormality in Hyalonema sieboldii} I am inclined to compare the smooth
processes forming the bell-shaped umbel at each end of the stem with the terminal rays
of other sjaicules, and to regard them therefore as simple tooth or spine structures. I
have never observed any trace of an axial canal within them, and find, with Marshall and
others, that they are not formed in the first rudiment of the amphidiscs but develop
subsequently. In Hyalonema sieboldii only two different kinds of amphidisc occur in the
skin. On the one hand, there are large strongly developed amphidiscs, 0"42 mm. or more
in length, with an umbel 0'13 mm. or more in breadth, the axial beam of which is 0'035
mm. in thickness, and more or less rough, with irregularly scattered, inconspicuously arched
prominences. At each end there are about eight slightly arched umbel-rays, each of
which, with a longitudinally directed basal piece about 0'07 mm. long and O'OIS mm.
broad, springs from the side of the terminal portion of the axial beam, and passes
externally into the ploughshare-like lateral plates, which are bent towards the surface
of the umbel, and therefore at right angles to the basal plate (PI. XXVII. fig. 7). These
curved lateral plates do not usually grow together laterally, but form each a free sharp
lateral border. They are continued at their free end into a lancet-shaped terminal point,
while at the other extremity, near the main axis of the spicule, they end in a rounded
prominence, and thus enclose between them a central terminal pit. The free margin
of the basal plate exhibits a rounded edge which becomes gradually raised from the
axial beam of the amphidisc, and is continued with a hook-like bend to the end of the
lancet-shaped, curved, external lateral plate (PI. XXVII. fig. 7).

The disposition of these large amphidiscs in the skin has not as yet been correctly
represented by investigators. They do not lie, as Marshall, 0. Schmidt, Bowerbank,
Carter and others have described, tangentially in the dermal membrane, but are radially
directed, so that one-half of the whole spicule projects freely from the external surface,
and the other extends internally into the subdermal space. I have found them
irregularly scattered, sometimes in comparative abundance, sometimes only sparsely,
l>ut could not distinguish any one region of the body as characterised by their
special frequency. They occur also on the external surface of the terminal sieve-
plate.

Besides these large amphidiscs with broad shovel-like umbel rays, numerous smaller
forms, measuring only O'Ol to 0"016 mm. in length, occur, irregularly scattered in the
derma] membrane, in no special direction, but for the most part tangentially disposed.
Their umbels are terminally hemispherical! y curv'^ed, and consist of eight umbel-radii

' Aim. and Mag. Nat. Hist., ser. 4, vol. sii. p. 372, anc\ pi. xiii. fig. 22.

REPORT ON THE HEXACTINELLIDA. 195

running out into straight terminal rays. The former vary greatly in length in the
diflferent amphidiscs, often hardly exceeding the length of the hemisphere, sometimes
almost touching the opposite raj^s (PI. XXVII. figs. 5, 6). The axial beam is sometimes
comparatively smooth, sometimes knotted, frequently exhibiting at the middle point a
cross or a ring of conspicuous knots. I have never found in the skin transitional stages
between the two above described very different forms of large amphidiscs, but such inter-
mediate structures are observed to occur in the parenchyma.

The gastral skeleton differs in many points from the dermal. The skin lining the
principal gastral and the four large basal diverticula is not supported by pentacts, but
by strands of large smooth diacts and monacts, which usually become gradually
narrower towards the extremity, terminating in a sharp apex or conical point, or
more rarely rounded off. The large amphidiscs, with broad, shovel-like, umbel-rays,
which occurred so abundantly in the external skin, are here altogether absent ; and their
place is occupied by irregularly scattered, medium-sized forms with eight narrow,
comparatively long rays, whose terminal portions are more nearly approximated to the
axial beam (PL XXVII. figs. 8, 9). The length of these rays is about 0'3 mm., the
breadth of the umbel O'l mm. The axial beam is comparatively narrow, and exhibits a
few irregularly disposed tubercles, and usually, at the middle point, four cruciate
strongly developed protrusions, which are probably to be regarded as traces of the
abortive rays (PI. XXVII. figs. 8, 9). The rays of the umbel resemble in general form
those of the large amphidiscs in the external skin ; they are, however, longer, measuring
usually about a third of the whole length of the amphidisc. Originating in a short
basal piece, they extend for the greater part of their length approximatel}' parallel to the
axis, or even with a slight external convexity. A comparatively large number of small
amphidiscs occur like those of the dermal membrane figured on PI. XXVII. figs. 5, 6,
as also some rather larger isolated forms of similar structure, represented in
fig. 4.

As in the external skin, pentact pinuli occur in abundance in the gastral membrane
and in its direct continuation into the efferent passages and canals, where the lining
membrane is covered with them. Compared with the dermal pinules, however, they
are somewhat different in shape and less abundantly present. The basal cross lying in
the limiting membrane consists of four weakly developed, and generally straight, rarely
somewhat curved rays, almost twice as long as the basal rays of the dermal pinules,
and beset with distally directed teeth. The ray which projects freely into the lumen
and varies greatly in length, is likewise but weakly developed, and usually bears
obliquely divergent, distally directed teeth.

The spicules which compose the basal coUar-pad of the sponge body demand special
notice. Besides the forms which have just been described in detail, structures occur
in the parenchyma and in the dermal membrane of this region, which are not found else-

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

where throughout the whole sponge. They have the form of thick hexacts, pentacts,
tetracts, triacts or cliacts, whose rays, uniformly thick, strong and cylindrical— exhibit
blunt or rounded ends, and are more or less abundantly covered with short, strong,
conical sjjines. These teeth or spines are either uniformly scattered over the whole
surface, or occur only on the distal ends of the several rays, the median and proximal
portions remaining smooth. As these peculiar spicules of the basal pad have been figured
in the case of Hyalonema sieboldii in Max Schultze's classic work on Hyalonematidse,
I have not devoted to them any special illustration.

Of all the skeletal parts of Hyalonema sieboldii, the long spicules of the basal tuft
have been longest and most intimately studied. These basalia form a long loosely
wound strand, which measures .30 to GO cm. or more in length, and breaks up towards
the foot in a brush-like fashion. To the detailed descriptions of former investigators J
have but little new to add, and therefore refer to their accounts, especially to that of
Max Schultze.' He found in one tuft 200 to 300 separate spicules, of which the shorter
lay nearer the axis. The shape of the inferior extremity was certainly determined only
in the shorter forms. He observed a simple pointed termination, similar to that of the
superior extremity hidden within the sponge. Besides perfectly smooth forms, numerous
spicules occur with a tooth-bearing spiral ridge. TJie free margin of the latter projects
obliquely outwards and upwards, so that a similar direction is given to the leaf or nail-
like, pointed or slightly rounded spines which it bears. Interruptions of the spirals
sometimes occur, and these are often so regularly disposed, that the remaining portions
of the projecting ridges are alternately opposed to one another.

On some very well-preserved (spirit) specimens collected by Dr. Hilgendorf in
Sagami Bay, Japan, I was able to study the very inferior extremity of many of the long
spicules of the tuft. It appeared that the spinose main portion of the spicule is followed
by a short, narrow, smooth neck, which bears terminally a solid, cap-like or hemispherical
thickening whose upper rounded margin exhibits eight recurrent teeth, forming an anchor-
like fiorure.

2. Ilyaloiu^ma gi-acile, n. sp. (PL XXVII. figs. 14-23).

This species of Hyalonema was dredged in the neighbourhood of the Philippine
Island Mindanao (lat. 8° 0' N., long. 121° 42' S., Station 211), from a depth of 2225
fathoms, and on a blue mud ground. Its body, which is approximately pear-shaj)ed,
measures 2 '5 cm. in length by 1'6 cm. in breadth. The upper end, which is somewhat
narrowed and transversely truncated, bears a delicate sieve-plate, while the lower,
conically pointed end runs out into a basal tuft, 1'5 mm. broad and about 3 cm. long.
The component spicules are not numerous, nor twisted, and they diverge but slightly

' Diu Hvfiloiiemeii, Taf. ii. tigs. 1-7.

KEPORT ON THE HEXACTINELLIDA. 197

inferiorly. Owing to the insufficient preservation of the single specimen, it -was
imiJossible to determine wlietlier a Pali/thoa-crnst enveloped the basal tuft, below the
inferior extremity of the body. Below the superior terminal sieve-plate, there is a flat
hollow space, from which four cruciately arranged wide passages, furnished with lateral
and terminal diverticula and canalicular prolongations extend into the parenchyma.
Nearly up to the sieve-plate, the centre is occupied by a columella, ending freely
in a conical prominence. From this central pillar the four cruciately arranged septal
plates radiate outwards, separating the four gastral spaces from one another (PI. XXYII.
fig. 14).

Of the external skin, as also of a delicate narrow cuflf-like fi'inge which surrounds the
sieve-plate and separates it from the skin, only a few pieces are preserved.

The spicules supporting the parenchyma consist of simple, flat oxyhexacts of
medium size, which are usually radially disposed at right angles to the surface, and
distributed with general uniformity over the whole body. The six rays are all of equal
length, and are very gradually narrowed towards their somewhat conically pointed
extremities. Besides these, numerous simple smooth oxydiacts occur, partly isolated,
partly disposed in strands. These sometimes exhibit a central swelling, either in the
form of a simple ring, more or less sharply marked off", or in the form of four cruciate, or
more rarely of two opposite roundish protuberances. In these well-developed central
portions an axial-canal cross can usually be seen. Less abundantly than these diacts,
triacts occur, which generally exhibit two long rays, lying in one axis, and a much
shorter thii'd ray, at right angles to the former and springing from a slight median
swelling.

Near the narrowed end of the body, and especially in the porous basal cushion,
hexacts, pentacts, tetracts, triacts and diacts occur, with cylindrical rays, which do not
run out to a point, but exhibit a truncated or even swollen end, and are terminally, and to
a greater or less distance inwards, thickly beset with conical tubercles. As an illustration
of the peculiarly modified spicules of the basal cushion, I have figured a triact on PL
XXVII. fig. 18. Tetracts are there, however, most abundant.

I have here and there found such a simple regular form of small oxyhexact, with
delicate narrow rays, as is represented in PI. XXVII. fig. 20. Very frequently, on the
other hand, and throughout the whole parenchyma, such forms occur as are seen in
PI. XXVII. fig. 23. The long narrow rays, covered with small, not oblique but
directly transverse protuberances and peaks, are more or less markedly bent round in
their distal portions, and the bendings of the two rays which lie in the same axis are
always in the same plane, but in opposite directions. The planes of curvature of the
three axes of the spicule form with one another equal angles of 120°. The representation
of these hexacts with curved rays in PI. XXVII. fig; 23 is so far unsatisfactory, since one
cannot recognise in it that three rays are approached by their ends, and their three

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

opposites likewise. This is, however, better expressed in the figures of corresponding
spicules in other species, e.g., Hyalonema apertum (PI. XXXVIII. fig. 5).

The dermal skeleton is composed of large, strong, liypodermal pentacts, of autodermal
pentact pinuli, and of autodermal amphidiscs of various kinds. The rays of the hypo-
dermal pentacts are quite smooth and are directed at right angles to one another. They
either run gradually to a sharp point or terminate conically. There is no hint whatever
of the missing distal ray. The proximal ray is usually larger than the four equal-sized
tangentials, but is sometimes smaller, and sometimes of equal length. The autodermal
pinuli are pentacts. The proximal ray is no longer recognisable, but the distal is dra-wTi
out into a slim fir-tree-like form, about 0'3 mm. long, and beset with obliquely inserted
and somewhat bent and pointed branchlets (PI. XXVII. fig. 21). The basal end of the
distal ray usually remains perfectly smooth. The four tangential rays, which lie at
right angles to one another and to the distal ray, are straight and of equal size. Only near
the somewhat blunt distal extremity do they exhibit some distally directed protuber-
ances, while the large proximal portion of each ray is quite smooth (PI. XXVII. fig. 21).

The very large kind of amphidisc which occurred in Hyalonema sieboldii is not
present in this species. The largest form of amphidisc has a length of 0"14 mm. and a
breadth of 0"06 mm. The axial rod is covered with protuberances, the highest of which
are in the middle, and arranged in a circle. The eight terminal umbel rays on each
side, which are disposed in a bell-like form, have a leaf- or spade-like shape, and
terminate in a somewhat rounded lancet-like point. The basal plate is comparatively
short. The individual terminal rays are sometimes laterally separated at their base by
grooves, but sometimes, on the other hand, they are laterally united. The terminal
arching of the beU-shaped umbel is not very broad, though strongly marked as the result
of the divergence of the umbel rays. The extremities of the ojiposite umbel-rays
approach one another to within a fourth of the whole length of the amphidisc. These
larger amphidiscs always have their median portion inserted in the dermal membrane,
so that the one half lies within the body of the sponge, while the other projects freely
above the surface of the skin; but, besides these, several other smaller forms occur as
represented in PI. XXVII. figs. 15, 16, 17. These all exhibit a relatively broad arching
of the much shorter terminal umbels, some of w^hich consist of more than eight umbel
rays. I have observed as many as thirteen umbel-rays in some of the medium-sized
amphidiscs (PL XXVII. figs. 16, 17). The axial rod either exhibits a few small,
irregularly distributed tubercles, or in addition to these four coarser, cruciately arranged,
median prominences, which may also occur alone. The umbel-rays have narrow bases,
they end in lancet-like fashion, and approach one another to the extent of not more
than one-third of the total length (0-03 mm.) of the amphidisc. I never found these
median and smaller amphidiscs except l>dng flat, that is, parallal to the surface of the
dermal membrane.

REPORT ON THE HEXACTINELLIDA. 199

The skeletal elements of the gastral membrane and lining of the efferent canals agree
in many respects with those of the dei'mal area. The hypogastral loentacts are, however,
wanting, being either replaced by diacts or not at all. The autogastral pentact pinnies
generally resemble the dermals in form and length ; they are, however, throughout more
delicate and slim (PL XXVII. fig. 22). They are at once thickest and strongest near
the margins of the openings, and are seen to be more widely separated and more weakly
developed the further the canals are traced inwards.

The long narrow diact marginalia in several layers, form a not very high, cuff-like
projecting marginal fringe. From the middle portion, which is distinguished by four
cruciate protuberances wdth axial canals, a smooth, gradually narrowing, terminally
pointed proximal ray extends inwards, while, in the opposite direction, an equally long,
or still longer distal ray, projects freely outwards in the same axis. The basal portion of
the latter is smooth, but the larger portion, on to the freely projecting terminal point, is
furnished with pointed, distally-directed lateral teeth, gradually decreasing in size.

In the basal pad, strong short spicules occur, with from six to two thick, cylindrical
rays, whose terminal portions are frequently somewhat thickened, and always covered
with radially disposed, pointed, conical teeth. Of the apparently smooth basal spicules
which form a simple, untwisted tuft, but few are present, and these all broken.

3. Hyalonema divergens, n. sp. (PI. XXVIII. figs. 1-11).

This species of Hyalonema was dredged in Mid-Pacific, to the east of Maldon Island
(Station 271), lat. 0° 33' S., long 151° 34' W., from a depth of 2425 fathoms, on a
Globigerina ooze ground. In shape it resembles an inverted bell, with a slightly widened
terminal portion ; it possesses a length of 8 cm. and a maximum breadth of 6 '5 cm. (PI.
XXVIII. fig. 1). The superior margin of the transversely truncated terminal surface
bears a delicate spicule-fringe. The terminal surface itself is formed by a transversely
,-tretched net, that is, by a typical sieve-plate. From the somewhat pointed lower end a
straight basal tuft projects. It is composed of parallel basal spicules, which are, for the
most part, broken somewhat short, and sometimes attain the thickness of almost a
millimetre.

The only specimen procured is somewhat compressed, and in some places rubbed
away, so that it is impossible to make any definite statements as to the characters
of the lateral surface. Only here and there does a patch of uninjured skin persist. At
the inferior extremity of the body, where the long, strong spicules of the root-tuft freely
project, the often mentioned thick basal pad is seen. There is no trace of encrusting
Pali/thoa, but the general state of preservation makes it impossible to say with certainty
that the Pahjthoa were absent during the life of the sponge.

The delicate net of the terminal sieve-membrane is raised from the marginal portion

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

of the upper surface of the thicker parenchyma of the body, and is also quite free from the
four cruciate septa and the central pillar, though it may possibly have coalesced with
the distal portions of the upper margins of the septa and with the end of the
columella. From the four large principal cavities of the efferent system, ramified lacunae
and canals lead to the sides, and downwards into the recesses of the sponge, so that, as in
Hyalonema sieboldii, the whole body-mass consists of a very wide-meshed system of
plates.

The great majority of the spicules of the parenchyma are straight or slightly bent,
smooth diacts, with or without central knot or swelling. Between these, smaller isolated
regular oxyhexacts occur, with straight, smooth rays (PI. XXVIII. fig. 8), and also
somewhat larger oxyhexacts with bent rays (PL XXVIII. fig. 6). The representation of
these forms in PI. XXVIII. fig. 6 is so far unfortunate, since each of the two opposite
trijjlets of rays should have the three ends of the rays a^jproximated.

In the dermal skeleton the strongly developed hypodermal pentacts of different sizes,
and with simple, smooth, somewhat pointed rays, are to be noted. The four tangential
rays of the latter are equipped with numerous pentact pinuli exhibiting strong,
short, slightly pointed or even blunted, smooth, straight basal rays, and exhibiting
a strikingly long, well-developed distal ray, furnished with somewhat closely set teeth
(PI. XXVIII. fig. 7). The freely projecting, toothed distal rays attain the length
of 1'3 mm. or more. The amphidiscs, which are present in considerable abundance,
exhibit considerable varieties of form and size. Most of them are 0"13 mm. in length,
and exhibit a long narrow axial rod with tubercles and frequently with larger cruciately
disposed protuberances in the centre. The eight tolerably broad, lancet-shaped umbrellar
rays attain a length of about a third of the whole amphidisc, and extend, with marked
divergence, in an incurved but oblique outward course (PI. XXVIII. fig. 3). In
reference to this striking divergence of the umbrellar rays in the largest amphidiscs, I
have bestowed on the species the title Hyalonema divergens. Among these amphidiscs
disposed at right angles to the limiting surface, others of a very large size occur here and
there. The latter attain a length of 0"5 mm., and have long umbrellar teeth curved
into the form of a long ovoid. From the middle of the delicate axial rod, four somewhat
long, cruciately disposed tubercles project. Among the numei"ous amphidiscs with
curved umbrellar rays, medium-sized forms, 0'06 mm. in length, are also found, with
knotted axial rods, and hemispherical arched umbels (PI. XXVIII. fig. 4). Small forms
are also abundantly present, measuring 0"03 to 0"04 mm. in length, and exhibiting short,
hemispherical terminal discs (PI. XXVIII. fig. 5). These smaller and smallest amphidiscs
are not disposed at right angles to the surface, but lie tangentially in the dermal
membrane.

The marginalia are long narrow diacts 2'2 mm. in length, with simple, smooth, or onl}'
inferiorly roughened surfaces. The proximal ray is uniformly narrowed and runs out to

REPORT ON THE HEXACTINELLIDA. 201

a point ; the free distal is at least twice as long, and bears short spines. At the boun-dar)-
of the two, in the centre of the diact, are four conical projections arranged in a circle
(PI. XXVIII. fig. 2). The gastral skeleton was not well preserved ; but it seemed not
to differ essentially from the dermah

4. Hyalonema toxeres, Wyville Thomson (PI. XXIX.).

Near the West Indian island St. Thomas (Station 24), lat. 18° 38' 30" N.,
long. 65° 5' 30" W.; several specimens of a Hyalonema were dredged from a Pteropod
ooze ground at a depth of 390 fathoms. The species has already been shortly described
by Sir Wyville Thomson in his preliminary account of the Voyage of the Challenger —
The Atlantic. I subjoin the account there given of this form.^

" Two examples of the sponge-body of a very handsome Hyalonema "were sifted out
of the coral mud. Unfortunately, in both cases, the sponge had been torn from the
central coil, and the absence of the coil might have thrown some little doubt upon the
form and mode of finish of the complete animal ; so that it was extremely fortunate that
a young specimen of the same species, about 40 mm. in length, was caught in the tangles
quite perfect."

" Hyalonema toxeres, Wyville Thomson, resembles closely the other known species,
Hyalonema lusitanicum and Hyalonema sieboldi, in general appearance and in the
arrangement of its parts. A more or less funnel-shaped sponge presents two surfaces,
covered with a network of different patterns formed by varying arrangements of large
five-rayed spicules. The upper concave surface shows a number of oscular openings
irregularly arranged, and the lower surface a more uniform network of pores, some of
which seem to be inhalent and others exhalent. The central axis of this sponge is
closely warped into the upper part of a coil of long and strong glassy spicules,
which, as in the other species, serve to anchor the sponge in the soft mud. Both
of the specimens dredged have the sponge more flattened and expanded than it
is in Hyalonema hisitanicum. In one of them it is nearly flat (fig. 5), forming
a uniform cake-like expansion, 80 mm. in length, by 70 mm. in width, and about 8
mm. in thickness. The upper or oscular surface is covered by an exceedingly close net-
work, with groups of large openings at nearly equal intervals. It is slightly raised
in the centre. The central elevation is followed by a slight depression, and the upper
wall then passes out nearly horizontally to a sharp peripheral edge, fringed with long,
delicate spicules, each consisting of a slender central shaft, with a cross of four
short transverse processes in the centre. The outer half of the central axis is delicately
feathered.

"The lower surface of the sponge (fig. 6) is protected by a singularly elegant

' Loc. cit., vol. i. p. 273.

(ZOOL. CHALL. ESP. — PAUT LIII.— 1887.) ^SS 26

202

THE VOYAGE OF H.M.S. CHALLENGER.

network of sarcode, with wide oval and round meshes radiating irreguhirly from a
central point. The membrane is transversed by irregularly radiating ridges of firmer
sul)stance, which unite in the centre in a projecting boss at the point, where in this
specimen the ' glassrope ' has unfortunately been torn out.

" In minute structure, Hyalonema toxeres corresponds in all essential respects with
Hyalonema sieholdii and Hyalonema lusitanicum. All the spicules are of the same
ground forms, with some little differences in detail, with the exception of one remarkable
spicule which enters largely into the structure of Hyalonema toxeres, and serves to
distinguish even the smallest portion of it. This is a large spicule, the largest above a
centimetre in length, and more than half a millimetre in width in the centre, shaped

i'lQ. i>.—Myaionemato:ceres, Wyville Thomson. Upper (after
W. Th.) surface, natural size.

Fig. 6.-

-Hyahmema toxeres, Wyville Thomson. Lower (after
W. Th.) surface of the sponge, natural size.

like a bow or boomerang. These spicules are distributed in all parts of the sponge, and
are particularly abundant near the insertion of the coil. No analogous form occurs in
the other species of Hyalonema.

" The large amphidiscs are much larger than in any other known sponge. They are
upwards of half a millimetre in length and visible to the naked eye, twice as large as in
Hyalonema lusitanicum. The feathered shafts of the five-rayed sj^icules which fringe
the openings are longer than in the other species, and the rays of the cross are much
shorter (fig. 7).

"The second specimen of the sponge body agreed with the one described in all

EEPOKT ON THE HEXACTINELLIDA.

203

essential points of structure, but was more conical in form. The young specimen
differed from the young of Ilyalonema lusitanicum at the same age in being
wider and more cylindrical, but the external wall, which afterwards becomes that of the
lower surface, showed the same arrangement in squares which we find in the young of
the other species, so that apparently the graceful, round-meshed, wide netting of the
under surface does not appear in the early stages.

"The coil is developed much in the same proportion and in the same way as in
Hyalonema lusitanicum, the filjres spreading out and incorporating with the sponge
substance. The characteristic bow-like spicules are abundant in the young sponge, and

Fig. 7. — Hyalonema toxeres, Wyville Thoinsou. Part of the membraue from the upper surface, x 40.

these, with the larger amphidiscs, place it beyond a doulit that it is the young of
Hyalonema toxeres.

" A quantity of loose spicules brought up in the dredge at the same time were referred
to this species. They were somewhat stouter than those of Hyalonema lusitanicum, and
less regular in outline. There was one coil nearly complete, involved in a calcareous
expansion of a branch of Diplolielia 'profunda. Two very young polyps, apj)arently of
Palythoa fatua, were commencing the formation of theii- investing crust at the top of
the coil of the young specimen, just below the sponge body."

Since the three specimens studied by Wyville Thomson are also at my disposal for

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

examination, though perhaps iu a less satisfactory state of preservation, I will add
to the above account some supplementary details.

The smooth cake-like specimen figured by Wyville Thomson (figs. 5, 6) was
not when I received it so weU preserved as it appears in the woodcuts referred to.
The specimen is represented in PI. XXIX. figs. 2, 3, of the natural size, after a photograph
from either side. It is evident that half the fringed margin (from which a rectangular
portion has been cut for investigation) is well preserved, while the rest of the
marginal portion seems compressed and torn. The side represented in fig. 3, covered
with a somewhat wide-meshed network of more or less thin strands, exhibits a tolerably
uninjured surface, and is almost smooth with the exception of the projecting knobs about
the centre of the (uninjured) disc. On the other side, however (fig. 2) the marginal portion
exhibits for about a finger's breadth an intact dermal layer, with a fine-meshed lattice-
work, through which were seen the large deep apertures or pits of the parenchyma,
occurring at uniform distances of about 5 to 8 mm., as figured in the woodcut (fig. 5)
of WyviUe Thomson's Atlantic. The middle portion of the flat sponge body is on the
same side much injured and apparently compressed (PL XXIX. fig. 2).

From Wyville Thomson's account this latter somewhat inbent side (PL XXIX.
fig. 2 of my figures, and fig. 5) is said to be the upper, while that covered by the
wide-meshed sieve-net and bearing a projecting central boss be the lower (PL XXTX.
fig. 3, and fig. 6). The projecting central protuberance of the latter is, according
to Wyville Thomson, the narrowed basal portion, from which the long tuft of needles
was torn away.

The results of my investigation force me, however, to another conclusion, namely,
that the side covered with the wide-meshed sieve-network is the upper, while the other
with the fine superficial lattice-work and the subjacent pits of the parenchyma is the
lower or outer. This is proved first of all by the character of the superficial layers.
The wide-meshed framework which covers the surface (fig. 3) exactly resembles
the sieve-plate present in the funnel-shaped second specimen (PL XXIX. fig. 1) of the
same species, while the fine-meshed dermal layer which covers the undamaged marginal
zone on the other side (PL XXIX. fig. 2) corresponds exactly iu structure to the dermal
membrane of all Hexactinellids and especially to that of the funnel-shaped specimen
(PL XXIX. fig. 1) of the same species.

The boss-like elevation iu the centre of the surface furnished with the wide-meshed
network does not represent the narrowed basal end of the body, but the conus centralis
which occurs on aU species of Hyalonema. The narrowed basal end of the sponge body
has, on the other hand, been torn off with the basal tuft of spicules, and is therefore not
to be seen on the lower surface of the body (PL XXIX. fig. 2). We have here apparently
a specimen of Hyalonema toxeres, which, after the tearing away of the basal tuft and
lower end, has been forced by oblique pressure into a flat compressed form. The normal

REPORT ON THE HEXACTINELLIDA. 205

form of the body as seen in the second specimen, which is comparatively well preserved,
but also destitute of basal tuft, is represented in natural size in PI. XXIX. fig. 1. We
have here an inverted conical form, with slightly bulging sides, and a broad upper
end not transversely truncated, Imt depressed in funnel-like fashion, and covered by a
narrow-meshed delicate network, while the external lateral surface of the body is enveloped
by a delicate dermal sieve-work — the dermal membrane, through which the more or less
regularly distributed large inhalent canals of the afierent system are seen as dark
spots (PI. XXIX. fig. 1). On the upper external margin, which projects as a sharply
defined ring, there is a continuous fringe of marginal spicules. On the lower, somewhat
incurved end, lies a round gaping opening, about 6 mm. in width. This leads into a
straight canal, and there can be no doubt that it represents the position of the torn off
basal tuft.

The third much smaller specimen, with a well-preserved tuft of basal spicules,
as also figured by Wyville Thomson, differs essentially in external form from the
above specimens. The body is not conical but oval, with truncated superior, and
somewhat narrowed inferior extremity. The structure, however, agrees so exactly with
that of both the other specimens, that one can have no hesitation in acknowledging the
correctness of Wyville Thomson's reference of this form to the same species, Hyalonema
toxeres. The specimen is unfortunately not well preserved, especially towards the upper
end. On the middle of the upper terminal surface the conus projects, and from it four
cruciate longitudinal septa with smooth, convex, superior margin extend through the
central space to the lateral wall into whieh they pass. The four cruciate cavities
thus formed are continued downwards and sideways into the efferent lacunar passages
of the parenchyma. Close beneath the narrowed inferior portion with the smaU basal
tuft thei'e is a ring of a few Palytlioa individuals (PI. XXIX. fig. 4).

Microscopic examination of the flinty spicules shows the intimate similarity of all the
three specimens, and the few variations which are present are referable to diflerences in
age and individual characteristics.

In the parenchyma, besides the familiar regular medium-sized oxyhexacts, and
numerous straight or slightly bent, weakly developed diacts with or without central
swellings or knobs, especially remarkable and characteristic diacts occur. These are thick
and spindle-shaped, curved or somewhat bent, with rounded terminal points, and without
central swelling or tubercles. These bent needles measure 3 to 8 mm. in length, and
exhibit in the middle a thickness of 0"3 to 0'4 mm. The concentric lamination is very
distinctly seen, and on the outer end the relation of the various layers of growth is
often most beautifully demonstrable (PI. XXIX. fig. 11). The distribution of these
strong, bent needles in the parenchyma seems to me quite irregular. I have not observed
their special abundance in any one region, nor on the other hand any definite disposition.
They frequently occur parallel to the skin near the external surfocc, but also in

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

the middle of the parenchyma, and in the most varied relations to the dermal surface.
In contrast to most other species of Hyalonema, small oxyhexacts with straight
(PI. XXIX. fig. 9) and with bent (PI. XXIX. fig. 10) rays are not abundant. The latter
occur indeed in very sparse isolation. Both forms are characterised by this, that the rays
are throughout their whole extent thickly beset with small barbs, turned obliquely
inwards, i.e., towards the middle.

The dermal skeleton is supported by medium or small-sized, simple, smooth, Ivy^o-
dermal oxypentacts without a trace of a distal sixth ray, and also by numerous medium-
sized hypodermal oxydiacts, which all lie tangentially. Above these hypodermalia, auto-
dermal pentact pinuli occur, which exhibit short, slightly toothed, basal rays of moderate
strength, and a narrow distal, 0'4 to O'o mm. in length, with somewhat long narrow
obliquely inserted, pointed lateral spines. Large dermal amphidiscs, 0'35 mm. in leno-th
(PI. XXIX. fig. 6) are tolerably abundant. In their breadth and in the somewhat flatly
arched short terminal umbels, with eight broad, lancet-shaped, umbel rays, they at once
recall the approximately equal, strong dermal amphidiscs of Hyalonema sieboldii. They
are all disposed at right angles to the surface, with their centre in the dermal membrane
(PI. XXIX. fig. 5). Besides these, numerous smaller amphidiscs with short hemispherical
terminal umbels (PI. XXIX. fig. 8) occur in tangential disposition. Less frequently
medium-sized amphidiscs occur, essentially resembling the greater above described (PI.
XXIX. fig. 7).

In the sieve-membrane covering the superior terminal surface of the sponge, and to a
large extent coalescent with the subjacent parenchymal layer, skeletal elements occur
similar to those in the external skin. The inner surface of the large lacunae and
passages penetrating the parenchyma, and also of the efferent canals is here smooth,
exhibiting no gastral or canahcular skeleton, containing neither hypogastral jjentacts
nor pinuli nor amphidiscs, in fact, solely supported by weakly developed, somewhat
curved diacts like those which occur so abundantly in the parenchyma.

The marginaha, which form a continuous projecting fi'inge on the sharp upper margin
of the body, are straight oxydiacts, whose proximal end inserted in the parenchjmia is
smooth, while the much longer and very gradually pointed distal end is beset with lateral
spines projecting obliquely downwards (PI. XXIX. fig. 5). At the boundary between the
distal and proximal portions, four cruciately disposed hemispherically arched knobs project,
being traces of the rudimentary transverse rays.

The basal pad of the smaller (PL XXIX. fig. 4) and of the large funnel-shaped
specimen (PI. XXIX. fig. 1) consists of six- to two-rayed, strongly developed spicules,
with cylindrical, externally truncated, or rounded straight rays. These agree thoroughly
with similar structures in the hitherto described species of Hyalonema. The proximal
portion of each ray is smooth ; the distal, on the other hand, is thickly beset with coarse
conical spines ; more rarely the whole surface of the spicule is uniformly spinose.

REPORT ON THE HEXACTINELLIDA, 207

The long basalia of the root-tuft which I examined in the small specimen (PI. XXIX.
fig. 4) exhibit opposite or spirally disposed, obliquely projecting ledges with marginal
teeth, like those which M. Schultze has figured in Hijalonema sieboldii in his well-known
work (Taf. ii. fig. 4). The extreme distal end of the long basalia was not preserved.

5. Hijalonema lentii, 0. Schmidt (PI. XXX. figs. 9-17).

In his Spongien des Mcerbusens von Mexico, 0. Schmidt describes (p. 65), under
the designation Asconema kentii, a Hexactinellid which was dredged at various localities
off Grenada, Martinique, Guadeloupe, and Bequia, from depths of 300 to 1500 fathoms.
This form occurred in two varieties, on the one hand, as a " flat or shallow bowl, either
rounded off inferiorly or furnished with a short or somewhat irregularly twisted pointed
stalk," and on the other, as " a saccular form with irregular margin, divided internally
into irregular pouch-like divisions and cavities, separated by thin ragged partitions."
"A delicate outer layer with fir-tree-like Sfiicules, extends like a fine ved over the
surface. The body is unusually rich in double verticils (Do^jpelquirrlen), occurring in
very varied dimensions and forms in the several parts of the body." As my esteemed
fellow worker in Strassburg was kind enough to give me for examination a dried
specimen of the goblet-type, I am able to include this beautiful form in the series of
species of Hyalonema. A figure, drawn from a photograph, represents the sponge of the
natural size. Both the external and the concave superior and internal surface of the loose
cup- or funnel-shaped body are covered by a delicate reticulate membrane. On the outer
wall this network is much finer and more uniform than on the concave upper surface,
where it seems to form a sieve with unequal, and in some case«, large meshes. It is
either attached to the subjacent body parenchyma, or extends freely across the large
cavities. On the superior external margin there is, as 0. Schmidt pointed out, a cuff-
like fringe of fine spicules.

The loose parenchyma of the sponge is supported by numerous straight or curved
smooth oxydiacts, with or without rudiments of the four abortive transverse rays, and
by smooth, simple, medium-sized oxyhexacts, which frequently exhibit a slight roughness,
and are furnished either with straight or with curved rays as represented in PI. XXX. figs.
14, 15. It is noteworthy that the oxydiacts are generally thin and often so long that
they become readily curved, not unfrequently appearing much bent or even coiled.

The h}'podermal oxj-pentacts are perfectly smooth. Their tangential rays are
mutually apposed and form a beautiful rectangular mcshwork. To these rays are
attached a great number of autodermal pinuli, with four straight, somewhat substantial,
basal rays which are comparatively short, conically pointed, and either quite smooth
or equipped terminally with minute distaUy directed teeth. The distal i-ay, on the
other hand, measures about 0*45 mm. in length, is smooth at the l^ase, but elsewhere

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

SO thickly beset with rather long, narrow spines, diverging obliquely outwards, that a
somewhat bushy appearance results (PI. XXX. fig. 16). An essentially different appear-
ance is presented by the pinuli which lie on the funnel-shaped concave upper surface, in
the strands of the sieve. The basal rays are here narrower and longer ; the weakly
developed distal ray has a decidedly greater length — 0'8 mm., and bears only a small
number of short ap250sed spines (PI. XXX. fig. 17). On the inner surface of the gastral
cavity and wider efferent canals, a special gastral or canalicular skeleton occurs. This
consists of moderately large, smooth, simple ox\i:)entacts, with the four transverse
rays inserted on the gastral or canalicular membrane, and sparsely furnished with
slim autogastral pinuli of various sizes. The four straight basal rays of the latter are
rather long and thin, and gradually pointed ; the distal ray is slim and beset with short,
somewhat apposed, lateral spines.

Among the dermal amphidiscs there frequently occurs a large, comparatively broad
form (0'5 mm. long by 0'2 mm. broad), with short, much arched umbels, and eight
broad lancet-shaped umbel rays. The strongly developed axial rod is for the most
part smooth, but generally exhibits at the middle point four cruciately disposed^
rounded tubercles, — the remnants of the four abortive transverse rays. In the superior
sieve-plate these large broad amphidiscs which occur all over in the external skins are,
curiously enough, absent. On the other hand, medium-sized (0"3 mm.) amphidiscs of
another form occur, with narrow, much arched, eight-rayed umbels, which occupy more
than a third of the total length, sometimes almost meeting in the middle. The slim axial
rod is beset with tubercles (PI. XXX. fig. 11). Besides these, somewhat smaller and
narrower amphidiscs, with eight medium-sized nmbel rays occasionally occur, and in
great numbers the decidedly smaller forms with short, broadly-arched many-rayed umbels
which occurred so abundantly in the external skin (PL XXX. fig. 13).^

In the basal pad the familiar strongly-developed six- to two-rayed spicules occur
abundantly. Their rounded ends are covered with tubercles, while the proximal main
portion of the ray is smooth. The bundle of needles in the basal tuft is unfortunately
torn away. At the inferior extremity of the body one can see the distinct defect left by
their removal.

Hyalonema poculum, n. sp. (PI. XXXIII. figs. 1-7).

In the neighbourhood of the island of Juan Fernandez, west of Valparaiso (lat.
33° 42' S., long. 78° 18' W., Station 300), a cup-shaped Hyalonema was dredged from
a Globigerina ooze bottom at a depth of 1375 fathoms. The inferior extremity and
tuft of basal needles is torn away. The upper funnel-shaped concave end of the
comparatively thin- walled cup has a transverse diameter of about 5 cm., while the lower
I Too few rays have been represented in PI. XXX. fig. 13. There are from thirteen to sixteen.

REPOET ON THE HEXACTINELLIDA. 209

broken portion measures about three. From this it may he inferred that the whole body
had a length of about 10 to 12 cm. On the superior somewhat sharply terminated
margin there is a continuous border-fringe of projecting spicules. The skin of the upper
surface of the funnel is gradually raised to a distance of about 1 cm. from the body-
parenchyma as an independent lattice- work (PL XXXIII. fig. 1). On the external
surface traces are seen here and there of the narrow-meshed rectangular network of the
dermal membrane.

In the parenchyma lie numerous long, thin, and somewhat pliable oxydiacts, on which
central tubercles or nodes are generally absent, or but slightly developed. Between these,
some medium-sized and numerous small, straight, regular oxyhexacts occur; the rays
of the latter exhibit a slight roughness (PL XXXIII. fig. 7). I have not found any small
oxyhexacts with bent rays. The dermal skeleton is characterised by the presence of
moderately large strongly developed hypodermal oxypentacts. The four tangential rays
bear autodermal piuuli with four short, strong, almost smooth basal rays, and a moderately
long, somewhat markedly spinose distal (PL XXXIII. fig. 6). Dermal amphidiscs are
represented by a somewhat large (0'5 mm. in length) form with broad, short, hemi-
spherically arched umbels, and with four or eight nodes at the middle of the axial beam.
The eight umbel rays are pointed like lancets, and not broadened out in shovel-like
fashion (PL XXXIII. fig. 4). I sometimes observed the sharp lateral margin of several
somewhat longer umbel rays uniformly beset with fine teeth in a saw-like fashion.
Besides these, the familiar small amphidiscs with short, flat, or hemispherically arched
umbels occur in great abundance, and also isolated medium-sized thinner forms with
eight to ten umbel rays.

The skeletal elements of the lattice-work covering the upper surface of the funnel
agree essentially with those of the skin.

The marginalia measure 0'6 mm. in length; the distal ray is covered with spines Hl^e
that of the pinuli; the proximal ray is smooth. In the centre are seen four cruciately
disposed, somewhat distally directed, bent tubercles with fine terminal points (PL XXXIII.
fig. 3).

Hyalonema conus, n. sp. (PL XXXIII. figs. 8-15).

South of Australia (lat. 50' 1' S., long. 123° 4' E., Station 158) a form of Hyalonema
was dredged from a Globigerina ooze bottom at a depth of 1800 fathoms. The basal tuft
was again torn away, but the rest of the body was otherwise tolerably intact. A portion
of a basal tuft with a small remnant of attached body was also found, and very probably
belonged to the same specimen. The body has the form of a cone, 6 cm. long by 5
cm. broad at the upper transversely truncated end, whUe the lower round end narrows
to about the thickness of one's little finger (PL XXXIII. fig 8). The somewhat incurved

(ZOOL. CHALL. EXP. — PART LUI. — 1887.) Ggg 27

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

upper terminal surface is covered by a terminal sieve membrane lying flatly over the
parenchyma of the body. Eound about this sieve-plate, on the somewhat sharp lateral
margin, an inconspicuous circular fringe of spicules projects. On the lateral surface of the
cone, remnants of the delicate dermal lattice-work are here and there recognisable (PI.
XXXIII. fig. 8), and through the latter large grooves and inhalent apertures are
visible.

The parenchyma is supported by medium-sized smooth oxyhexacts and by numerous
thin and pliable oxydiacts of various sizes, wdth or without central swelling. More rarely
diacts occur with one or two club-shaped swollen ends. Straight, smooth, small
oxyhexacts occur in moderate abundance.

The dermal skeleton consists of medium-sized smooth h}^odermal oxypentacts. The
autodermal pentact pinuli are relatively long (about 0"4 mm.), and bear four slim medium-
sized basal rays, and a distal beset with somewhat long spines (PL XXXIII. fig 15).

The superior sieve-network is supported by reticulate strands of strongly developed,
moderately long, smooth oxydiacts, which, as a rule, exhibit at the middle only two
hemispherical or flatly apposed nodes. Between these, simple, strong, smooth oxy-
tetracts occur here and there, and also ^jentacts witli rays of a simdar character. The
autodermal pentact pinuli of this terminal sieve-plate are, for the most part, like the
dermal, though in some regions longer (PL XXXIII. fig. 14).

Since the abundant amphidiscs of the external skin and those of the terminal sieve-
membrane are exactly alike, I will describe them for both localities at once. Most
strikiog is the largest form (about 0"2 mm. long), which exhibits a strongly developed axial
rod, smooth with the exception of four cruciately disposed central protuberances, and
with very gently arched terminal umbel rays, extending only over about one-third of a
sphere, and provided with eight teeth. These short umbel rays are very broad, and
terminate in a paddle or leaf-like point (PL XXXIII. fig. 9). Besides these, amphidiscs
of medium size occur with irregularly toothed axial rods, and with terminal umbels,
hemispherical or longer, with eight slender rays (PL XXXIII. fig. 11). Similar forms
sometimes occur with longer and more divergent umbel rays (PL XXXIII. fig. 12). As
in almost all species of Hyalonema numerous small many-rayed amphidiscs occur, as
figured in PL XXXIII. fig. 13.

The marginalia exhibit a long toothed distal ray, a smooth proximal, and four
cruciate hemispherical rudiments of the tangentials.

The spicules of the isolated basal tuft, which was found beside the body of the sponge,
and most probably belongs to it, exhibit on their inferior extremity a very remarkable
anchor-like structure. On the somewhat thickened terminal portion, four twisted, strongly
developed rays project transversely in a circle, and then bend upwards at right angles
along the equally long terminal portion. The latter portion is somewhat curved, and
runs gradually to a point (PL XXXIII. fig. 10). On the prolongation of the axis there

REPORT ON THE HEXACTINELLIDA. 211

is a small flat tubercle-like process, representing the abortive sixth ray. The lono- stalk
exhibits, like the four anchor teeth, a distinct axial canal.

Subgenus 2. Stylocalyx, n. subgen. (PI. XXXI.; PI. XXXIV. figs. 1-11;
Pis. XXXV., XXXVIL, XXXVIIL, XL., XLL).

The superior aperture of the gastral cavity is not covered by a sieve-plate, but
remains quite open. The gastral cavity is divided into four chambers by a central cone
and four cruciate radial septa.

1. Hyalonema {Stylocalyx) thomsoni, Marshall (PI. XXXIV. figs 1-11).

In his memoir on the HexactineUidre^ W. Marshall describes a species of Hyalonema
obtained by WyviUe Thomson on the "Porcupine" Expedition to the north of the Shetland
Islands, at a depth of 550 fathoms. This form measured 7 cm. in length, 3 "5 being
occupied by the body proper, and 3 "5 by the basal tuft, and about 8 mm. in thickness.
Through the kindness of 0. Schmidt I have been able to examine this di-ied specimen.
The body ^ has a spindle-like form, the upper end not being transversely truncated,
but extending to a somewhat pointed apex, in the form of a projecting cone. The
basal tuft is surrounded, just below the lower pointed end, by an encrustation of
Palythoa, 3 mm. in length, and with four or five polypes. The number of
slightly twisted basal spicules is estimated by Marshall at fifty or sixty. On some
of them he observed a ridge with nail-like teeth, which was absent from others; the
lower ends are all broken off. In the still partially preserved external dermal
skeleton Marshall found regular "four-rayed" spicules which were mutually apposed
by their limbs, and also several five-rayed forms. In the fir-tree forms situated on these
spicules of the dermal network the spinose principal ray projects, according to Marshall,
into the lumen of the rectangular dermal meshes in the plane of the skin, that is, lies
tangentially. Besides these, amphidiscs occasionally occur, 0"03 to 0'08 mm. in length,
and with six long, narrow, pointed anchor-teeth, as well as very small (O'OOB mm. in length)
double anchors. Within the body of Hyalonema thomsoni, Marshall describes fom-
large cavities, into which large and small parenchymal passages open. On the wall of
these cavities long curved uniaxial spicules occur, besides isolated smooth six-rayed forms
and amphidiscs with small double anchors. In the parench}Tiia proper he found,
besides the long uniaxial spicules, four-, five-, and six-rayed forms, and most frequently
small hexacts.

In the covering membrane, which extends from the truncated superior and external
margin to the conical knob projecting freely in the centre, Marshall observed four cleft-

1 Zeitschr.f. leiss. Zool., Suppl.-BJ. xxv. p. 225, 1875. '^ Loc. cit., Taf. xvii fig. 84.

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

like openings whicli led into the foni' subjacent cavities. In this lid, and in the conical
terminal knob, the same fir-tree-like amphidiscs occurred as in the external skin.

From my own examination of the specimen in Professor 0. Schmidt's possession, which
was originally investigated by Marshall, I am able to corroborate the essential results of
Marshall's research, and I shaU therefore only confine myself to extending and completing
the latter.

The parenchyma proper in the above-described Hyalonema is supported by medium-
sized, smooth, regular oxyhexacts, beside which numerous smooth, straight, or curved
diacts occur, with or without central swellings. These have been already figured in
detail by Marshall. Small oxyhexacts with straight roughened rays — and these only —
occur in moderate abundance (PL XXXIV. fig. 4). In the dermal skeleton the familiar
smooth hypodermal pentacts occur, as also autodermal pentact pinuli with somewhat long
terminally toothed basal rays, and with a freely projecting pointed distal, about 0"24 mm.
in length, and beset with long oblique curved spines (PI. XXXIV. fig. 6). Eoundish
amphidiscs of various form and size, of which the largest have an average length of 0"2 mm.,
also occur. They are characterised by the possession of a flat, short, terminal umbel on
either side, with six simple hook or claw-shaped, narrow rays, not terminally broadened.
The moderately strong axial rod bears isolated hemispherical nodes, four of which, usually
somewhat more strongly developed, are cruciately disposed in the centre (PI. XXXIV.
fig. 2). An amphidisc of this type, in process of formation, is figured in PI. XXXIV.
fig. 7. Besides these typical six-rayed large amphidiscs, others half as large occur, with
long, campanulate, narrow, eight-rayed umbels, the slender rays of which almost meet,
diverging somewhat laterally towards one another. The narrow axial rod has a central
nodule (PI. XXXIV. fig. 3). Still smaller forms whose umbels with eight or more rays
are short and hemispherical (PL XXXIV. figs. 5, 8), occur abundantly. In the basal pad
of the lower end of the body strongly developed spicules occur, variously provided with
rays from six to two in number. These truncate rays are thickly studded with spines or
teeth from the ends to within a varying distance of the centre (PI. XXXIV. figs. 10, 11).
The tuft of spicules is composed solely of long spicules, which bear toothed ringed
ridges. These are either broken oS" inferiorly, or continued into a club-shaped or
hemispherical four-toothed anchor with four strong, backwardly bent, narrow, pointed
teeth (PL XXXIV. fig. 9).

Among some Hexactinellids of the "Porcupine" Expedition which I obtained through
the Challenger Ofiice, I found a specimen with characters exactly corresponding to those
above described. This was obtained from Station 47 of the " Porcupine " Expedition,
from a depth of 542 fathoms. The extended spindle-shaped form had a total length of
4 "5 cm. The body proper, which was about 5 mm. in thickness, occuj^ied 3 cm.; and
was continued superiorly into an apex, while the narrowed lower end bore a colony of
three or four Pcdythoa polypes, embracing the basal tuft for 2 mm. ' The spicules both of

REPORT ON THE HEXACTINELLIDA. 213

tlie parench}Tna and of the basal tuft were exactly similar to those above described. The
skin was unfortunately wholly absent, so that nothing further could be ascertained as to
dermal skeleton.

A third specimen, which in my opinion belongs to this species, Hyalonema
thomsoni, is deposited in the Berlin Zoological Museum. It is labelled as Hyalonema
lusitanicum, Socage (?), was collected by Wyville Thomson off the West Hebrides,
and is numbered 405. This splendid specimen has a total length of 21 cm., and the
body proper, which occupies 9 cm., has a maximum breadth of 3 cm. I have figured it in
its natural size on PI. XXXIV. fig. 1. The apical cone projects — 2 cm. in length — from
the centre of the truncated terminal surface, which exhibits four cruciately disposed radial
septa and interjacent apertures, and resembles exactly the superior surface of the speci-
men figured by Marshall. Below the somewhat sharply angular projecting lateral margins
of this smooth terminal surface, which is not covered by a dermal sieve-network, the
body exhibits a slight annular constriction but curves outwards again below the middle,
and is finally conically narrowed towards the lower end. Below the somewhat rounded
off" annular basal jjad, there is an encrustation (6 cm. long) of Palythoa, and from this the
basal tuft is prolonged downwards (PI XXXIV. fig. 1).

In the parenchyma, besides medium-sized, smooth, regular oxyhexacts and numerous
smooth, straight, or gently curved oxydiacts, with or without central nodes or tubercles, a
large number of small oxyhexacts with straight, somewhat roughened rays occur, similar
to those which we have already described and figured in the smaller specimens (PI. XXXIV.
fig. 4).

The dermal skeleton exhibits the same smooth hypodermal oxypentacts and somewhat
long, narrow pinuli, with moderately long, slightly spinose, basal rays, as we have already
described in the smaller specimens. The amphidiscs also essentially agree in size and
form with those above described and figured (PL XXXIV. figs. 2, 3, 5, 7, 8,), and differ
only in this, that the largest amphidiscs with short, narrow umbel rays have, as a rule, not
six but eight rays. The substantial toothed spicules of the basal pad and the long spicules
of the root-processes essentially resemble those of the above forms ; occasional, slight,
and inconstant differences seem to me to be simply individual, conditioned by the large
size, and in no way of specific importance.

The superior terminal surface is in this specimen much better preserved than in the
smaller forms, so that the essential agreement between the covering membrane and the
external skin is demonstrable, except that the autodermal pentact pinuli are somewhat
narrower and longer than in the former. The same enveloping layer extends for some
distance into the system of efferent tubes.

Near the basal pad numerous more or less long diacts occur in the parenchjTna,
and do not run out at their ends into smooth points, but are somewhat thickened
and spinose before finally terminating in conical points. At the middle point of

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

these diacts with spinose terminal portions four cruciate hemispherical tubercles usually
occur.

This species, Hyalonema thomsoni, differs from Hyalonenui lusitanicum, Bocage, not
only in general form and in the long projecting cone, but more especially in the large
amphidiscs with short narrow umbel rays, and in the absence of smaU oxyhexacts with
curved rays.

Though I have no hesitation in including the three forms described above within
the one species, Hyalonema thomsoni, Marshall, I am in doubt as to a specunen of
Hyalonema dredged near the Azores (Station 73, lat. 38° 3' N., long. 31° 14' W.), at a
depth of 1000 fathoms, from a Pteropod ooze ground. This fragment, which has a total
length of only 3 cm., represents the lateral portion (about one-third) of a pear-shaped body,
about 8 mm. in breadth at the upper end, while the inferior portion narrowed to 2 mm.
projects as a broken tuft of few spicules. The superior extremity of the body is not
preserved.

The parenchymal spicules consist of smooth oxydiacts of various lengths, with or \dth-
out central tubercles, and of moderately large smooth oxyhexacts and smaller forms ^-ith
straight somewhat roughened rays, — all exactly agreeing with those already described in
Hyalonema thomsoni. The hypodermal smooth oxypentacts and the somewhat slim auto-
dermal pinuli of the skin do also not differ essentially from those of Hyalonema thomsoni,
and the same may be said of the substantial spicules of the basal pad (PL XXXIV. figs. 15,
17), or of the long spicules of the root-tuft, distinguished by theu- four-toothed terminal
anchors (PL XXXIV. fig. 16), which exhibit an axial cross.

Only the amphidiscs, still discoverable in the small and isolated remnants of the skin,
are somewhat divergent from those of the specimens of Hyalonema thomsoni, in the
apparent absence of the large form bearing short hemispherical terminal umbels with six or
eight narrow hook-shaped umbel rays, and in such slight differences in the form of the small
amphidiscs, as may be discovered by inspecting PL XXXIV. figs. 12, 13, 14. Since the
apparent absence of the large amphidiscs may very probably be referred to the incom-
pleteness of the specimen in which the skin was almost gone, and since besides the small
divergent amphidiscs exactly congruent forms also occur, there seems no reason to erect a
separate species for this fragment, and I therefore content myself with designating it
Hyalonema thomsoni, var. exiguum.

2. Hyalonema (Stylocalyx) apertum, n. sp. (Pis. XXXVII., XXXVIIL).

In the Sagami Bay, west of Yokohama in Japan (Station 232, lat. 35° U' K,
long 139° 28' E.), from a depth of 345 fathoms and a green mud bottom, several
specimens of Hyalonema were dredged. In some of these the body is still well preserved,
but in most only the basal tuft and the Palythoa incrustation remain. The body of the

REPORT ON THE HEXACTINELLIDA. 215

sponge is cup-shaped, more or less curved outwards laterally, and truncated transversely
or obliquely above. The superior terminal surface is sometimes almost flat, usually,
however, markedly concave, and always j^rovided with a sharp-angled projecting marginal
fringe, and with a central more or less markedly projecting cone. From the inferior
lateral portion of this central cone four radial longitudinal plates extend as septa through
the gastral cavity, broadening out laterally to unite with the thick body-wall. By these
four cruciately disposed septa, four cavities opening widely on the upper terminal surface
are enclosed, and these are continued laterally and inferiorly into the tree-like branched
system of efi'erent canals. The fine terminal canals extend to within a short distance
of the external skin, and there end blindly in those diverticula of the membrana
reticularis which are characteristic of the Hyalonematidse (PI. XXXVII. figs. 1-3 ;
PI. XXXVIII. fig. l). Close below the inconspicuous annular basal pad, surrounding
the upper portion of the long much twisted root-tuft, there is an incrustation of
Palythoa, which embraces in some forms only a small portion, and in others two-thirds or
more of the long tuft.

The total length of the sponge varies between 18 and 50 cm., of which the body
itself occupies, on an average, from 5 to 8 cm. The maximum diameter sometimes occurs
just below the upper end, sometimes about the middle, and varies from 3 to 6 cm. The
average thickness of the basal tuft, just below the lower end of the body, is 3 to 4 mm. ;
further down, however, it becomes greater until the brush-like divergence of the spicules
begins. Apart from some circular apertures from 1 to 4 mm. in diameter, the superior
convex surface of the body and the whole surface of the freely projecting conus are seen to
be fairly smooth. On the other hand, the whole lateral surface distinctly exhibits, even to
the naked eye, the rectangular lattice-work of the dermal membrane (PI. XXXVII. fig. 3).

The parenchyma contains a few medium-sized superficially smooth oxyhexacts, and
numerous oxydiacts, disposed in strands or lying singly, with or without central swellings
or with two or four cruciate nodes (PI. XXXVIII. fig. 3). The small oxyhexacts
which occur in great abundance are all rough, and almost thorny, and have more or less
markedly curved rays (PI. XXXVIII. fig. 5). Isolated very small but strongly developed
oxyhexacts, with distally directed teeth (PI. XXXVIII. fig. 4), occasionally occur.

The dermal skeleton, which appears on surface view as a conspicuous rectangular lattice-
work (PI. XXXVIII. fig. 2), consists of simple smooth medium-sized hjrpodermal oxypen-
tacts, and of abundant autodermal pentact pinuh, about 0'2 mm in length, with strongly
developed short-toothed basal rays, and somewhat strongly developed distal ra}^ with
short obliquely inserted, spines (PI. XXXVIII. fig 10), and finally of irregularly scattered,
large, medium-sized, and small amphidiscs. The former, measuring from 0'15 to 0'17 mm.
in length, have a strong knotted axial rod, and broad umbels with eight paddle-
shaped rays (PI. XXXVIII. fig. 7) ; the middle-sized forms have from ten to twelve long
narrow umbel rays (PI. XXXVIII. fig. 6) ; while the very abundant small forms bear

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

hemispherical umbels, with many (usually twelve) rays, and thin axial rods, either smooth
or with median nodes (PI. XXXVIII. figs. 8, 9).

The gastral skeleton exhibits smooth hj^jogastral pentacts similar to those occurring
in the skin, the same form of pinuli as autogastralia, and similar or somewhat smaller
amphidiscs (PI. XXXVIII. figs. 1, 12). While the smooth pentacts and the amphidiscs
are confined to the gastral membrane and to the lining of the largest efi'erent passages,
the pinuli have a somewhat wider distribution, and, becoming scarcer as the four basal
rays are more prolonged, are continued into the eiferent system of canals even to the fine
terminal branches, where they finally entirely cease (PI. XXXVIII. fig. 1).

The projecting fringe of spicules on the external sharp margin of the oscular region
consists of marginal oxydiacts at least 0'5 mm. in length, with long distal ray bearing
minute obliquely disposed spines. The proximal ray, which is about a third shorter, bears
very small backwardly directed teeth. At the boundary between the two rays four
cruciately disposed hemispherical or somewhat larger knobs protrude (PI. XXXVIII.
fig. 11).

In the lower narrowed end of the body, and especially in the basal pad, compact six
to two-rayed spicules with blunt, toothed, or spinose ends occur, like those in the same
position in most species of Hyalonema. The spicules of the basal tuft vary greatly in
lencfth, and resemble in structure those of Hyalonema sieboldii, to which Hyalonema
apertum presents a close resemblance in the general structure of its spicules. Whde the
upper portion of these long-stalked anchors is c^uite smooth, the lower bears oblique rows
of barbs on projecting ridges, and, on the very end, borne on a smooth narrow neck, a
hemispherical or helmet-shaped terminal knob with four to eight shovel-shaped marginal
anchor teeth, which are directed obliquely upwards and outwards. The intersection of
the axial canals in these long anchor spicules always lies in the lower swollen end, in
the head of the anchor.

In a rich collection of Japanese forms brought home by Dr. Hilgendorf, and deposited
in the Berlin Zoological Museum, are several dried and well-preserved spirit specimens of
Hyalonema apertum. On closer examination of these I found two small dried specimens
of this species, which had been temporarily designated by Professor W. Marshall as
Hyalonema affine, Marshall. I refer on this point to the brief description which Marshall
gives, in his researches on Hexactinellids (1875), of Hyalonema affine, a species which
he had erected for a form which had been already examined by Max Schultze and noted
by him as " B " and " C."

The diagnosis which Marshall gives {loc. cit., -p. 234) is as follows: — " Hijalonema
affine, if not, as I believe, a distinct species, is at least a specially well defined variety of
Hyalonema sieboldii. The tuft is 47 cm. long and only 8 mm. broad; the sponge body
proper has a length of 9 cm. Other specimens exhibit, with an axial strand of about the
same length, though of greater breadth, a larger sponge-body, which measures for instance

REPORT ON THE HEXACTINELLIDA. 217

in the specimen figured by Max Seliultze on Taf. i., 13 cm., while the root-tuft is 48 cm.
long and 1'5 cm. broad. I wish, however, to lay special emphasis on this, that in
Hyalonema affine the dermal skeleton (partly rubbed off however, though not to such an
extent as Max Schultze describes) exhibits peculiar characteristics. From the lower end
of the sponge body, from the position where the axial tuft enters, numerous longitudinal
strands, 0'5 mm. in breadth, arise, dividing repeatedly and anastomosing with one another,
becoming gradually narrower in so doing, and I did not observe these longitudinal strands
in any other specimen of Hyalonema, although I examined a considerable number in
various states of preservation. The strands are composed of uniaxial, decidedly long
(8 cm.) pliable spicules which frequently intersect. At the point of intersection there
is a tolerably constant occurrence of a spicule with dimensions which neither I nor
Max Schultze have ever found in Hyalonema sieboldii, namely, with an axial length of
fully 3 mm. The intersecting bundles lie in the direction of the axes. Between these
strands there is a further dermal skeleton, but this in no way diflfered from that of the
other specimens. Similarly the afferent apertures, and indeed the rest of the sponge-
tissue, agreed both in macro- and microscopic characters with Hyalonema sieboldii, though
the apertures were not so abundantly present. It seemed to me, however, that the great
differences in size, and especially the peculiar character of the dermal skeleton, justified
the erection of a new species, closely related to Hyalonema sieboldii."

According to this description, which was of course based on a single, dry, and probably
not very well preserved specimen, it seems to me possible that the form above described
as Hyalonema apevtum may be identical with Marshall's Hyalonema affine, and I have
indeed hesitated long whether I should adopt the older designation. There is, however,
in Marshall's diagnosis no mention of some of the peculiarities which are t}^ical and
characteristic of Hyalonema apertum, especially the complete absence of the terminal
sieve membrane, while the character especially emphasised by Marshall, namely, the
presence of intersecting bundles of long uniaxial spicules below the skin of the dried
specimen, occurs also in various other species of Hyalonema (though not in Hyalonema
sieboldii, owing to the presence of countless commensal polypes), so that no certain and
exclusive character seems to remain to Hyalonema affi^ne. The dimensions of the body
or of the basal tuft cannot be regarded as in any way determinative.

Since, then, it is not probable that Hyalonema affine, Marshall, is identical with
my Hyalonema apertum, I prefer to retain for the time the latter title.

3. Hyalonema {Stylocalyx) deptressum, n. sp. (Pis. XXXV., XXXVI. ).

In the Mid-North Pacific, north of the Mellish Islands (Station 246, lat. 36° 10' N.,
long. 178° 0' E.), from a depth of 2050 fathoms and a Globigerina ooze ground, several
specimens of Hyalonema were dredged, which in general form differ markedly from the

(ZOOL. CHAl.L. EXP. — PART LIII, — 1887.) Ggg 2S

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

above-described species, in exhibiting a body broader than long (PL XXXV. figs. 1, 2).
The breadth bore to the length an average proportion of 2 : 1, so that the whole body of
the sponge appeared as compressed from above downwards. The lateral margin is,
however, never sharp. From the middle of the superior gently convex surface, an
irregularly roundish, narrow and sharp, oscular fringe projects, consisting of a circle of
parallel rod-like spicules (PI. XXXV. figs. 1, 2). Through the roundish oscular opening,
one sees the bluntly rounded conus centralis projecting to the level of the aperture,
while from the sides of the former four cruciately disposed main septa extend laterally
and inferiorly. Between these cross septa the ramified efferent canals penetrate into
the parenchyma (PI. XXXV. fig. 2). The surface of the under side is but rarely
simply convex, as in PI. XXXV. fig. 1 ; it usually appears somewhat flattened or even
pressed slightly inwards. The portion from which the tuft of basal fibres springs is
however usually somewhat protruded. The spicules of the basal tuft are rather strongly
developed, in large specimens almost attaining the thickness of pins. They are, however,
not preserved throughout their whole length. Sometimes the bunch of spicules attains
the thickness of one's little finger (PI. XXXV. figs. 1, 2). The largest specimen measured
12 cm. in breadth, and about 6 cm. in height; the smallest was 5 cm. broad, and 3 cm.
high. Ten perfect specimens and some fragments were procured.

The parenchyma of the sponge is supported by the usual medium-sized smooth
oxyh exacts, and by numerous smooth, straight, oxydiacts with tolerably large central
nodes. More rarely a terminal swelling occurs (PI. XXXVI. fig. 2). Between the above
forms lie a great number of small rough oxyhexacts, with straight or curved rays
(PI. XXXVI. figs. 9, 10 ; PL XXXV. figs. 9, 10).

In the dermal skeleton somewhat large, smooth hypodermal oxypentacts occur, in which
the four tangential rays are not disposed at right angles to the radial proximal in a single
plane, but are bent somewhat inwards (PL XXXVI. fig. 8). The proximal ray is generally
three or more times longer than the tangentials. Externally the skin is thickly covered
with numerous autodermal pentact pinuli, with four somewhat short and thick basal rays
bearing small firm tubercles, while the moderately long, strong distal exhibits somewhat
strong lateral spines directed upwards (PL XXXV. fig. 7). The largest of the dermal
amphidiscs are about 0 '2 mm. in length, and occur somewhat sparsely; they bear a very
narrow, smooth, or slightly roughened axial rod, with four or eight lateral teeth. In
some cases these middle teeth are curved forward towards the distal umbel of rays
(PL XXXVI. fig. 3). The terminal umbels are somewhat long and bell-shaped, and consist
of eight to twelve lancet-shaped, but usually quite narrow rays (PL XXXV. fig. 4).
Besides these, medium-sized amphidiscs of similar form occasionally occur, and likewise
small forms with many-rayed hemispherical terminal umbels (PL XXXVI. fig. 6).

Of similar structure is the gastral skeleton on the free surface of the central cone,
on the main septa, and in the large gastral cavities, while it gradually assumes a difi'erent

REPORT ON THE HEXACTINELLIDA. 219

character in the tubular eflfereut canals, especially through the degeneration and final
disappearance of the supporting hypocanalicular pentacts, as also through the shortening
of the autocanalicular peutact pinuli, which further exhibit a less thickly spinose free ray
and longer roughened basal rays. Lastly, in the finer canals the pinuli gradually become
further and further distant from one another, until just before the diverticula
of the membrana reticularis only isolated very delicate pinuli occur (PI. XXXVI. fig. 1).
Numerous amphidiscs lie in the gastral-canalicular membrane, but do not penetrate as
far as the layer of chambers. The large form of amphidiscs (PI. XXXVI. fig. 3) is
here altogether absent, but the medium-sized, and especially the smaller are often present
in great abundance.

The marginalia form a closed ring on the free sharp edge of the oscular aperture.
They have the form of moderately long oxydiacts with perfectly smooth proximal, and
somewhat thickly spinose distal rays. On the boundary between the two rays, two
opposite, or rarely four cruciate tubercles or bosses almost always project. These are
usually hemispherical or somewhat longer than broad, or else completely roimded off.

The basal pad contains the familiar firm spicules with six to two rays, the ends of which
are beset with spines, while the inner portion remains smooth (PI. XXXV. fig. 8). The
strongly developed spicules of the basal tuft, almost as thick as pins, are in part smooth,
and in part studded with numerous nail-like barbs, which are not inserted on annular or
spiral cross ridges, but are disposed in isolated fashion in indistinct and: irregularly
developed spiral rows (PI. XXXVI. fig. 7).. Viewed from the surface, the freely pro-
jecting points of the teeth appear to be marked ofi" by a line of basal tubercles. Such a
bounding line does not really exist above, as may be readdy seen when viewed in profile,
but is merely the optical expression of the upper margin of the origin of the barbs on the
surface of the spicule. Towards the lower end the long spicules become markedly thinner,
the lateral barbs diverge further from one another, and the spicule ends below in a
peculiar hemispherical anchor sti'ucture with four double hooks cruciately disposed
(PL XXXVI. fig. 11), or in a hemispherical terminal knob from which four minute simple
cruciate pointed barbs project backwards, as in the- anchors of Hi/alonema tJwmsoni,
var. exiguum (PL XXXIV. fig. 16).

Several forms of Hyalonema, difiering only in subordinate characters from that above
described, and whoUy agreeing with it both in general organisation and in the structure of
the spicules, and therefore undoubtedly belonging to the same species, were dredged in
the Mid-Pacific (Station 271, lat. 0° 33' S., long. 151° 34' W.) from a similar depth of
2425 fathoms and from a Globigerina ooze bottom. There are two perfect specimens,
destitute only of the freely projecting portion of the basal tuft, and also the fraginent of
a third form, all procured at the same locahty. All the three seem somewhat smaller
than the above, and exhibit either a conical form with an inferior point, or a flat cake-like
shape. The breadth varies from 5 to G cm; the height from 3 to 4, and in the latter form

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

measures only 2 cm. In other respects the form and structure are not essentially
diflerent. The somewhat lirighter colour is rather striking, being in these three forms light
greenish-yellow, while those from Station 246 were dark dull greenish-brown. The
section in PI. XXXVI. fig. 1, is a diagrammatic representation of a longitudinal section
through the external marginal portion of one of the specimens from Station 271.

4. Hyalonema {Stijlocalyx) davigerum, n. sp. (PI. XLL).

In the neighbourhood of the Penguin Islands (Station 147, lat. 46° 16' S., long. 48° 27'
E.), from a depth of 1600 fathoms, and from a Diatom ooze ground, a fragment of a
Hyalonema was dredged, which is represented in PI. XLI. fig. 1. The fragment represents
about one-eighth of the whole body, and is without any trace of the basal tuft, but still
exhibits a portion of the oscular margin and one of the radial septa. The consistence of
this sponge fragment, which is about 5 cm. long, and represents a probably truncated
oval body, is not so great as that of Hyalonema globus. The whole external surface
distinctly exhibits a well-developed rectangular dermal framework.

The parenchymal skeleton consists again of somewhat large smooth and radially dis-
posed oxyhexacts, and of simple oxydiacts which are either smooth, or have a median
swelling or four projecting median knobs. Between these, thin forms equipped at both
ends with delicate barbs (PL XLI. fig. 6) occasionally occur, like those found abundantly
in the parenchyma of Hyalonema globus (PI. XL. fig. 5).

Besides these, minute delicate oxyhexacts occur in some abundance, exhibiting fine
straight pointed rays, on which a gentle curvature may, here and there, be detected
(PI. XLI. fig. 7).

Those oxyhexasters which are represented in large numbers in the diagrammatic
section of PI. XLI. fig. 2, and under high power in PI. XLI. fig. 5, do not really belong
to this sponge, nor indeed to the genus Hyalonema, but are forms which have been
accidentally imported.

The dermal skeleton is mainly composed of somewhat large smooth hypodermal
oxypentacts, with oblique, inwardly directed tangential rays. On the outer surface
numerous autodermal pentact piuuli occur, with rather long, smooth and pointed, basal
tangential rays, while the relatively short distal ray bears, as in Hyalonema globus,
long, fine, upwardly bent lateral spines, and a knob-like external terminal portion with
thick, somewhat pointed axial end, and so has a tufted appearance (PI. XLI. fig. 4).

The somewhat large amphidiscs which are irregularly scattered in radial disjaosition
within the external skin, exhibit a firm smooth axial rod with several (four or eight)
radially projecting tubercles in the centre. The bell-shaped terminal umbels are rather
broad, and measure about one-fourth of the total length. They consist of eight broad
shovel-shaped rays with lancet-like pointed ends (PL XLI. fig. 9). There is a much sparser

REPORT ON THE HEXACTINELLIDA. 221

occurrence of rather small elongated forms, with narrow umbel-rays (PI. XLI. fig. 10).
Very frequent, on the other hand, is the occurrence of the familiar small amphidiscs with
ten- or twelve-rayed, short, hemispherical umbels (PI. XLI. fig. 11).

In the gastral membrane and in the lining of the larger eff"erent canals the hypogastral
oxjrpentacts are absent, but strands of oxydiacts of various thickness, with or without
central thickening and projecting nodes occur abundantly. The autogastral pentact pinuli
are strikingly distant from one another, and bear long, pointed, somewhat spinose basal
rays, and a rather long spindle-shaped, thickened proximal with short, almost scale-like
teeth (PI. XLI. figs. 2, 8). Small amphidiscs with twelve- or more rayed, short, hemi-
spherical umbels occur abundantly in the gastral membrane. The cufi'-like limiting fringe
round the sharply defined oscular margin consists of oxydiacts, almost 1 mm. long, in
which the proximal ray is quite smooth and simple, and uniformly narrowed into a sharp
point, while the free distal ray, close above the smooth inner portion, is beset in fir-tree-
like fashion with scale-like or pointed spines, projecting oblic[uely outwards, and also
exhibits a spindle-shaped thickening (PI. XLI. fig. 3).

5. Hyalonema {Stylocalyx) globus, n. sp. (PI. XL.).

Near the Banda Islands, in the Malay Archipelago (Station 194a, lat. 4° 31' S., long.
129° 57' 20" E.), from a depth of 360 fathoms and volcanic mud ground, a specimen
of Hyalonema was dredged which difiered in some points from the species hitherto
described. The body which is almost globular measures 23 mm. in length and 22 mm.
in breadth. On the superior pole there is a round oscular aperture, 4 mm. in diameter,
with a sharp tangentially directed margin bounded by a thin fringe of spicules (PL XL.
fig. 1). Through this terminal opening the gastral cavity can be seen with the somewhat
pointed conus rising from below, and with four cruciately disposed radial septa, extending
from the lower portion of the conus to the margin, towards which they become broader.
The lower pole of the spherical body bears an apparently much damaged basal tuft.
The scanty spicules do not exceed 0*2 mm in thickness (PL XL. fig. 1). The whole
uniformly curved external surface of the body is covered with a dermal lattice-
work, the distinct meshes of which are approximately or perfectly square, and formed of
strands which are arranged like the lines of longitude and latitude on a globe.

The parenchyma is supported by strongly developed smooth medium-sized oxyhexacts,
and by oxydiacts of variable length and thickness disposed in strands. Besides these
smooth oxydiacts, forms occur in some abundance, in which not only both ends, but the
central portion, and rarely the whole spicule, are more or less thickly covered wath
centrally directed barbs. Four cruciately disposed projecting nodes occur in the centre,
and these are sometimes continued into curved hooks (PL XL. fig. 5). The parenchyma
further contains a considerable number of very slim, smooth oxyhexacts of variable size.

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

but usually about 0"2 mm. in diameter (PI. XL. fig. 10). The spicules represented
on PL XL. figs. 3, 11, 13, also occur here and there in the parenchyma, but do not really
belong to this species, being extrinsic importations from other forms, perhaps from a
Crateromorpha.

The basis of the dermal skeleton consists of strongly developed smooth oxypentacts.
The apposition of the tangential rays of the latter forms the square-meshed lattice-work,
which can be recognised even with the unaided eye (PI. XL. fig. 2). This gives this
sponge a greater compactness than is possessed by any other species of Hyalonema. The
radially directed proximal ray is longer than the tangentials. Adjacent to the latter
lie strands of smooth oxydiacts with central swelling or nodes. The autodermal
pentact pinuli exhibit comparatively long, straight, spinose basal rays, about 0"05 mm. in
length, while the somewhat short (about O'l mm.) projecting distal is characterised by
the long lateral spines, directed upwards and outwards, which begin at some distance
from the base, are somewhat distantly inserted in the middle portion, become more
closely apposed in the upper part of the ray, and finally form a kind of bud, in the
middle of which lies the end of the ray, in nowise thin or gradually pointed, but rather
thick and ending in a conical point (PI. XL. fig. 16).

The abundant and characteristic amphidisc in the skin is a large, strongly developed
form, 0"4 mm. in length, in which the thick smooth axial rod is centrally beset by eight
slim tubercles or bosses. The somewhat broad, but not very long terminal umbels are
not uniformly hemispherically arched, but are at their outer end slightly truncated. The
eight umbel rays with broad, blade-like, longitudinal bases, have a broad paddle-like form
and a rounded end (PL XL. fig. 7).

Besides these, there is a frequent occurrence of those small amphidiscs with hemi-
spherical, twelve- or more rayed terminal umbels, and with slim delicate axial rod, which
occur in the skin of all species of Hyalonema (PL XL. fig. 15). In the gastral membrane,
however, the hypogastral pentacts are either whoUy absent, or of very sparse occurrence,
while the strongly developed smooth oxydiacts and the associated pentact autogastral
pinuli are very abundant. The latter exhibit rather long slightly spinose basals, and a
free fir-tree-like ray, which runs out into a thin terminal point and bears short curved
lateral spines. I did not find any large amphidiscs in the gastral membrane, but the small
forms occur in great abundance.

The marginal fringe of the superior oscular aperture is formed of oxydiacts which
attain a length of only 1 mm. The freely projecting distal ray is beset with hook-shaped
externally curved spines, while the proximal bears only small pointed tubercles. At the
boundary between the two, on the thickest portion of the spicule, four cruciately disposed
hemispherical bosses project (PL XL. fig. 6).

In the basal pad numerous strong spicules with six to two rays occur, in which the
blunt ends are thickly beset with spines. It may be frequently observed, especially in

REPORT ON THE HEXACTINELLIDA. 223

the numerous tetraets, that the rays are somewhat curved, even in the cross plane (PI. XL.
fig. 12). The spicules of the basal tuft, which vary greatly in size, are smooth superiorly,
but inferiorly beset with isolated barbs, while the helmet-shaped thickened lower end
bears four long cruciate, flat or roundish, anchor teeth of variable length, directed upwards
and outwards (PI. XL. figs. 8, 9).

G. Hycdonema [Stylocalyx) elegans, n. sp. (PL XXXL figs. 1-7).

In the Mid Pacific, to the south-east of the Christina Islands (Station 271, lat. 0° 33' S.,
long. 151° 34' W.), from a depth of 2425 fathoms and a Globigerina ooze bottom, an
extremely delicate and soft Hyalonema was trawled. The compressed roundish inferiorly
pointed body has a thickness of about 2 cm., and bears a basal tuft of a few projecting
spicules which measure 12 to 14 cm. in length (PL XXXL fig. 1). At the upper end of
the body a central cone slightly projects, and from it four cruciately disposed radial
plates extend, dividing the central cavity into four divisions. The superior openings
of these cavities are seen as irregular roundish gaps on the upper surface. It is
improbable that a terminal sieve plate extended over these apertures; no trace of such
a structure persists. On the upper half of the body, indeed, the whole outer skin seemed
to have been rubbed away, while traces of the same are still recognisable on the lower
somewhat bulging conical portion.

The loose parenchyma contains not only simple smooth medium-sized oxyhexacts,
and numerous smooth (or rarely terminally roughened) (PL XXXL fig. 3) oxydiacts of
medium size and inconsiderable length, but also a great number of small smooth oxy-
hexacts with curved rays.

The dermal skeleton (PL XXXL fig. 2) consists of somewhat strongly developed,
smooth oxypentacts, on which numerous autodermal pentact piuuli are apposed. The
latter exhibit four moderately long, terminally slightly toothed basal rays, and a more
or less spinose distal (PL XXXL fig. 4) which may attain a length of 0"5 mm. Besides
these, numerous medium sized eight-rayed amphidiscs occur, bearing on their axial rod
several somewhat irregularly distributed boss-like, projecting, lateral tubercles, and
exhibiting straight, bell-shaped terminal umbels which are so long that their flat lancet-
shaped rays almost meet one another (PL XXXL fig. 6). There are also numerous
minute amphidiscs with slightly arched short umbels, 0"02 to 0'04 mm. in length
(PL XXXL figs. 5, 7). Isolated large broad amphidiscs also occur, but I am doubtful
whether they really belong to this species. From their scarcity and irregular disposition it
seems probable that they owe their origin to one of the other species of Hyalonana which
lay in the same glass and which contain these broad amphidiscs in abundance. The
basal tuft exhibits numerous firm spicules with six to two rays, which are straight and
terminally beset with strong teeth, or else truncated or rounded.

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

The long spicules of the tuft are quite smooth and are all broken at their
lower end.

7. Hyalonema {Styloccdyx) tenerum, n. sp. (PI. XXXI. figs. 8-20).

In the South Pacific (Station 289, lat. 39° 41' S., long. 131° 23' W.), from a depth of
2550 fathoms and a red clay bottom, a delicate and very loose Hyalonema was dredged,
which exhibited a pear-shaped body 4 cm. ia length, and 25 cm. in maximum breadth.
In the middle of the upper, very injured extremity a pointed cone projects, from
the sides of which four cruciately disposed longitudinal septa extend to the side wall,
separating the central cavity into four chambers. From the lower, somewhat pointed
end, a narrow basal tuft projects for about 16 cm. It is composed of about fifty spicules
which are wound spirally together, and which exhibit only at the very extremity a
somewhat straighter, and more separate course. The general form of the body, as repre-
sented on PI. XXXI. fig. 8, has obviously been very considerably altered by damage
done to the upper half. The latter appears to have been much rubbed away, and the
superior external margin of the upper terminal surface has been wholly destroyed.

The spicules of the parenchyma consist for the most part of medium-sized smooth
oxydiacts, beside which there occasionally occur, especially near the outer wall, smooth
knobbed or terminally toothed diacts, and sometimes even monacts (PL XXXI. fig. 10).
Besides these long rod-like spicules, medium-sized straight oxyhexacts occur in consider-
able abundance in the parenchyma, while small oxyhexacts with curved smooth rays
(PI XXXI. figs. 9 and 17) are very prevalent, and exactly resemble the small curved
oxyhexacts of Hyalonema elegans.

The dermal skeleton contains the often mentioned, moderately large, smooth hj^o-
dermal oxypentacts without a trace of a (sixth) distal ray. Numerous large autodermai
pentact piuuli also occur — with four rather long basal tangential rays — strongly developed
and with toothed pointed ends, while the extraordinary long (1*6 mm. and more), radially
projecting distal is beset with short teeth except in the smooth basal portion
(PL XXXI. fig. 11).

The dermal membrane contains numerous short broad amphidiscs, which exhibit a
perfectly smooth, strongly developed axis-rod, bearing hemispherically arched, broad,
terminal umbels with smooth, paddle-shaped rays almost meeting in the middle, and
occurring in variable number from eight to twelve, but usually ten or eleven (PL XXXI.
figs. 12, 13, 20). The interior side of the smooth, terminally rounded, paddle-shaped
umbel rays exhibits an inward projecting median ridge, ending in an elevation before
the extremity of the ray. In PL XXXI. fig. 18 one of these abnormal spicules is
represented, with four rays, as occasionally occurs in addition to the form with six.
Besides these large almost globular amphidiscs, in which the opposite paddle-shaped umbel

REPORT ON THE HEXACTINELLIDA. 225

rays arc frequently not exactly opposed, but more or less regularly alternating (PI. XXXI.
fig. 12), more elongated medium-sized forms also occur, which are sharply distin-
guished from the above, by the fact that the umbel is not hemispherical but termin-
ally truncate, and furnished with more straightly stretched, obliquely inserted, terminal
rays (PI. XXXI. fig. 14). A somewhat smaller form also occurs, in which the umbel
rays are not broadened out in paddle-like fashion, but exhibit a narrow point, and in which
the axial rod is more or less thickly beset with lateral spines. Finally, the familiar small
amphidiscs, with narrow, smooth, axial rods and hemispherical many-rayed (sixteen to
twenty) umbels (PI. XXXI. fig. 16), also occur in tolerable abundance, varying in the
different regions of the body. In the lowest terminal portion of the body, in the so-called
basal pad, strongly developed spicules with six to two rays occur as in all species of
Hyalonema. The rounded or truncate ends of the rays are beset with conical teeth or
spines, extending to a greater or less distance interiorly (PI. XXXI. fig. 15).

The long spicules of the basal tuft exhibit spiral or annular, obliquely projecting ridges
and rows of teeth, which are inserted on the projecting edge of the ridges.

There now follow those species of Hyalo7iema, the upper end of which was not
sufficiently preserved for deciding the question, whether there is a sieve-plate or not.

1. Hyalonema lusitanicum, Barboza du Bocage (PI. XXVIII. figs. 12-17 ;
PL XXXIX. figs. 16-18).

The specimen which Barboza du Bocage obtained in the neighbourhood of Setubal in
Portugal, and which he described and figured ^ under the title Hyalonema lusitanicum,
consisted of a spirally twisted tuft of spicules, the narrower end of which was for a length
of 16 cm. and also at the terminal point, completely covered by a continuous layer of
Palythoa, while in the longer free portion the spicules diverged in bushy tufted
fashion. Neither on this specimen, nor on others which Bocage afterwards obtained
from the same locality, could any portion of sponge-body be detected. The following
diagnosis was given : — " Hyalonema polypario elongato, fibris setaceis hyaliuis spiral-
iter tortis, corio pol}q)igero ad apicem usque ad f longitudinis totius involutis ;
polypis dilatatis, ellipticis, valde aggregatis, parum elevatis, per series longitudinales
ac spirales regulariter digestis." Barboza must afterwards have found more perfect
forms in which the sponge-body was preserved, for I found in the British Museum in
South Kensington, London, a spii-it specimen with the label Hyalonema lusitanicum,
Barboza du Bocage, and described as a gift from the discoverer. The specimen was 56
cm. in length, and exhibited on the much damaged upper end a projecting tuft.

1 Proc. Zool, Soc. Land., 18C4, p. 205.
(ZOOL. CHALL. EXP. — PART LIII. — 1887.) Ggg 29

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

I was able to procure several small fragments of this original specimen, and the results of
my investigation seem to warrant me in amplifying the specific diagnosis of Hyalonema
limtanicum. From my own notes and from some subsequent observations which Mr.
Ridley was kind enougli to make for me, the length of the much damaged body is 12
cm., the breadth only 27. The much twisted basal tuft of spicules has a breadth of
15 mm. The Palythoa crust is not present having been probably separated off.

In the absence of a marginal fringe and of the probably originally present terminal
sieve-net, it is impossible to determine whether the tuft of spicules, projecting for about
1 cm. on the upper end, and doubtless representing a damaged conus, really projected in
the intact form.

Another form in the British Museum, which is said to have been brought from Porto
Rico, exhibits a body transversely truncated above, with a well-developed terminal sieve-
net, and with an indication of the conus centralis. The basal tuft here attains a length of
2G cm. and is covered with the Palythoa crust. Whether this second specimen really
belongs to Bocage's Hyalonema lusitanicum, I am not able to determine, since I have not
been able to analyse the structure of its component spicules.

The following description of the skeleton refers only to the first mentioned original
specimen of Hyalonema lusitanicum, Bocage.

The supporting spicules of the parenchyma consist for the most part of diacts of
variable size and thickness. These are somewhat pointed at either end, and usually
exhibit a simple, spindle-shaped, perfectly smooth course, or occasionally bear a single
tubercle, or there may be four or two, projecting from the middle point. They are generally
somewhat curved, though frequently perfectly straight. Between these diacts medium -
sized oxyhexacts occur, though on the whole not very abundantly. They exhibit six
simple, somewhat pointed, straight and smooth rays, which are generally of equal
length. Sometimes, however, four rays cruciately disposed are of uniform length, while
the other two lying in the third axis, that is to say, radially disposed, are somewhat
longer. I also found isolated monacts witlt one end thickened into a knob or rounded
off like a button, while the other ended in a point. Scattered through the whole
parenchyma in relative abundance there are small, smooth or slightly roughened
oxyhexacts, with raj^s which are seldom perfectly straight (PL XXVIII. fig. 16), but
usually bent in the familiar fashion, so that three approach each other, and the three
antipodals likewise (PI. XXVIII. fig. 15).

The middle-sized hypodermal oxypentacts which serve to support the skin are quite
smooth, their rays run terminally to a somewhat marked point. The autodermal pentact
pinuli, which are usually inserted on the latter, bear four short, strongly developed, tan-
gential basal rays, with blunt, slightly toothed ends, and a strong distal, about 0"3 mm. in
length, which has an almost bushy appearance due to the long lateral processes (PI. XXVIII.
fig. 17). The majority of the numerous, radially disposed, dermal amphidiscs have a

REPORT ON THE HEXACTINELLIDA. 227

lenctli of O'l mm., a long, narrow, axial rod with central teeth, and at each end, a narrow
bcll-shaj^ed umbel, occupying about a third of the total length, and provided with eight
rays, which have no great breadth, and which have terminally a rather straight course
(PL XXVIII. fig. 13). More rarely larger amphidiscs occur, of similar form, but with a
length of 0-2 mm., and a proportionate breadth (PL XXVIIL fig. 14). Besides these
numerous small amphidiscs occur from 0'03 to 0'02 mm. in length, exhibiting the same
form as the above, but lying tangentially in the dermal membrane (PL XXVIIL fig. 12).

I was not able to examine the spicules of the marginal fringe. The pinuli of the
gastral surface and of the associated diverticula difi'er from those of the dermal region in
this, that the distal ray is slimmer, shorter, and not tufted, and that the four basal rays
are both relatively and absolutely longer than those of the dermal pinuli. In the basal
pad at the lower end of the sponge body, there are numerous strongly developed
spicules with a variable number of rays which exhibit toothed rounded ends (PL XXVIII.
figs. 10, 11.)

Among the Challenger material there is a basal tuft of strong spicules much encrusted
with Palythoa, which bears a striking resemblance to the specimen figured by Bocage
{Joe. cit.), and very probably belongs to the species Hyalonema lusitanicum. I have
figured it in its natural size in PL XXXIX. fig. IG, and have investigated the few
persisting remnants of the sponge-body. The form was obtained at Station IV. to the
south of Cape St. Vincent, lat. 36° 25' N., long. 8° 12' W., from a depth of 600 fathoms
and a blue mud ground. From the upper end of the straight cylindrical form, which
measures 20 cm. in length by 10 mm. in breadth, some broken basal spicules project. From
the lower end shreds of a Palythoa encrustation hang down, but the bundle of spicules is
broken off. That the somewhat conical point turned upwards in the figure, is really
the superior pole, which supported the sponge-body, and had a few projecting spicules
inserted in it, is best demonstrated by the examination of the few small remnants of
the lost body, which persist loosely bound up with the Palythoa crust. The siliceous
elements recognisable in the traces of the basal pad are firm spicules with six to two
terminally rounded, cylindrical rays (PL XXXIX. fig. 17), which are beset with conical
elevations of variable size either throughout their whole extent or only at their ends.
On the larger conical elevations, small secondary teeth very frequently occur (PL XXXIX.
fig. 18).

Unfortunately neither the results attained by Barboza du Bocage in his original
research on Hyalonema lusitanicum, nor those of my investigation of the above described
form, which is probably to be identified with Bocage's, are sufiicient to establish any
definite diagnosis of the form and structure of the species. Whether the much injured
specimen which I examined in the British JNIuseum is really the same as the root-tuft
described by Bocage in 1864 is likewise doubtful, although the label certainly referred
the specimen to the discoverer himself.

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

I ought not, however, omit to mention that in the Berlin Zoological Museum there is a
well preserved Hyalonema form, obtained from Sir Wyville Thomson, and labelled as
Hyalonema Itisitanicum, Bocage. This specimen figured in PI. XXXIV. fig. 1, does not,
however, in its form and in its spicules, resemble the Hyalonema of the British Museum,
but agrees in all essential points except size, with the Hyalonema thomsoni, Marshall.
I was therefore compelled to describe it under the latter designation.

2. Hyalo7iema cebuense, Higgin.

The species of Hyalonema which Higgin described in 1S75,* under the title Hyalonema
cebuense, is not represented among the numerous forms of the Challenger collection.
According to Higgin's account the form described must be a distinct species, and not — as
Carter seems inclined to make out in his remarks on Higgin's article — a mere variety of
Hyalonema sieholdii. The most important characteristics of the species are: — (l) the
peculiar sculp tor's-mallet-like form of the large body, which measures more than 1 4 cm.
in length and 12 cm. in breadth; (2) the strikingly thick and short, straight, smooth oxy-
diacts, which are said to belong to the dermal skeleton;" (3) the peculiar small parenchy-
mal oxyhexacts, which bear hook-shaped lateral spines at the ends of the rays ;^ (4) the
slim double-harpoon-shaped oxydiacts, bearing all along both rays curved teeth turned
towards the centre,* and finally (5) the presence of four blunt anchor teeth on the long
spicules of the tuft.^

It is not evident from Higgin's otherwise accurate and detailed description whether
a terminal sieve-net covers the gastral cavity, or whether the latter remains open, nor
does he note the presence or absence of a fringe of marginal sj^icules. From the fact
that no polype encrustation was observed on the upper end of the fibrous tuft, one
cannot of course certainly infer its absence during life, since it might readily have been
separated ofi".

3. Hyalonema tenue, n. sp. (PI XXX. figs. 1-8).

To the east of the mouth of the Rio de la Plata (Station 323, lat. 35° 39' S, long.
50° 47' W), from a depth of 1900 fathoms and a blue mud ground, two fragments of a
Hyalonema were dredged, the siliceous elements of which appear to me sufficiently
characteristic to justify the erection of a distinct species. One of the fragments has the
form of a thin irregular plate about 2 cm. square in the middle of which a normal oval
space, 6 mm. long by 4 broad, can be detected (PI. XXX. fig. 1). This opening is

^ Ann. and Mag. Nat. Hist, ser. 4, vol. xv. p. 377. '^ Loc. cit., pi. xxii. fig. 1.

3 Loc. cit, pi. xxi. fig. 5. « Loc. cit, pi. xxi. fig. 3.

' Loc. cit., pi. xxi. fig. 9.

REPORT ON THE HEXACTINELLIDA. 229

surrounded by a circle of radially disposed straight spicules, and as examination witli a low-
power at once reveals, represents an oscular opening. The other fragment consists of a
somewhat thicker (about 3 mm.) plate, with irregular ragged edges. On one surface
smooth-edged oval or roundish pits occur, from which narrow diverticula pass towards the
side and towards the interior (PI. XXX. fig. 2). The other surface is rough and
apparently damaged. This second specimen undoubtedly forms part of the internal wall
of the body, and exhibits the surface limiting the gastral cavity and bearing the openings
of the canals.

An examination of the siliceous elements shows, in the first place, that they are all
strikingly slim and delicate, whence my choice of the specific designation tenue. Among
the long spicules of the parenchyma, between the numerous slim oxydiacts, one remarks
an unusual abundance of oxyhexacts of medium and small size (PL XXX. fig. 8), with
thin, gradually pointed smooth rays. Besides these, though not so abundantly, small
oxyhexacts occur with perfectly smooth curved rays (PI. XXX. fig. 5). In the external
skin, on the smooth hypodermal oxypentacts, and on the tangential oxydiacts, numerous
autodermal pentact pinuli are inserted, with four smooth ba-sal rays which are relatively
long (0"2 mm.), while the slim, finely pointed distal, beset with minute obliquely inserted
teeth, attains a length of almost 1 mm. The amphidiscs of the skin are of medium size
(from 0"2 to 0"! mm. in length) and with simple, straight-rayed, eampanulate, medium-
sized umbels, occupying one-third of the total length. The umbels never have eight,
but always thirteen or twelve smooth, lancet-shaped, umbel rays. The axial rod is beset
with nodes, which attain especially in the middle a considerable prominence and a
cruciate disposition. The marginalia, which are radially grouped round the oscular
opening, attain a length of 1^ to 2 mm. Their pointed proximal end is smooth, and
not so long as the likewise pointed distal, which has a length of at least 1 mm. and
is beset with oblique lateral teeth. On the boundary between the two rays, four
cruciately disposed, globular tubercles project, the rudimentary representatives of the four
tangential rays, as is readily demonstrated by the presence of a well-developed axial canal
(PI. XXX. fig. 7).

4. Hyalonema robustum, n. sp. (PL XXXII. figs. 1-10).

In the west of the North Pacific .(Station 241, lat. 35° 41' N., long. 157° 42' K),
from a depth of 2300 flithoms and a red clay bottom, the fragment of a Hyalonema
was dredged, which had the form and size of half an apple, and evidently represented
about half the entire body of a sponge. On the convex external surface of the hemi-
spherical specimen, the dermal membrane is well preserved as a fine-meshed network,
while the torn internal surface exhibits a loose, wide-meshed parenchyma, ■with cavities
increasing in width towards the interior.

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

Although I am not usually inclined to bestow specific distinction on such an imper-
fect specimen, it seemed to me justifiable to make an exception in favour of this form,
on account of the marked breadth and strength acquired by the larger almost spherical
amphidiscs, and also because of other peculiarities among the spicules.

The large supporting spicules of the parenchyma are more or less long diacts
of variable strength, besides which isolated medium-sized, strong oxyhexacts also
occur, as also very aljundaut slimmer medium-sized oxyhexacts, in which the rays
are covered with slender spines inserted at right angles (PI. XXXII. fig. 8). Between
these a host of small smooth oxyhexacts with curved rays also occur (PI. XXXII.
fig. 6).

The dermal skeleton contains strongly developed hypodermal oxypentacts with
proximal rays of variable length. On the tangential rays are inserted rows of autodermal
pentact pinuli, with four short, strongly developed tangential basal rays, beset with short
distaUy directed teeth, while the distal ray, about 0'3 mm. long, has a more bushy appear-
ance, due to somewhat long obliquely inserted spines which are specially weU-developed
on the median and external portions (PI. XXXII. fig. 7). Especially striking in the
dermal membrane are the very broad, almost spherical, strongly developed amphidiscs
(PI. XXXII. figs. 1, 2), with broadly arched terminal umbels which almost meet in the
middle. Each umbel consists of eight to twelve broad, paddle-shaped rays, which exhibit
a median ridge projecting inwards and rounded off terminally. They do not pass into a
terminal plate at the pole, but end sharply with a projecting margin round a central
circular pit (PI. XXXII. fig. 2). The opposed umbel rays are usually not exactly
ojjposite but alternating (PI. XXXII. figs 1, 2).

Besides these, the dermal membrane contains other amphidiscs, with eight slender
umbel rays of smaller, and sometimes of much smaller size (PL XXXII. figs. 4, 5). The
small umbels only occupy a third of the total length of the amphidiscs, or even less in
the smallest forms (PI. XXXII. fig. 5).

On the inner surface of the large internal cavity, in the gastral and canalicular mem-
brane, numerous amphidiscs occur of medium and small size. They exhibit in part the
same structure as those last described in the external skin (PI. XXXII. fig. 4), and in
part somewhat larger forms (PL XXXII. fig 3). Those large, almost spherical amphidiscs,
which occur so abundantly in the dermal membrane, are here altogether absent, nor have
I found any hj^ogastral pentacts or autogastral pinuli in the gastral or canalicular
membrane.

In the lower portion of the body, which forms an annular pad, strongly developed
spicules with six to two cylindrical rays occur. Their rounded ends are beset for a
variable distance with teeth and spines (PL XXXII. figs. 9, 10), which often appear on
the rudiments of undeveloped rays (PL XXXII. fig. 10).

REPORT ON THE HEXACTINELLIDA. 231

Hyalonema species diversse indefinitai (PI. XXXII. figs. 11-lG; PI. XXXIX.
figs. 1-15).

Of the above-described species of Hyalonema, more or less perfectly preserved
specimens were at my disj^osal. so that an approximately accurate investigation was
possible, and diagnoses sufficient for the determination of the several species were
obtained. But besides the above, the Challensrer collection included several fragments
which could be certainly enough referred to the genus Hyalonema, but not with any
definiteness to any of the species pre\aously known, or now established by my investiga-
tions. They belong to new, not yet described species. Following the frequent precedent
of many investigators, I might have selected the most prominent characteristics of these
forms, and used them for the erection of distinct species. I have, however, preferred to
refrain from designating these imperfect and torn fragments, of which I should be unable
to frame a description or diagnosis sufficient for their re-identification. I have, nevertheless,
figured several of the more remarkable spicular structures of sponges from these localities
for the use of future investigators, and by way of extending our knowledge of the variety
of form within the genus Hyalonema.

Several large but insufficiently preserved fragments, referable to a sponge form
about the size of one's fist, were dredged from the Mid-Pacific, near the Equator (Station
271, lat. 0° 33' S., long. 151° 34' W., where Hyalonema dejyressum was found), from a
depth of 2425 fathoms and a Globigerina ooze ground. The smooth oxyhexacts and
oxydiacts of the parenchymal skeleton do not exhibit any marked peculiarities. Small
oxyhexacts are only represented by somewhat abundant forms with straight or slightly
curved pointed rays (PL XXXII. fig. 15). The complete absence of small curved
oxyhexacts essentially distinguishes this form from the species Hyalonema depressum
found in the same locality. The hypodermal oxypentacts and oxydiacts of the
dermal skeleton resemble, in form and size, the similar forms of Hyalonema clavigerum.
The autodermal pentact pinuli are, however, peculiar, and resemble most closely
the external pinuli of Hyalonema rohustum. Like the latter, they exhibit short,
strongly developed, and minutely spinose basal rays, and also a free, spindle-shaped,
somewhat bushy distal, almost 0"4 mm. in length, beset with numerous rather long
spines disposed like scales. The very end of the ray frequently exhibits a strongly
developed, thick point, which is not very satisfactorily represented on PI. XXXII. fig. 16.

The much extended, larger and medium-sized amphidiscs of the skin, are distinguished
by the slender, delicate character of the slightly toothed axial rod, and also by the long-
umbels, which bear eight slender rays, and exhibit a campanulate form, somewhat trun-
cated terminally. These forms of amphidisc differ greatly in size, as may be seen from
some of the modifications represented in PI. XXXII. figs. 11, 12, 13. Occasional
giant-forms (fig. 11), 0'4 mm. or more in length, occur. Besides these, there is an abun-

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

dant occurrence of tlie familiar small ampliidiscs with hemispherical many-rayed terminal
discs (PI. XXXII. fig. 14).

No hypogastral pentacts occur in the gastral skeleton, but their place is taken by
strands of supporting diacts. The autogastral pentact pinuli are essentially different from
the autodermal. They are not only decidedly shorter, and in no way so bushy as the
latter, but they exhibit longer, less tuberculated basal rays, and are continued into a
slender, pointed, terminal ray.

The oxydiacts, which form the long marginal fringe, bear a smooth, proximal ray,
and a distal, beset with obliquely inserted spines, directed downwards and outwards.
From the centre two or more, rarely four, rounded, or slightly pointed, lateral tubercles
project.

A much torn fragment with isolated, broken tuft spicules, was trawled to the west of
Cape York (Station 184, lat. 12° 8' S., long. 145° 10' E.), from a depth of 1400 fathoms,
and a Globigerina ooze ground.

From the size of the shreds, and the thickness of the tuft spicules, it was to be
inferred that the intact body had the size of an average apple. There are no definite
indications as to the original form -of the body.

In the parenchyma, besides long, narrow, smooth oxyhexacts and oxydiacts, small
lank oxyhexacts occur, some with straight, and others with bent, smooth rays, though
neither can be said to be abundant.

The dermal skeleton exhibits large, smooth, hypodermal oxj^^entacts, with tangential
rays turned somewhat inwards, and long, slim, slightly spinose autodermal pentact pinuli
w^th moderately long, and terminally spinose basal rays (PL XXXIX. figs. 5, 7). Amphi-
discs of very various sizes and forms also occur. The larger exhibit a strongly developed
axial rod beset with spines, and bear rather broad, campanulate umbels, composed of
eight broad, paddle-like, terminally broad and rounded rays (PI. XXXIX. figs. 2, 3).
The length of these rays varies in relation to the total length of the amphidiscs, so that
they sometimes attain to only a third of the whole length, at other times to about a half,
and sometimes almost meet in the middle. The abundant medium-sized amphidiscs
(PI. XXXIX. figs. 4, 9) are slimmer, and bear narrower umbel rays in variable number.
Each umbel usually consists of eight rays, liut ten or twelve rays also occur. Finally,
as in all species of Hyalonema, small amphidiscs with hemispherical many-rayed
terminal umbels occur in great abundance. Some oxydiact marginalia were observed
with smooth proximal, and shortly toothed, pointed, distal rays, while on the boundary
between them, two or four rounded or slightly pointed tubercles projected (PI. XXXIX.
fig. 1).

In the South Atlantic, to the west of Tristan da Cunha (Station 333, lat. 35° 36' S.,

REPORT ON THE HEXACTINELLIDA. 233

long. 21° 12' W.), from a depth of 2025 fathoms, and a Globigerina ooze bottom, a small
irregularly torn fragment of Hijalonema was trawled. A small portion of the marginal
fringe and of the superior sieve-plate was fortunately preserved. The parenchymal
skeleton includes smooth, medium-sized oxyhexacts, smooth oxydiacts of various sizes,
and small, somewhat roughened oxyhexacts, with straight rays. The hypodermal pentacts
are strongly developed, of medium-size, and with smooth tangential rays, directed some-
what inwards. The autodermal pentact pinuli are slim and of moderate size. Their basal
rays are rather long, and slightly spinose (PI. XXXIX. figs. 13, 15).

The rather sparsely distributed larger amphidiscs of the skin are not very long (0-23 mm.
and less), but with strong smooth axial rods and flatly arched, broad, short umbels, which
consist of eight or seven broad paddle-shaped rays, pointed at the end like a Gothic arch
(PI. XXXIX. fig. 10). Medium-sized amphidiscs of similar structure are more abundant,
while the small forms (PL XXXIX. fig. 11) with somewhat longer and more campanulate
umbels also occur, and finally the familiar small amphidiscs with hemispherical umbels of
ten to thirteen rays.

To the west of Luzon (Station 205, lat. 16° 42' N., long. 119° 22' E.), from a depth
of 1050 fathoms and a blue mud ground, a ragged fragment of a Hyalonematid was
obtained, which is in many respects very divergent from the hitherto described forms of
Hyalonema. It is improbable, in fact, that it belongs to the genus, but rather to a type
not yet investigated. From the nut-like fragment obtained, it was, however, impossible
to draw any definite conclusions as to original shape and size of the sponge, except
that it was probably about the size of an average apple.

It is very remarkable that the supporting spicules of the parenchyma are exclusively
long slim pliable ox}i;riacts, with two long rays lying in straight line, and a third much
smaller disposed at right angles accross the middle. Besides these, I have found only
small oxydiacts and some oxytetracts, whose rays did not, however, lie in one plane, A
large number of minute, very delicate oxyhexacts, with rather long, uniformly slender,
straight rays also occur.

The dermal skeleton with its hypodermalia I have not been able to find, probably
because of the absence of that layer. Autodermal or autogastral pinuli are, however,
present in abundance. They are characterised by their medium-sized and somewhat
bush}^ distal, and by the moderately long slender basal rays, which are somewhat
thickened and toothed towards the end, though terminating in a sharp point. The
amphidiscs vary greatly in size. Giants of 0'5 mm. occur, medium-sized forms one-
third to one-fifth as large, and the ordinary small type. Most of the medium-sized
amphidiscs have deeply campanulate terminal umbels with nine to four rays. I
observed some large forms in which the opposite umbel rays had fused.

(ZOOL. CHALL. EXP. —TART LIII. — 1887.) Ggg 30

23-4 THE VOYAGE OF U.M.S. CHALLENGER,

Genus 2. Pheronema, Leidy (Pis. XLII.-XLVL).

Holtenin, Wyv. Tliomson.

1868. Leidy, Proc. Acad. Nat. Sci. Philad., p. 9 {Pheronema annse).

18C9. Wyville Thomson, Ann. and Mag. Nat. Hist., vol. iv. p. 112.

18G9. WyviUe Thomson, Phil. Trans., vol. clis. p. 707 {Holfenia carpenteri).

1870. Leidy, The American Naturalist, vol. iv. p. 17 (March) {Pheronema annx).

1870. Kent, Monthly Micr. Journ., p. 243.

1870. Kent, Ann. and Mag. Nat. Hist., vol. vi. pp. 182-186. f Pheronema carpenteri and

1870. Gray, Ann. and Mag. Nat. Hist., vol. vi. pp. 309-312. ( Pheronema grayi.

1870. Schmidt, Die Spongien des atlantischen Gebietes {Holtenia pourtalesii).

1871. Carter, Ann. and Mag. Nat. Hist., vol. vii. pp. 112-141.

1871. Ehlers, Zeitschr. f. wiss. Zool., Bd. xxi. p. 540.

1872. Gray, Ann. and Mag. Nat. Hist., vol. ix. pp. 442-461.

1873. Carter, Ann. and Mag. Nat. Hist., vol. xi. p. 234.
1873. Carter, Ann. and Mag. Nat. Hist., vol. xi. p. 275.

1873. Carter, Ann. and Mag. Nat. Hist,, vol. xii. p. 349.

1874. Meyer, Ann. and Mag. Nat. Hist., vol. xiii. p. 66.

1874. Gray, Ann. and Mag. Nat. Hist., vol. xiii. p. 284.

1875. Carter, Ann. and Mag. Nat. Hist., vol. xv. p. 113.
1875. Carter, Ann. and Mag. Nat. Hist., vol. xvi. p. 1.

1875. Marshall, Zeitschr. f. vriss. Zool., Bd. xxv., Suppl., p. 142.

1876. Marshall, Zeitschr. f. -wiss. Zool., Bd. xxviL p. 113.

1877. Zittel, Abhandl. d. baier. Akad. d. "Wiss.

1877. Marsliall and Meyer, Mittheil. d. Zool. Mus. in Dresden.

1877. Wyville Thomson, The Atlantic.

1880. Schmidt, Spongien des Meerbusens von Mexico.

1880. Norman, Ann. and Mag. Nat. Hist., vol. vi. p. 430.

1881. Milne-Edwards, Comptes rendus, vol. xciii.; Ann. and Mag. Nat. Hist., vol. ix. p. 37.
1881. Duncan, Journ. Roy. Micr. Soc, ser. 1, p. 173.

1881. Carpenter, The Microscope.

1885. Filhol, La vie au fond des mers, p. 285 {Pheronema parfaiti).

History. — In 1868^ Leidy made a report on a sponge from Santa Cruz, which presents
an oval form about 3 inches in size, and bears on one extremity a tuft of siliceous spicules
about 2 inches in length, and differing from that of Hyalonema only in its smaller
size. Afterwards he described this new form more accurately under the name
of Pheronema annw,^ and he repeated his communication in 1870 in a paper
illustrated by woodcuts and entitled Remarks on some Curious Sponges.^ The latter
description ran'* thus : — " The body of the sponge is oblong ovoidal, with one side more
protuberant than the other. The narrower extremity, which I suppose to be the upper,
is conical, and its truncated apex presents a single circular orifice, the third of an inch in
diameter. The opposite extremity is rather cylindrical, with a broad, slightly rounded

1 Proc. Acad. Nat. Sci. Philad., 1868, p. 9.

2 Proc. Acad. Nat. Sci. Philad. (1868).

3 American Naturalist, vol. iv. pp. 17-22, fig. 1. * Loc. cit., pp. 21, 22.

REPORT ON THE HEXACTINELLIDA. 235

extremity, from which project numerous fascicles of siliceous threads. The sponge body
is of a light brown hue, and rigid to the feel. Its surface exhibits an intricate interlace-
ment of the sponge tissue, which appears mainly composed of stellate, siliceous spicules
of various sizes. The coarser spicules of the surface have five rays. Four of these
together are irregularly cruciform, while the fifth projects in a direction opposite to all the
others. They appear to be so arranged that the crucial rays interlace with those of the
contiguous spicules forming a lattice-work on the surface of the sponge, while the odd ray
opposed to the others penetrates the interior of the sponge. The finer tissue, seen
through the intervals of the latticed arrangement on the surface of the sponge, appears to
I)e made up in the same manner of finer stellate spicules. Some of the longest stellate
spicules of the surface have a spread of half an inch.

" The fascicles of siliceous threads projecting from the body of the sponge are upwards
of twenty in number, and over 2 inches in length. They resemble in appearance tufts
of blonde human hair. The individual threads are nearly like those proceeding from the
lower end of Euplectella. Where thickest, they are less than the ^^^ of an inch in
diameter, and become attenuated towards the extremities. At first, as they proceed from
the body of the sponge, they are smooth and then finely tuberculate. The tubercles
are gradually replaced by minute recurved hooks, which become better developed
approaching the free end of the threads, which finally terminate in a pair of longer
opposed hooks, reminding one of the arms of an anchor. The objects of the tufts of
threads, with their lateral booklets and terminal anchors, would appear to be to maintain
or moor the sponge in position in its ocean home."

Without knowing of these publications by Leidy, Wyville Thomson had in 1869
given a detailed and thorough account of a sponge designated Holtenia carpenteri,
several specimens of which he had collected to the north-west of Scotland, from a
depth of 530 fathoms in the course of his deep sea investigations in the " Lightning."
The form of these specimens, which were about as large as one's fist, was " globular,
elliptical or subcylindrical." A spacious and tolerably smooth central cavity, half the
diameter of the sponge, diminishes superiorly to a round space of perfectl}'' uniform
breadth, while from the closed under side a tuft of numerous individual clusters of
tolerably long, bow-shaped siliceous spicules doubtless rooted in the sand, runs out
in a radial direction. On the lateral wall, and most abundantly near the upper surface,
long radially directed spicules project freely, while the boundary of the wide superior
osculum is formed by a compact wreath of vertical spicules measuring about 1 cm. or
more in length. Both on the outer and inner surfaces of the tolerably firm cup-shaped
sponge body, a firm rind is formed of strong compact siliceous spicules which lie in
the skin, and are grouped together in stellate fashion, while the parenchymal mass itself
is penetrated by a system of numerous cavities and canals. The siliceous spicules which
penetrate and cover the entire organism resemble generally those of Hyalonema. The

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

form presents, however, so many peculiarities that the erection of a special genus seems
to be justified, even apart from the different form of the body and its microscopic
structure.

In a Notice of a New Vitreous Sponge — Pheronema grayi ^ — Saville Kent first
justified the change of the generic title Holtenia, which had been applied by Wyville
Thomson to his Holtenia carpenteri, into Pheronema (Leidy), on the ground of priority,
since a generic agreement between the two species Pheronema annae, Leidy, and Holtenia
carpenteri, Wyville Thomson, cannot be doubted. He now brought forward a third
species — Pheronema grayi — belonging to the same genus Pheronema. Of this several
specimens had been procured during the expedition of the " Noma " off the coast of
Portugal and in the neighbourhood of Setubal, from depths varpng from 400 to 600
fathoms. This new species was further described and figured by Kent in the Monthly
Microscopical Journal (1870, p. 243). Pheronema grayi is distinguished by the
Portuguese fisherman as " Nidos de Mer," or " the sea bird's nest," being so named chiefly
on account of the very broad and depressed form of the body, which is indeed very like
the nest of a chaffinch, and also on account of the uniform distribution of the hair-like
siliceous spicules over the whole surface, but especially on the inferior arched portion.
These spicules are not arranged in bundles, but are isolated, and project for a greater or
less distance, while in some specimens they are prolonged to form a very long basal tuft.
Saville Kent also called attention to the fact that the shaft of the amphidiscs or
" recurvate birotulate spicula" appeared to him to be rougher or "more profusely
echinate," than in the case of the similar spicules of Pheronema carpenten. " The
sarcode investing and constituting the sponge body was of a brilliant orange colour."

Under the name of Holtenia pourtalesii, Oscar Schmidt described and figured in the
same year 1870,^ several sacciform sponges, some of which are provided mth a superior
oscular opening, while others are entirely closed. The form of these types, however,
makes it difficult to refer them to this genus, or even to Pheronema; they are rather to
be related, as Carter noted in 1875,' to Rossella (or Lamiginella, Schmidt). It is
impossible to say whether a form designated by 0. Schmidt Holtenia saecus belongs to
Pheronema. It is described by 0. Schmidt as "sacciform, the wide opening having
very thin walls and irregularly projecting spicules." In the sarcode, according to
Schmidt, " innumerable small hexradiate spicules occur, besides many five-rayed spicules
with a projecting fifth ray, while whorl-like double anchors (amphidiscs) here and there
occur. The sparsely distributed large hexradiate spicules and the long spicules are
disposed on the incomplete meshes."

In some Notes on Anchoring Sponges,* Gray announced his inclination to unite

^Ann. and Mag. Nat. Hist., ser. 4, vol. vi. pp. 182-186.

- Spongien des atlantischen Gebietes.

^ Ann. and Mag. Nat. Hist., ser. 4, vol. xv. p. 118.

*Ann. and Marj. Nat. Hist., ser. 4, vol. vi. pp. 309-312, 18T0.

REPORT ON THE HEXACTINELLIDA. 237

Pheronema graiji, Leidy, Holtenia carpenteri, Wyvillc Thomson, and other forms in a
single family — the Pheronemada3, which might be characterised by the " ovate globular or
purse-like body, with a large internal cavity and outer walls formed of hexradiate spicules
placed side by side, producing a tesselated surface formed of stars." For Pheronema
(jrayi. Gray moreover, proposed the new generic name Callisphaera, and for Holtenia
saccus, 0. Schmidt, the generic name Vazella.

A somewhat different diagnosis was given by Gray in 1872' for his family
Pheronemadas in the words : — " Sponge oblong ; outer surface formed of hexradiate
spicules, lower surface with elongate filiform spicides ending in three recurved lobes."
In this family he distinguished (a) those forms with " anchoring filaments arising in a
circle of tufts around the base of the sponge," such as Pheronema, Leidy and Kent =
Holtenia, Thomson, and (h) those with anchoring spicules arising from all parts of the
sponge, such as Callisphaera = Pheronema grayi, Kent, and Vazella = Holtenia, 0.
Schmidt.

Under the designation of Laharia hemisphserica, Gray described, in 1873,^ a sponge
from Cebu, one of the Philippine Islands, sent through A. B. Meyer to the British
Museum, with the following brief diagnosis : — " It is hemispherical, about 2 inches in
diameter, and rather more than 1 inch high, mth a rather smooth outer surface, and a
rather deep regular concavity on the upper surface, which seems formed of interlacing
spicules, leaving considerable spaces between them. The outer surface and its margin are
scattered with distant, but rather regularly placed cylindrical perforations, from the
centre of which are emitted tufts of elongated filiform spicules, diverging in all directions
from the surface of the sponge. The middle of the underside deeply concave, with a
well-defined edge, from which is emitted a very large tuft of very numerous crowded
spicules, forming a kind of bi'ush, each filament when perfect ending in three short
recurved spines."

A detailed description of the same specimen was afterwards given by Carter' in
which he also described the form and distribution of the various siliceous spicules, whUe
the insignificant points of diff"erence between these and the corresponding spicules of
the genera Hyalonema, Holtenia, and Pheronema were pointed out.

In his great work on the Hexactinellida which Carter published in 1873,* and which
contains a particularly detailed account of the form of individual siliceous spicules,
Pheronema annw, Leidy, Pheronema grayi, Kent, Holtenia carpenteri, WyvUle Thomson
(as well as Meyerina claviformis, Gray, which is further-referred to below) are united into
one group, which is characterised as follows : — " Species more or less globular, excavated,
provided with anchoring spicules, and characterised by the birotulate flesh spicule above

^Ann. and May. Nat. Hist, ser. 4, vol. ix. p. 450.

^ Ann. and Mag. Nat. Hist., ser. 5, vol .xi. p. 235.

' Ann. and Mag. Nat. Hut., ser. 4, vol. xi. pp. 275-288.

* Ann. and Mag. Nat. Hist., ser. 4, vol. xii. pp. 349-472.

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

mentioned. Anchoring spicules of two kinds, viz., smooth and spiniferous, the hitter
terminating at its free extremity in a recurved double hook, termination of the former
unknown." On the other hand, Laharia hemisplnBrica, though closely allied, is so
peculiar as to constitute a section by itself, with the following characteristics : —
" Species possessing the birotulate flesh spicule, in which the termination of both kinds of
anchoring spicules are known. Free termination of spiniferous anchoring spicules much
the same as in the above mentioned ; termination on the smooth anchoring spicule con-
sisting of a double hook or arm, opposite, compressed, slightly recurved, and twice the
size of the head of the spiniferous form." Carter placed Oscar Schmidt's Holtenia
jMurtalesli at a great distance from both of these groups, between Crateromor-pha and
Rossella, and characterised it in the following terms : — " Eosette many-rayed ; rays
multitudinous, of unequal length, straight and capitate, pappiform."

As the result of a thorough investigation which Higgin^ was able to carry out on a
well-preserved specimen of the Laharia hemisphwrica, Gray, from Cebu, some errors
in Carter's earlier description were corrected — errors which were mainly attributable
to the fact that the specimen which Carter had studied was mixed up with foreign
spicules belonging to another sponge. This was confirmed'* by Carter himself, for he
remarks, " that the anchoring spicules with spined shaft are abnormal." Carter also
observes, that " while the shafts of the anchoring spicules of Laharia hemisphserica and of
the genus Eossella are all smooth, those of Hyalonema, &c., are all spined; and that the
latter only appear to be sometimes smooth from the spines being continued upwards
from the free end for a less distance in some than in others."

In his notes on the affinities of the Hexactinellida, Marshall observes, ^ that " Labaria
and Pheronema are clearly very closely related to one another, and perhaps scarcely
separable generically ; " and further, that ''Holtenia saccus, 0. Schmidt, is also related
to Pheronema, though it does not belong to this genus."

Among the sponges which were collected by the French Expedition of the " Travailleur "
in the Bay of Biscay, Norman also records Holtenia carpenteri.* According to the report
of Oscar Schmidt,^ a typical specimen of Pheronema annas was found in the neighbourhood
of Santa Cruz, at a depth of 180 and 248 fathoms. A special genus LeiohoUdium has
been established by Oscar Schmidt for a spherical sponge measuring 1 5 mm. in diameter,
soft and smooth to the touch, resembling in general habit and softness a delicate
Eeniera, and exhibiting the same microscopic component elements as Pheronema. " At
the one pole there is an osculum surrounded by a rim, over against this there is a
small irregular depression from which a root-tuft seemed to have been torn."

' Ann. and Mag. Nat. Hist., ser. 4, vol. xv. p. 385.

'^Loc. cit., p. 389.

^ Zeitschr. f. urns. Zool., 1876, vol. xxvii. pp. 113-136 ; vide p. 130.

* Ami. and Mag. Nat. Hist., ser. 5, vol. vi. p. 436.

^Die Spongien des Meerbusens von Mexico, 1880, pp. 64, 65.

REPORT ON THE HEXACTINELLIDA. 239

The specimen was found at Bequia, at a deptli of 1507 fathoms, on muddy
ground.

Monsieur Filliol, in his preliminary review of the chief results of the voyages of the
" Travailleur" and of the " Talisman," ' gives the following brief notice in regard to the
genus Pheronema: — "Les Pheronema paraissent etre r^pandus dans tout I'Atlantique, dont
elles habitent en certaines points de tres grandes profondeurs. Communes sur la c6te du
Portugal, elles apparaissent encore plus nombreuses au large des cotes du Maroc et du
Senegal. Nous les avons prises a partir de 600 metres jusqu'a 2200 metres. Dans un
dragage execute h, 4789 metres le chalut a rapporte des debris d'une eponge brisee, qui
semblait avoir dii etre a un Pheronema. Les esp^ces nous ont paru devoir etre varices.
Certaines d'entre elles sont remarquables par un enorme developpement, alors que d'autres,
telles que les Pheronema Parfaiti {loc. cit. fig. 92), se font remarquer par leur transparence
et I'absence de collerette des spicules autour de I'oscule. La coloration des Pheronema,
que nous avons captures, etait brunatre, et nous n'avons jamais eu I'occasion d'observer la
belle coloration d'un rouge orange, dont M. Saville Kent a fait mention, h- propos d'une
espece de Pheronema qu'il avait capture dans les envii'ons de Gibraltar."

Characters of the Genus. — The lower end of the cup- or goblet-shaped, thick- walled
body, is continued into a broad, not always sharply defined root-tuft of basal prostalia,
while from the lateral surface tufts of pleural prostalia project. The circular oscular
margin is surrounded by a cuS'-like fringe of freely projecting marginal prostalia, which
seem, however, to be absent in one species, — Pheronema annse, Leidy. The parenchyma
of the body contains uncinates. The basal tufts consist of tolerably elongated
hidentate anchors. The marginalia and many of the pleuralia end in club-shaped
thickenings.

1. Pheronema annae, Leidy (PI. XLIL).

The kindness of my esteemed fellow-worker 0. Schmidt of Strassburg, placed at my
disposal for examination a dry specimen of Pheronema annse, Leidy. Like Leidy's
original specimen, this form was got at Santa Cruz, W. L, from a depth of 180 or 248
fathoms. It is represented in natural size, from a photograph, on PI. XLIL fig. 1. The
body is about 11 cm. in length, and in the middle 5 cm. thick, exhibiting on the whole a
long, oval form, though the upper half is narrowed to a transversely truncated end, 2 cm.
in breadth. The broad lower portion is also somewhat truncate, with a breadth of about
3 cm., and from it there projects a number of slim, isolated tufts of basal spicules (5 to 6
cm. in length). Minute, slender bundles of shorter (2 to 3 cm.) straight, rod-hke spicules
are disposed radially, but in irregular distribution, over the whole lateral surface of the
sponge, without however anywhere forming a closed wreath. Nor is there any special

' La vie au fond des Mere, 1885, p. 85.

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

fringe of spicules on the somewhat sharpened, oscular margin of the gastral cavity. On
the outer surface of the body, between the radial groups of spicules, one can see the
characteristic network formed from the tangential rays of the hypodermalia, while on the
inner surface of the gastral space irregular groups of closely adjacent, minute, circular
apertures (0'5 to 1 mm. in diameter) penetrate, like sieve pores, the somewhat uniform,
curved, smooth wall. In the parenchyma, I have only been able to find quite isolated,
laro-e and middle-sized hexacts. Beside the long, radial rays of the large hypodermal or
liypogastral pentacts, a great number of various diacts occur, and small oxyhexacts in
special abundance. The latter are for the most part long and slender, but in part also
comparatively short. Only a few are smooth ; most of them bear somewhat closely
apposed spines turned in the same direction. They may perhaps be thus best described
as uncinates. Others occur in which the long, median portion is smooth, but the ends
beset with spines.

Some of the very long, slender uncinates, which usually occur in bundles, bear several
simple, obliquely projecting, or almost apposed bars ; while others exhibit, below each
slender pointed, almost appressed spine, a minute, longitudinal furrow or rib, as has
been figured by Wyville Thomson in Pheronema carpenteri (loc. cit., pi. Ixviii. fig. 4).
The short and rather thick uncinates have on an average a length of only 0"5 to 1 mm.,
and bear spines of very varied length, and occasionally quite rudimentary (PI. XLII.
fig. 6).

The small oxyhexacts have an average diameter of 0'14 mm., and are characterised by
transverse, rather blunt, externally directed, slender and somewhat bent spines, longest
in the middle of the rays, and decreasing towards the ends (PI. XLII. fig. 12).

The very strongly developed dermal skeleton exhibits strong hypodermal oxypentacts
with long smooth rays, which are not however always at right angles to one another.
The proximal radial ray is indeed usually at right angles to the four tangentials, but these
frequently form acute or obtuse angles, or are bent near the point of intersection into
the tangential plane. The rays of these large smooth oxjqientacts are as a rule straight,
but slight curvatures not unfrequently occur, either in simple or in 3- shaped fashion.

The pinuli of the external skin are moderately large oxypentacts, with four rather
long, slender, tangential bases, which are rarely quite smooth, in fact, usually beset with
more or less long, oblique, externally directed spines. The oblique somewhat externally
bent spines of the more or less long free distal ray vary in abundance. They sometimes
produce an almost bushy appearance, and are longest in the middle of the ray (PI. XLII.
fig. 10). The dermal ampliidiscs vary greatly in size and form. The largest are
between 0'2 and 0"3 mm. in length, and exhibit a strongly developed axial rod with
hemispherical tubercles and expanded campanulate umbels with eight smooth lancet-
shaped umbel rays (PI. XLII. fig. 2). Besides these, numerous amphidiscs of medium size
occur, measuring 0'05 to O'l mm. in length, with a slightly tuberculate axial rod and eight

REPORT ON THE HEXACTINELLIDA. • 241

slender umbel rays, wliicli vary considerably in length in different anipliidiscs, and are
often markedly divergent (PL XLII. fig. 3). The smaller and smallest amphidiscs,
varying from 0"04 to 0"02 mm. in length, bear short, approximately or perfectly
hemispherical terminal umbels, with eight, twelve, or more slender umbel rays (PL XLII.
figs. 4 and 5).

The firm skin which lines the long cylindrical gastral cavity is supported by strong
smooth hypogastral pentacts, which entirely resemble the already described hypodermalia.
The bushy autogastral pentacts are also exactly similar to the autodermals. I did not
find here the largest form of amphidisc, with broad paddle-like or lancet-shaped umbel rays,
but the medium-sized and small forms were present in abundance, as in the skin. On the
internal surface of the numerous large ramified efferent ducts, which occur in the gastral
walls, hypocanalicular oxypentacts with canalicular pentact pinuli occur, but the latter
are different from the autodermal and autogastral forms in this, that their four,
sometimes smooth, sometimes spinose basal rays are on an average longer, and the freely
projecting ray less thickly beset with lateral curved spines (PL XLII, figs. 8, 9, 12, 13).
The further the efferent camd system is followed tow'ards the chamber parenchyma, the
less conspicuous, the more slender and sparse do the pinules become, and the weaker is
the development of the spines on the freely projecting ray (PL XLII. fig. 11) until tliey
finally disappear just before the orifices of the chambers.

The spicules, which project radially in long tufts from the lateral surface, are diacts
or uncinates, either smooth or covered with barbs. The long bundles, projecting at the
lower end consist of spicules, which are smooth at the upper pointed end, while the lower
is beset with spines, bent obliquely upwards and outwards, and usually distinctly disposed
in two opposite rows. Somewhat above the extremity the spines disappear, and the
spicule ends in an anchor structure, which exhibits two, more or less long, sliglitly re-
curved, opposite teeth (PL XLII. fig. 7). Three such teeth exceptionally occur. If two
anchor teeth are formed, as is usually the case, they lie in the same plane as the
biseriate alternately disposed barbules of the shaft (PL XLII. fig. 7).

2. Pheronema carpenteri (W^'ville Thomson) (PL XLIIL).

On the "Lightning" Expedition in the north of Scotland, Wyville Thomson found,
at a depth of 530 fathoms, some beautiful sponges which he described and figured in
a masterly fashion^ under the title Holtenia carpenteri. Of these I obtained for examin-
ation some spirit and dried specimens captured by Wyville Thomson himself on the
"Lightning" and "Porcupine" Expeditions, and other specimens well-preserved in
alcohol, which were dredged by Mr. John Murray on the "Triton" Expedition. On
the Challenger Expedition several much torn portions of this Hexactinelhd were trawled on

■ P7ii7. Trans., 1859, p. 701.
(zOdL. CHALL, EXP. — PART LIU. 1887.) Ggg 31

242 THE VOYAGE OF H.JI.S. CHALLENGER.

tlie Brazil coast, off Macio, to the east of the mouth of the San Francisco (Station 124,
lat. 10° 11' S., long. 35° 22' W.), from a depth of 1600 fathoms, and a red mud ground.
Several portions of the basal tuft, and of the lateral wall, with vestiges of the external
skin, are alone preserved.

For my study of the tissue and the disposition of the various spicules, the specimens
collected by Mr. John Murray were especially satisfactory, owing to their excellent
preservation in absolute alcohol. These specimens exhibited an ellipsoidal form, and
the body measured, exclusive of the basal tuft, 5 cm. in length, and 4 to 5 cm. in
maximum breadth. The circular oscular opening at the superior pole is surrounded by
a wreath of vertically projecting, marginal spicules, lias a diameter of 12 mm., and
leads into a smooth- walled, cylindrical, inferiorly truncated gastral cavity, 25 mm. in
depth. From the whole external lateral surface, fine, pointed, radial spicules project in
loose bundles, 2 to 3 cm. in length. These are, however, irregularly disposed, with the
exception of certain closely set, somewhat longer spicules, which form an annular zone, a
few mm. in breadth, about 10 mm. below the marginal fringe. On the lower basal end
of the sponge, there is a large number of slender (1 to 2 mm. in breadth) tufts of long
flexible spicules, 30 to 40 cm. in length, which interlace abundantly in the thick felt-work
of the basal tuft. The indi\'idual slender bundles are however distinctly separate as
they issue from the sponge-body. The external surface of the sponge exhibits a delicate
dermal network, marked, especially in the dried specimens, by stellate knots at tolerably
uniform intervals (PI. XLIII. fig. 1).

The ellipsoidal shape may be regarded as characteristic (jf the general form of the
body {cf. PI. XLIII. fig. 1, and W^-ville Thomson, loc. cit., pi. Ixvii., pi. Ixix. fig. 1),
although a comparative survey of all the forms before me reveals a slight difference
in this, that the larger, and therefore probably older specimens, are somewhat more drawn
out longitudinally, the smaller younger forms are approximately spherical, while the very
smallest, less than a pea in size, often exhibit, as Wyville Thomson has shown, the form
of a hen's egg with an inferior pointed pole (Wyville Thomson, loc. cit., pi. Ixxi.).

As is noted in Wyville Thomson's careful investigation of the siliceous spicules of
Pheronema carpenteri, the parenchyma, which is on the whole only slightly massive
but is penetrated by countless lacunge and wide passages, contains among the larger
independent strictly parenchymal spicules, medium-sized, rarely large oxyhexacts, while
long, slender, flexible oxydiacts predominate. These are either smooth or thickl}^ beset
with appressed uniformly directed barbs, and are for the most part radially disj)osed
towards the external or'gastral surface, with the distal pointed end projecting for a variable
distance beyond the surface. Besides these, there are in the parenchyma a large number
of short, strongly developed uncinates, 0"2to 0"3 mm. in length, with but short spines or
barbs (PI. XLIII. fig. 5). Finally, there occur in the parenchyma small simple oxy-
hexacts, with smooth straight rays. That the strong and somewhat large smooth

REPORT ON THE HEXACTINELLIDA. 243

oxypentacts, which form the main support of the dermal, gastral, aud canalicular
skin, and -which here and there appear to lie quite within the parenchyma, are really to
be regarded as true parenchymalia, does not seem to me probable. As a rule I could
readily connect them with the above named surfaces, and thus regard them as hypo-
dermal, hypogastral, and hypocanalicular elements respectively.

The dermal skeleton has for its supporting basis large, strong oxypentacts, in which the
four long smooth tangential rays are usually indeed all but straight and crossed at right
angles, though not unfrequently somewhat curved, and in their inclination to one another
more or less diverted from a right angle. They are frequently disposed not exactly
tangentially, but slightly inclined inwards. The likewise long and strongly developed,
smooth, straight, proximal ray is, as a rule, at right angles to the surface ; it may, however,
in certain cases, for instance in the immediate neighbourhood of the marginal boundary,
deviate from this exactly radial disposition. The tangential rays of the adjacent
hypodermal pentacts lie for the most part in long stretches close to one another, and thus
form a strong, quadratic, dermal lattice- work, in the meshes of which the sieve network of
the dermal membrane is spread out. The quadratic skeletal meshwork is, however, in no
way uniformly composed, but numerous deviations and displacements occur all over.
The skin is externally beset with autodermal pentact pinuli which are more or less rough-
ened. Their tangential basal rays are of considerable length and almost always smooth,
but rarely tuberculate towards the end, while the moderately long (0"15 to 0"2 mm.)
radially projecting distal is smooth below, but on the outer three-fourths of its length
beset in fir-tree-like fashion with strongly developed lateral spines (PI. XLIII. fig. 4).
It is noteworthy that these dermal pinuli are in no way so abundantly or thickly present
as is usual in the other Hyalonematids ; in certain positions, indeed, they are sparsely
present or have a quite isolated occurrence. Especially near the oscular margin they are
only to be found here and there among the especially abundant pleural prostalia
(PL XLIII. fig. 2).

As to amphidiscs, I found, in the external skin of the North Atlantic specimens,
only the small (0-03 to O'OS mm.) eight-rayed form (PI. XLIII. figs. 6, 8) with short
hemispherical umbels. In the remains of the Brazilian specimens, however, besides the
above forms, larger eight-rayed amphidiscs of similar form (PL XLIII. figs. 9, 10)
occurred. I hesitate, however, before proposing to erect a distinct species on the strength
of this single deviation.

In the compact gastral membrane the skeletal elements are disj^osed in essentially the
same way as in the external skin. The following deviations are, however, of some
interest. In the first place, the whole free surface is much more closely and uniformly
beset with pinuli, and these pinuli differ from the autodermal forms in this, that
they are for the most part, especially in the neighbourhood of the oscular opening,
bent by the stream of water towards the efferent aperture (PL XLIII. fig. 2). The

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

small amphidiscs here are equal in abundance, and similar in form to those in the
outer skin.

Even the larger afferent and efferent canals are furnished with pinuli, though in
rapidly decreasing size and strength. The small amphidiscs also extend for a consider-
able distance as canaliculana.

In regard to the numerous radial pleuralia which project to a greater or less distance
beyond the surface of the sponge, the uncinates proper demand special notice. These are
long oxydiaets which are thickly surrounded by appressed barljs. They occur chiefly at
the upper end in the neighbourhood of the oscular margin, and are always radially
disposed so that the pointed ends of the barbs are directed inwards, and the anterior
point of the whole spicule outwards (PI. XLIII. fig. 2). Wyville Thomson has figured
a specimen of such a long (up to 8 mm.) pleural uncinate [loc. cit., pi. Ixviii. fig. 4).
Besides these a large number of cylindrical spicules of various calibre occur, frequently
in the form of thin, coiled threads, which gradually decrease in thickness, towards
both ends from a maximum diameter about the region of the external surface of the
sponge. They terminate internally in a simple point, while the outer extremity generally
evades observation owing to breakage. They are usually smooth, either altogether or
for the greater portion of their length. Although these long, smooth, pleural prostalia are
for the most part broken, and the external extremity only uncertainly distinguishable, a
few seem to be intact. I was unable to decide whether there are any, or if so how many,
pleuralia which run out externally into a point. Most of the long, projecting, smooth
spicules gradually become thinner and thinner towards the extremity, but the fact that the
terminal portions themselves are broken off, suggests that they do not simply form points,
but terminate like the shorter forms about to be described, in which the outer ends were
preserved. On several of the thin fibre-like pleural spicules, which are quite smooth on
their internal pointed portion,' one notices further outwards the ajipearance of small teeth,
at first very inconspicuous, but afterwards gradually more prominent and spine-like.
These are directed transversely or somewhat obliquely outwards. The outer or most
external spinous portions of these pleuralia finally terminate in a club-shaped swelling
with four cruciately disposed lateral elevations and a single terminal point. The
lateral points either project transversely, or are directed oblic[uely outward's ; more rarely
they are bent somewhat inwards. In some instances it was possible to recognise in this
terminal knob a hint of an axial canal intersection, corresponding to the five conical teeth.
These spinose monacts occur especially on the upper portion of the sponge-body. I
have repeatedly found, even in the neighbourhood of the oscular margin, simdar monacts
of smaller size (2 to 3 mm.), which projected but slightly from the surface of the skin,
and were throughout their whole extent, from the internal pointed end to the external
terminal knob, uniformly thickly beset with small .spines directed transversely or
obliquely outwards (PI. XLIIT. fig. 2).

. REPORT ON THE HEXACTINELLIDA. 245

Finally, the pleuralia include isolated two-toothed, very rarely threc-tootlied, anclior
forms with a strong shaft bearing barbs. These occur more abundantly towards
the base, and seem to be wholly absent from the upper end of the sponge-body. The
inner portion of these spicules, which vary greatly in length, is smooth, and ends in a
simple uniform point. On the outer j)ortion inconspicuous tubercles first appear, these
are succeeded by larger forms, which finally pass into strongly developed spines pointed
backwards and slightly recurved. In the portion of the anchor shaft which is surrounded
by barbs the diameter of the spicule rod decreases gradually outwards. Just in front of
the end which bears two simple anchor teeth the barbs disappear, and the diameter
again increases on to the thickened end with its slightly arched apex. From the latter
the two strong roundish (about 1 mm. in length) anchor-teeth originate laterally in a
distal plane. They extend opposite one another — smooth and slightly curved — and are
directed obliquely outwards and backwards. In abnormal cases three similar anchor
teeth occur, forming equal angles with one another, or the number may be reduced
to one.

The basalia forming the numerous slender tufts' — about 2 mm. in thickness — which
project from the lower rounded end of the sponge, are for the most part of great length,
and may in the larger specimens, as Wyville Thomson has shown, measure several
decimetres. They are, on the whole, stronger than the pleuralia, and form in spirit or
dried specimens a thick feltwork, between the fibres of which, portions of the substratum
and all kinds of foreign bodies are included (PI. XLIII. fig. 1). In their upper portions
all the fibres are smooth, while externally they exhibit, for the most part, perhaps with-
out exception, the curved barbs described above in connection with the shaft of the
pleural anchors. Like the pleural anchors, further, they pass by a smooth cyKndrical
neck to a thickened terminal portion with two more or less large recurved anchor teeth.

The marginalia surrounding the circular margin of the oscular aperture form a
closed wreath — a few spicules in breadth — and consisting either of long, strongly
developed uncinates like those which occurred among the pleuralia, or of perfectly
smooth, internally pointed, externally narrowed spicules of various calibre, which
probably end in a point or in a spinose portion with terminal knob.

"While on the internal surface of the gastral cavity of the Hyalonematids no spicules
projected except the pinules, in this form numerous uncinates occur in the neighbourhood
of the marginal fringe, arranged in tufts directed inwards and upwards, and projecting
freely for about half theii- entire length into the lumen of the gastral cavity (PI. XLIII.
fig. 2).

' Oscar Schmidt has, in his Spongien ties Meerbuseiis von Mexico (p. 65), maintained that the Pheronema figured by
Saville Kent (Monthly Micr. Journ., 1870, pi. Ixiii. fig. 1) is not AVyrille Thomson's Pheronema carpenteri, hut
Pheronema annse, Leidy. With this opinion I cannot agree. The hasalia do indeed protrude as in Plieronema annse,
but the same occurs in Pheronema carpenteri, though to a less marked degree. Besides, as Saville Kent notes, the
specimen was much macerated, so that the individual Imndles of basalia were naturally isolated.

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

As to the young specimens, from 10 to 5 mm. in size, so thoroughly studied by
Wyville Thomson, it is to be noted further that besides the already mentioned ovoid
form of the body, the slender development of the fir-tree-like distal ray of the dermal
pinules is striking.

3. Pheronema grayi, Sav. Kent.

This Portuguese species is distinguishable, even on external macroscopic inspection,
by its likeness to the nest of a chaffinch, and by the scattered disposition of the lateral and
basal prostalia, which are not grouped in bundles. Though there was no specimen at my
disposal for more intimate study, I shall attempt to sum up the characteristics on the
strength of the descriptions and figures before me, and as the result of notes which I
made on some specimens in the British Museum. The breadth and height of the
two-third spherical form are approximately equal, and measure 10 to 12 cm. The loose
beard-like root-tuft has a length of 26 to 60 cm. or more. The circular oscular aperture
measures 5 cm. in diameter, and bears on its margin a continuous annular fringe of
perpendicularly projecting needles, of which the freely projecting j)ortions are about
10 em. long. While the prostalia lateralia over the rest of the external surface are
uniformly but sparsely scattered, there is here, as in Pheronema carpenteri, a thicker
zone, about 12 mm. below the marginal fringe. The zone consists of pleuralia, projecting
radially, or directed somewhat oblic|uely outwards and upwards. When the sponge was
brought on board " the sarcode investing and constituting the sponge body was," according
to Saville Kent {loc. cit., p. 284), "of a brilliant orange colour."

In regard to the spicules, it is to be noted that they essentially agree with those of
Pheronema carpenteri, and from the report which I have given above in regard to the
prostalia of Pheronema carpenteri, it will be seen that even " the long attenuate spinu-
late forms and others with straight spines " which Saville Kent figures as characteristic
of Pheronema grayi, are not absent from the former. It is possible that the spinose
needles with a five-pointed terminal and external knob, which I found in Pheronema
carpenteri projecting as lateral prostalia from the neighbourhood of the oscular margin,
are absent in Pheronema grayi, but this supposition requires to be corroborated by direct
investigation, which I have not been al)le to accomplish.

4. Pheronema hemisphssricum (Gray).

As W. Marshall has noted, the genus Laharia erected by Gray (with a single S2:)ecies,
Laharia heynisphsericum. Gray), cannot be separated from PJieronema, Leidy. This form
(from the Philippine Island, Zebu) has been described by Gray, Carter, and with special

REPOKT ON THE HEXACTINELLIDA, 247

tliorougliness by Thomas Higgin, and the result is to show that it must shiiply
be included in the genus Pheronema. Since I have not myself been able to examine
Pheroncma hem{s2')hiericiim, I will here content myself with citing the last and most
thorough description, that of Higgin/ and with shortly summarising the most essential
characteristics.

According to Higgin, the sponge is in form " like a small bird's nest, the bottom of
which is flat, mth a well-defined edge : the sides are rounded ; and the sponge attains its
greatest diameter about one-third of the way down from the edge of the hollow of the
nest, towards the base." According to his figure on PL XXH. fig. 3 the shape is like
that of a prehistoric Greek clay urn, and exhibits a convex surface above and below the
annular pad which occupies about the median third of the lateral wall. The maximum
transverse diameter measures about 10 to 12 cm., and the height about 9. The upper
flatly spherical cavity is at the margin 8 to 9 cm. in width, and is 4 cm. deep. The
length of the anchor needles in the basal tuft is 8 to 9 cm. On the superior, sharply
defined margin there is a wreath of varying length and 1 to 3 cm. in breadth, composed
of long straight pointed spicules, which project at right angles. The lateral usually pro-
jecting annular zone is covered with " whisker-like bundles of long spicules which issue from
circular holes the edges of which are slightly raised, each bundle consisting of a dozen or
more spicules. Around the circumference of the base are arranged loose fascicles of
anchoring spicules, from 7 "5 to 8 cm. in length, and having a diameter of about 12 mm.
measuring along the edge of the base, by 6 mm. to 10 mm. A few scattered short
spicules project here and there from the base generally, but there are no bundles other than
those around the edge. The anchoring spicules are of one kind only, viz., smooth, fusiform,
terminating at the free end in two opposite hooks; there are no spined forms. The spicules
of the whisker-like tufts are plain, fusiform ; the spicules of the erect fringe round the labrum
are also fusiform and smooth throughout, but there is an appearance of spines on some
towards the free end." In the surface reticulation T. Higgin found four kinds of spicules —
" (l) pentacts, whose four tangential arms are equally smooth and opposite or at right
angles to each other, inclined slightly downwards or inwards ; the shafts of the larger
spicules are 12 mm. long; (2) long, slender, acerate spicules, thickly covered with short
sharp spines, all jjointing towards one and the same end of the shaft ; (3) smooth acerates
with the cross on the central canal ; (4) plumose spicules, of shapes intermediate between
one with a very thick shaft, short and bushy looking, with long, strong, bluntly ended
arms, and another -with small, short, fine arms and a long feather-like shaft ; the crucial
arms of which, thickly studded with short obtusely pointed spines, are bent downwards,
as if to embrace or fit to the arms of the large spicules on which they rest. The strongly
woven together basket work of the interior, as seen through the investing network,
is composed of: — (1) smooth spicules of the sexradiate type (that is, acerate with simply

1 Ann. and Mag. Nat. Hist., rer. 4, vol. xv. p. .377

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

a central cross indicating their Hexactinellid character) ; (2) acei^ate with four tubercles
at the middle of the shaft ; (3) sparsely spined acerates, the spines bent towards the middle
of the spicule ; (4) four-rayed, (5) five-rayed, and (6) six-rayed spicules, the long arms of
which are bent together in all varieties of ways ; among these are (7) large and small
eio-ht-armed birotulates with dome-shaped heads, and some very minute ones ; (8) small
sexradiate spiuules, the arms of which are furnished towards the free end with three,
four, or five long spines projecting in the direction of the free end ; also (9) a small
acerate spicule in great abundance peculiar to the species, furnished with fine spines not
very close together, all of which are bent towards one end of the spicule, increasing in
lenoth along one-third of the spicule (viz., from the end from which they look), and then
gradually diminishing again from this point to the other end of the shaft; and (10)
plumose spicules in great variety."

5. Pheronema glohosum, r\. sp. (PL XLIV.).

Besides several other Hyalonematids, some specimens of a Pheronema were ob-
tained from Station 192 of the Challenger Expedition, near the Little Ki Island (lat.
5° 49' 15" S., long. 132° 14' 15" E.) from a depth of 129 fathoms and a blue mud ground.
At first I was inclined to regard the latter as identical with the above described
Pheronema hemisphwricum, Gray, from the Philippine Island, Zebu. The specimens before
me, however, which are for the most part well preserved in spirit, differ so markedly both
in external appearance and in the structure of many of their spicules from the latter
species, that a separation is certainly necessary. While the smaller specimens, with a
diameter of 20 to 25 mm., have an almost spherical appearance, and only exhibit at the
superior pole a hemispherical depression with a strong-margined aperture, 10 cm. in
diameter, the larger forms, 9 to 10 cm. in breadth and 6 cm. high, have the form of a
superiorly truncated and somewhat depressed three-quarter sphere. The circular terminal
opening of the shallow hemispherical gastral cavity, 3 to 4 cm. in depth, measures 5 to
6 cm. in breadth, and exhibits a somewhat sharply defined edge, under which a gentle
protrusion inwards is distinctly visible. From this oscular margin a wreath of pointed
spicules projects perpendicularly for 12 to 15 mm. In the external dermal layer there
is a tolerably firm irregularly shaped network, which bears at intervals of 1^ to 2 cm.
small knots with slight boss-like projections (PL XLIV. fig. 1). From each such boss a tuft
of eight to twelve blunt spicules projects radially, attaining a length of 3 cm. and more.
The whole uniformly convex basal surface is furnished with these tufts of spicules, at
approximately equal intervals, and that all over, not merely forming a circle, or leaving
the central portion free (as Higgin described in Pheronema hemisphaericum). The spicules
of these basal tufts far exceed the pleural in number, and still more in length. The
normal length of the basal tuft, which is in all the specimens much thickened

REPORT ON THE HEXACTINELLIDA. 249

aud felted, is difficult to estimate. In the largest forms it cannot be less than 10
to 20 cm.

The concave internal surface of the upper gastral cavity exhibits a tolerably firm and
solid portion, about 10 mm. in breadth, and arched slightly inwards, while aU the rest of
the surface forms a tolerably well-differentiated quadratic lattice-work, the strands of
which enclose meshes, 2 mm. or less in width.

From this description, and from the figure in PI. XLIV. fig. 1, drawn from a
photograph, the external specific differences between Pheronema globosum and Pheronema
hemisphsericum, Gray, must be evident enough. As to parenchymal spicules, mention
must be made of (1) the numerous rather large oxyhexacts, with long, straight, or slightly
curved smooth rays, which are sometimes reduced in number so as to result in pentacts,
tetracts, or triacts ; (2) medium-sized smooth oxydiacts, which are for the most part
somewhat bent ; (3) long radial uncinates, reaching as far as the surface or even further ;
(4) very numerous short uncinates, which though especially abundant round about the
efferent canals, occur in varied disposition, and are distinguishable from the long uncinates
not only by their smaller size, but also by this, that the strong laterally inserted spines
are curved, and somewhat markedly divergent ; (5) slender oxydiacts of rarer occurrence,
with strongly developed, or almost wholly reduced spines (PI. XLIV. figs. 4, 7).

The numerous medium-sized and small amphidiscs (PI. XLIV. fig. 3) which occur in
the parenchyma, do not, in my opinion, really belong to it, but owe their origin to the
outer skin, or to the membrane of the gastral or canalicular cavities, whence they have
been forcibly pushed or floated into the parenchyma.

The dermal skeleton includes strong hypodermal oxypentacts with long, smooth,
straight or slightly bent rays, measuring 10 mm. or more in length (PI. XLIV. figs. 8, 9).
The four tangential rays are inclined slightly inwards (PI. XLIV. fig. 9). Numerous
strongly developed autodermal pentact pinuli also occur, exhibiting a somewhat thick,
bushy, free distal ray, thickly beset with strong, bent, lateral spines, and four moderately
long, blunted basals, slightl}^ inclined inwards, and thickly beset with short spines,
(PI. XLIV. figs. 3, 5). In various places the distal ray of the pinuli is somewhat more
slender, of greater length, and slightly curved.

As to amphidiscs, I observe a large form, 0"2 mm. long, with campanulate but
rather short terminal umbels, in which the eight, or more rarely six, umbel rays are
tolerably smooth, and moderately rounded in paddle-like fashion (PI. XLIV. fig. 6).
The rather thick stalk is richly beset with roundish tubercles. Besides these, somewhat
small amphidiscs occur, with exactly similar form, and finally very small forms, 0'02 mm.
or less in length (PI. XLIV. figs. 12, 13).

The dermal skeleton of the gastral cavity resembles that of the outer skin, except in
this, that the pinules are furnished with a much longer and more slender distal ray.

The pleural prostalia consist of the long, more or less markedly projecting uncinates,

(ZOOL. CHALL. EXP. — PART LIII. — 1887.) Ggg 32

250 THE VOYAGE OF H.M.S. CHALLENGER,

but for the most part of long spicules, of which the internal portion buried in the body
is wholly smooth and runs out into a point, while the outer freely projecting portion
(2 to 3 cm. in length), though, indeed, for the greater part of its length also smooth, is in
the neighbourhood of the outer termination beset with teeth, directed obliquely outwards,
and either runs out into a point, or exhibits a club-shaped thickening, with one terminal,
and four cruciately disposed lateral points. Since most of the pleuralia are broken off, it
is difficult to determine which mode of termination is the more abundant, and whether
there are not here also spicules which are quite smooth on to the very end.

The marginalia, which project in a perfectly continuous series for 12 to 15 mm., and
become narrowed towards their outer end, are always quite smooth in their inner pointed
portion. Towards the exterior they bear small spines projecting obliquely outwards, and
end in a sort of lance-like point.

The basalia (PI. XLIV. fig. 11) are very long and for the most part quite smooth.
While the inner end always runs out into a point, they terminate externally in a double-
toothed anchor, in which both teeth diverge almost transversely, are very gently curved,
and end in a point. The shaft becomes gradually narrower to within a certain distance of
the end of the anchor, and then increases in thickness on the anchor itself. The
inferior margin of the whole anchor is a simple continuous are, like that of Pheronema
giganteum (PI. XLV. fig. 9).

On young specimens I have sometimes seen small anchor forms buried in the body,
and in these the shaft was, for the most part, beset with lateral barbs projecting towards
the inner end.

6. Pheronema giganteum, n. sp. (Pis. XLV., XL VI.).

At the same station where the almost spherical specimens of Pheronema globosum
were obtained, i.e., near Little Ki Island (lat. 5° 49' 15" S., long. 132° 14' 15" E.), the
trawl brought up from a depth of 129 fathoms and from a blue mud ground, a single
well-preserved specimen of another Pheronemxi form. This is well figured in PI. XLV.
fig. 1, in half its natural size, and in PI. XL VI. fig. 1, one-third reduced, in longitudinal
section.

It differs from Pheronema globosum in its large size, to which it owes its title of
giganteum, while as to shape it presents a long ovoid and somewhat obliquely depressed
form, the peculiarity being due to a slight one-sided flattening from the side of the
inferior pole. It measures 48 cm. in length and 20 cm. in maximum diameter, and ex-
hibits an irregularly tuberculate surface. At the superior pole there is a circular terminal
aperture, 6 cm. in width, and surrounded by a somewhat sharply angular projecting margin,
enclosing the almost cylindrical gastral cavity, which is more than 20 cm. in depth.

From small boss-like (2 to 5 mm. broad) elevations of the external body surface.

REPORT ON THE HEXACTINELLIDA. 251

which occur irregularly at intervals of 5 to 15 mm. over the whole skin, roundish tufts of
radially disposed, pointed spicules, 3 to 6 cm. or more in length, project. From the
somewhat protruding oscular margin, there projects a continuous, cuff-like, annular fringe
of (3 to 4 cm.) long pointed spicules, with slight outward curvature. The whole down-
ward directed surface, which is markedly enlarged by the above described oblique
flattening of the sponge, bears a thick basal tuft, 10 cm. in length by 20 to 24 in
breadth, and composed of much felted spicules. As in Pheronema annas, Pheronema
carpenteri, and Pheronema globosum, this tuft is composed of numerous individual and
separate bundles of spicules, springing from the whole of the lower surface (PI. XLVI.
fig. 1). The external surface of the skin, as seen between the laterally projecting tufts
of spicules, appears to the naked eye very uniform and even. Here and there, through
the somewhat thick covering of the subdermal cavities, the afferent canal system may
be recognised (PI. XLVI. fig. 3). The internal lining of the gastral cavity, which is also
formed from a firm uniform layer, is penetrated by groups of pores, with the exception of
a somewhat protruding thick zone, 2 cm. in breadth, close beside the oscular margin.
The groups of pores consist of three to five roundish exit openings of the efferent canals,
and are covered by a delicate gastral network (PI. XLVI. figs. 1, 4, 10, 11).

The wall of the sponge, which measures 4 to 6 cm. in thickness, is penetrated by
roundish lacunas and ducts of the afferent and efferent canal system, which, in their
widest portions, especially below the dermal and gastral walls, attain a diameter of 10
mm. and more (PI. XLVI. fig. 2). Between these wider passages there are tolerably
thick (up to 10 mm. and more) tissue layers, which are penetrated by much narrower
canals, opening laterally with narrow apertures into the larger (PI. XLVI. fig. 2).

The spicules which belong especially to the parenchyma are of the following main
types : — first of all, numerous strong pentacts, usually with a greater radial ray, while
the four others do not always lie quite in one plane, and therefore not exactly at right
angles to one another, but are slightly bent in various ways. It seems as if these
parenchymal pentacts were originally hypogastralia and hypodermalia, which became
secondarily involved in the parenchyma proper (PI. XLVI. fig. 7). A second form of
skeletal element, occurring abundantly in the parenchyma, represent the small oxyhexacts.
The straight, regularly disposed, moderately strong rays of these forms are beset with
somewhat distant, longer or shorter spines, projecting somewhat transversely (PI. XLVI.
fig. 8). In the third place, there are in the parenchyma, and in special abundance near the
outer skin, uncinates of various length. The shorter forms, 0'6 to 1"0 mm. in length,
are comparatively thick, and furnished with strong compressed barbs. They occur in
irregular distribution throughout the parenchyma, while the longer forms, some mm.
in length, project radially in bundles to the outer skin, or penetrating the latter, pass
into the tufts of the pleural prostalia. On these long uncinates there are numerous barbs
all round, longer and more slender than on the shorter forms, and further closely

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

appressed. They spring from a projection and cover over a groove-shaped depression.
Finally, we have to note the occurrence (here and there very abundant) of small thin
oxydiacts, which exhibit on their surface fine tubercles or spines, or only irregular
rouo-hnesses. It seems to me improbable that the amphidiscs found in the parenchyma
really belong to it. I believe rather that they are present only by secondary disloca-
tion, having originated in the dermal, gastral or canalicular skin.

The dermal skeleton includes strongly developed hypodermal oxypentacts of variable
size. "While the long proximal ray has usually a perfectly radial disposition, the four
tangentials extending below the dermal membrane do not always form right angles with
one another, and frequently exhibit near their origin a marked curvature to the side or
inwards. The rays may exceed 1 mm. in diameter, and that strength is indeed common
enough, though not at the point of origin, but at some distance up the ray. Between
pentacts of this sort, tetracts occasionally occur, arising by reduction of the proximal
radial ray. Not unfrequently, in some or in all, the ends are more or less markedly
rounded. The autodermal pentact pinuli, present in great number and in strong develop-
ment exhibit great variations in the size and form of the cjrpress-like or fir-tree-like free
distal ray. The four cruciately disposed moderately long and thin basal rays lie in the
same plane. Near their origin they are smooth, but bear on the larger outer end small
teeth and spines. They usually end in a point, but in the larger specimens the ends are
sometimes more or less bluntly rounded off (PL XLV. figs. 2, 3). The distal ray of the
larger dermal pinnies attains a length of 0'3 mm. and more. In form it resembles a
slender cypress, so closely lie the uniformly long, oblique, lateral spines (PI. XLV. fig. 2).
Between these long pinnies numerous smaller forms occur, which are usually much more
slender, and run out into a single long terminal point (PI. XLV. fig. 3). Less frequently
compressed, broad and bushy forms occur (PI. XLV. fig. 5). The tall cypress-Uke
pinuli are usually associated with the strong tangential rays of the large hypodermal
pentacts, while the slender and smaller forms occur on the sieve-like perforated dermal
meshes, which are enclosed by large hypodermal pentacts.

The larger amphidiscs, which do not occur in what could be called abundance, attain
a length of 0"18 mm., exhibit a moderately thin, tubercled, axial rod, and short cam-
panulate, almost hemispherical terminal umbels, with eight smooth, terminally slightly
pointed rays (PI. XLV. fig. 8). Less frequently, somewhat broader ovoid amphidiscs
occur, in which the umbel rays almost, or actually meet the opposites (PL XLV. fig. 7).
Finally, a large number of small amphidiscs are found, with longish form, and almost
hemispherical umbels composed of eight uniformly slender rays (PL XL VI. fig. 5).

The gastral skeleton exhibits the same strong and somewhat slimmer oxypentacts
which occur under the external skin. The numerous pentact pinuli otherwise resemble
the bushy forms of the outer skin, but have a much shorter bushy radial ray. On the
sieve-network in the gastral skin, which covers the large grouped apertures, there are

REPORT ON THE HEXACTINELLIDA. 253

delicate pinules with thin pointed basal rays, and with slender pointed free radial, as we
have already described in the outer skin (PI. XLVI. figs. 10, 11).

There is in the gastral skin a very abundant occurrence of large and long amphidiscs,
with short, hemispherical, eight-rayed umbels, and tubercled, moderately slender, axial
rods (PI. XLVI. fig. 11), such as occurred sparsely in the dermal membrane. More abun-
dant are the small, long amphidiscs, with slender, roughened axis rod, and short, hemi-
spherical, eight-rayed terminal umbels.

The canalicular skeleton of the larger efi'erent passages differs essentially from that
of the afferent ducts of the lacunar subdermal or subgastral spaces. Even macroscopic
inspection of the internal surface of these canals and lacunae, one notices a marked
difference in the character of the surface, which is in the efferent canals quite rough and
villous, while that of the afferent canals, and of the subdermal or subgastral lacunae,
appears comparatively smooth, or only exhibits a fine uniform roughness. On all the
larger canals and lacunae, the walls are supported by strongly developed hj^ocanalicular
oxj'pentacts, with rays varying in strength, according to the size and wid.th of the canals.
Between the larger hypocanalaria, smaller forms always occur. The peculiarly rough
nature of the afferent canals is conditioned by especially long, though not particularly
broad autocanalicular pentact pinuli, in which the freely projecting cypress-like ray
attains a length of 0*4 mm. and more. Of course between these long pinules shorter
forms occur, with slender, slightly spinose, free radial rays (PI. XLVI. fig. 6). Isolated
amphidiscs of the larger sort and numerous representatives of the smaller type occur in
the skin of the wider efferent canals (PI. XLVI. fig. 6). Both the number and the size
of the canalicular pinules and amphidiscs gradually decreases with the width of the
canals, until they finally disappear in the neighbourhood of the chambers.

The lining layer of the large afferent canals and of the subdermal spaces is supported
by smooth hypocanalicular oxypentacts entirely similar to those of the efferent canals.
Both the autocanalicular pinuli and the amphidiscs are, however, entirely absent
(PL XLVI. fig. 7, left).

The pleural prostalia, which project for 3 to 6 cm. from the lateral wall of the
sponge, are strong smooth needles, which attain the thickness of a millimetre. They run
to a point within the body, but the external termination unfortunately eluded distinct
observation, since they were almost all broken off. Doubtless, however, they are
either smoothly and simply pointed, or are terminally toothed, and end in a small
five-pointed club. The same is true of the marginalia, which project for about
3 to 4 cm. I have abeady mentioned that numerous long uncinates with pointed
external extremities project between the pleural prostalia.

The long basalia, which issue like the pleuralia in bundles, but become woven into a
felted basal tuft, begin like the pleuralia and marginalia in an internal pointed end, swell
out beyond the body into smooth cylindrical beams, which again decrease in thickness

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

towards the outer end, before whicli, however, they become again enlarged and end in a
bideutate crescent-shaped anchor. The two teeth of the latter rise from the gently bow-
shaped, rounded and thickened terminal portion (PI. XLV. fig. 9), and end in simple
conical points. The distance of these two terminal points, i.e., the total breadth of the
anchor, is about 0*5 mm.

Genus 3. PoUopogon, Wyville Thomson (Pis. XLVII.-L.).

Literature and History. — In his preliminary account of part of the Challenger
Expedition ^ Wyville Thomson described a large sponge — PoUopogon amadou — which
was found to the south of the Canary Islands at a depth of 1525 fathoms. This
he regarded as type of a new genus — PoUopogon. This sponge forms an oblique and
upward directed, semi-involute plate, with sharp upper and lateral margins, having the
general shape of a tree fungus. From the transversely truncated base a strong beard
of long siliceous fibres projects, and these fibres bear on their extremities two
widely extended anchor teeth, which serve for fixing the sponge. An uniserial
fringe of fine, straight, parallel, projecting siliceous spicules adorns the sharp free
side and upper margins. The concave inner and the convex outer surfaces are
covered by a fine network with quadrate meshes. " The sponge when brought up was of
a delicate cream colour ; it was necessary, to steep it in fresh water to free it from salt,
and the colour changed to a leaden grey."

Character of the Germs. — The body has the form either of a thick-waUcd goblet, or of
an ear-shaped involute plate. It exhibits a broad basal tuft, and an oscular fringe of
marginalia, but no laterally projecting pleuralia. The parenchyma contains small,
extremely rough or spinose, oxyhexacts and uncinates, and in one species even small
smooth oxydiacts of varying size and in varying abundance. The two teeth of the basal
anchors are disposed approximately at right angles to the long, almost smooth shaft.
The marginalia end externally in club-shaped thickenings.

1. PoUopogon amadou, n. sp. (Pis. XLIX., L.).

South-west from the Canary Islands (Station 3, lat. 25° 24' N., long. 20° 14' W.), from
a depth of 1525 fathoms and hard ground, a beautiful PoUopogo7i form was dredged.
This type, which was in 1877 figured and shortly described under the above title in
Wyville Thomson's Atlantic, has not the cup shape characteristic of most other
Hyalonematidse, but exhibits rather the form of an ear, or that of a leaf rolled up into a
semi-funnel (PI. XLIX.). There is thus no gastral cavity but only a^concave gastral, and
a convex external surface, both of which are quite smooth, without radially projecting

1 The Atlantic, vol i. p. 174-176.

REPORT ON THE HEXACTINELLIDA. 255

spicules from the delicate quadrate mcshwork, and are distinctly marked off from one
another by a somewhat sharp-edged, spicule-bearing border, extending down to the bushy
basal portion. Towards the external margin, the flat body which measures 3 to 5 cm. in
thickness, becomes gradually sharper ; the maximum height is 40 cm., and the breadth
about as much. The somewhat irregular knobbed base is continued into a bushy beard-
like basal tuft of spicules 10 to 12 cm. in length. By this the sponge is fixed among
the coral and other detritus (PI. XLIX.).

The parenchymal skeleton consists of large or medium-sized smooth oxjr|[Dentacts,
which probably had their four tangential rays originally inserted in some bounding
surface, while the fifth stood radially. Afterwards, however, they came to be
embedded in the parenchyma. The angles of the five rays are generally, though by no
means always, right angles, and one or other of the rays not unfrequently exhibits a
simple curvature near its origin. The individual rays usually have a length of 10 to 20
mm. Throughout the whole parenchyma irregularly scattered, small, lank oxyhexacts
also occur, with rays of about equal length, straight or slightly curved, and usually some-
what roughened, i.e., beset with small pointed tubercles, which are occasionally longer, and
project obliquely outwards, as represented in PI. L. fig. 6. Less frequently uncinate
forms occur, but only near the two limiting surfaces, and usually in radial disposition.
Some uncinates only attain a length of 2 to 4 mm. (PI. L. fig. 3), but most are much
longer. The short, smooth, spindle-shaped oxydiacts, which occur so abundantly in the
parenchyma of Poliopogon gigas, to be described below (PI. XL VIII. figs. 3, 7), are here
wholly absent.

The supporting spicules of the gastral skeleton are, as in all Hyalonematids, strongly de-
veloped, smooth oxypentacts of varied dimensions. Their radially directed ray may attain
a length of 1 to 2 cm., while the four tangentials, crossed approximately or exactly at right
angles, and but rarely bent, may be as long or longer, and are apposed to one another in
twos or threes to form the familiar quadratic lattice-work. The autodermal pinuli are
somewhat small pentacts about 0"4 mm. in length, with straight spines directed obliquely
upwards and outwards. The outer end of the distal ray passes into a long thin point ;
the basal portion is smooth. The four moderately long (O'l mm.) basal rays, are internally
smooth, but are on their outer halves beset with short, distant, outward directed teeth,
and end in slightly conical points. While disposed at right angles to the distal ray, they
do not form right angles with one another, but two opposite obtuse and acute angles,
with a slight curvature in the two acute angles so that the form of the central
portion of a 8 results (PI. L. fig. 5). Numerous eight-rayed amphidiscs of various size,
but of similar form occur in the dermal membrane, and appear to penetrate thence into
the parenchyma. The larger have the middle portion of their axis rod inserted in the
dermal membrane, while the one end projects freely to the exterior, and the other into a
subdermal space. Some of these have a length of 0"2 mm. and a moderate thickness, are

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

tolerably smooth, exhibiting four to eight protruding bosses on the middle of the axis rod,
and bear bell-shaped umbels with somewhat divergent rays of a flat paddle-like form
(PI. L. fig. 8). Others, measuring O'l mm., have a thinner, knotted axis rod, and umbel
rays with somewhat pointed ends, while small forms also occur, measuring 0'06 to 0'04
or even 0-03 mm. in length (PI. L. figs. 10, 11, 12).

Very similarly constituted are the spicules which form the skeleton of the other, or
gastral side of the curved body. Here also the same oxypentacts occur as hypogastralia,
the same pentact pinuli as autogastralia, and amphidiscs of similar form and equally varied
dimensions (PI. L. fig. l).

In the lateral wall of the larger efferent canals there are also oxji^entacts similar to the
hj'podermal and hypogastral forms, though less strongly developed and less long. Pinuli,
however, are absent, but the very abundant amphidiscs exhibit the same size and form as
those of the skin.

The long lank marginalia which form the marginal fringe round the edge of the whole
sponge, usually measure some mm. (up to 1 cm. or more) in length. They are for the
most part quite smooth, form internally a long thin point, while externally they bear
small, somewhat distant lateral teeth, which project obliquely upwards and outwards.
They finally end externally in a slight club-shaped or bud-like swelling (PI. L.
figs. 1, 4). The long strongly developed spicules (5 to 10 cm. in length), which
project like a beard from the thickened lower portion of the sponge body into the mud,
are quite smooth. They end internally, i.e., in the sponge body, in a simple point, while
towards the lower outer free end they first decrease gradually in thickness, and then again
slowly increase, finally forming a double toothed very gently curved anchor. The two
teeth of the latter stand out almost at right angles from the shaft, are only slightly bent,
and end in a somewhat blunt point (PI. L. fig. 7).

It is especially interesting to note that in the larger ca^dties of one of the specimens of
Polwpogon amadou some small, approximately spherical sponges were found, measuring
about 3 mm. in diameter, and undoubtedly young forms of Poliopogon. Since their
tissue was still in tolerable preservation, it was possible to make sections thi'ough the
small bodies. These results were important not only for the species but for the whole
genus, and indeed for the family of Hyalonematidse.

I have given in PI. L. fig. 2 a diagrammatic representation of the structure exhibited
in successful axial sections of the young sponge forms. The characters more or less
distinctly prominent on various sections have been combined in a synthetic figure.
A central cavity with which all the diverticula of the folded chamber-layer directly
communicate, opens to the exterior at the upper pole of the somewhat transversely oval
section. This external aperture is not, however, completely free and open, but seems still
to be covered by a delicate membrane. The membrana reticularis which forms the
chamber-like diverticula, is perfectly continuous, forming a much folded and puckered

IlEPORT ON" THE HEXACTINELLIDA. 257

membrane, the diverticula of which all project outwards against the smooth distended
outer skin. On the inner side of the larger diverticula, a delicate network extends, while
outside of the reticularis, between it and the reticulate dermal membrane, there is the
external trabecular network.

The external layer of the dermal membrane is supported by smooth simple, hypodermal
oxypentacts, with four tangential rays, intersecting parallel to the surface and somewhat
curved, whUe the proximal ray, at right angles to the latter, is inserted radially in the
parench}Tna. Long slim spicules project radially on all sides for a greater or less distance
from the sponge body. They are either simply pointed at both ends, and thus in part to
be described as incipient uncinates, or exhibit on the freely projecting portion minute
teeth or spines directed obliquely outwards, and on the outer end a slight club-shaped
toothed swelling with a narrow terminal point, while the internal proximal portion
remains smooth. Both in the external skin and in the parenchyma there are numerous
amphidiscs of the same form and size as those in the adult Poliopogon amadou. On
the lining of the gastral central cavity similar forms occur, which have been overlooked
in the figure (PL L. fig. 2). At the inferior pole, opposite the oscular opening, the anchor
forms above mentioned occur, in part buried in the parenchyma, in part more or less
protruded. They resemble those of the adult but are of much smaller size. It is
striking that the parenchyma is wholly destitute of the small hexacts and of the
pinuli on the external and internal sides of the bounding surface.

2. Poliopogon gigas, n. sp. (Pis. XLVIL, XLVIII.).

An immense sponge form, which turned out to be a Poliopogon, was trawled between
the Raoul and Macaulay Islands, to the north of New Zealand, (Station 170, lat. 29°
45' S., long. 178° 11' W.) at a depth of 630 fathoms, from a volcanic ground. The
roundish cubical mass has a diameter of 50 to 70 cm. The lower wholly torn surface
still exhibits the remains of bundles of spicules, which have undoubtedly formed, along with
others now lost, a thick broad basal tuft. The lateral rounded external surface is
tolerably smooth, without projecting bundles of spicules or isolated radial prostaha, and
covered by an incompletely preserved fine quadratic dermal lattice-work. The upper
much injured terminal surface bears a median hemispherical depression, from 15 to 20 cm.
in breadth and the same in depth, representing the gastral cavity (PI. XLVIL). The
marginal fringe of the oscular aperture is unfortunately not preserved.

The spicules of the parenchyma, which is penetrated by lacunae and canals of vary-
ing width, essentially resemble those of Poliopogon amadou. Here too we find smooth
oxypentacts of variable size, with long straight or slightly curved rays. The original
disposition of these pentacts seems to have been not so much in the parenchyma, as in
the external or internal limiting membrane or in the walls of the larger canals. Small

(ZOOL. CHALL. EXP. — FART LIII. — 1887.) Ggg 33

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

oxyh exacts with slightly curved slender rays occur, similar to those which were found
so abundantly in the former species (PI. XLVIII. figs. 1, 10). In certain regions, for
instance somewhat abundantly below the outer skin, a skeletal element, not represented
in Poliopogon amadou, occurs — in the form of small spindle-shaped smooth oxydiacts,
(0'15 to 0*25 mm. in length) which have a maximum thickness either in the neighbour-
hood of their centre (PI. XLVIII. fig. 3) or more rarely nearer one end (PI. XLVIII. fig. 7).
Here and there, especially in the neighbourhood of the external surface, long uncinates
occur, with narrow appressed pointed barbs covering minute depressions. I have not
been able to determine whether the amphidiscs of varied form and size, which occur
scattered in great abundance in the parenchyma (PL XLVIII. fig. 1), are really true
parenchymalia, or have grown in from the bounding surfaces of the external skin, of
the gastral cavity, or of the canals.

The dermal skeleton consists of rather strongly developed smooth oxypentacts.
The autodermal pinules are indeed for the most part pentacts, but others not unfrequently
occur, in which the sixth proximal radial ray is more or less distinctly developed
(PI. XLVIII. figs. 8, 11). The four basal rays of the ordinary pentact pinules are some-
what long, and bear externally short spines projecting obliquely outwards, while the
extreme outer end is rounded off or even truncated, though rarely pointed. As a rule
the basal rays are like those of the dermal pinuli in Poliopogon amadou, bent like the
middle portion of a figure 8 (PI. XLVIII. figs. 5, 9). The freely projecting, radial distal
ray is not so long and slender as in Poliopogon amadou, but rather broad and only
about 0"15 mm. in length. It is beset with lateral spines, which are curved somewhat
markedly outwards, and terminally apposed in bud-like fashion. Less frequently the distal
ray ends in a point (PI. XLVIII. fig. 5). The large amphidiscs lying in the dermal
membrane have a length of about 0'2 mm. Their axial rod is somewhat uniformly beset
with small tubercles. The campanulate umbels have usually eight, less frequently
more rays, which are of considerable length, often almost meeting, and but slightly
divergent (PL XLVIII. fig. 2). Besides these, there is an abundant occurrence of small
amphidiscs with short, hemispherical, eight-rayed terminal umbels and slender axial rod
(PL XLVIII. figs. 4, 6).

Whether the gastral skeleton surrounding the cavity closely resembles the dermal I
was unable certainly to determine, owing to the want of sufiicient material from that
portion of the giant sponge. It is, however, extremely probable, from analogy with the
related species of Poliopogon amadou, that the structure of the canalicular skeleton
lining the larger ducts and lacunae essentially resembles that of the skin. Pinules are
not wholly absent, but occur as slender, scattered, canalicular pentacts or hexacts, with
long, narrow, pointed basal rays, and a slender free radial, which ends in a delicate point,
and bears isolated, short, straight, oliliquely projecting lateral spines. In the larger ducts
and lacunse, numerous large and small amphidiscs occur of the form described in the

REPORT ON THE HEXACTINELLIDA. 259

dermal skeleton. No marginalia are preserved. In the thick beard-like shreds which here
and there still project from the base, numerous anchors occur, of the same form as those of
PoUopogoti amadou, with a shaft which may measure 10 cm. or more, while the slightly
forward-curved ends of the two transverse teeth are separated from one another by about
0'5 mm. (PI. XL VIII. fig. 14, a). The pointed upper end of the anchor shaft, which is
buried in the sponge body, frequently bears several short, scattered, downward-directed,
lateral barbs (PI. XLVIII. fig. 14, h). In PL XLVIII. fig. 13, I have figured a remark-
able abnormal form of anchor structure.

Subfamily 2. Semperellin^, F. E. Schulze.
Genus Semperella, Gray (Pis. LL, LIL).

1868. Semper, VerhandJ. der Wiirzb. Gesellsch., vol i. p. 29 (Hyalonema schuUzei).

1868. Gray, Ann. and Mag. Nat. Hist., vol. ii. p. 373 (Semperella).

1868. Herklots and Marshall, Arch. Neerland. d. sc. nat., iii. p. 435 {Hijalothauma lu/lekingii).

1872. Carter, Ann. and Mag. Nat. Hist, vol. x. p. 110.

1872. Gray, Ann. and Mag. Nat. Hist., vol. x. p. 131.

1872. Gray, Ann. and Mag. Nat. Hist., vol. x. p. 76 {Meyerina claviformis).

1873. Carter, Ann. and Mag. Nat. Hist., vol. xL p. 275.

1873. Carter, Ann. and Mag. Nat. Hist., vol xii. p. 349.

1874. Gray, Ann. and Mag. Nat. Hist., vol. xiii. p. 284.

1875. Carter, Ann. and Mag. Nat. Hist., vol. xvi. p. 200.

1875. Marshall, Zeitschr. f. wiss. ZooL, Suppl., Bd. xxv.

1876. Marshall, Zeitschr. f. wiss. Zool, Bd. xsvii.

1877. Miers, Journ. Linn. See. Lond. (Zool), vol. xiii. p. 506.

History. — In 1868 Semper described under the title Hyalonema schuUzei a Philippine
sponge, resembling in size and form Euplectella aspergillum.^ Semper's preliminary
note was to the following effect : — "Hyalonema SchuUzei, S. resembles in form and size
Euplectella aspergillum. The root-fibres, which are either smooth or serrate, di^^de
towards the sponge body into single bundles. They traverse the latter both internally
and superficially, becoming connected with the main framework in the same way as do
the long root fibres of Euplectella. Cruciate spicules of many kinds are united with
these long strands of fibres, forming sometimes a very dense and sometimes a loose net-
work, which is penetrated in all directions by the large canals of the sponge. The com-
paratively wide excurrent appertures occur irregularly over the whole sponge, and are
frequently associated with tufts of fine, almost silk-like fibres. On several regions of the
somewhat injured surface, a fine network with wide rectangular meshes may be observed.
The whole sponge framework is, as in all true Hyalonemata, composed of free fibres or

'■ VerhandL phys.-Tned. Gesdhck. Wiirzhurg, vol. i. p. 29, 1868.

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

cruciate spicules. Sometimes, however, several spicules are fused together, and the origin
of the connected siliceous framework of EuplecteUa is thus suggested. The shapes of
the numerous free siliceous bodies recall those of Hycdonema sieholdii, Gray, from Japan."

In the same year (1868) Gray proposed to designate those sponges, which were in
every way so different from the known species of Hyalonema, by the new generic title
Semperella} Moreover, to a specimen of the same species from the island of Ceram the
name Hyalothauma ludekingi was given by Marshall and Herklots. Another sponge
belonging to the same species, was sent from the Philippine Island of Zebu through
Dr. A. Meyer to the British Museum, and was shortly described by Gray in 1872,^ under
the title Meyerella claviformis. Carter gave a detailed analysis of the same specimen,'
and he changed the generic name Meyerella which had been given by Gray, into
Meyerina because the former had already been applied to one of the Lepidoptera. In
regard to the forms of the spicules. Carter regarded the new species as a combination
of Carteria, Hyalonema, Holtenia and Pheronema.

Gray now erected * for this sponge a special family — the Meyerinidse — and character-
ised it in the following manner: — " Sponge elongate, tubular, covered with a cobweb-
like netted coat, with a circle of tufts of anchoring fibres at the base which extend more
than half way through the length of the body, and then by repetition of a shorter kind,
are continued on to the apex, where they also form a circle of tufts round the margin of
the apical aperture."

In his first systematic catalogue of the known Hexactiuellida Carter * united Meyerina
claviformis, Gray, with Holteyiia and Pheronema into one group. On pi. xiv. he
has compared the extremities of the tuft spicules, which are very like the anchors of
Holtenia and Meyerina.

In his Classification of the Spongidfe, published in 1875,'' Carter formed the group of
" Birotulifera " in the family of the Sarcohexactinellida, of the genera Hyalonema,
Holtema, Meyerina and Laharia.

The specific agreement of the variously designated forms was first recognised by
Marshall, who, in his researches in the Hexactiuellida' in 1875, pointed out that, with
exception of the generic name Hyalonema used by Semper, the oldest designation is that
of Semperella schultzei, and that this name is therefore entitled to be retained as the
proper one. In his accurate and careful description of the two specimens at his
command — one of which was found at Ceram and the other at Zebu — Marshall pointed

1 Ann. and Mag. Nat. Hist., ser. 4, vol. ii. pp. 373-377.

* Ann. and Mag. Nat. Hist, ser. 4, vol. x. p. 76, 1872.

2 Ann. and Mag. Nat. Hist., ser. 4, vol. x. p. 110, 1872.

* Ann. and Mag. Nat. Hist., ser. 4, vol. x. p. 134

5 Ann. and Mag. Nat. Hist., ser. 4, vol. xii. p. 162, 1873.
^ Ann. and Mag. Nat. Hist, ser. 4, vol. xvi. pp. 199, 120.
' Zeitschr.f. leiss. Zool, Bd. sxv., Suppl., pp. 212-225.

REPORT ON THE HEXACTINELLIDA. 261

out that the dense sieve-plate openings which occur abundantly on the truncated lateral
borders of the five-sided prismatic body, and are surrounded by delicate spicules project-
ing in a cuff-like manner, are the oscular openings of the main anastomosing canal system
running longitudinally in the walls of the sponge. The larger hollow spaces occurring
in the axis and opening above he designated pseiidogasters, and their terminal lattice-
work closing plates as pseudo-sieve-plates. A second system of passages and canals, which
lie between these exhalent oscular openings, separated from them by sponge tissue, and
covered towards the outside by fine dermal lattice-like networks were referred by
]\Iarshall to the subdermal spaces of Haeckel, or to the intermargincd cavities of
Bowerbank. In consequence of these results of his examination, Marshall,^ in 1876,
characterised the genus Semperella in the following manner : — " Polyzoic with
pseudogasters. Anchor bundles anastomosing with each other throughout the whole body
wall. Dermal skeleton of cross spicules separated by the tissue of the body and spread
over large subdermal longitudinal spaces into which the internal canals open. Oscula of
the individuals in rows, with peristome wreath and sieve-plate. The gastral skeleton
formed of large four-rayed spicules, provided with meshes occlusible by means of fir-tree-
like spicules, and in direct connection with the dermal skeleton. The cavities of the
pseudogastral system are covered internally by six- and five-rayed spicules."

Semperella schultzei, Semper.

Near the Philippine Island, Zebu, the Challenger Expedition obtained a beautiful
specimen of Semperella schultzei, 38 cm. in length, and 5 to 7 in thickness. This form
having been well preserved in spirit, remains almost uninjured. The cylindrical body,
which measures 30 cm. in length exclusive of the basal tuft, has inferiorly a cylindrical
form, but becomes gradually wider upwards, forming an irregular pentagonal prism with
truncated edges. The latter are from 5 to 8 mm. in breadth, and from the middle of
the body upw\T,rds do not extend in exact longitudinal dii-ection, but extend on the one
hand obliquely, and further divide and anastomose, till they finally unite in the
superior flat cone, which occupies the terminal region of the upper end. The root-
process, which extends perpendicularly downwards from the lower end of the sponge,
becomes widened out inferiorly into a loose brush, and penetrates by means of its
diverging spicules into the very varied detritus-substratum. There was no trace of
commensal Anthozoa.

It is noticeable, even with the unaided eye, that a striking difierence can be observed
in the structure of the external layer of skin on the above mentioned truncated lateral
edges, and that on the intervening flat, or even shghtly concave, lateral surfaces (10 to 30
mm. in breadth). For while the latter exhibit a very delicate narrow-meshed quadratic

I Zeitschr. f. roiss. Zool., Bd. xxvii. p. 131.

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

lattice-work, witli meshes which do not attain a width of a ^ mm., and are separated by
yet thinner strands, the skin on the rib-like projecting truncate edges exhibits a much
more irregular network with polygonal or roundish meshes, 1 to 2 mm. in diameter, and
with firm whitish strands passing laterally into a compact marginal fringe about 1'5 to
2 mm. in breadth, which forms the boundary between the narrow-meshed quadratic net-
work of the sides, and the wide-meshed firmer sieve-network of the rib-like edges (PI. LI.
figs. 1, 16). Through both these networks, which difi"er so much from one another, the
subjacent cavities are seen, shining through as a labyrinth of united passages. On the
superior, somewhat obliquely pyramidal or flatly conical, truncate extremity of the sponge,
the -ttdder network, strengthened by firm solid junction-plates, is terminally expanded and
united, while the fine quadratic dermal lattice-work is confined to the sides, with the
exception of several angular or rounded terminal prolongations, which extend beyond
the lateral terminal margin on to the summit. Where some of this lateral dermal
sheath has been rubbed off by accident or design, the labyrinthine passages and spaces
are directly exposed, and it can be seen how the canals and cavities beneath the narrow
C[uadratic dermal network form a connected, continuously distinct anastomosing system
of canals, a little finger's breadth thick, which are separated by a wall of 1 mm. thick-
ness from another adjacent canal system, which likewise forms a connected network of
anastomosing spaces and passages. This second system of passages extends directly under
the wide-meshed, more irregular lattice-work, which we have noted on the rounded off
lateral edges. The two canal-systems are everywhere separated only by a thin partition,
and their canals extend side by side, not only under the external skin, but penetrating
inwards, traverse the whole internal body, so that the broad, irregular, longitudinal canal
in the axial region is an integral part of that system of passages, which extends below
the wider iiTegular network of rounded side margins, and which opens at the superior
terminal plate. • This remarkable presence of two completely separated canal systems
is represented in PI. LI. fig. 16, in the lateral portion of the sponge from which the
external wall has been partly removed, and also in the diagrammatic cross section of
the whole sponge in PL LII. fig. 1.

Through the fine quadratic dermal network which covers the slightly convex lateral
surfaces of the body, the water passes first into the subjacent canal system, by which it
is carried to all parts of the body, penetrating at length through the partition wall (which,
though only about 1 mm. thick, contains the memhrana reticularis and forms the proper
parenchyma) into the second system of canals which is in connection with the central
longitudinal canal or gastral cavity. Thence the water reaches the exterior by the
efferent ducts, namely, either by the wide-meshed oscular sieve-network of the longitudinal
side edges, or by the superior terminal region.

That the relation of these two adjacent, but perfectly distinct, labyrinthine canal
systems to the flow of water is as above described, is demonstrable not only from their

EEPORT ON THE HEXACTINELLIDA. 263

relation in reference to the skin, or from the connection between the second and the
main central canals, but also from the structure of the soft sponge body, and especially
from the form and disposition of the chamber layer, or more correctly of the system of
irregular diverticula from the memhrana reticularis. As in all Hyalonematidse the latter
extends in the parenchyma between afferent and efferent canal-system in such a way,
that the convexity of all the diverticula is directed outwards against the entering
stream of water, that is, against the afferent system of canals. Every longitudinal or
transverse section of the sponge body shows that the memhrana reticularis found in the
partition between the two systems is so disposed or manifoldly bent outwards, that the
convexity of each small protrusion is against the processes of those canals which lie under
the fine-meshed quadi-atic dermal network on the flat or slightly convex sides (PI, LIL
fig. 3).

The canals below the dermal membrane, which extends in the form of the fine-meshed
quadratic network, represent wide subdermal spaces, and those penetrating inwards are
somewhat uniformly wide afferent canals which do not break up into branches, but form
an anastomosing labyrinth (PI. LIL fig. 1).

Large supporting spicules are represented in the parenchjrma by a few medium-sized
oxyhexacts, and by numerous oxypentacts, some of them with very long rays. Long
uncinate spicules also occur (PL LL fig. 3), disposed in brush-like strands or groups, at
right angles to the skin, where the parenchymal canal-wall is inserted on the external
skin. Besides these, we note smaller spicules of the following types : — Firstly, small
oxyhexacts with slim rays of equal or diverse length, beset to a variable extent with
somewhat distant, straight or slightly curved spines, which are inserted either at right
angles or approximately so (PI. LL fig. 15; PI. LIL fig. 5); secondly, oxydiacts of a
similarly spinose character, and with four short, smooth, conical median rays or spines
intersecting at right angles, and representing the survival of the other four degenerate
rays of the hexacts (PI. LIL fig. 4); and thirdly, isolated short uncinates with smaU short
barbs (PI. LI. fig. 6).

The dermal skeleton, which supports the fine-meshed quadratic lattice-work, consists
of very varied, strongly developed, hypodermal oxypentacts, with long smooth tangential
rays, which are often somewhat strongly curved at the base, though sometimes but
slightly developed, and which are apposed to one another to form the quacbatic lattice-
work, while the more or less long, straight, proximal tangential ray, extends to a variable
distance inwards, and often projects freely into the hypodermal canal (PI. LII. fig. 3).
Only where the internal wall of the canal is connected with the skin does the tangential
ray lie throughout its whole — often considerable — length in the parenchyma. On the
tangential strands of hypodermalia, and on the bands of the dermal network extending
between the former, autodermal pentact pinuli occur, usually disposed in rows. The four
thick and moderately long, externally spinose, terminally somewhat conical or rounded basal

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

rays of these pinuli exhibit a median curvature like that of the figure 8, as we formerly
saw in the dermal pentact pinuli of Poliopogon. The free strongly developed distal ray
has usually a length of about 0"15 mm., and is beset with strong, though not long, lateral
spines bent like hooks upwards and outwards. The very upper end forms a free point
(PL LI. figs. 5, 13).

As to amphidiscs, I find a large isolated form about 0'25 mm. long, with thick knotted
axial rod, and short, broad, terminally somewhat transversely truncate umbels. The eight
broad rays are paddle-like and terminally rounded (PI. LI. fig. 10). Beside these, though
also but sparsely, a similar form of medium-size occurs (PI. LI. fig. 8), and somewhat
more abundantly the familiar small type with hemispherical terminal umbels, 0"02 to 0"03
mm. in length (PL LII. fig. 7). More frequent than in the stretched, freely exposed,
dermal lattice-work, is the occurrence of these various amphidiscs in the portions of the
skin which lie over the insertion of the parenchymatous canalicular wall (PL LII. fig. 3).

The same kinds of spicules, but with somewhat different development and dimensions,
compose the skeleton which forms the wide-meshed sieve-network in the oscular regions
of the bevelled sides and of the superior extremity of the whole sponge body. The large
supporting pentacts exhibit indeed the same fundamental form and dimensions as those
of the dermal membrane, but reductions of individual rays very frequently occur, with
the production of rounded ends, or with the formation of slight, terminal, club-shaped
swellings. The downward curvature at right angles, exhibited by some of the tangential
rays, is also of very common occurrence. Especially remarkable, however, is the con-
siderable length attained by the free distal ray of the pinuli. As a rule, this thick ray,
beset with short hook-like lateral spines, measures 0*5 mm. in length, and ends in a
projecting point. The four basal rays are moderately long, less curved, and terminally
somewhat spinose and rounded (PI. LI. fig. 4). Besides these, shorter pinuli occur with
slimmer distal ray. The strong eight-rayed amphidiscs (figured in PL LI. fig. 10) occur
here more abundantly than in the dermal lattice-work, and always exhibit in successful
preparations the characteristic disposition represented in PL LII. fig. 3. Numerous
eight-rayed amphidiscs of similar form, but of medium size occur, and also quite minute
forms with hemispherical umbels.

It is noteworthy that neither pinuli nor amphidiscs occur on either of the bounding
surfaces of the parenchymatous septa between the afierent and efi"erent canals, but only
the familiar parenchymalia, including numerous medium-sized oxypeutacts, with numerous
bent, or with more or less reduced rays. I have, however, frequently found — instead of
the familiar small oxypentacts with several laterally projecting, somewhat bent spines,
— pentacts of similar structure, in which the sixth ray was either wholly absent, or
represented only by a- small conical elevation.

The freely projecting portion of the basal tuft exhibits spicules up to 20 cm. in length,
which terminate superiorly in a very gradual point, while the lower end forms a strongly

EEPOET ON THE HEXACTINELLIDA. 265

developed bifurcate anchor. The simple tooth- or paddle-shaped transverse arms of the
latter exhibit a much thickened base rising from the club-shaped swollen terminal portion,
and in the larger forms are connected by a distinct lateral fringe on either side. AVhile
the upper half of the long anchor spicules, which is for the most part buried in the sponge
body, is perfectly smooth, gradually increasing in thickness downwards, the lower half,
which becomes gradually narrower, bears small barbs increasing in height and breadth.
These are spirally disposed round the shaft, and decrease again in size towards the very
end, disappearing entirely a little above the anchor. Thus the much narrowed, and just
above the anchor yet slimmer, terminal portion appears quite smooth (PI. LI. fig. 14).

Among the many Hexactinellids which were trawled by the Challenger near the
Little Ki Islands (Station 192, lat. 5° 49' 15" S., long. 132° 14' 15" E.), from a depth of
192 fathoms and a blue mud ground, there was a beautiful specimen of a Semperella,
which agrees in so many points with Semperella schnltzei, Semper, that I regard it as
a young form of the latter. There are indeed some differences, but these may be
explained as characteristic of the young form. The specimen has a total length of
10"5 cm., of which only 5"6 go to the stretched cylindrical body, and 4*9 cm. to the
inferiorly much broadened root process, which is almost wholly enveloped in a Palythoa
encrustation (PL LII. fig. 2). At its broadest the body measures 12 mm., and the
somewhat bevelled longitudinal edges, which measure about 2 mm. in breadth, anastomose
abundantly by means of cross processes. They project somewhat more markedly than
in the adult specimen, and are bounded by a fringe of perpendicularly inserted, straight,
rod-like spicules (PI. LII. fig. 3), which may be called marginalia, and are almost wholly
absent in the large specimen. These slim spicules, which in many places form a con-
tinuous garniture and in others are not discoverable, are quite smooth on their internal
pointed half, while the free external portion is for the most part beset with small lateral
teeth (turned upwards and outwards), and ends in a delicate point, below which there
is usually a small club-shaped swelling with two or four lateral teeth. Similar needles
also project in the above described young specimens of Poliopogon amadou, both on the
oscular margin and on the lateral surface, but are not discoverable on the side of the adult
form. This fact seems to make it especially probable that we have here to do with a
young specimen, and not with another species, and that the more, since all the other
spicules closely resemble in form and disposition those of the larger specimen, as is equally
true of the structure of the soft tissue.

Suborder II. DICTYONINA, Zittel (Pis. LXXI.-CL, CIIL, CIV.).

Hexactinellida in which the principal hexacts are already at an early stage united
into a connected and compact (dictyoual) framework in a more or less regular fashion.

(zOOIi. CHALL. EXP. — PART LIII. — 1887.) Gfgg 34

,_j

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

Tribe I. UNCINATAEIA, F. E. Scliulze (Pis. LXXI.-XCVIIL).

Dictyoniua with uncinates.

Subtribe I. Clavularia, F. E. Schulze (Pis. LXXI.-LXXVL).
Besides the pentact hypodermalia and hypogastralia radially disposed clavuljB occur.

Family F A E R E i D ^. (Pis. LXXI.-LXXVL).

The dictyonal framework forms in the youngest regions of the body a single-layered
network with quadrate meshes, from the nodes of which conical protuberances project on
either side towards the dermal and gastral surfaces.

Genus Farrea, Bowerbank (Pis. LXXL-LXXVL).

1857. Owen, Trans. Linn. Soc. Lond., vol. xxiii., jjI. xxi. figs. 8, 9.
1862. Bowerbank, Phil. Trans., vol. clii. p. 747, Tab. xxxii. fig. 7.
1864. Bowerbank, Mon. Brit. Spong., Ed. i. ]). 204.
1864. Kblliker, Icones bistologicas.

1867. Gray, Proc. Zool. Soc. Lond., p. 492, Tabs, xxvii., xsviii.

1868. Bowerbank, Proc. Zool. Soc. Lond., p. 118.

1869. Bowerbank, Proc. Zool. Soc. Lond., p. 66.

1869. Wyville-Thomson, Phil. Trans., vol. clix. p. 701.

1870. 0. Schmidt, Grundziige einer Spongienfauna des atlantischen Gebietes, p. 16.
1870. Saville Kent, Monthly Mior. Journ., p. 241.

1873. Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. xii. pp. 17, 349.

1875. Bowerbank, Proc. Zool. Soc. Lend., p. 272.

1875. Marshall, Zeitschr. f. wiss. Zool., SuppL, Bd. xxv. p. 142.

1876. Bowerbank, Proc. ZooL Soc. Lond., p. 535.

1876. Marshall, Zeitschr. f. wiss. Zool., Bd. xxvii. p. 113.

1877. Sollas, Ann. and Mag. Nat. Hist., ser. 4, vol. xix. p. 1.
1880. 0. Schmidt, Spongien des Meerbusens von Mexico, p. 43.
1885. Carter, Ann. and Mag. Nat. Hist., ser. 5, voL xv. p. 387.

History. — In the root-tuft of that beautiful siliceous sponge which was on one
occasion presented to Captain Etheridge by the king of the Seychelles, and which, after
it passed into the possession of Dr. Arthur Farre, was accurately studied by Richard
Owen, and described under the name of Euplectella cucumer,^ fragments of the sUiceous
skeletons of other sponges were entangled. Some of these skeletal fragments, which had
already been observed by Dr. Farre, were subjected by Owen, along with the Euplectella,
to a closer examination, and he even figured two of them, both in the natural size and

1 Trans. Linn. Soc. Lond., vol. xxiii., pi. xxi., 1857.

REPORT ON THE HEXACTINELLIDA. 267

also under a magnifying power of ten and twenty diameters, in his pi. xxi. figs. 8, 9.
In the course of his paper he describes the one piece in the following words : — " An
irregular network, more or less bent, with subquadrate meshes, sometimes crossed by
oblique threads," while, with regard to the other he said : — " The meshes of the network
are on nearly the same plane, and of a more regular square form, with a short pointed
spiculum projecting from one side of each decussation of the threads like the teeth of
a harrow."

Some years later Bowerbank also communicated the results of his examination of the
same object. He considered the two skeletal fragments which had been described
and separated by Owen to be parts of one and the same sponge. The harrow-like
network of siliceous beams represented by Owen in his figs. 8 and 9, with its regular
quadrate meshes and pointed teeth projecting perpendicularly from the crossing points,
he regarded as the dermal skeleton, under which only the irregularly formed inner body
skeleton was said to have lain. After more precise examination of both pai-ts by the
aid of stronger magnifying powers, Bowerbank saw that the beams of the fragment
regarded by him as belonging to the inner body skeleton were completely perforated by
a manifest axial canal, and were richly covered externally with fine spines, while the
beams of the (dermal) skeletal network which formed quadrate meshes appeared solid and
smooth. Only the teeth projecting at right angles to the knob points appeared rough,
Bowerbank, moreover, drew attention to the fact that the rough teeth stand at right
angles to the plane of that quadrate network not only on its outer side, but on both
surfaces.

To this form, the body and dermal skeletons of which he had examined in the above
two fragments, Bowerbank gave the generic designation Farrea, in honour of the fortunate
possessor of that specimen of Etiplectella cucumer in whose root-tuft the forms were
found, and he added to this the specific name of occa on account of the great similarity
of one of the specimens to a harrow.^ In his Monograph of the British Spongiadse
(part i. p. 204, 1864), Bowerbank referred his Farrea occa to the sponges with a " canali-
culated siliceo-fibrous reticulate symmetrical skeleton," the "fibres" of which being
" composed of concentric layers of solid sUex, mth a continuous central canal," and he
added : — "The fibres in Farrea occa are rather coarse, abundantly tuberculated, and the
mode of reticulation is rectangular." The inner body skeleton referred to in this latter
description is figured in pi. xv. fig. 277, and in the description of the plate is referred
to as "simple fistulose siliceous fibre, spinulated"; while the solid and smooth network
of beams which, in Bowerbank's opinion, belongs to this sponge and constitutes its
dermal skeleton, is figured on pi. xxi. fig. 311, and in the description of the plate is
referred to as a "quadrilateral siliceo-fibrous network showing the double series of
entirely spined spicular organs projected from its angles."

1 Lat., occa.

Mas

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

Finally Bowerbank again gave, in 1869/ a more minute description of his
Farrea occa, and supplied three new figures on pi. xxiv. figs. 1, 7. In this latter
very detailed communication the siliceous network with its Cjuadrate meshes, regarded as
dermal skeleton and compared in form to "a harrow, is represented by Bowerbank just as
formerly ; on the other hand, in addition to the more irregular and rough network of
beams regarded as belonging to the inner framework, a number of variously formed isolated
siliceous spicules are figured and described as accessory parts of the skeleton. These
have four, five, or more rays, and do not present the right angles of the Hexactinellidan
spicules. The generic characters of Farrea were summarised by Bowerbank in
the same paper (p. 76) in the following manner: — "Skeleton siliceo-fibrous. Fibres
canaliculated, canals continuous. Eete symmetrical ; interstices rectangulated." The
view first announced in the well-known paper by Wyville Thomson On the Vitreous
Sponges,^ is noteworthy ; it is to the effect that the framework of beams which in
the skeleton of Farrea forms a system with exactly square meshes has arisen by an
amalgamation of regular hexradiate spicules.

Among the deep-sea sponges collected by Count Pourtales in the Caribbean Sea
Oscar Schmidt found in 1870^ several irregular dichotomously branched tubes from
2 to 6 mm. in diameter. These were attached by a plate-like expansion, were thick-
walled at the base, and became towards the wide open upper extremity gradually
thin-walled and fragile, till finally on the outermost and doubtless youngest parts of the
little tubular tree only a single layered network of siliceous beams and square meshes was
found. From the intersections of the latter rough slender conical teeth projected on
both sides. In addition to this siliceous network, which, in its youngest parts at least,
presents a certain resemblance to the harrow-Hke siliceous network of beams of Farrea
occa which Bowerbank regarded as a dermal skeleton, Oscar Schmidt also found and
described the following free siliceous spicules : — (1) long spindle-spicules beset with barbs ;
(2) long spicules which run out at one extremity to a point, and are provided on the
other with a hemispherical or slightly convex and marginally toothed umbel-like roof or
cap; (3) hexradiate spicules, in which each of the rays is beset on its extremity with three
pronged, thin teeth, with minute convex terminal umbels ; (4) thin spicules which run
out to a point at one extremity, and are provided on the other somewhat expanded
end with five to eight bristle-like narrow prickles which project in a brush-like manner.
These various spicules, which occurred in special abundance in the neigbourhood of
the surface, were regarded by Oscar Schmidt as sufiiciently characteristic of a new
species distinct from the Farrea occa of Bowerbank, and this he designated Farrea
facunda.

In the same year (1870) Saville Kent described,* along with several other Hexac-

1 Proc. Zool. Soc. Loud., p. 339. 2 ^,„[_ and Mag. Nat. Hist., 1868.

* Spongien des Atlaatisohen Gebietes, p. 16. * Monthly Micr. Journ., November 1870.

EEPORT ON THE HEXACTINELLIDA. 269

tinellicla collected on the coasts of Spain and Portugal, a full-grown siliceous skeleton
which he identified as Farrea occa, Bowerbauk.

This form consists of a continuous branched tube exhibiting anastomoses here and
there, and measuring from 5 to 8 mm. in diameter. The tube opens by a slightly funnel-
shaped, expanded and projecting cup about 1 cm. in width. To conclude from the figures
given * all the tubes consist only of a simple framework with quadrate meshes. Although
in this the specimen described by Saville Kent quite agrees with the network of siliceous
beams, regarded by Bowerbank as the dermal skeleton of his Farrea occa, yet it difi"ers
from the latter in the circumstance that the beams of the framework are not smooth
externally, but are beset with pointed tubercles. In this character it rather resembles
the siliceous network regarded by Bowerbank as part of the inner body skeleton of his
Farrea occa.

On the same Lophohelia stock upon which the specimen which was determined as
Farrea occa, Bowerbank, had been found, Saville Kent also observed some " small
fistulose ramifications bridging over the minor interspaces between the branches of the
coral." He regarded this sponge, in spite of the great similarity with the adjoining Farrea,
as a species belonging to another, and even a new genus, which he named Aulodictyon.
From the genus Farrea this is said to difi'er in the following points : — " In Farrea
the basal skeleton is composed of a single reticulated lamina ; in Aulodictyon the
basal skeleton consists of a complex reticulate tube between, and continuous with
the primary meshes, of which an abundant network of coalescing simple hexradiate
stellate spicula occurs." Moreover " the minuter spicula of the sarcode" are also said to
be " of an entirely different type." And since Oscar Schmidt had, in the case of his
Farrea facunda, described no single-layered dictyonal framework in the basal part, but a
complicated network of several layers, and isolated spicules, similar to those found by
Kent in his Aulodictyon, the latter was of opinion that Schmidt's Farrea facunda ought
to be referred as Aulodictyon facundum to his new genus.

As I shall show further on, however, all tubular species of Farrea have in their
basal portion a dense dictyonal framework of several layers, and also the " long attenuate
spicules inflated at the extremity and reflecto-peltate, with a dentate margin or with
a simple series of recurved hooks," which Kent looks upon as characteristic peculiarities
of his genus Aulodictyon. Consequently the genus Aulodictyon must be entu-ely
abandoned.

In the survey of all the above described HexactineUida which Carter gave in 1873,^
this accurate observer first sharply distinguishes the two sdiceous networks found on the
already often mentioned original specimen oi Fiq)lectella cucumer, Owen. These had been
described by Bowerbank as dermal and body skeleton of one and the same sponge, while

1 Loc. cit., pi. Ixiv. fi-s. 12, 13, 14.

^ Aim. and Mag. Nat. Hist, ser. 4, voL xii. p. 3G0.

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

Carter distinguished them as body skeletons of tivo dij^erent s2Jecies which he named
Farrea occa, Bowerbank, and Farrea densa.

Carter confined the name Farrea occa, which had been applied to both parts (taken
together), to the siliceous network (regarded by Bowerbank as a dermal skeleton), which
consists throughout of one layer, surrounds exactly square meshes, and bears at the
intersections rough conical pegs on both sides. In justification of this, Carter notes that
only this skeletal portion in reality resembles a harrow (occa), and that it had therefore
been specially considered by Bowerbank in applying the name.

In his examination of this species {Farrea occa, Bowerbank), besides using some of
the fragments collected during the " Porcupine" Expedition, and the specimens studied
by Saville Kent, Carter also employed the fragments which had been obtained from the
basal tuft of the beautiful Euplectella cucumer studied by Owen. He was, however,
only able to procure completely macerated and greatly eroded specimens, and accordingly
could not study the spicules that occur freely on the soft parts. Yet he believed that
some spicules which were casually included here and there in the continuous framework
of beams might with probability be interpreted as belonging to Favrea occa. With
regard to the formation of central canals in the beams of the siliceous framework Carter
was convinced from the direct examination of these numerous remnants of Farrea occa,
that the entire network of beams, with its rectangular meshes, had not originally possessed
a continuous canal system, but, as this arose by the amalgamation of isolated hexradiate
sjiicules, it at first consisted merely of the separate hexradiate canals, which terminated
blindly at the six extremities, and belonged, of course, to the individual hexradiate
spicules. These separate axial canals corresponding to the individual hexradiate spicules
usually became very manifest after the death of the animal. On account of internal
absorption or solution they are specially wide and striking, and this points to a previous
more prolonged maceration of the dead sponge.

A sponge described by Carter under the designation of Farrea infundibidiformis,
from the Caribbean Sea, presents a small funnel-like body, with a much widened thin
margin, having an opening of about 2^ cm. in diameter, and with a solid round stalk of
about 1 cm. in length and ^ cm. in thickness. The sponge is attached to its stratum
by means of the somewhat expanded inferior extremity of the stalk. The skeletal frame-
work of the solid stalk consists of a dense lattice- work, with more or less distinctly
defined rectangular meshes, and is continuous wdth the skeleton of the funnel-like j^late,
so that in the middle of the plate a rectangular lattice-work remains, while more irregular
networks of fibres extend over the two surfaces. The beams of the rectangular lattice-
work are beset with small spines, those of the irregular network of fibres exhibit still
finer spines and bear numerous simple, hexradiate, lateral spicules. In such soft i^arts
between the foramina of the siliceous network as were visible in the di'ied condition,
Carter found numerous floricomes with minute laterally spinose terminal knobs.

REPORT ON THE HEXACTINELLIDA. 271

During 1875 and 1876, in his Monograph of the Silico-fibrous Sponges, parts iii.-vi.,^
Bowerbank ascribed no fewer than fourteen new species to his genus Farrea. These
he names Farrea gassioti, Farrea pocillum, Farrea Jistulata, Farrea Isevis, Farrea
parasitica, Farrea valida, Farrea spinosissima, Farrea spinifera, Farrea spinulenta,
Farrea aculeata, Farrea rohusta, Farrea inermis, Farrea perarmata, and Farrea
irregularis. Since, however, the description of these new species was usually based
only on a small fragment without characteristic form, and withal more or less macerated
and injured, it is impossible, in most cases, to determine (notwithstanding the perfect
figures given under a magnifying power of 36 or 80 diameters) to which species
the fragment in question properly belongs. It is much to be regretted that, in
almost all these Bowerbankian species of Farrea, the free spicules of the dermal
system were not preserved or are not sufficiently clearly figured, and accordingly the
words used by Bowerbank at the close of the introduction to the monograph referred
to must be employed; he says: — "When the expansible dermal system is present,
wholly or in part, in specimens under examination, we are enabled to establish specific
characters of external form and structural peculiarities of the most satisfactory
description, but when that important portion of the organic structure of the sponge
is absent the characters derived from the form and surface of the rigid skeleton
are necessarily provisional, and can maintain their ^jlaces in its description only until
a specimen in a natural and perfect state can be procured." Since Bowerbank,
moreover, lays the very greatest weight on the width of the axial canals and regards
these — as his generic diagnoses "fibres canaliculated, canals continuous" indicate — as
at least essentially in a condition of perfect continuity, whereas, as Carter has already
observed, in the skeletal framework of the Hexactinellida, the axial canals of the
individual hexradiate spicules are at first unconnected, while the width depends chiefly
on the condition of the skeleton as regards maceration or solution, it is conceivable
that the specific characters and differential features given by Bowerbank have often
little value.

The two types indicated by Bowerbank {loc. cit., p. 272) as Farrea gassioti, and
Farrea pocillum agree so thoroughly in form, size, and structure that it is really only
the different breadths of the central canal of the network of beams that form the
difference. But this distinction is seen to be insignificant when it is observed that,
even according to Bowerbank's own statement, some greatly macerated specimens of
Farrea gassioti have the central canals remarkably wide and clear. Further, in the
case of the two specimens of Farrea pocillum, which have the soft parts dried and
richly provided with the included isolated spicules, and which were accordingly,
without doubt, secured as fresh specimens, the canals are especially delicate, and in
some places even almost imperceptible. Moreover, the great similarity in form, size,

' Froc. Zool. Soc. Lond., 1875, 1876.

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

and structure, which these two Bowerbankian species bear to the Farrea infundihxdi-
formis, Carter, which was described two years previously by Carter,^ but was entirely
unknown to Bowerbank, is certainly striking. Bowerbank, indeed, says nothing of
the elegant floricomes, which Carter's specimen possessed in rich abundance, but it is
quite possible that these fine forms may have escaped Bowerbank in the course of
his examination, and that accordingly all three species may perhaps l^e united into one.

The fragment which Bowerbank has designated as Farrea Jistulata consists of a
bent tube of 23 mm. in length and 6 mm. in transverse diameter. It is not only open
at both extremities, but further exhibits a lateral orifice to which a short tube of equal
width is attached. The skeleton with its two or three square lattice-layers consists
of smooth, cylindrical beams, with wide central canals. From the intersections of the
outer as well as of the inner surface tolerably long tuberculated teeth project.
Bowerbank also describes a thin dermal membrane with numerous spicules ; but since
the latter are called " acerate " and " contort bihamate," and since, moreover, a sj)ecimen
of Hymedesmia johnsoni, Bowerbank, which is distinguished by "acerate" and " trenchant
contort bihamate " spicules, was found to have settled in the interior of the tube, there
can be no doubt that those free spicules of the outer skin-like casing do not belong to
the Hexactinellid any more than the spicules of those Desmacidonids which are firmly
fixed in the interior. The whole fragment may have lain dead for a long time on the
bottom of the sea before it was brought up, and the width of the central canal of the
beams is on that supposition readily explained.

Farrea Iwvis is the name given by Bowerbank to a new species, of which he
possessed only a small fragment, consisting of a tube-like siliceous network from 4 to 5
mm. in breadth, and scarcely 1 cm. in length. Since the entire system of beams which
encloses approximately square meshes only forms a single layer, Bowerbank looks
upon this as a " dermal network." The beams of the network are smooth throughout,
and so too are the long conical teeth which project outwards and inwards at right angles
to the intersections. The central canals, of which two, or sometimes even three, lie close
to one another, are very wide, and become confluent at the angles.

Since the small fragment found in dredged sand was provided only in one small
place with a thin skin consisting of a dried soft mass without any spicules, it may be
inferred that the skin did not belong to the sponge, and here too, the breadth of the
central canals was caused by longer solution in sea-water. Bowerbank finally notes
the great resemblance between his Farrea lievis and the branched, partially anastomos-
ing, tubular network, figured by Saville Kent^ and designated Farrea occa. He also
calls attention to the fact, that that branched tube-like form agrees with a certain
" Farrea tubulata." I have carefully endeavoured, but to no purpose, to discover in

1 Ann. and Mag. Nat. Hist., vol. xii. p. 448, 1873.

2 Monthly Micr. Journ., pi. Ixiv. fig. 12, 1870.

REPORT ON THE HEXACTINELLIDA. 273

the literature of the group something about tliis Farrea tubulata, which, so far as I am
aware, is referred to only in this place.

As to the irregular network of beams with central canals of various widths, which
adhered to the inner side of a Farrea gassioti, Bowerbank, and which is spoken of by
Bowerbank as Farrea parasitica, one can scarcely say more than that it belonged to a
Hexactinellid, for the fact of its connection with the genus Farrea is not even plausible
from Bowerbauk's own description and figure.

The Farrea valida of Bowerbank was established upon a small fragment of a
reticulated skeleton. The fragment, which is regarded by Bowerbank as representing
the dermal skeleton, is distinguished by the possession of uniform square meshes, by
thick, smooth, tubular beams in the network, and by short, tuberculated, conical teeth
and prickles. This dermal skeleton is associated with an inner body skeleton formed of
distinctly thinner, smooth beams, but of this skeleton little is preserved. The numerous
hexradiate crowded spicules, which are placed upon, and have one ray at right angles to
the beams of the framework, are noteworthy.

Bowerbank designates by the name of Farrea spinosissima a fragment of a skeleton,
consisting of several reticulated layers, and provided with more or less square meshes.
From the tubular beams, which are of various thickness, though generally slender,
numerous long pointed prickles project at right angles, and are either smooth or slightly
beset with spines.

Farrea spinifera was erected by Bowerbank on the basis of a skeletal fragment,
whose beams, which are of various thickness and provided with a central canal, surround
more or less regular meshes, and are distinguished by long, transversely disposed, or
slightly conical prickles.

A small skeletal fragment, which Bowerbank has named Farrea spinulenta, differs
from most of the others hitherto referred to in the very convenient circumstance, that in
the present case a portion of the soft tissues, along with the free spicules that occurred
in them, was preserved in the dried specimen. It is true that here also it is not quite
certain whether the dried mass in question really belonged to the same sponge as the
siliceous skeleton, but it is in the highest degree probable. The framework consists
of a simple network with square meshes and long rough teeth, which project on both
sides at right angles to the knots of the network.

As was to be expected, Bowerbank failed to find any central canals in the beams
of the siliceous network of this fragment, which was certainly taken from the sea bottom
in a living condition. The skeleton was not yet macerated and dissolved internally
like the majority of the other specimens of various species of his genus Farrea, which
have been referred to above. These apparently solid beams of the network are, however,
not quite smooth, but are beset with small pointed tubercles or with parallel longitudinal
rows of fine prickles, which stand at equal distances from, and alternate with one

(ZOOI. CHALL. EXP. — PAKT LIU. — 1887.) Ggg 35

274 THE VOYAGE OF H.M.S. CHALLENGEE.

another. Among the free spicules which are present in the soft parts, and are especially
abuudnnt in the skin, Bowerbank notes simple hexradiate spiciiles, with smooth or finely
fringed rays, and also floricomes, with small terminal plates on each of the four
S-shaped terminal rays of every principal ray. Bowerbank notices the great resemblance
between this species and Farrea occa, but he at the same time calls attention to some
small differences, such as the slightly roughened surface of the beams of the network, the
slender form of the teeth, and the peculiar floricomes.

Of Farrea aculeata, Bowerbank, the author possessed only a fragment, the form of
which pointed to a cup-like form for the animal. The usually two-, more seldom three-
layered lattice-like framework consists of strong, canaliculated, siliceous beams, which
surround tolerably regular square meshes, and are irregularly beset with conical prickles
and thorns of various sizes. Here and there hexradiate spicules which have grown
upon the above also occur. This species has a striking resemblance to Farrea spinifera,
as Bowerbank himself recognised.

The skeletal framework of Farrea robusta, Bowerbank, has a close resemblance to
that of Farrea aculeata. It differs from it strictly only in the greater size of the beams
and in the corresponding narrowness of the square meshes, as also in the greater
thickness of the fully developed rough hexradiate spicules — differences which everywhere
occur between older and younger parts of the same sponge. The soft parts and
IMonactinellid spicules found by Bowerbank in the framework are quite independent, and
belong to a Desmacidonid which had settled in the skeleton after the death of the
Hexactinellid.

Another fragment of a siliceous framework with several lattice-like layers, and provided
on lioth surfaces of the somewhat bent plate with meshes which are uniformly square
but more irregular in the interior, Bowerbank has named Farrea inermis, because the
beams are entirely or almost entirely smooth. From the knots of the network slender
smooth teeth project at right angles. Similar prickles also project here and there into
the inner meshes.

Bowerbank has designated as Farrea perarmata a flat, slightly bent, skeletal fragment
in which the framework of beams, which forms square meshes, is everywhere richly beset
with spines and prickles of various sizes, and exhibits no recognisable central canals.
Long rough conical teeth project outwards and inwards. Numerous thorny hexradiate
spicules also occur with one ray in each directed at right angles to the beams of the
network.

Another bent skeletal fragment, consisting of a tolerably irregular framework
of beams which here and there exhibit square meshes, Bowerbank names Farrea
irregularis. The individual beams have a very varied thickness, and for the most part
perfectly smooth, isolated, small hexradiate spicules also occur.

If we review these fourteen species established by Bowerbank during the years

REPORT ON THE HEXACTINELLIDA. ' 275

1875 and 1876, we can obtain only in four cases any adequate conception of their
form; namely, on the one hand, in the case of Farrea gassloti and the very closely
related Farrea 2)0cillum, both of which have the same cup-like form ; and on the other
hand, in the case of Farrea Jistulata and Farrea lievis both of which are tubuliform
and are also closely related to one another. Which of the other ten species belong
to the one or the other of these two groups of forms, or perhaps to neither, it is
scarcely possible to say on account of the small size and imperfect condition of the
fragments examined. According to Bowerbank's own representation his Farrea
spinulenta seems to agree for the most part with the old Farrea occa, Bowerbank, so far
as regards the lattice-like framework. If this agreement were indeed more intimate
it would be of great importance for the establishment of the generic character of
Farrea, because some free siliceous spicules were found and described by Bowerbank
in the dried soft parts, and these are absent in the other cases. We must regret that
Bowerbank has communicated no definite information in regard to the free spicules
which he mentions as having been found in "various forms" in an original fragment of
his Farrea occa (and indeed in the so-called dermal skeleton); he only says (page 561)
that in Farrea spinulenta one may find that "the large simple rectangulate sexradiate
interstitial spicula with spinous radii, a few of which are entangled in the inner surface
of the dermal rete, also form efficient specific characters, none such having hitherto been
found in Farrea occa."

In his description of the sponges of the Gulf of Mexico — 1880 — Oscar Schmidt
maintains his previous diagnosis of the sjDccies Farrea facunda. To the spicules
then described as characteristic he adds a hexradiate form in which every individual ray
divides at the extremity into from two to four (usually three) fine bristle-like teeth,
and he conjectures that in the previously examined specimens certain spicules represented
on his pi. V. fig. 9, which bear on each of their thickened extremities four transversely
arranged and somewhat recurved hooks, might be present as well as in the specimens
subsequently studied. He further reports that in the majority of his numerous specimens
those peculiar broom forms (Besengabeln) figured on pL i. figs 18, 19,' and in some the
umbelled spicules represented in fig. 2 of the same plate, were not to be found. He is,
however, by no means inclined to erect a specific difference on that account, and that
the less since the familiar fir-tree-like spicules occurred here and .there instead of the
absent " brooms." Only one form — in which the absent umbel and hook spicules are
represented by the so-called knob- broom form, that is to say, by forks with several teeth
which do not run out to points but terminate in terminal knobs, and which were in the
meantime described by Carter as Eurete farreopsis — should be excluded from his Farrea
facunda, and at any rate become the representative of a special species. On the other
hand, the genera Eurete and Aulodictyon, which were placed near Farrea hj Semper

' Spongien des atlantischen Gebietes, 1870.

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

and Kent, be regards as untenable, and would without further delay unite them to
Fai'rea. According to Schmidt's view all the forms ascribed to these two genera, with
the single exception of Carter's Eurete farreopsis, belong to the circle of foiTQs round
his Farrea facunda.

In a Report on a Collection of Marine Sponges from Japan, made by Dr. T.
Anderson, Carter gives a detailed and accurate description of the skeleton of a Japanese
Farrea, which he designates Farrea occa, Bowerbank.^ The conical and branched
tube exhibits in its upper portion "an extremely thin and delicate reticulated wall,
rendered denser lower down (towards the base) by additional matter of a similar
kind. Extremities of the branches open and dilated. The points of intersection of
the quadrangular reticulum are marked on each side by a long, curved, spiniferous process
or spur, which is directed upwards, thus supporting the sarcode or soft parts and its
spicular contents, both externally and internally, that is, on each side of the skeletal
Avail." The loose dermal spicules exhibit five forms, viz., (l) hexradiate with outer ray
aborted ; (2) acerate, straight, unsymmetrically fusiform, sharply pointed at each end,
spinose at short intervals throughout (" barbulae ") ; (3) nail-shaj^ed forms, with a long
straight shaft, slightly swollen and pointed at the free end, and expanded horizontally at
the other into a circular head, more or less spinose at the edge, very abundant and
varied ; (4) hexradiate rosettes with four straight arms without central swelling, ending
in four divergent rays disposed round a central one, rays smooth and simply pointed, or
else more or less capitate and beset with small spines ; (.5) smaller hexradiate forms, in
which all the four rays are equally developed, straight, pointed and spinose.

Since Carter has designated this Farrea from Japan (off Misaki) by the title Farrea
occa, and described it with sufficient clearness to admit of its being recognised,
Bowerbank's name may be retained for this definite form. Pis. LXXL, LXXIL, LXXIIL,
and LXXVL were printed before I received Carter's paper, so that I was unable to
change the specific designation Farrea haechelii into Farrea occa (Bowerbank)
Carter.

Generic Cliaracter. — A simple tubular stalk, attached by its compact expanded base,
bears a dichotomously branched anastomosing tube-work, the walls of which are supported
by a dictyonal framework with rectangular meshes. In older specimens the dictyonal
framework is below of more than one layer, but it becomes gradually thinner towards the
upper end, and finally forms a single-layered meshwork. From each of the intersections
of this network a stout rough spine projects at right angles both inwards and outwards.
Besides the pentact hypodermalia and h5q30gastralia other dermalia occur as clavulse with
a nail-like, clavate, or verticillate shape. Besides uncinates, the parenchyma contains
oxyhexasters, and in some sj)ecies discohexasters.

Many species of Farrea have been already erected and described, but mostly with

1 Ann. and Mag. Nat. Hist., ser. 5, vol. xv. p. 387, 1885.

REPORT ON THE HEXACTINELLIDA. 277

reference to macerated specimens. Only one of tliese can I regard as sufficiently diagnosed
to permit of certain recognition, namely, the form lately described by Carter, and desig-
nated by him as Farrea occa, Bowcrlxank.

1. Farrea occa (Bowcrbank) Carter (Pis. LXXL, LXXIL, LXXIII. ; PI. LXXVI.
figs. 1-3).

Both the material of the Challenger Expedition and the collection of the
Hexactinellida brought by Dr. Doderlein from the Sagami Bay, Japan, include
numerous specimens of this species of Farrea. These are partly spirit specimens with
the tissue preserved, and partly dried forms. They exhibit considerable diflerences both
in size and form.

It is unfortunate that not one of all the specimens is quite intact. The outermost
ends of the tubes are generally broken off for a greater or less distance. I hope, however,
that the representation given in PI. LXXII. fig. 1, of a macerated skeleton in lateral
aspect, and those in PI. LXXL figs, 1 and 2, from photographs of spirit specimens in lateral
and superior asjiect, will give a correct conception of the general habit of this sponge.
It is frequently richly branched, forming composite masses sometimes 12 cm. in height.

The simple hollow stalk is attached by a flat expansion to a more or less compact
substratum, sometimes consisting merely of a crumbly mass of clay. This expanded
portion consists of an irregular tuberculate plate, which is closely appressed to the
substratum. In its centre it has a thickness of 0'3 mm. or more, but becomes gradually
thinner towards the irregularly frilled edge, forming a delicate smooth margin. On the
free upper surface of this compact basal plate there are usually some radially disposed
furrows, from 1 to 2 mm. in breadth, whicli sometimes divide externally into two or three
narrower branches. From the lumen of the round tubular stalk which rises from the
middle of the plate, a round excurrent aperture leads either through the plate itself
straight dowmwards, or just above the plate through the wall of the tube (PI. LXXII. fig. 2).
The stalk, which usually stands erect at right angles, has a diameter of 5 to 10 mm. and
an equaUj' short length. It passes immediately by division and gradual expansion into
the crowded anastomosing tube-work of the stock. The ultimate ends of the tubes form
thin-walled smooth margined cups projecting freely on the somewhat uniformly arched
convex surface of the whole complex. The dichotomous division of the single tubular
ends takes place in the following fashion. The cup-shaped expansion extends transversely,
usually at right angles to the last plane of bifurcation, the two long parallel margins
thus formed approach one another in the middle and fuse, so that the cup becomes
divided into two independent and diverging tubes (PI. LXXII. fig. 3). After a more or
less prolonged growth these again experience a dichotomous division of the same sort.
The length of the tubular portion seems to vary greatly in different stocks and also in the

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

diflferent regions of one stock. In PL LXXII. fig. 4 I have represented a fragment of a
specially elongated tubular system. This was trawled to the west of Manila (Station 207,
lat. 12° 21' N., long. 122° 15' E.) from a depth of 700 fathoms and a blue mud ground,
while in contrast to this the specimen trawled in the Bay of Sagami by Dr. Doderlein,
from about 150 fathoms and a clay bottom (PI. LXXI. fig. 2), exhibited quite short
tubular branches of at most 1 cm. in length. In the largest specimen of this lot,
which is figured in PI. LXXII. fig. 1, the tubes have a length of only 1"5 to 2 cm.

In the spirit specimens, in which the whole soft tissue remains perfectly intact
with the dermal and gastral membrane, both the external and the inner surface are
uniformly marked by fine dark points arranged in transverse and longitudinal rows
crossing at right angles (PI. LXXI. fig. 2). Where the soft tissue has been removed by
maceration or by mechanical means, the beautiful quadratic network of the dictyonal
framework is more or less distinctly seen (PI. LXXII. figs. 1, 3).

The diameter of the tubes increases continuously from the simple basal portion to the
freely projecting, widely open, terminal cups. On quite young specimens only a few
centimetres in height, the stalk measures hardly 3 mm. in breadth, just above the plate
of attachment, while in older specimens external thickening results in a transverse
diameter of 5 to 10 mm. While the cup-shaped ends of the tubes in young specimens
have usually a width of only 5 mm. or so, the terminal openings of the large stocks are
frequently 15 to 20 mm. in diameter.

Since the dichotomous ramification begins at a very early stage at variably distant
intervals and not in the same plane, but in planes often disposed at right angles to one
another, no long straight main trunk results, but an irregular much twisted system of
tubes, expanding on all sides. The anastomosis seems to occur partly as the result of
direct union between newly formed tubes which meet one another in their growth, and
partly as the result of the apposition of adjacent tubes which are at first externally united,
and afterwards come into direct communication by the absorption of the intervening
partition. Now and then, but on the whole rarely, I have observed one or other of the
tubes, bent downwards or to the side, and closed by a transverse porous plate, while all
the others remained open. I can only regard such a closure as an abnormality, which
has probably resulted as a protection against the injurious influence of mud or the like.

In regard to the microscopic structure of the dictyonal framework, I have in the first
place to notice, in opposition to the previous statements of other investigators, that the
greater part of the whole tubular system, and especially the superior youngest portion, is
supported by a single-layered network, but that in the lower, and therefore older
portions, and also at the base of all the stocks which are not very young, the framework
exhibits two la3-ers or more (PI. LXXIII. fig. 2 ; PI. LXXVI. fig. 3). The network en-
closes approximately cubical meshes. This many-layered framework has undoubtedly
arisen in this way, that to each distal ray of the single-layered lattice-work the proximal

REPORT ON THE HEXACTINELLIDA. 279

ray of a hexact of the new outer layer is apposed, and both become enveloped in a
common concentrically laminated siliceous sheath. It is obvious that in this way the
older and lower portions of the tubes may gradually increase in thickness.

The thick, firm, spongy mass, which forms the pedicel on the older stocks and the
flatly expanded basal plate, has had another origin. Besides the apposition of a new
layer of the cubical lattice-work, we have here to deal with the interposition of numerous
smaller hexacts in the already formed meshes. And while the external apposition of
new layers leads in general to the regular formation of rectangular or cubical meshes, the
interposition of small hexacts does not usually occur in the radial tangential direction,
but at very various angles to the main strands, so that an irregular narrow meshed
network results (PI. LXXVI. fig. 2). Where the basal plate is in direct contact with
the substratum, as in all other Hexactinellids fixed on a solid basis, a narrow meshed
bounding plate is formed, which arises mainly by the abundant development of
synapticula between the bands of spicules.

While the strands of the single-laj'ered dictyonal network which lie parallel to the
surface of the tube are cylindrical and smooth, the intersections of the net exhibit
conical prominences projecting at right angles both outwards and inwards, and always
beset with small tubercles or rough elevations. In the many-layered dictyonal frame-
work, the radial beams extending between the la5'ers, which lie parallel to one another
and to the tube wall, exactly resemble the tangential beams in their cylindrical shape
and in their smooth surface, while here also the prominences projecting from the outer
and inner surface of the whole lattice-work are always tubercled and rough. The length
of these freely projecting conical prominences varies as much as their form, and that
between tolerably wide limits. They are generally straight or only slightly curved ; are
longer in the younger portions, especially in the single-layered framework, and shorter on
the surface of the many-layered framework of the older regions. On the surface of the
stalk and on its basal expansion they are either altogether absent, or represented only I33'
small knob-like elevations.

At the points of intersection of the dictyonal framework there is a slight thickening
gradually increasing with age, but this is not in any way marked off from the strand, nor
different from the latter in the character of its surface.

The composition of the whole framework from separate hexacts may be most readily
recognised on those specimens which have remained dead in the sea for some time and
are thus macerated. The axial canals of the individual hexacts have through the
dissolution of the loose inmost layers become more or less markedly -nidened, and are
thus in their disposition and extent readily observed (PI. LXXVI. fig. 5). In fresh
specimens, however, they can usually be detected as very fine canals (PL LXXI. fig.
3, &c.)

The length of these axial canals is often surprising. While they usually extend only

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

to tlie proximity of the axial cross of the next spicule, where they are rounded off and
terminate, they are sometimes continued right along to the middle of the second spicule ;
thus three axial canals may not unfrequently be seen running alongside in one strand.

In his last description of Farrea occa ^ Carter states that the conical prominences
which project freely on either side of the narrow meshed lattice-work are bent outwards.
I cannot, however, admit the accuracy of this as a general statement. As a rule the
prominences are straight and directed at right angles to the surface. As Carter correctly
notes, they gradually decrease in length towards the base of the stock.

As to the loose parenchymalia lying between the beams of the dictyonal framework,
I shall first discuss the uncinata. These have been already well described and figured
by Carter (/oc. cit.). His description is as follows :— " Acerate, straight, unsymmetrically
fusiform, that is, the outer or j)rojecting part being thicker than the inner two-thirds
of the spicule, which is thus rendered long and whip-like ; sharp-pointed at each end,
spined at short intervals throughout, the spines long, smooth and slender, respectively
supported on bracket-like processes, which, being spirally arranged around the shaft, give
it an iri-egular zig-zag appearance, all sloping in the same direction, that is backwards, or
towards the sponge, at a very slight angle upon the shaft, whose outer or thickened end is
extremely sharp." He calls them barbulas,^ a diminutive of the English barb, and says
that they are " common to almost all the Hexactinellida." This assertion I cannot, how-
ever, admit. Among the numerous Lyssacina which I have examined, structures like
these uncinates only occur in the genera Pheronema, Polio])ogon, and Semjyerella.
Among the Dictyonina I have not found them in the genera Myliusia, Gray ; Dacty-
localyx, Gray, Scleroplcgma, 0. Schmidt, Ilargaritella, 0. Schmidt, Euryplegma,
F. E. S., Joanella, 0. Schmidt, and Aulocystis, F. E. S., and these may therefore be
designated as Inermia, in opposition to the Dictyonalia with uncinates — the Uncinataria.

As Carter notes, the uncinates exhibit in Farrea occa an extraordinary variability
in length, though this depends to some extent on the age of the individual uncinates and
of the regions of the sponge in which they occur. Further, in the form, number, and
arrangement of their inturned spines, there are in one and the same sponge very marked
difierences. Earely, and only in the younger portions of the tubes, I find the spines
distant and very divergent from the stem as represented in PI. LXXI. fig. 4, and
PI. LXXIH. fig. 3. Sometimes a spiral arrangement of the spines can be detected; in
other cases this is impossible, and the distribution appears to be quite irregular. The
expansion of the distal third portion of the uncinates as contrasted with the interior is
very marked, especially on the large and strongly developed forms, where, too, the inner
end is not unfrequently distinguished from the outer by the development of weak spines.

' Arm. and Mag. Nat. Hist, ser. 5, vol. sv. p. 388.

^ Since I had used the Latin term uncinatum (a rod beset with barbs) before I received Carter's memoir, and since
Latin or Greek terms seem to me to have the preference as being cosmopolitan and not national, I have adhered to my-
original designation.

REPORT ON THE HEXACTINELLIDA. 281

All the uncinates in Farrea occa are disposed at right angles to the surface, and thus
traverse the wall of the tube in a radial direction. The sharp external point does not,
however, penetrate the dermal membrane, but at most reaches it. The inner weaker,
and always pointed ends remain at some distance from the gastral membrane, varying
according to the relation between the length of the uncinates and the diameter of the
tube-wall.

Besides the uncinates, parenchymalia are represented by simple oxyhexasters in more
or less abundant occurrence. In these the long smooth principal rays divide into two to
four ■^\ddely diverging, straight or gently curved terminals, half as long as the principals
(PI. LXXI. fig. 7). These oxyhexasters occur most abundantly in the subdermal and
subgastral trabecular space, but more sparsely in the middle layers of the tube-wall, even
between the chambers, without, however, being altogether absent (PI. LXXI. fig. 3 ;
PL LXXIII. fig. 2). They are embedded in the trabecular network itself, and I am
inclined to think that the formation of the trabeculse and their interconnection is of the
greatest importance in relation to the form of these spicules. In Carter's specimen of
Farrea occa the terminal rays of the rosettes were as a rule smooth and pointed, though
occasionally microspined and more or less " capitate." In the specimens which I
examined, on the other hand, there were never any traces of terminal knobs or discs on
the terminal rays of the rosettes, that is to say, of transitional forms between oxyhexasters
and discohexasters. I have, however, found specimens of Farrea in which the rosettes
were exclusively discohexasters, and these of an unusual type, but these were representatives
of a second species, namely, Farrea sollasii. Other specimens in which, besides the
oxyhexasters, discohexasters also occurred, belonged to a third species, Farrea vosmaen.
But if the different forms of rosettes as above noted really occur in one and the same
sponge, with distinct transitional types, then the forms which I have designated as two
distinct species, Farrea sollasii and Farrea vosmaeri may perhaps be regarded as
varieties of Farrea occa. I have not, however, as yet discovered such transitional forms,
and shall therefore meanwhile continue to regard these divergent specimens as distinct
species, until then- relation to Farrea occa be indisputably established.

Besides the uncinates and the above mentioned rosettes, the parenchyma of Farrea
occa contains simple hexacts, of the same size as the rosettes, with smooth or rough,
straight, pointed rays, which usually exhibit a knob-like thickening on their point of
intersection. While these simple oxyhexacts have only a rare and isolated occurrence in
the upper younger portions of the Farrea stocks, they are abundant in the lower portion,
and in older stocks even numerous in the pedicel, and in the basal plate, within the
meshes of the dictyonal framework, where they are abundantly fused to the dictyonal
beams, so that by the gradual thickening of the rays, and their union with adja-
cent strands, they contribute to the thickening and compacting of the connected
skeleton.

(ZOOL. CHALL. EXP. — PART LIII. — 1887.) Ggg 36

282 THE VOYAGE OF H.M.S. CHALLENGER,

The dermal membrane of Farrea occa is supported by the four rectangularly
intersecting tangential rays of simple pentacts, whose fifth rays penetrate into the
parench}Tna at right angles to the surface of the tube, while here and there a tubercle or
knob-like distal elevation remains as the persisting trace of a sixth (distal) ray. While
the proximal ray is always perfectly straight and slightly narrowed into a conical form
towards the pointed end, the four tangential rays are usually bent gently inwards, and end
in a slight usually blunted thickening, or in a somewhat knob-shaped swelling (PI. LXXI.
fig. 6). The proximal ray usually exceeds the tangential in length, is always cylindrical,
and generally smooth with the exception of a slight roughness at tlie point. The tangential
rays, on the other hand, are somewhat flattened on their outer side, and beset with
smaller or larger tubercles which are sometimes pointed, sometimes rounded and boss-like.
These tubercles or spines are usually strongly developed only on the gently convex and
somewhat flattened side ; on the lateral edges they are more weakly developed, and are
altogether absent on the inner side of the tangentials. In some cases they have an
isolated occurrence on the internal side. Towards the end of the rays the spines
increase in height, and occur more densely and all round.

The disposition of these pentacts in the dermal membrane is frequently after this
fashion, that the quadratic framework, which is formed from their apposed tangential
rays, corresponds exactly to the subjacent dictyonal framework, each proximal ray being
apposed to a distal prominence of the latter, and the tangential rays lying parallel to the
beams of the dictyonal lattice-work. The meshes of the dermal lattice-work thus of
course correspond in size to those of the dictyonal framework.

In other cases, as for instance in that on which the figures on PI. LXXII. fig. 5 and PI.
LXXIII. fig. 1 were based, each distal prominence of the dictyonal framework corresponds
indeed to the proximal ray of a dermal pentact, but besides this there is above the middle
of each mesh of the dictyonal lattice-work a dermal pentact; and the tangential rays of
all the dermal pentacts do not lie parallel to the beams of the dictyonal lattice, but
diagonally to the quadratic or rectangular meshes. Often enough, however, no definite
relation between the dermal pentacts and the dictyonal lattice can be determined except
this, that above most of the distal rays of the latter there occurs a pentact of the
dermal skeleton. The extent to which the corresponding opposite tangential rays of two
adjacent dermal pentacts may overlap one another, varies in the difi"erent regions of the
sponge. Each tangential ray frequently reaches almost to the nodes of intersection of
the neighbouring pentact (PI. LXXII. fig. 5).

Besides the pentacts above described, the dermal skeleton of Farrea occa contains
clavulee, which are so far in connection with the former since they always lie close beside
the proximal ray. While they form an externally divergent tuft, the inferior pointed ends
of the long stalk are usually in contact with the proximal ray of the related pentact, and
the heads with their convex terminal discs lie about the level of the dermal membrane or

i

EEPORT ON THE HEXACTINELLIDA. 283

project more or less beyond it, and fill up to a greater or less extent the four angles of
the cruciate tangential rays of the pentact. Contrary to Carter's report, according to
which six clavulaj usually accompany each pentact, I generally found eight round each
in the best preserved specimens, and disposed so that two lay in each of the angles of the
tangential rays. This number does not seem, however, to be in any way constant. Not
unfrequently I have found ten or more clavulaj in one bundle, and very frequently fewer
than eight, and sometimes very few. It is difficult to determine whether they may be
sometimes wholly absent; where this appears to be the case, it is quite conceivable that
they may have fallen out or have been torn away. Nor does the fact that they frequently
project more or less beyond the surface of the skin necessarily imply that they have this
position in the living sponge. On the contrary, in the best preserved specimens
the normal position seems to be that in which the hemispherical terminal discs occur
within the dermal membrane. Deformation of the body-waU or pressure of apposed
foreign bodies has forced them outwards, as may be inferred indeed from their
shape.

The form of the dermal clavulse in Farrea occa varies considerably, as Carter has
already noted, and that between somewhat wide limits, from the long-toothed umbels to
the smooth club (PL LXXII. fig. 8), nevertheless one definite form is so predominantly
frequent, that it may be regarded as the normal type, of which all the others are only
exceptional modifications (PL LXXI. figs. 5, 9). This typical dermal clavula exhibits a
club-shaped or poppy -head-Like terminal swelling, which is externally and laterally
roofed over by a hemispherically arched umbel, frec^uently raised in the centre into a
projecting boss. The overhanging margin of the umbel is toothed. The teeth, which
number eighteen to twenty-two are either simple marginal teeth resulting between
notches of the margin, or are more or less sharply defined off from the edge and laterally
compressed even at the base. The cylindrical stalk is gradually narrowed towards the
lower, simply rounded extremity, and is usually straight or gently bent, and, like the
poppy-head-like extremity, is beset with small irregularly disposed tubercles, which
increase somewhat in size and number towards the narrowed lower end (PL LXXI.
fig. 5).

Various deviations from the above described typical form often occur both in different
regions of the same sponge and in specimens obtained from difi"erent localities. The
variation is especially associated with the terminal portion which we have shortly
designated the " head." In thickness and length, for instance, it varies greatly. In many
cases the stalk shows towards its upper end at most a gentle thickening, which bears
terminally a thin toothed umbel like that of a toad-stool, while in others there is a broad
club-shaped terminal portion from which the umbel region is but slightly, or even not
at all marked off", so that a perfectly smooth club-like swelling may result. While the
terminal portion of the umbel is sometimes quite flat and without any central elevation

284 THE VOYAGE OF H.M.S. CHALLENGER

(PI. LXXI. figs. 5-9), it is not unfrequently markedly arched, and frequently bears
a more or less sharply defined central papilla. The toothed outer margin extends for a
varying width beyond the centre of the head, or may on the other hand be reduced to a
simple smooth fringe (PI. LXXII. fig. 8), or even wholly disappear. The number and
size of the marginal teeth are very variable. The normal number, twenty, is rarely
greatly exceeded, and frequently not attained. I may, however, call attention to the
fact that amoDo- the dermal clavulse I have never found the anchor-like forms with eisfht
to four long curved rays, which occur so frequently on the gastral side. This marked
diff"erence may sometimes be utilised to distinguish the two sides on broken fragments.
The portion of the head below the umbel is either a simple conical or trumpet-shaped
expansion of the stalk, or is a thickening sometimes cylindrical or even narrowed in
its superior portion (PI. LXXI. fig. 9). In the latter case this swollen lower portion of
the head may bear a special second row of teeth, which lie parallel to the toothed margin
of the umbel. Not only do the length and strength of the stalk vary greatly, but the
shape of the lower portion, which is sometimes quite gradually pointed, sometimes
conically truncated or even rounded ofl". The degree of roughness is also as variable on
the lower as on the upper portion.

The gastral skeleton agrees essentially with the dermal, so that I may simply refer to
the above description of the latter, and content myself with noting the principal
deviations. These difi'erences are indeed notable enough to keep us from ignoring them,
as Carter has done in his otherwise excellent description of the skeletal elements in
Farrea occa. The large gastral pentacts agree almost perfectly in shape, size, and
disposition with those of the dermal skeleton. It seemed to me, however, that there was
a more frequent occurrence of the remnant of the lost sixth ray in the form of a conical
or rounded boss. Often enough, however, in the gastral membrane, pentacts may be
seen without any trace of this remnant.

The difference between the gastral and dermal skeletons consists chiefly in the form of
the clavulse, which occur here also beside the pentacts, corresponding in number, position,
and arrangement to those above described, but differing somewhat in form. In the
majority of the gastral clavulge, four or eight long, slender, more or less markedly
recurved hooks or teeth are formed on the very diflerently shaped head portions, and
produce a certain resemblance to an anchor. The head at least, and the greater part of
the stalk, are destitute of those roughnesses which are so frequent in the dermal clavulse
(PL LXXI. figs. 8, 10; PI. LXXIL fig. 9).

This difference in shape between gastral and dermal clavulse, is not of course constant
nor everywhere marked to the same extent. While in some specimens, and in certain
regions of a single specimen, almost all the clavulse of the gastral skeleton exhibit this
anchor-like structure, and only a few jjrovided with the flat terminal umbels occur in the
dermal membrane, in other regions and in other specimens the umbel-bearing forms may

REPORT ON THE HEXACTINELLIDA. 285

predominate, and indeed the anchor-like forms may decline to such an extent, that they
are only found here and there.

I have taken special pains to try to elucidate the characters of the soft body and
especially the system of chambers in those specimens which are best preserved. Neither
the dermal nor the gastral membrane exhibits any special peculiarities. Where the
dermal network has been in any way preserved, it exhibits meshes of variable width,
sometimes very narrow as represented in PI. LXXI. fig. 3 and PI. LXXII. fig. 5,
and sometimes so wide, that the whole rectangular region between four adjacent
pentacts exhibits only one large round aperture. The gastral membrane exhibits similar
relations.

In the younger portions, which are supported only by a single reticulate layer of the
dictyonal framework, the chambers form a simple, or slightly folded layer of large sac-
cular or thimble-shaped diverticula of variable size. Below the rectangular meshes of the
dictyonal framework, they unite to form a wide excurrent space (PI. LXXI. fig. 3), so
that, on looking down from the gastral surface, below each mesh, one finds a large
excurrent aperture. When the thickening of the wall towards the base of the stock
results in a dictyonal framework of several layers, broad canals traverse this thick wall.
The chambers, which are directed obliquely, then appear rather as lateral and terminal
diverticula of the wide principal canals, and seem to be continuous with them without
any marked boundary (PL LXXIII. fig. 2).

The number of chamber-wall pores varies greatly in the different regions of the body
and in different specimens. The thin layer of connective substance which forms the walls
of the chambers bears internally the epithelial cells, arranged in rectangular fashion as in
Euplectelta, and also united by the same lateral processes into a rectangular network.
The trabecular framework which extends between the two bounding membranes and
the chamber layer consists of delicate strands with fine expansions at the points of
insertion and union.

In many cases the external surface of the chamber wall exhibits numerous groups of
small, crowded cells, with nuclei which stain with special readiness. It is possible that
these groups of six to twelve cells are concerned with reproduction ; I have at least
remarked their total absence in several specimens which contained numerous sperm balls
at various stages. I have unfortunately found no ova or larvae. It is true indeed that
in one fragment from Sagami Bay, the subdermal trabecular space contained numerous
blastulfe in some regions, but whether these belong to the Farrea or to some
commensal, or perhaps to some quite unconnected form, it is difficult to determine, since
the state of preservation of the specimen was by no means favourable. The minuteness
of the larvse (0'03 mm. in diameter) seemed, however, noteworthy.

__l

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

2. Farrea soUasii, n. sp. (PI. LXXIV. figs. 1-G).

On the much macerated and half-eaten skeleton of a dead Japanese Farrea
occa (in Dr. Doderlein's collection), I found a system of dichotomous tubes about 3 cm.
hio-li, and evidently belonging to a Farrea, which in external appearance so closely
resembled Farrea occa that any special description is superfluous. It is, however, at
least probable that we have here the representative of another species. The difference
lies especially in the structure of the discohexasters scattered in the parenchyma.
They differ markedly from those of Farrea occa, so that I felt compelled to erect a
distinct species.

I shall not enter into any detailed description of the strongly developed dictyonal
framework, with its rough internal and external conical radial bosses, nor of the uncinates
of various strength usually very much narrowed at the gastral extremity, nor of the dermal
and gastral framework as a whole, but refer simply to the distinct representation given in
PI. LXXIV. fig. 1. The length of the anchor-hooks in many of the gastral clavulae
(PL LXXIV. fig. 5) is rather remarkable. It is, however, necessary to call attention to
the peculiar discohexasters which occur in all parts of the dry skeleton, in more or less
abundance, beside the quite isolated oxyhexasters which occur in the form constantly
represented in Farrea occa. The basal portion of the uniformly developed rays of these
discohexasters seems rather shor,fr, and divides into three long thin rather outwardly bent
terminal rays, which bear on their outer ends a small transverse terminal plate, with four
to eight gently recurved marginal teeth (PL LXXIV. fig. 6). Besides these delicate and
weakly developed discohexasters, similar smaller forms here and there occur with stronger
terminal rays, bearing discs without marginal teeth, and on the whole more resembling
thickened knobs (PL LXXIV. fig. 1, left, below).

3. Farrea vosmaeri, n. sp. (PL LXXIV. figs. 7-13).

Among numerous specimens of Farrea occa, Carter, which Dr. Doderlein
dredged in Sagami Bay, Japan, from a depth of 100 to 200 fathoms, and preserved
immediately in spirit, there are some well-preserved portions of a Farrea, which m
microscopic appearance and in the structure of the individual siliceous elements differs
from Farrea occa, and so markedly from Farrea sollasii, that the erection of a distinct
species seems inevitable. The spicules are very like those of Aulodictyon woodwardii,
Sav. Kent, but they include oxyhexasters which are not represented in the latter.
The saving clause must again be noted, that it is possible that the divergent spicules have
been intruded from outside. As can be inferred from PL LXXIV. figs. 7, 10-13, neither
the general structure of the dictyonal framework nor the majority of the isolated

REPORT ON THE HEXACTINELLIDA. 287

spicules differ essentially from those of the corresponding portions in Farrea occa. The
anchor-hooks in many of the gastral clavulse seem to be especially long (PI. LXXIV.
fig. 12), and the hexasters are in part very divergent. Though most of the latter have
a general resemblance to the oxyhexasters of Farrea occa, they differ' from the latter in
the greater thickness of their principal rays (PI. LXXIV. fig. 8). Besides these, some-
what larger hexasters occur with weakly developed principal rays, which divide into four
short S-shaped terminals, disposed in a perianth-like whorl and tipped with a small knob.
(PI. LXXIV. fig. 9).

4. Farrea clavigera, n. sp. (PI. LXXV.).

Near the Banda Islands (Station 194, lat. 4° 34' S., long. 129° 57' 30" E.), from a
depth of 200 to 360 fathoms and volcanic mud ground, a Farrea was captured which in
external appearance differed essentially from all hitherto described forms. This sponge
forms a straiglit tubular stem about as broad and as long as a finger, with walls 1*5 to
2 mm. in thickness. It becomes gradually wider towards the upper end, attaining
a diameter of 20 mm., and rising at right angles to a compact smooth basal expansion,
which is directly attached to the solid substratum. Just above the solid base the internal
lumen has a width of 3 mm., and is continued obliquely downwards into an external
groove-like keel. At a height of 12 mm. the first lateral branch is given off, with a
diameter of 8 mm., while further up several branches fork off in irregular distribution, in
a somewhat transverse direction to the main stem. The latter seems to divide superiorly
into two large branches which are unfortunately broken ofi". While some of the transverse
tubes are broken ofi" quite close to the main stem, others exhibit dichotomy into two very
divergent branches. It is interesting that a very similar form has been described by
Oscar Schmidt under the title Farrea facunda from among the Hexactinellids of the
Bay of Mexico. It is figured in the work already referred to, pi. vii. fig. 1a.

The dictyonal framework does not vary essentially from that described in Farrea occa.
Inferiorly it increases, on the one hand, in thickness, and on the other very markedly in
the fineness of its meshes. The basal portion becomes a thick, stone-hard, very finely
porous mass, in the meshes of which countless small hexacts occur, in part free, and in
part fused to the adjacent framework. Isolated parenehymalia are further represented by
uncinates and hexasters. The former seem to be comparatively long and narrow. Among
the latter I have observed the form which occurs so abundantly in Farrea occa, but in
much more sparse distribution. Whether a second form of hexaster, characterised by
short principals and very long terminals, and occurring abundantly in some regions (PI.
LXXV. fig. 6), really belongs to this Farrea, or has been intruded from some other
Hexactinellid, I cannot unfortunately determine with certainty.

The dermal pentacts resemble in general the corresponding dermalia of Farrea occa.

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

Lut exhibit more frequently than in the latter remnants of the atrophied sixth ray,
persisting in the form of a protruding, rounded, more or less elongated distal tubercle.
They also differ in the strong development of rounded lateral teeth on the four tangential
rays (PI. LXXV. fig. 8).

Very unique and characteristic of this species are most of the dermal clavulse, of
which the common form in Farrea occa has only an isolated occurrence. Two pre-
dominant forms occur, of which one exhibits a thick, short, superiorly rounded head
with a central terminal jjapilla, bearing two j^arallel circlets of teeth (PL LXXV. fig. 3),
while the other has a perfectly smooth, long, club-shaped distal end, which is not in any
way sharply marked off from the stalk (PL LXXV. figs. 2, 5). These two forms of
clavulse are either quite irregularly distributed beside one another, or are distinctly
restricted to certain regions of the sponge body, or finally disposed in alternate bundles
in the manner represented in PI. LXXV. fig. 2.

The strong pentacts of the gastral skeleton are distinguished by the large blunt
remnant of the atrophied sixth ray and also by the specially large and distant teeth on
the tangential rays. These teeth are not conical but rather cylindrical, and are rounded
ofi" at their ends (PI. LXXV. fig. 8). The gastral clavulaj are all developed in anchor
form, and usually possess only four strongly recurved long and narrow anchor teeth
(PI. LXXV. fig. 4).

Corresponding to the strong development of the wall in the still preserved portion of
the single specimen captured, cross sections reveal a deep folding of the chamber layer.
The single chambers are somewhat smaller and more irregular than in the tube of Farrea
occa (PL LXXV. fig. 2), nor did the latter exhibit an equal strength of development.

Farrea sp. (?) (PL LXXVL figs. 4, 5).

Among numerous macerated Farrea skeletons in the Challenger collection, which
could not be satisfactorily identified in the absence of all isolated spicules, there was one
fijrm which, though small and crumbling, seemed worthy of more notice than could of
course be given to the majority. This specimen, which is figured in PL LXXVL fig. 4,
essentially differs in its whole configuration from any of the hitherto here described
species of Farrea. It has the form of a slightly convex funnel, somewhat broken at the
sides, and 1 to Ig cm. in height. It consists of a single lattice-work layer with well-
developed rectangular meshes, and with strongly developed beams which resemble those
of the dictyonal framework in undoubted species of Farrea. I would not give so much
attention to this form, if Carter and Bowerbank had not previously described and figured
as species of Farrea certain funnel-shaped Hexactinellids, of which it is possible that
some belong to the genus, though diflering from the tubular type in more respects than
in form. These forms are Farrea infundibuliformis, Carter (identical with Farrea

REPORT ON THE HEXACTINELLIDA. 289

infundibularis, Carter), Farrea gassioti, Bowerbank, and Farrea pocillum, Bowerbank.
The last of these three forms [Farrea pocillum, Bowerbank) certainly does not belong to
the genus Farrea, as Bowerbank's figures ' and description clearly show. On the other
hand, Farrea infundlbuliformis, Carter, and Farrea gassioti, Bowerbank, may quite
possibly belong either to the genus Farrea itself or to some related genus within the
family Farreida^. I only desire to direct the attention of specialists to this funnel-shaped
Farreid, for such at least the fragment figured in PI. LXXVI. certainly is.

Subtribe 11. Scopularia (Carter), F. E. Schulze (Pis. LXXVII.-XCVIIL).
Besides the pentact hypodermalia and hypogastralia, radially disposed scopulae occur.

Family L Euretid^, F. E. Schulze (Pis. LXXVII.-LXXXIL).

Branched and anastomosing tubes, which either form an irregular framework with
tubes of almost uniform width, or else the wall of a cup. The dictyonal framework
exhibits from the very first more than one layer, so that a single layered network of
strands never occurs at the ends of the tubes.

Genus 1. Eurete (Semper), Carter (Pis. LXXVII.-LXXIX.).

1868. Semper, Verhandl. d. Wiiizburg. phys.-med. Gesellsch., Sitzungb. vom July 18.

1875. Marshall, Zeitsclir. f. wiss. Zool., Suppl., Ed. xxv. p. 181.

1876. Marshall, Zeitschr. f. wiss. Zool., Ed. xxvii. p. 113.

1877. Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. xix. p. 112.

1877. Zittel, Abhandl. d. k. baier. Akad. d. Wiss., Bd. xiiL Heft 1, p. 1.
1880. 0. Schmidt, Die Spougien des Meerbusens von Mexico.

History. — In the Transactions of the Physico-Medical Society of Wiirzburg, 1868,
there occurs, in the report of the session held on 18th July 1868, the following notice : —
" Mr. Semper showed some new siliceous sponges from the Philippines. One is a new
species of the genus Hyalonema, and another may be regarded as a type of a new genus
Eurete." " The genus Eurete was established for a sponge, having the form of a
coral, the cylindrical and hollow branches of which are everywhere united to one another.
The wide openings at the extremities of the branches seem to be exhalent, the fine
pores between the network, which constitutes the walls of the tube, are apparently
inhalent. The tissue of the wall of the tube — which measures about 1 mm. in thickness
— is composed of a tolerably dense network of fine siliceous tubes, which are sometimes

I Proc. Zool. Soc. Land., pi. xxxix. fig. 48, 1875.
(zool. chall. EXP.— part uii. — 1887.) Ggg 37

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

fused together irregulaiij', thougli occasionally also crossing one another in a regular
manner, so that a network surrounding rectangular meshes becomes formed. It seems
nowhere to show actually free cross spicules, yet very frequently the cavities of the
amalgamated cross fibres remain independent of one another, so that often two or three
hollow spaces lying close to one another, yet not connected, are cemented together by a
common siliceous mass. In these siliceous tubes the uncommonly wide cavity of the
axial cord is remarkable. It is often so wide that its diameter exceeds the thickness
of the wall about six times. It is to be regretted that the single specimen is much
bleached and macerated, so that it is not possible to say whether the almost complete
absence of all free siliceous bodies should be looked upon as a peculiarity distinguishing
this genus. To judge from the structure of the tissue, this sponge probably belongs
to the same genus as Farrea occa, Bowerbank. Since, however, only fragments of the
latter are known, it is still possible that they belong to the Euplectella cucumer, Owen,
on whose roots they were found, so that provisionally Farrea occa and Eurete simpli-
cissima, Semper, may be regarded as distinct from one another. An accurate examination
of the tissue of Euplectella cucumer would throw light on this question. Detailed
descriptions will shortly be given in the Zeitschrift fiir wissenchaftliche Zoologie."

The description here promised was given by W. Marshall in 1875.^ Marshall first
confirms the general descrijDtion of Semper, and calls attention to the occurrence of
transverse divisional walls close beneath the orifices of some tubes. These he designates
by the name of sieve-plates, and is inclined to regard them as indicating the regular limit
of every " individual" properly so called. Of free spicules Marshall, like Semper, found
but a mere trace, and he therefore concluded that even the living sponge had none.
This view has again been definitely expressed by Marshall ^ after re-examination of
Semper's original specimen.

The sponge minutely described by Carter under the name of Eurete farreopsis ^ was,
like the majority of objects of this nature that have to be dealt with, much macerated,
yet in some remnants of the dried soft parts, a number of free siliceous spicules could
still be found. From Carter's minute description I make the following excerpt as being
of very great importance : — " General form bush-like, composed of many tube-branches
anastomosing clathrously. Branches short, thick, cylindrical, hollow, formed of a delicate,
thin, reticulated wall, thickening from the growing margin towards the base or oldest
part. Orifices of branches respectively circular at first, then expanded, afterwards
funnel-shaped, becoming elliptical and contracted in the centre, where by the union of
the approximated parts of the margin, two circular orifices are formed, which grow into
two short, round, tubular branches in opposite direction, to divide again after the same

1 Zeitschr.f. vdss. Zool., Bd. xxv. p. 181.

2 Mittheil. K. Zool. Mus. Dresden, ii. p. 272.

^ Ann. and May. Nat. Hist, ser. 4, voL xix. p. 122, 1877.

REPORT ON THE HEXACTINELLIDA. 291

manner, and so on — or to anastomose with other neighbouring branches. The meshes of
the reticulated tissue are subquadrangular."

" Spicules of three kinds, namely, skeleton, subskeleton and flesh spicules. Skeleton
spicules sexradiate arms spined throughout, pointed in the smallest, inflated at the
extremities in the largest specimens. Subskeleton spicules of two forms: — (1) acerate,
straight, fusi/oi'm, attenuately pointed, spined throughout, spines all inclined one way,
and more or less closely applied to the shaft ; (2) scopidine spicule, consisting of a shaft
and head ; shaft cylindrical, abruptly pointed at the free end, quadrangularly inflated
at the other, micorspined throughout, most evidently towards the free end ; head con-
sisting of four arms, respectively supported by the four angular projections at the end of
the shaft, at first remaining parallel or slightly curved towards each other and then
expanded ; arm much thinner than the shaft, inflated globularly at the extremity, micro-
spined throughout, especially towards the inflation, where the spines are long and
inclined backwards, leaving the convexity of the inflation smooth or bald. Flesh spicule
a Hexactinellid rosette, each arm bearing four capitate rays, expanded enjleur-de-lis, or
without extended arms, the latter being reduced to a central point, from which the rays
radiate in all directions so as to present a globular form."

" Vitreous fibre smooth between the knots, which are globular and spino-tuberculated
all over, except where interupted by their union with the fibre, or by the projection of
one or more arms of the sexradiate spicule in the form of large spines, thickened or
elongated, pointed or inflated at the extremity and spinulated throughout."

According to Carter the following are the peculiarities of Eurete farreopsis : — " The
globular tuherculated knots of vitreous fibre, which, with the centrally developed spine,
looks like a bossed omphalic shield, and the globular inflations respectively at the ends of
the scopuliue arms very much like a ' bald head.'"

Character of the Genus. — A system of multifarious dichotomously branched and
richly anastomosing tubes of approximately equal diameter, fixed to its substratum by
means of several solid or hollow supports which are inferiorly expanded in a plate-like
manner. The margin of these oscular openings is not attenuated as in Farrea. The
dictyonal framework of beams forming the skeleton consists in all parts of the sponge —
even in the latest formed margins of the oscular openings — of several layers, and
surrounds more or less regularly formed, in many cases almost square, in others irregularly
quadrangular or triangular, meshes with simple or knot-like thickened intersections. On
the dermal and ventral surfaces of the framework of beams conical pegs of various
length arise from the intersections, and are directed at right angles to the bounding
surface.

The dermal and gastral skeletons consist of pentact hypodermalia and hypogastralia
respectively, and of numerous scopul^ provided with knobbed or pointed teeth.

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

lu the parencli}Tiia harpoons directed at right angles to the surface and hexasters of
various kinds occur in addition to simple hexacts.

1. Eurete semperl, n. sp. (PL LXXVIL).

Near the Little Ki Island (Station 192, lat. 5° 49' 15" S., long. 132° 14' 15" E.),
from a depth of 140 fathoms and a blue mud ground, the trawl brought up, among
numerous other Hexactinellids, that form of Eurete represented in PL LXXVII. fig. 1. It
was abundantly beset with small Actinise and exhibited in spirit a dull light grey colour.
Several solid basal pedestals, 5 to 8 mm. in diameter, partly united in a common basal
plate, bear a system of irregular reticulated and anastomosing tubes, 8 to 15 mm. in
diameter, and 1 to 2 mm. in wall's thickness. The free end is unfortunately broken off,
or more or less seriously injured.

The basal pedestals are dead up to a level of about 10 mm., otherwise the stock
is tolerably well preserved. The strongly developed, somewhat irregular, dictyonal
framework, which only rarely exhibits square or rectangular meshes, consists of strong
beams beset with a few small spines, and united in thickened, spherical, swollen nodes of
intersection more or less thickly covered with strongly developed but low-set teeth
(PL LXXVII. fig. 2). The freely projecting spherical spinose bosses on the dermal
and gastral surfaces are remarkably short and stunted. In the inferior regions of the
stock the meshes of the dictyonal framework are very much narrowed by numerous
small apposed hexacts, which, becoming thickened and united all round, contribute to
strengthening the already existing framework. The result is the formation of a firm,
stony, finely porous mass. Superiorly, however, the meshes become wider. Afi'erent
canals traversing the wall at right angles to the surface, and corresponding efferent canals
are seen in the dictyonal framework as round passages which run alternately from the
outer and the inner bounding surface, and either end blindly or divide into lateral twigs.
The thicker the wall of the tube, the more is the canalicular system developed within the
dictyonal skeleton. The free parenchymalia are represented especially by small simple
regular oxyhexacts, present in extraordinary abundance, and thickly beset in every
region with minute pointed tubercles, so that they appear rough even under low power
(PL LXXVII. fig. 7). In almost any region of the dictyonal framework they become
reacbly fused to a ray perpendicular to the surface of the framework, or to one of the thick
nodes of intersection (PL LXXVII. fig. 8). There they become thickened by the apposi-
tion of concentric siliceous lamellae and also unite with other adjacent strands, thus
leading to the growth and thickening of the whole dictyonal skeleton. The size of these
rough hexacts varies greatly, but they rarely exceed 0'17 mm. in diameter.

The uncinates, which are always disposed at right angles to the bounding surfaces,

REPORT ON THE HEXACTINELLIDA. 293

are only sparsely present, somewliat slender in form, and thickly beset with long, closely
appressed, narrow barbs.

Besides these, I find in the parenchyma, between the strands of the dictyonal frame-
work, isolated minute diseohexasters with six divergent, comparatively straight, terminal
rays of equal length, on each of the six short, slender principals, which exhibit small
knob-like terminal plates (PL LXXVII. fig. 9).

The dermal skeleton exhibits medium-sized pentacts with much thickened rough and
rounded ends on the four gently incurved tangentials, a somewhat longer proximal ray
also with a rough rounded off" end, and lastly, a simple inconspicuous distal tubercle in
place of the atrophied sixth ray. Besides these hypodermalia which are united by their
tangential rays into a quadratic network, there are abundant, very characteristic — indeed
specifically distinctive — dermal scopulas in which the smooth stalk with a rough gradually
pointed proximal end is expanded distally in a conical fashion, and bears four, or rarely
five or six, cruciately disposed prongs. The thin stalks of these slightly divergent
prongs are usually slightly bent (PL LXXVII. fig. 5), and less frequently straight.
They are continued outwards into a smooth, spindle-shaped, thickened end, which
occasionally exhibits a somewhat sharply defined point and an internal sharp edge.

The pentacts of the gastral skeleton resemble those of the dermal, but the scopulas
which occur beside them diff"er in some essential points from those in the dermal region.
The four to six thin prong-stalks, which are curved gently outwards, bear club- or pear-
shaped terminal thickenings which are thickly beset round about with short obliquely
disposed barbs. Only the outermost pole of the terminal prong-swelling remains free from
these minute teeth (PL LXXVII. figs. 6, 10).

The soft tissue exhibits here, as in all species of Eurete, a continuous dermal and
gastral membrane with pores of very varied width, and a delicate subdermal and subgastral
trabecular framework, in which the usually simple and straight afferent and eS"erent
passages are seen as roundish canalicular spaces. The system of chambers forms a deeply,
but somewhat simply folded layer of small chambers of the ordinary type, and somewhat
sharply separated laterally from one another (PL LXXVII. fig. 2).

2. Eurete schmidtii, n. sp. (PL LXXVIII. figs. 1-G).

Among the numerous specimens of Eurete now at my command, there are three
which agree both in their microscopic appearance, and in the minute structure of the
skeletal parts. These belong to a new species which I have named Eurete schmidtii,
in honour of the illustrious spongiologist, Professor Oscar Schmidt of Strassburg.
Two of these are preserved in alcohol, and were brought home by the Challenger
Expedition, having been trawled in the neighbourhood of the Philippines at Station 201,

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

lat. 7° 3' N., long. 121° 48' E., from a depth of 102 fathoms and on stony and gravelly
ground. The third was bought in a dried condition by Dr. Doderlein in Enoshima.

The transverse diameter of the tolerably thick-walled tubes, which form a dense
basket-work, varies between 5 and 7 mm. The oscular openings are frequently some-
what contracted on the extremity of the freely projecting, short ends of the tubes. The
Japanese specimen shows one irregularly bounded basal plate, which formed a flat
expansion on the stony substratum. From this plate arise the thick, partly solid, partly
hollow supports of the plexus of tubes. It is probable that these relatively large
basal plates have resulted from the confluence of the several basal expansions of the
originally independent supports. In the case of the two Challenger specimens, one of
which is represented on PL LXXVIII. fig. 1, the basal portion is entirely absent. On
the Japanese specimen, as well as on both those from the Philippines, there were numerous
small irregularly scattered Actiniae with a diameter varying from 1 to 1*5 mm.

The dictyonal framework exhibits a tolerably regular structure with predominantly
square meshes, smooth or but slightly spinose beams, and scarcely thickened, moderately
spinose nodes of intersection (PL LXXVIII. fig. 2). Between the beams of the dictyonal
framework small rough oxyhexacts occur in variable abundance, most plentifully in the
basal region. They lie freely, or fused with single-rayed parenchymalia. In the latter
case the dictyonal framework is, in several regions of the stock, further strengthened by
the tolerably abundant occurrence of a second kind of parenchymalia, viz., medium-sized
oxyhexasters with short principal rays, each bearing twci long moderately divergent
terminals (PL LXXVII. fig. 6). Besides these, numerous uncinates occur, which are
remarkable in their structure, and to some extent also in their disposition. For while the
uncinates are otherwise almost always disposed at right angles to the bounding surfaces,
in this case many of them not only penetrate the wall transversely, but occur abundantly
disposed obliquely or parallel to the bounding surfaces. Forms so directed are much
longer than the others. Both exhibit a somewhat marked flattening on both sides. The
somewhat thickly apposed pointed barbs are not uniformly distributed all round, but
sometimes occur on two slightly spiral longitudinal rows on the nari'ow edges of the
uncinate, while the two flatter surfaces remain smooth.

The pentacts of the dermal skeleton vary in size, and bear rough cylindrical rays of
medium length, which are simply rounded ofi" at the extremity. The four tangentials,
which are tolerably straight, are not disposed quite at right angles to the proximal ray,
but more or less inclined towards it. A knob-like protuberance represents the abortive
distal ray.

Besides these pentacts there are numerous scopute, represented by two difierent
forms. The less frequent form is provided with four knobbed terminal prongs, whUe
the much more numerous second type bears, on a knob-like thickening of the inferiorly
pointed stalk, two to four slightly divergent barbs, which are rather strongly develoj^ed

REPORT ON THE HEXACTINELLIDA. 295

at the bfise, but gradually decrease in diameter towards the pointed extremity, and are
laterally compressed like the blade of a knife, with the edge turned inwards and the
back towards the exterior (PI. LXXVIII. figs. 3, 5).

The pentacts of the gastral skeleton generally resemble those of the dermal skeleton,
but more frequently exhibit a gently arched or even markedly protruding knob-like
elevation in place of the undeveloped sixth ray. The scopulse are tolerably abundant ;
and resemble the dermal forms of the first type. The expanded end of the posteriorly
pointed stalk bears a knot-like swelling with four to five slightly diverging prongs. The
slender stalk of the prongs exhiljits a club-shaped extremity beset with small barbs (PI.
LXXVIII. fig. 4).

The structure of the soft parts does not differ in any essential feature from that
described in Eurete semperi.

3. Eurete farreopsis, Carter (PL LXXIX. figs. 5-8).

The species of Eurete figured in PI. LXXIX. fig. 5 is a tubular feltwork, which in
several places is firmly attached to the solid substratum. The tubes have a comparatively
small diameter (of 5 to 8 mm.) and only a moderate peripheral thickness. The form was
obtained near the Little Ki Island (Station 192, lat. 5° 49' 15" S., long. 132° 14' 15" E.)
from a depth of 129 fathoms and a blue mud ground. It was thus found in the same
locality as Eurete seynperi. In the structure of its skeletal elements it resembles very
closely the Eurete farreopsis described by Carter in 1877,^ and represented in a very
efi'ective figure (pi. ix. figs. 1-7). The identity of the two forms is unquestionable.
The dictyonal framework, in which the square form of meshes predominates, consists of
smooth or only slightly tubercled beams, and of more or less markedly thickened and
roughened nodes of intersection. Simple, rough, minute oxyhexacts occur in the
parenchyma, especially in the older and lower portions, and are in part fused by one ray
to the general framework. Besides these, the parenchyma includes small discohexasters
with terminal knobs on the four divergent, moderately long, often perianth-like, curved
terminal rays, and in sparse occurrence the familiar uncinates beset round about with
appressed barbs.

The dermal skeleton agrees perfectly with the gastral. The pentacts are provided on
both sides with slightly curved tangential rays, somewhat rough at the rounded off"
extremities, and a somewhat longer straight radial, in which the narrowed, roughened,
terminal portion is moderately pointed or rounded ofi".

Just as the dermal and gastral pentacts resemble one another, so do the peculiarly
shaped scopulse which occur beside them ; but the latter are characterised by the sharp
break-like bend of the terminal rays. Each of these thin bent stalks, w^hich diverge

1 Ann. and Mag. Nat. Hist, ser. 4, vol. xix. i^. 122.

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

considerably after the curvature, bears a more or less sharply defined terminal knob, with
a smooth apex, but with sides bearing several rows of fine recurved barbs.

4. Eurete carteri, n. SIX {VI LXXVIII. figs. 7-12).

Two specimens of Eurete — one of which was trawled by the Challenger Expedition in
the vicinity of Little Ki Island (Station 192, lat. 5° 49' 15" S., long. 132° 14' 15" E.),
at a depth of 129 fathoms and upon blue mud ground, while the other was dredged by
Dr. Doderlein in Sagami Bay, Japan, at a depth of 150 fathoms — are, on account of the
agreement of their structure, united into a single species which is very closely allied to
Eurete far reopsis, Carter, from the Moluccas. I name these Eurete carteri in honour of
the famous Nestor of spongiologists — Mr. Carter.

The Japanese specimen exhibits a narrow-meshed tubular framework which was fixed
to a piece of rock by means of a few compact pedicels. The constituent tubes are from
5 to 7 mm. in breadth, and open out by means of short projecting terminal branches.
In the wall of several of the tubes there are circular holes measuring from 3 to 4 mm. in
width — but whether these are to be ascribed to accidental injuries, or are to be regarded
as normal structural features, appears to me to be doubtful. I would lay little weight on
the fact, were it not that similar circular perforations of the wall of the tube are also
found in the other specimen, which is a fragment of a network of wider tubes.

The dictyonal network of beams exhibits but little regularity. The meshes are
occasionally perfectly square, but are as a rule triangular. The beams are never quite
smooth but are more or less richly beset with small pointed tubercles. The intersections
are usually somewhat thickened, though here and there they appear but slightly
diilerentiated. They are always thickly covered with small tubercles. The pegs which
project on the dermal and gastral surfaces are tolerably thin, either conical or provided
with a knob-like thickening on the extremity, and are always rough and tuberculated.

The hypodermalia and hjqsogastralia are rough pentacts with slightly bent transverse
rays, while each is provided with a straight radial ray which varies considerably in
length. The extremities of the transverse rays are as a rule somewhat swollen, but
they are sometimes simply rounded. The radial ray is in most cases simply rounded at
the extremity, though occasionally provided with a slight swelling.

The scopulae of the outer skin resemble those of the gastral surface. Both possess
smooth or cjuite insignificantly rough stalks which terminate in the parenchyma by simple
rounded extremities or become slightly attenuated ; on the other side, however, from four
to six terminal prongs spring from a small conical thickening, and these are provided on
the outer extremities with a knob-like rough swelling. The thin, usually smooth, but
here and there also somewhat rough stalk of these terminal prongs is always straight or
shghtly bent (PI. LXXVIII. figs. 9, 10), but it is never sharply dislocated like the prong

EEPORT ON THE HEXACTINELLIDA. 297

stalk in the scopulfe of Eurete farreopsis. The terminal knob-like thickening may, aa
shown in figs. 9, 10, be pear-shaped, or be more sharply differentiated from the thin stalk,
so as to approach nearer to the spherical form.

The uncinates are feeble and not very long, but beset with fine barbs.

The small discohexasters which are scattered quite irregularly in the parenchyma
bear, on each of their principal rays, three or four straight terminals, which are
irregularly disposed, or arranged hke the petals of a lily (PI. LXXVIII. figs. 11, 12).

Moreover, at certain spots here and there, but in special abundance in the basal regions
there are simple hexacts, which are either smooth or provided with scattered tubercles,
and run out to fine points. These originally lie freely in the parenchyma, but subse-
quently amalgamate with the dictyonal framework of beams, or with one another, and
so contribute to the thickening or strengthening of the continuous skeleton.

5. Eurete marshalli, n. sp. (PI. LXXIX. figs. 1-4).

A species closely related to Eurete farreopsis and Eurete carteri, was found near the
Little Ki Island (Station 192, lat. 5° 49' 15" S., long. 132° 14' 15" E.) at a depth of
140 fathoms, on a blue mud ground. Two specimens were obtained, both covered with
small Actinise. The smaller is represented in PL LXXIX. fig. 1. In its smooth, or only
slightly toothed strands, and thickened spinose nodes of intersection, the dictyonal frame-
work closely resembles that of Eurete farreopsis. On the other hand the parenchyma
includes numerous oxyhexacts with long divergent terminals (PL LXXIX. fig. 3) very
different from the corresponding spicules in the above species. Between the beams of
the dictyonal framework small simple hexacts are present, and also uncinates beset
round about with slender pointed barbs; neither of these forms, however, exhibit any
striking peculiarities. The spicules of the dermal and gastral skeleton are so closely alike
that a separate description is quite unnecessary. The slightly curved tangential rays of
the pentact hypodermalia and hypogastralia are rough at their rounded ends, and the
same is true of the long proximal. The tubercle which frequently occurs as a persistent
trace of the undeveloped sixth ray is usually inconspicuous. A noteworthy fact, and
distinctive of this form as distinguished from Eurete farreopsis, is this, that the somewhat
markedly divergent terminal rays of the scopulse, which are provided -with pear-shaped
barbed terminal knobs, exhibit no bend or break in their thin stalks, but are quite straight
throughout their entire length.

6. Eurete howerbankii, n. sp. (PL LXXIX. figs. 9-13).

Among the Japanese specimens of Eurete collected by Dr. Doderlein in Sagami Bay
by aid of the dredge and trawl, and well preserved in spirit, there occurs a tree-like form

(ZOOL. CHALL. EXP. PART LIII. 1887.) Ggg^ 38

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

with a fuunel-sliaped expanded principal tube, and several lateral tubes 3 mm. in breadth.
The dictyoual framework of this sponge exhiljits beams which have a slightly tuberculated
surface and no sharply diflfcrentiated or slightly thickened crossing knots. In addition
to the hypodermalia and hypogastralia, which do not differ essentially from those of
Eurcte farrcopsis, scopulse occur. These have straight, unbroken, rough terminal rays
which are of a thick club- or knob-like shape. The uncinates are of various thickness,
but in most cases they are quite slender, and do not always stand at right angles to the
surface of the tube. Simple hexacts occur which are partly free and partly fused together
in various numbers. In contrast to Eurete carteri and Eurete marshalli, the great
abundance of oxyhexasters is particularly characteristic of the species. They possess
simj)le cylindrical principal rays, which are twice as long as the four greatly diverging
conical terminals, which are arranged in a cruciform manner on the extremity of each of
the principals (PI. LXXIX. fig. 13).

I shall further add a short note in regard to a species described by Semper, which
presents some peculiarities worthy of notice.

Eurete simplicissima, Semper,

The beautifully developed network of tubes which Semper procured from Zebu, and
established as the basis of the genus Eurete, has been carefully described and figured
(two-thirds natural size) by Marshall.^ This specimen was kindly lent to me for
comparison by Professor Semper, but the remnant was unfortunately too utterly
macerated. With the exception of some parenchymalia (oxyhexasters with long thin
terminals), no free spicules were to be found. The dictyonal framework is composed of
smooth or slightly spinose beams, without any marked thickening at the nodes of
intersection, and exhibits so few characteristic peculiarities that it is difficult or
imjDossible to distinguish it from that of the other species.

I am not in a position to say whether one of the species described may not be
identical with Eurete sim2)Ucissvma, Semper. It was suggested by the structure of
the oxyhexasters found in the base of an attached Actinia-like Anthozoon, that Eurete
$imiilicissima resembles my Eurete marshalli, which is figured in PI. LXXIX. figs. 1 to 4.
But since the dictyonal framework of this last sisecies is provided with thickened and
spinose nodes of intersection, the resemblance is not complete.

1 Zeitschr. f. wiss. ZooL, Bd. xsv., Suppl., pi. xii.

REPORT ON THE HEXACTINELLIDA. 299

Genus 2. Perii^hrageUa, MarsliaU (Pis. LXXX., LXXXL).

1875. Marshall, Zeitschr. f. wiss. Zool., Bd. xxv. Suppl., p. 177.

1876. Marshall, Zeitschr. f. wiss. Zool., Bd. xxvii. p. 113.

History. — Among the Hexactinellida which Marshall described in 1875 in the
Zeitschrift flir wissenschaftliche Zoologie, there was a cup-shaped sponge about 15 cm. in
height from the Moluccas. On the lateral walls of this " there are curious tubuliform
individuals, which have coalesced and communicate with one another, and are provided
with a labyrinth of internal canals. They are sometimes broad and flat, sometimes
tubuliform, but always very slender." Marshall has applied to this form the name
Periphragella elisse.

The network of siliceous beams, which is very dense beneath, but more diffuse above,
presents a tolerably regular structure with permeating square meshes, and very small
plain, round knobs on the surface of the beams. Of free spicules Marshall found (1)
smooth hexradiate spicules ; (2) simple five-rayed spicules, the unpaired principal ray
being twice as long as the others ; (3) small hexradiate spicules with concave terminal
discs, with five circularly disposed teeth on the margin of each ; (4) somewhat larger
hexradiate forms in which the rays are divided about the middle of their length into from
two to four pointed prongs; (5) broom-shaped spicules with a smooth shaft, on the
expanded extremity of which there are four somewhat diverging clubs, whose circular
terminal knobs are provided with zones of plain recurved tubercles. All these spicules
occurred exclusively in the middle and upper parts of the sponge ; in the base only small
smooth hexradiate spicules were found between the lattice-like network of beams.

In his statement in regard t© the afiinities of the Hexactinellida,^ Marshall summa-
rised the characters of the genus Periphragella in the following words : — " Polyzoic, indi-
viduals conical, mouth naked, with conspicuously developed goblet-shaped pseudogasters.
Spicules sometimes hexradiate or simple dentate rosettes. Broom spicules with four
regular teeth. Special closing spicules for the dermal pores. Lattice-work very regular."

Peri2)hrageUa elisai, Marshall (PI. LXXX., PI. LXXXL).

Among the dried Hexactinellida which Dr. Doderlein acquired in Enoshima (Japan),
there were three specimens of Perijyhragella elisw, Marshall. One of these, which is
figured on PI. LXXX. fig. 1, has the form of a slightly bent cup or funnel 15 cm. in
length. This rises with a round hollow stalk of 12 mm. in diameter, from an irregularly
formed basal plate from 30 to 40 mm. broad, and gradually expands upwards towards
the round terminal opening, which is 50 mm, in breadth. The thin plate which directly
surrounds the lumen of the funnel is continued, above the stalk region, into numerous
radially disposed tubes of various width (from 3 to 15 mm. or more). These in most

' Zeitschr. f. vnss. Zool, BJ. xxvii. p. 113.

300 THE VOYAGE OF H.M.S. CHALLENGER,

cases branch, lie apposed to cue auotlier, and anastomose freely. The narrow (from
2 to 4 mm. in diameter) tubular branches, which project externally, open outwards
by means of simple circular terminal openings. Here and there, but especially in the
neighbourhood of the upper extremity, there are also broad tubes, which are curved
outwards in a trumpet- or funnel-like fashion, and terminate with a wide irregularly
rounded margin.

The entire system of these simple or branched and anastomosing tubes, which pass
out laterally from the central space, increases in thickness upwards, and is from 1 to
4 cm. broad. It forms a covering to the outer side of the cup, on the upper
terminal margin of which a certain outward curvature exhibits the structure of the
t ubework. If one looks through the large terminal opening into the lumen of the cup
— the gastral cavity — one sees the oval or circular inner openings of the tubes, which
increase in diameter upwards, and have a radial direction.

In another specimen represented on PI. LXXX. fig. 2, the stalk and basal plate are
absent, and the axis of the cup is unbent. The greatly developed system of branched and
anastomosing lateral tubes increases in extent towards the top, but is much injured and
broken towards the thin margin of the cup, so that the latter are clearly marked in the
figure, and disclose the transverse beams of the often very obvious square-meshed
dictyonal framework.

Close above the stalk the wall of the cup measures 2 "5 mm. in thickness, in the
middle of the body 1*5 mm., and on the upper margin only about 1 mm., i.e., the thickness
the wall in the individual wall-tubes.

A third cup belonging to the same species was greatly weathered, and only its inferior
part was preserved. It exhibits, like the first mentioned, a slight bending of the axis, and
seems to have belonged to a very strong specimen, as it was borne upon a pedicel more
than 2 cm. in thickness.

A plate-hke fragment with the well-preserved soft body, from 1 to 1'5 mm. in thick-
ness of the wall, exhibiting an irregular tubular form, and probably referable to the greatly
expanded upper lateral tubes of alarge cup, occurred among fragments of Farrea, Eurete,
and Aphrocallistes, in the materials collected by Dr. Doderlein in the Bay of Sagami, and
preserved in alcohol. I have figured this piece on PI. LXXXI. fig. 1, and hav« used it
especially in the study of the soft parts as well as of the more minute structural relations
of Periphragella. The beams enclosing the somewhat regular square meshes of the
dictyonal framework are either quite smooth, or beset more or less richly with small
simple spines. The free terminal bosses are always thickly spiuose. The nodes of inter-
section exhibit no marked thickenings.

I have found uncinates only sparsely, and not always at right angles to the bounding
surface. They vary in length and thickness, and are surrounded by barbs. The
parenchyma includes a large number of scattered hexasters of two difierent types. Of less

REPORT ON THE HEXACTINELLIDA. 301

frequent occurrence are the medium-sized oxyhexasters with simple principal rays of
medium length, and with two to four outward bent medium-sized terminals (PI. LXXXI.
fig. 5). The terminal rays are usually twice as long as the principals which bear them.
In the second place and much more abundantly, sometimes indeed remarkably crowded,
are small discohexasters, with four to six simple or S-shaped, short, terminally knobbed
terminal rays on each of the simple medium-sized principals (PI. LXXXI. fig. 4). The
dermal and gastral peutacts are very like one another. Both the slightly incurved
tangentials and the radial ray projecting into the parenchyma are somewhat rough, and
end ill club-shaped, rounded, or more rarely sharpened swellings (PI. LXXXI. fig. 8).
The sixth ray is absent. Its position is either wholly unoccupied, or is marked only
by a gentle elevation. Besides the pentact hypodermalia, there are a great number
of dermal scopulse, usually pushed far forward, and are generally arranged in tufts on
each pentact. The shaft is rough, swollen into a club-shape at the lower end, somewhat
conically expanded above. It bears four rather markedly diverging terminal rays,
frequently bent somewhat outwards. Each terminal forms a club-shaped, often almost
spherical, and tolerably distinct terminal portion, which gives off laterally several
transverse rows of barbs, while the somewhat flattened terminal surface appears to be
smooth. The thin stalk of the terminal ray is beset with very minute pointed barbs,
which are also directed backwards.

The gastral scopulse, which occur much more sparsely, have in general the same
structure. The terminal rays seem to be more slender, and their terminal knob is smaller
than in the dermal scopulse.

In the character of its soft body, Periphragella elisie does not essentially differ from
Eurete. In the subdermal trabecula, in some regions, groups of small cells occur, which
I regard as immature sperm-balls.

Genus 3. Lefroyella, Wyville Thomson (PI. LXXXIL).

Lefroyella decora, Wyville Thomson.

History. — In Wyville Thomson's preliminary account of a part of the voyage of the
Challenger Expedition, entitled The Atlantic, 1877, p. 401, we read : — " On the following
day we sounded in 2500 fathoms, and on the 29th in 1075 fathoms, in sight of the Ber-
mudas, with a bottom of coral mud. The dredge was put over and veered to 1600 fathoms.
It came up at noon with the pasty mortar-like lifeless contents which we find almost
constantly on the slopes of coral reefs ; the lime sediment was mixed with a large pro-
portion of the shells of Pteropods and Heteropods. Two fine specimens of a HexactineUid
sponge were hanging to the tangles, both unfortunately dead and slightly water worn.
The largest specimen, which seems to be nearly complete, is 170 mm. in height, and
shaped somewhat like an old fashioned tall champagne glass. It rests on a very solid

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

hard base of attachment, it then contracts to a kind of stem, and then gradually expands
upwards to a width at the top of 40 mm. A deep cavity passes from the upper open
end down to the stem-like constriction. The outer surface of the sponge is raised into
spiral ridges somewhat as in Euplectella, and under the ridges are irregularly spiral lines
of large holes. The interior of the cup presents a very remarkable character, which
reminds one at once of many of the chalk Ventriculites. The inner layer is deeply fluted,
thrown into a series of alternating vertical grooves and ridges, so that the outline of the
cavity in a transverse section is deeply sinuous. The substance of the sponge through-
out is composed of a close anastomosing network of siliceous fibres ; towards the outside
the network much resembles that of AphrocaUistes ; while on the inner wall the structure
is trellis-like, and the form of the meshes square and more regular. The spaces of the
network are crowded with small regular Hexactiuellid spicules, some free, some cemented
to the continuous skeleton by an attachment of silica. For this beautiful sponge, which
I have every reason to believe is undescribed, I propose the name Lefroyella decora. I
have associated with this species the name of our kind friend, His Excellency Major-
General Lefroy, C.B., F.E.S., Governor of Bermudas."

Oscar Schmidt says ' : — " Under the name Lefroyella decora, W. Thomson describes a
sponge dredged in the vicinity of the Bermudas in 1075 fathoms, which is very probably
identical with our Syringidiimi. Since, however, even the best preserved specimen
figured is greatly macerated and eroded, so that the more minute details cannot be
recognised, I have deemed myself justified in regarding the sponge, from which I was
able to give a description in accordance with reality, as new."

In Leuckart's Jahresbcricht (edited with the co-operation of Marshall)^ it is sug-
gested, in reference to Lefroyella decora, Wyville Thomson, that this form is perhaps
identical with Peripliragella elisse.

Lefroyella decora, Wyville Thomson (PI. LXXXIL).

In the neighbourhood of the Bermudas Islands, the Challenger Expedition dredged
several specimens of this beautiful form, but these were unfortunately completely
macerated and much rubbed. One specimen, with a narrow cup-shaped basal portion
(6 cm. in length, 4 in breadth above and 2 below), was dredged at Station 33 (lat.
32° 21' 30" N., long, 64° 35' 55" W.), from a depth of 435 fathoms and a coral mud
ground. The other form, which Wyville Thomson himself described (see above), was
found hanging to the tangles of the dredge at Station 56 (lat. 32° 8' 45" N., long.
46° 59' 35" W), and brought up from a depth of 1075 fathoms and a coral mud ground.
This last specimen is represented on PL LXXII. fig. 1, in its natural size and from a

Spongien des Meerbiisens von Mexico, p. 47.
2 Ueber die Fortschritte der wissenschaftl. Leistungen in der Naturgeschichte der Spongien, wahrend der Jahre
1876-1879, Archivf. Naturgesch., 1883.

REPORT ON THE HEXACTINELLIDA. 303

photograph. It has the form of a slender cup, attached to a broad basal plate, and is
12 cm. long by 4 broad in its upper portion. It is not, however, intact. Since Sir
Wyville Thomson speaks of two specimens from this locality, of which the larger was
17 cm. in length, it is evident that the latter is not in the collection before me.

In Wyville Thomson's description, and in fig. 1 on PI. LXXIL, attention is directed
to the oblique or transverse lidges projecting on the external surface of the cup and
alteruatino- with somewhat broader erooves. These transverse ridges consist almost
wholly of a row of closely apposed and externally fused, short tubular stumps. The
circular external end measures 4 to 5 mm. in breadth, and is directed radially outwards.
In the grooves, on the other hand, there are numerous, irregularly arranged, round or
elongated, oval openings of variable size, which lead into more deeply situated ducts.

The internal gastral surface of the cup presents another appearance. Here one ob-
serves a number of longitudinal ridges from 2 to 3 mm. in breadth, which project inwards
and are separated by deep longitudinal grooves of equal breadth. Since these longi-
tudinal ridges with arched roof extend from the lower blind end of the slender cup-shaped
gastral space to the upper, and here and there part, their number gradually increases from
below upwards, and in the upper broken end amounts to twenty. Each of these longi-
tudinal ridges consists of two plates, which pass into one another at the free inner
edge, and thus enclose an elongated slit-like space. Internally this is shut off from the
gastral cavity by the skeletal fold referred to, while it communicates externally
with those cavities and clefts which occur at the foot of the external transverse
grooves. From the longitudinal furrows, however, which occur on the internal surface
of the cup between the gastral ridges, there is a direct communication into the lumen of
the tubular stumps which project radially on the external transverse ridges. When this
is compared with the structure of the not very remotely allied Perijihragella, it seems
clear that the lateral tubular stumps of the outer transverse ridges, whose cavities com-
municate directly with the gastral cavity and w^hich project radially outwards, represent
the efferent lateral tubes of Periphragella, and like the large oscular terminal aperture of
the whole cup serve for the exit of the water, while the apertures and cleft lying in the
external transverse grooves, together with the associated, but internally closed, longi-
tudinal ducts or slits, belong to the afferent system.

The dictyoual framework consists of smooth beams enclosing predominantly square or
rectangular meshes, without thickening of the nodes of intersection. The freely projecting
conuli are on the other hand beset with tubercles and teeth.

In the compact portions of the lattice-work, a few isolated spicules here and there
occur, but it is, of course, doubtful whether they really belong to Lcfroyella decora or
are only intrusions. I have, therefore, simply mentioned their occurrence without giving
any figure. Besides simple, smooth hexacts of small size and uncinates of medium length,
there are scopulai with four straight, terminally knobbed prongs.

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

As appendix to the Euretidse I must here mention two small sponges, which in form
and structure undoubtedly belong to this family, though they neither agree with any of the
species described above, nor yet were sufficiently differentiated to admit of adequate and
distinct specific diagnosis.

They form two small thin-walled funnels, 1 and l^ cm. in height, fixed on a small
basal plate, Avith a terminal aperture above, 5 to 7 mm. in width, and with several short,
tubular, stump-like lateral openings. They thus appear to be young individuals from
which a Eurete or Periphragella or even Lefroyella might quite well develop. They
agree with one another so thoroughly, both in macroscopic and microscopic characters,
that they unquestionably belong to the same species. They were both trawled off Little
Ki Island, at Station 192, from a depth of 140 fathoms and a blue mvid ground.
The dictyonal framework generally resembles that of species Ezirete. The beams are
thickly beset with coarse and fine tubercles in irregular disposition, and exhibiting a
distinct thickening of the nodes of intersection, and also short freely projecting conuh
(PI. LXXII. fig 4). The parenchymalia are of variable size, and lie at right angles to
the two bounding surfaces. They are represented by rmcinates, and by oxyhexasters
with short principal rays, and rarely with two longer slightly curved terminals on
each principal (PL LXXII. fig. 7), and also by discohexasters in which the short
principals bear a number of long slightly curved terminals with marginally toothed
terminal discs (PL LXXII. fig. 8).

The dermal skeleton consists of hypodermal pentacts, in which the four tangentials
and also the radial proximal ray are at their ends somewhat roughened and slightly
pointed or rounded at the tips. Beside the radial ray of these hyjoodermalia there lie
radially directed scopulse, in which the shaft is pointed at the inner end, while the outer
exhibits a thickening with four or five slightly curved or even geniculate, thin, smooth,
terminal rays, in which the club-shaped terminal knobs are equipped with barbs
(PL LXXII. figs. 5, 6).

On the gastral surface I could only recognise pentact hypogastralia, which resemble
the hypodermalia (PL LXXII. fig. 4).

Family II. M e l i t t i o n i d ^, Zittel (Pis. LXXXIII.-LXXXVL).

Scopularia which have the form of a ramified tube or of a cup mth lateral blind
diverticula. The dictyonal framework forms somewhat regularly hexagonal open radial
spaces. In each of these a funnel-shaped, outstretched continuation of the reticular
membrane of the ciliated chambers extends across the lumen. This is covered
further along the external surface by the dermal membrane and on the internal surface by
the gastral, stretching flatly over the various apertures. The gastral skeleton includes
no scopulse.

EEPORT ON THE HEXACTINELLDIA. 305

Genus Aphrocallistes, Gray (Pis. LXXXIII.-LXXXVI.).

1858. Gray, Proc. Zool. Soc. Lond., vol. xxvi. (Ann. and Mag. Xat Hist., ser. 3, vol. ii. p.
224) {Aphrocallistes heatrix).

1867. Gray, Proc. Zool. Soc. Lond., p. 492.

1868. Wyville Thomson, Ann. and Mag. Nat. Hist., ser. 4, vol. i. p. 114.
18G8. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. i. p. 161.

18G9. Wyville Thomson, Phil. Trans., vol. clix. p. 701.

1869. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. iii. p. 192.

1869. Bowerbank, Proc. Zool. Soc. Lond., p. 66 (75).

1870. 0. Schmidt, Grundziige einer Spongienfauna des atlant. Gebietes.
1870. Wright, Quart. Journ. Micr. Sci., vol. x. p. 77, pi. i.

1870. Kent, Monthly. Micr. Journ., vol. iv. p. 241.

1871. Gwyn Jeffreys, Proc. Roy. Inst., No. 54, p. 258 (Aphrocallistes bocagei).

1872. Carter, Quart. Journ. Micr. Sci., vol. xii. p. 450.

1872. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. ix. p. 442.

1873. Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. xii. p. 349.
1875. Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. xvi. p. 1.
1875. Willemoes Suhm, Zeitschr. f. wiss. Zool., Bd. xxv.

1875. Marshall, Zeitschr. f. wiss. Zool., Bd. xxv.

1876. Marsshan, Zeitschr. f. wiss. Zool., Bd. xxvii.

1877. Zittel, Abhandl. d. Baier. Akad.

1877. Wyville Thomson, The Atlantic.

1878. Zittel, Zur Stammesgeschichte der Spongien.

1879. Zittel, Handbuch der Pateontologie.

1880. 0. Schmidt, Spongien des Meerbusens von Mexico, p. 48.

1881. Milne-Edwards, Comptes rendus, vol. xciii. pp. 876, 931 ; Ann. and Mag. Nat. Hist.,

ser. 5, vol. ix. pp. 37-41-46.

1882. Weltner, Beitrage zur Kenntniss der Spongien.

History. — The genus Aphrocallistes was established by Gray in 1858,^ for the
species Aphrocallistes heatrix, from a skeleton obtained at Malacca, and was characterised
in the following manner: — "The sponge cylindrical, tubular, branched, the end of the
main tube closed with an open network formed of spicula; branches cylindrical, simple,
rarely bifid, rounded and closed at the end; the inner surface of the tube with large un-
equal-sized concavities placed in longitudinal series, having a large roundish oscule near
its lower edge. The sponge hard, close, calcareous, with uniform, close, equal, regular
hexangular pores on the surface, and large round ostioles in series on the sides of the main
tubes. The outer surface formed of intertangled transparent spines, which inosculate and
unite with each other at their intersections, forming a hard rather brittle crust. The
inner surface lined with a coat of fusiform transparent spicula, which are placed in
bundles parallel to each other in the spaces between the roundish internal apertures of
the crowded small superficial pores."

In his spongiological system,^ Gray founded for this new form a special family —

1 Ann. and Mag. Nat. Hist., ser. 3, vol ii. p. 224 ; Proc. Zool. Soc. Lond., vol. xxvi.

2 Proc. Zool. Soc. Lond., p. 507, 1867.

(zool. CHALL. EXP. — PAET LIII. — 1887.) Ggg 39

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

the Aplirocallistidoe — which he placed in the order of the Corallispongia between
the families Dactylocalycid^ and Euplectellidse ; and he characterised it as follows: —
" Sponge tubular, tubes closed with a reticulate lid ; parietes formed of agglutinated
siliceous spicula, with round horizontal lateral pores ; inner surface strengthened with
clustered longitudinal bundles of elongated spicula."

The generic diagnosis of Aphrocallistes ran thus : — " Sponge tubular, closed with a
lid, with smaller lateral tubular branches, which are generally open at the ends." The
remark contained in the first description given in 1858, to the effect that the skeleton of
Aphrocallistes beatrix is " calcareous," Gray now speaks of as a mere " slip of the pen,"
and particularly notifies that it should have been called " siliceous."

A figure of several of the forms of spicules, together with a portion of the reticulate
framework of siliceous beams belonging to Aphrocallistes beatrix, occurs in Wyville
Thomson's communication on the vitreous sponges,* where special attention is called to
the peculiar spicular form, " which consists of a lengthened shaft, ending in a small
expansion, from which spring four equal branches, each terminated by a little knob."

In his paper on Holtenia,^ Wyville Thomson refers not only to Ap>hrocaU{stes
beatrix, Gray, but to a second species, Aphrocallistes bocagei, Wright MS. (p. 713). As
a very frequently occurring spicule of Aphrocallistes he describes a "regular six-
rayed star, with the principal axis longer than the transverse rays and one half of it
feathered."

In 1869^ Bowerbank referred Aphrocallistes beatrix, Gray, to his genus Iphiteon, and
named it Iphiteon beatrix. This generic name — Iphiteon — Bowerbank has, it is true,
ascribed to Valenciennes, but upon no other ground, it would seem, than that he found
the form in the Museum of the Jardin des Plantes, Paris, noted on the labels as
Iphiteon panicea, Valenciennes. Bowerbank has characterised this genus by the following
diagnosis {loc. cit., -p. 76): — "Skeleton siliceo-fibrous, fibres solid, cylindrical, reticula-
tions symmetrical, areas rotulate, confluent."

Perceval Wright has described and figured a second species of the same genus, under
the name of Aphrocallistes bocagei} Various specimens of this form, brought from the
Cape Verde Islands, were preserved in the Museum of Lisbon, in the British Museum, and
in his own collection, and had also been dredged by Wyville Thomson during the
" Porcupine " Expedition, off the south-west coast of Ireland, in deep water. The short
but clear description given by Wright runs as follows : — " Sj^onge fistulous, erect,
branching somewhat irregularly ; skeleton siliceo fibrous, more or less symmetrically
radial ; radii short and stout on the outer surface of the skeleton, forming a series of
hexagonal spaces, which are nearly all of the same dimensions, central umbo of the ray
giving origin on its inner surface, often on both surfaces, to a long spine. These spines,

1 Ann. and Mag. Nat. Hist., ser. 4, vol. i. p. 123, 1868. ^ p/ij^. Trans., vol. clix. pp. 701-720.

3 Proc. Zool. Soc. Land., p. 75. * Qiuirt. Jomii. Micr. Sci., January 1870, pp. 77-7S.

REPORT ON THE HEXACTINELLIDA. 307

generally long, sliarp-pointed, sometimes knob-headed ; spicules acerate, retentive,
verticillately spined, attenuated, rectangulated, liexradiate, and subfusiform cylindrical,
entirely spinous. Main tube closed by an irregular siliceous network, which is deeply
concave. Pores aud dermal system unknown."

Among the differences between Aphrocallistes hocagei and Aphroccdlistes heatrix to
which Wright called attention, but which hardly seemed to him to have more than a
varietal significance, was the entire absence of the "porrecto multiradiate spicules" in
the case of AjjhrocaUistes hocagei, — a distinction all the more essential, as these spicules
were found to be quite constant in Aphrocallistes heatrix. As further points of
distinction, he noted that "the areas forming the skeleton in Aphrocallistes hocagei are
much more regularly hexagonal than those in Aphrocallistes heatrix. The spines on
the bosses are very much longer in the former than in the latter species ; in it, too, the
central cavity is larger. The reticulated network-like lid is much more radiate in its
composition than in AphrocaUlstes heatrix. The bosses on the rays of the body skeleton
are often knobbed." He also announced that Alexander Agassiz had recognised, from
the description placed before him, the agreement between these and one of the sponges
dredged by Count Pourtales to the south-east of Florida. This was soon afterwards
confirmed by Oscar Schmidt, who examined the sponges dredged by Pourtales, and
described them in his Grundziigen einer Spongien-fauna des Atlantischen Gebietes.
He discovered the transversely disposed plate, described by Wright as a "reticulated
network-like lid," of the wide exhalent opening of the main tube, not merely at the
extremity, but also at severttl places in the interior of the tube, forming transverse
septa.

Special attention was directed by Oscar Schmidt to certain spicules, which were very
similar to the forms described by Wyville Thomson in Aphrocallistes heatrix. Each of
the latter was provided with, a straight shaft and four knobbed terminal teeth, while
the former exhibited only three such knobbed terminal teeth provided with fine prickles.

In 1870 Saville Kent also found, in the sponge material obtained by him during the
" Noma" Expedition, ofi" the coasts of Spain aud Portugal, Ap>hrocallistes hocagei, Wright.
From his own examination he established its specific distinctness from Aphrocallistes
heatrix, and he added, as further diff'erential characters, that the whole skeleton is much
more slender, and is wanting in that echinate aspect of the bosses and shafts of the radii,
characteristic of Apihroccdlistes heatrix. The " porrecto-multiradiate" spicules, which seem
in fact to be characteristic and typical of the genus, are not wanting, as Professor Wright
imagined ; but there are none of the verticillate spined forms so abundant in
Aphrocallistes heatrix ; while, on the other hand, Aphrocallistes hocagei is at once
recognised by the presence of abundant hexradiate spicula, having one extremity of the
shaft profusely spinous, and accordingly bearing a close resemblance to those that occur
in Pheronema grayi.

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

On the Portuguese coast Gwyn Jeffreys also found A'phrocallistes hocagei.

Carter, in bis system of the Hexactinellidse, ^ unites the characters which are common
to the two then known species of Aphrocallistes in the following manner: — "Species
tubular branched, branches closed at their free extremities, wall thick, formed of
polyhedrally reticulated fibre possessing a scopuline shaft."

Of Aphrocallistes heatrix, Gray, he says : — " Rosette (small) with elongated shaft-like
axis, many-rayed, rays straight, pointed or capitate, thorn-like chiefly situated in the
middle and at the terminations of the shaft, arranged more or less verticillately ; or
(large) with microspined rays slightly curved and not capitate. Scopuline shaft headed
with four rays of equal length, slightly everted, microspined, and terminating in small
globular heads."

With respect to Aphrocallistes hocagei, Wright, on the other hand, he notes the
following as peculiar : — " Rosette many-rayed ; rays of equal length, straight, capitate ;
or with long shaft-like axis, like that of the small form in Aphrocallistes beatnx, with or
without heads.

" Scopuline shaft headed with four rays of equal length flexed outwards, e7i fleur-de-lis,
microspined, and each terminating in a large conical end."

In the more minute description^ given of the two species, which are especially related
by the form of the spicules, Carter notes the following different spicular forms as being-
characteristic of Aphrocallistes hocagei : —

(1) " Linear fusiform spicules with inflated centre and extremities.

(2) " More delicate linear fusiform spicules, spiued throughout, all the spines being
in the same direction.

(3) " Hexradiate spicules whose arms are more or less unequal in length, five being
smooth at the commencement and conieally inflated and spined at the termination, and
the sixth spined, feather-like, rounded, the spines increasing in length from the fixed
end to the free.

(4) " Scopuline spicules, consisting of a long shaft and four rays terminating iu
conical heads surrounded by recurved spines.

(5) " Rosettes with five-rayed capitate arms.

(6) " The same rosettes with the axis stretched out linearly, shaft-like, and the rays
arranged round it more or less spirally.

(7) " The same rosettes, with the rays of the shaft more confined to its centre and
all simple (that is, not capitate, but pointed)."

In Aphrocallistes heatrix, Carter found the spicules generally similar to those in
Aphrocallistes hocagei, yet some forms exhibited typical differences, namely, "the
hexradiate spicules whose pointed arms are sparsely and irregularly covered throughout
with smooth spines curved outwards: scopuline spicules whose four rays were quad-

' Ann. and Mag. Nat. Hist, ser. 4, vol. xii. p. 359. ^ Loc. cit., pp. 449-452,

KEPORT ON THE HEXACTINELLIDA. 309

rangularly based on a hand-like expansion of tlie end of the shaft; a straight large shaft
more or less beset with long thorn-like spines, most numerous towards the centre where
they are vertical, and at the extremities wdiere they are divergent, each slightly curved
and microspined ; and a smaller kind in which the rays are straight smooth and capitate."

In 1875, in his Classification of the Spongida,^ Carter erected within the famUy of the
Vitreohexactinellida a special group — the Scopulifera — in which he noted, as type,
Aphrocallistes bocagei, Wright.

Marshall (1876)" ranked the genus Aphrocallistes in his group of Pleionacidfe, and
characterised it in the following words : — " Polyzoic, walls with prismatic anastomosing
radial tubes; individuals more or less tubular or ball-shaped, astomate, arranged into
groups by partition walls. Framework-tissue possessing an apparent regularity.
Spicules do not throughout constitute the groundwork of the siliceous beams. The
latter are often strangely bent."

ZitteP (1877) based his family Mellitionidse on the genera Ajihrocallistes, Gray
Fieldingia, Saville Kent, Stauronema, SoUas, and noted the following characters : —
"Sponge body branched, spherical or plate-like. Wall completely perforated by
numerous tubular water, canals and thus divided into honeycomb-like chambers.
Skeletal spicules with thick intersections. Surface (naked 1 or) overspread by a delicate
meshed or porous siliceous skin, which also covers the openings of the canals. Root
absent."

Oscar Schmidt found Aphrocallistes abundantly among the sponges of the Gulf of
Mexico.^ He believed that the peculiar structure of the six-sided prismatic parietal
meshwork could be explained by a modification of the fundamental hexradiate spicules —
in which all the six rays do not cross at an angle of 90°, but two at an angle of 120°. He
compares the lattice-like retiform transverse walls to the sieve-plate of Eu^ilectella and
suggests that they had been formed during pauses in the growth. The shaft provided
wdth prongs on both ends and on the middle, which was proposed as a charac-
teristic feature of the species Aphrocallistes heatrix, 0. Schmidt declares to be an
accidentally intruded element, and expresses the belief that this species is not specifically
distinct from Aphrocallistes bocagei.

During the " Porcupine " Expedition a cup-shaped sponge fragment, 1^ cm. in height,
was dredged off the south-west coast of Spain from 1095 fathoms. This formed the
swollen base of a Hexactinellid and was carefully described by Duncan in 1881^ as a new
species of Ap>hrocallistes. If this sponge belongs to the genus Aphrocallistes — which,
however, according to Duncan's description of the continuous skeletal framework, can

1 Arm. and Mag., Nat. Hist., ger. 4, vol. xvi. p. 199.

2 Zeitschr. f. vriss. Zool., Bd. xxvii. p. 124.

3 Studien iiber fossile Spongien, Ahhandl. d. baier. Akad., ii., vol. xxii., div. 1. p. .3G.
' Spongien des Meerbusens von Mexico, 1879-80, p. 48.

5 Journ. Linn. Soc. Land., vol. xv. pp. 324-328.

310 THE VOYAGE OF H.M.S CHALLENGER.

hardly be regarded as certain— it would represent, to judge from the peculiarity of
the isolated spicules, a new species differing at once from Ai^hrocallistes bocagei and
Aphrocallistes beatrix.

According to the report of Milne-Edwards,'' Aphrocallistes bocagei was also dredged
by the French " Travailleur" Expedition in the Atlantic Ocean, off the coasts of France
and Portu(!;al.

After Zittel, in his Studies on the Hexactinellida," had mentioned that in the living
species of Aphrocallistes a very delicate framework extended over the outer surface and
the branches of the canals, Weltner described^ (1882) a continuous covering both on the
inner side and outer side of Ajjhrocallistes bocagei. In the case of the outer surface
he noted the presence of a delicate covering with large inhalent pores, which passed
into the characteristic parietal meshes and also spread over the ostia on the inner surface.
The outer membrane contained hexradiate spicules with an externally projecting fir-tree-
like ray and forming a regular mesh work; in the inner lay " the characteristic Aphrocallistes
rosettes, the rod-like spicule (thorn spicule), and the greatly reduced form which, though
only exhibiting four medium knobs, was yet recognisable as hexradiate, and lastly,
a large well-developed hexradiate form which is often fused to the parietal skeleton."
Moreover Weltner observed in his specimen " a third plasma layer between these
coverings of the outer and inner surfaces in the interior of the meshes (canals). This
was for the most part well preserved and was interrupted only in tlie middle, while it
was continued inwards to form an outer covering for the cavity of the sponge. In it the
free spicules of the internal covering were but seldom met with, although it stood in
manifold connection with the latter by means of plasma-threads. This third layer,"
continues Weltner, " may indeed be an artificial product. I would not, however, leave it
unnoticed merely on that account, because it seemed to me to show that the free spicules
(flesh spicule) do not as a rule extend into the interior of the plasma."

Character of the Genus. — The wall of the cup- or tube-like body is supported
by the honeycomb-like framework of the dictyonal skeleton. The radial canals which
traverse it, and are from 1 to 2 mm. in breadth, appear almost regularly hexa-
gonal and prismatic. The septa between the latter form a network of beams, with
irregular but predominantly three-sided meshes, from the margins and surfaces of which
conical pegs project, which are sometimes swollen and knob-like at their extremities.
Where three such bounding plates meet laterally the networks of beams usually fotnl
three-sided prismatic interspaces, and thus the margins of the six-sided prismatic canals
or honeycomb-like mesh-spaces become somewhat truncate or rounded.

The strongly developed reticulate dermal membrane is continued without inter-

1 Comptes rendus, 1881, vol. xciii. pp. 876-931 ; Ann. and Mag. Nat. Hist., vol. is. p. 46.

» Ahlmndl. d. haier. Alcad., 1877, p. 49.

3 Zur Kenntniss der Spongien, 1882, p. 32.

REPORT ON THE HEXACTINELLIDA. 311

ruptioii over the whole outer surface of the sponge, and extends over the external
openings of all the prismatic mesh-spaces. In like manner the gastral membrane, which
has a perfectly similar structure, extends internally parallel to the former, and covers
the inner openings of the prismatic honeycomb-like meshes. Between the two
perforated bounding membranes a fine skin extends deeply folded in funnel-shaped
fashion; this forms a continuation of the wall of the chambers, which are shaped
like the finger of a glove, and lie around and open into the common central space. The
latter is provided with an internal large exhalent orifice. This remarkable funnel-shaped
expansion of the chamber wall may be supposed to have arisen by amalgamation of
some specially large external chambers whose dividing walls projected to some extent
even into the excurrent central space (PL LXXXIV. fig. 1 ; PI. LXXXVI. fig. 2).

The dermal skeleton consists of hexacts, in which each of the projecting outer rays
has usually a tree-like appearance, though sometimes reduced to a prickly peg or knob, or
even in many cases to a small tubercle, so that finally the spicule in question is no longer
to be regarded as h exact but as pentact. Besides these dermal hexacts or pentacts
scopulse with knobbed or pointed terminal rays also occur.

The gastral skeleton consists of diacts which are completely enclosed in the gastral
membrane.

Among the loose parenchymalia, there are present in addition to the uncinates which
project at right angles to the outer surface, small hexacts and hexasters of various kinds,
in variable abundance and irregular distribution.

It is doubtful whether Aphrocallistes hcatrix, Gray, and Aphrocallistes bocagei,
Wright, are distinguished by suificiently marked and sufiiciently constant peculiarities to
be regarded as distinct species. On the other hand, the new forms which, on account of
their figure, I have named Aphrocallistes vastus and Aphrocallistes tubulosus, I regard
as sharply defined " good species. "

1. Aphrocallistes heatrix. Gray (PI. LXXXIV. figs. 9, 10).

Since I had an opportunity of examining in the British Museum the original specimen
from Malacca on which this species was established by Gray in 1858, I can, in the first
place, bear testimony to the correctness of Gray's description and figures of the micro-
scopic structure, and can also confirm those of Wjrville Thomson, Bowerbank, and Carter,
in regard to the minute structure of the dietyonal framework and the isolated spicules.
I may, therefore, refer the reader to these representations.

Among the comparatively insignificant, and hardly qualitative distinctions which
have been noted in regard to the microscopic structure of this form as compared with
that first described by Perceval Wright (in 1870) from the Atlantic basin, special
emphasis is laid on a spicular form which has been spoken of by aU observers since Wyville

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

Thomson as characteristic of the species AphrocaUistes heatrix, Gray. This, which
has been figured by me in PL LXXXIV. figs. 9, 10, occurs irreguhirly scattered in
great numbers throughout the whole parenchyma. While the one axis of the spicule
has two rays greatly prolonged, and while, each of these two long rays divides into
four diverging, pointed terminals, the rays of the two other axes, which cross the first in
the middle, remain simple short principals, which end in sharp points. The entire form
may be therefore described as a longitudinally extended oxyhexaster, in which the four
short princij)al rays remain undivided, while each of the principals of the long axis, which
are sometimes provided with lateral prickles, divides into four diverging, pointed terminals.

Since parenchymalia of this kind are not found in any other form of AphrocaUistes,
it becomes possible to determine the separation of this form as a distinct species ; I must,
however, draw attention to the fact that in AphrocaUistes hocagei, which is also very
similar in external appearance, I found widely scattered parenchymal oxyhexasters, and
similar forms were also figured by Oscar Schmidt in the Spongien des Meerbusens von
Mexico (pi. vi. fig. 3). They do not indeed completely agree with the above peculiar
spicules, but they approach them, and evidence at least a close affinity between the two
forms. The view which Oscar Schmidt has expressed,^ to the effect that the spicule
w^hich is characteristic of AphrocaUistes heatrix is an accidentally introduced foreign
element I cannot accept.

For purposes of comparison with the other species of the genus AphrocaUistes more
carefully described below, I will here give a short summary of the most important
microscopic skeletal characters, based upon my examination of the original specimen of
AphrocaUistes heatrix, Gray.

The dictyonal framework is formed, as represented by Wy\dlle Thomson ^ and by
Bowerbank,' of a tolerably irregular, narrow-meshed network with strongly thickened
crossing knots. The beams are almost entirely and more or less thickly beset with small
tubercles, but these tubercles are stronger and more abundant on the spherical nodes of
intersection, and on each of the strong conical pegs which project freely both on the
dermal and gastral surfaces, and also in the interior of the radial prismatic mesh-spaces.
The pegs, projecting into the lumen of the mesh-spaces, seldom stand exactly at right
angles to the surface of the reticulate partition from w^hich they spring, but are directed
obliquely inwards towards the gastral cavity of the entire sponge. The middle portion
of all the septa between adjoining prismatic radial canals consists only of a single layer
of irregularly fused hexacts, and an irregularly triangular prismatic interspace is formed
where three such septa meet.

The dermal skeleton consists of delicate hexacts in which the distal ray bears
numerous narrow, curved, fir-tree-like, lateral prickles, while the five other straight

' Spongien des Meerbusens von Mexico, p. 50. ^ ^^yi_ g^^yi Mmj. Nat. Hist., ser. 4, vol. i. p. 123.

^ Proc. Zool. Soc. Land., 1869, pi. xxi. figs. 2-4.

REPORT ON THE HEXACTINELLIDA. 313

and similar pointed rays are beset only with short lateral prongs. Besides these
hexacts, which form with one another a quadrate dermal lattice-work, the dermal
skeleton contains numerous scopulse with their four or five knobbed prongs at right
angles to the surface, while the long slender smooth stalk runs out to a point.

In the case of the gastral skeleton unfortunately nothing certain could be discovered
in the small dried fragment at my disposal ; on the other hand, the loose parenchymalia
were well preserved in great numbers. There were numerous uncinates arranged at
right angles to the bounding surface, and therefore parallel to the radial prismatic canals.
The greatest breadth of these does not occur at the middle, but in their outer third part,
while the attenuated gastral extremity gradually runs out to a fine point. The thin
pointed barbs of the uncinates are tolerably densely apposed. The peculiar elongated
oxyhexasters, which have already been described, occur scattered irregularly and in great
numbers throughout the whole parenchyma.

2. Aphrocallistes hocagei, Wright (PI. LXXXIII. ; PL LXXXIV. figs. 1-8).

Both among the Hexactinellida of the Challenger Expedition and among the others
purchased by Dr. Doderlein in Enoshima, there are numerous representatives of this
form. Some of these are well preserved in alcohol. The fully developed typical form is
a tube gradually widening upwards, with numerous radial glove-finger-like swellings
on the lateral walls. The axis of the entire tube, which may attain a length of 20 cm.
or more, exhibits as a rule a slight curvature. The inferior extremity, which is firmly
attached to the substratum, has the form of a small cup, the wall of which shows diver-
ticulum-like swellings only a few mm. above the basal plate, which is from 3 to 5 mm. in
breadth. These are at first quite low, but further upwards they gradually increase in
length, and finally attain a length of 5 cm. or more. The breadth of these diverticula,
which always end blindly, measures on the under end of the tube in most cases only from
3 to 5 mm., but gradually increases in the middle and upper parts to a diameter varying
from 1 to 2 cm. Very frequently much elongated diverticula occur here and there at a
short distance above the base. These are bent obliquely downwards, reach the firm
substratum or some laterally adjacent solid body, and become supports for the entire
sponge (PI. LXXXIII. fig. 1). In many cases the diverticula are arranged in more or less
longitudinal rows, which in the inferior part of the entire tube are usually four in
number and arranged in a cruciate manner. Superiorly this arrangement becomes
indistinct or is no longer to be seen. A well-marked whorled disposition of the diverticula
I have not been able to observe ; on the other hand, I now and again saw certain varia-
tions from the normal conditions which are perhaps of importance as to the relation of
this form to the others which, though separated off as distinct species, are at the same
time closely related forms. On the one hand, cases are not unfrequent in which a lateral

(ZOOL. CIIALL. EXP. — PART LIII. — 1887.) Ggg 40

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

communication is established between two or more adjoining diverticula, while they are
united at the base into a common simple tube or are in open connection throughout
their length. If this union occurs between the diverticula belonging to the longitudinal
row, it may result in extreme cases in the formation of a longitudinal fold, on which
the individual diverticula are only indicated as short boss-like swellings. On the other
hand, the long diverticula W'hich occur here and there are sometimes forked, and in
this there lies the tendency to form branches.

Where the upper terminal opening with its natural margin is preserved quite uninjured
it is closed, just as in Aphrocallistes beatrix, Gray, by a transversely stretched narrow
meshed lattice-like plate. The latter is usually somewhat concavely incurved and becomes
united to the honeycomb-like lateral wall in a com^^act, somewhat tuberculate margin.
As already reported by Oscar Schmidt and Marshall, several such thin lattice-like trans-
verse partitions usually occur in the interior of the tube, but I would call attention to
some points of distinction between these internal diaphragms and the terminal sieve-plate
of other Hexactinellids. While the narrow-meshed terminal sieve-plate of other Hexac-
tinellids is united all round to the body-wall so that (apart from the sieve-like meshes)
a complete closure of the tube results, in this case, a semicircular marginal portion of
the internal diaphragms remains unclosed wherever a lateral diverticulum opens into
the large lumen of the tube (PI. LXXXIII. fig. 2). With regard to the occurrence,
number, and arrangement of these transverse septa I have found great diflFerences. While
some specimens well preserved in other respects possess, apart from the terminal plate,
no trace of septa, others show three or more internal diaphragms, but no constant relation
to the whorls of diverticula can be recognised, so as to suggest the reduction of the
entire tube to series of metameres. I regard it as most probable that during the growth
of the tube a temporary provisional occlusion is effected by a transverse sieve-net, and
that only after growth has ceased is a terminal regularly constructed lattice-work formed
which entirely closes the lumen. While the latter consists of tolerably similar thick round
beams, varying from 0'3 to 0"5 mm. in diameter, which surround rounded polygonal meshes
of tolerably uniform size, and while freely terminating rays only project here and there
into the lumen of the meshes, the transverse septa in the interior of the tube have
a somewhat different character, inasmuch as they consist of beams of very various
thickness which meet one another to form a network at very diverse angles, in which the
mesh-spaces are not rounded but have sharp angles (PI. LXXXIII. fig. 2).

The ihicroscopic structure of Aphrocallistes hocagei agrees essentially with that
of the corresponding skeletal parts of Aphrocallistes beatrix, as represented by Bower-
bank in his excellent figures.^ The dense network of beams which forms the di\4diug
septa of the six-sided prismatic meshes of the wall consists, as in the case of every
dictyonal Hexactinellid framework, exclusively of amalgamated hexacts. These do not

' Proc. Zool. Soc. Loncl, 1869, pi. xxi. figs. 2-4, and pi. x.xii. fig. 2.

REPORT ON THE HEXACTINELLIDA. 315

however by any means form a regular lattice-like framework with cubical mesh-spaces.
A framework with predominantly three-sided meshes is in fact formed, on the one hand
by the diversion of some rays from their original position at right angles to one another,
on the other hand by quite irregular fusion of the rays of adjacent hexacts. In older
portions these meshes appear distinctly compressed and rounded. The surface of the
beams may be completely smooth, or may be roughened by the development of more
or less numerous small pointed tubercles. The pegs projecting outwards towards the
dermal membrane, or inwards towards the gastral membrane, are always beset with
numerous tubercles and are frequently swollen in a knob-like manner. While the freely
projecting dermal pegs are usually straight, that is to say, usually stand at right angles
to the surface, the projecting pegs on the gastral side are found to be mostly incurved or
thickened into short knobs (PL LXXXIV. fig. 1). The ends of the free rays of the
dictyonalia which project from the surface of the septa into the lumen of the meshes
also exhibit rough conical pegs which are seldom directed quite at right angles to
the wall of the canal, but project as a rule obliquely.

The dermal skeleton consists of hexacts which form a regular quadrate network and
are provided with a fir-tree-like distal ray (PI. LXXXIV. fig. 8), and very variously
formed scopulse. The four equally long transverse rays, as well as the usually distinctlj^
shorter proximal ray of the dermal hexacts, are either smooth or somewhat rough on the
truncated extremities.

In the dermal scopulse I observe that the shaft, which is of variable length, runs out
at the inner extremity to a point, while the outer extremity, which is moderately swollen,
or more rarely provided with a knot-like thickening, gives ofi" four (more seldom three or
five) prongs which either terminate in a point or are provided with a knob-like or club-
like terminal expansion. Both on the pointed extremities and also on the terminal
clubs or knobs small lateral barbs usually occur (PL LXXXIV. figs. 4, 5). The branches
of most of the scopulae possess the usual length of about 0'07 mm. Scopulae also occur
here and there which have the branches twice as long, and either terminate in j^oints with
small smooth knobs, or are beset beneath the extremity with small barbs (PL LXXXIV.
figs. 3, 5).

In contrast to the Dictyonina hitherto described, the gastral skeleton of Aphrocallistes
ftocagrei diff'ers very essentially from the dermal. On the surface of the gastral wall I finil
neither hexacts nor pentacts, nor any trace of scopulse, but only diacts of variable length,
more or less rough, sometimes even pronged, with rounded extremities and central knots
variously developed. These diacts lie properly in the gastral membrane, but they also
extend into the subgastral space, and seem not unfrequently to fuse with the projecting,
but frequently incurved and club-like swollen pegs of the dictyonal framework.

The loose parenchymalia include the uncinates, which are disposed at right angles
to the bounding surface, and vary greatly in length and strength. The inner extremity

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

is frequently very feebly developed (PI. LXXXIV. fig. 2). In addition to the uncinates,
the parenchyma contains hexasters of various form and most irregular distribution.
Sometimes one has to search long for one of these hexasters in a section, while in other
cases the whole preparation appears to be permeated by them. Perfectly regular
oxyhexasters, in which each of the principals runs out into four strong and moderately
long, diverging, secondary rays, are comparatively rare. Usually one finds one or more
principal rays terminating in simple points, while others run out into three or four ter-
minal rays. In another frequent form the two principal rays belonging to one axis are
very strongly developed, and divide into four strong, diverging terminals, while the
four remaining rays, which are cruciately disposed, and belong to the two other axes, run
out into perfectly simple points, or are only partially divided into two tei'minal rays (PI.
LXXXIV. fig. 7). These latter spicules seem to be related to the forms characteristic of
Aphrocallisies beatrix, Gray, and represented in PI. LXXXIV. figs. 9, 10, while, on the
other hand, small discohexasters sometimes occur, which agree in form and size with
certain discohexasters, which occur very frequently in Aphrocallisies vastus. These
have a diameter of 0"03 mm., and each of the short principal rays bears four to six
somewhat curved diverging terminals, each of which is tipped by a small end plate, or
merely by a spherical terminal knob (PI. LXXXV. figs. 8, 9).

In regard to the soft parts, I may state that the structure of the dermal and gastral
membrane hardly varies from the ordinary Hexactinellid type, and the same ma}'' be
said of the trabecular network which extends between these two membranes and the
chamber layer. It is different, however, with the chamber layer itself. The most
important peculiarity has been already referred to in the generic diagnosis ; it claims,
however, more attention, especially since I was able to investigate carefully some
comparatively well-preserved specimens.

In each of the honeycomb-like hexagonal spaces, which are quite open on both sides
in the macerated skeleton, and penetrate the wall in a radial direction, there is a peculiar
system of chambers, which opens on the gastral surface through a single wide round
aperture above the lattice-work of the gastral membrane. Into the wide, canalicular
space above this aperture there open laterally a number of thimble-shaped chambers of
medium size, in the form of simple, closely-apposed diverticula, while from the dermal
surface, three to five large blind diverticula also open into the same. These latter
diverticula closely surround the wall of the tubular skeletal space, and are laterally so
closely apposed to one another, that they form between them a wide, median, funnel-
shaped space. The inner wall of the large diverticula which surround this funnel-shaped
space is simply smooth, while their external wall adjacent to the surrounding skeleton
is distended into chamber-like diverticula, just as we previously saw on the surface of
the w^ide general excurrent space (PL LXXXIV. fig. 1). In this way a funnel-shaped
membrana reticularis is so stretched in the hexagonal honeycomb space, that the

REPORT ON THE HEXACTINELLIDA. 317

circular incurrent aperture of the funnel lies under the porous dermal membrane, while
the blind pointed apex lies somewhat in the middle of the space, and turned towards
the gastral membrane.

While a section at right angles through the wall, which shows longitudinal sections
of the radial canals, exhibits the funnels in longitudinal or lateral section, that is to
say, affords a lateral view of a chamber system (PL LXXXIV. fig. 1), a view from
the dermal skeleton reveals in each of the hexagonal spaces, on the inner surface of the
skeletal enclosure, a circle of chambers which surround a simple, central, funnel-shaped,
space (PI. LXXXIII. fig. 4), and a view from the gastral surface into the wide,
excurrent tube, shows the septa which arise by the coalescence of the large diverticula
forming the funnel (PI. LXXXIII. fig. 3).

3. Aphrocallistes vastus, n. sp. (PI. LXXXV.).

The specimen represented on PI. LXXXV. fig. 1, in its natural size, was collected
by Dr. Doderlein in the Sagami Bay (Japan), from a depth of 180 fathoms. There
was also a small fragment firmly fixed to a Coral, and apparently of similar structure,
probably belonging to the same specimen. They represent parts of the lateral wall
of a large cup. Instead of the glove-finger-like sacculations which occur in Aphro-
callistes beatrix and Aphrocallistes bocagei, there is here a simple folding of the wall.
Whether the tolerably irregular, bulging folds, which are here and there attached to
the Coral branch were directed longitudinally or transversely to the axis of the entire
cup covdd not be certainly determined, though I am inclined to believe that they were
longitudinal. In this connection it is interesting that in another specimen of Aphro-
callistes bocagei, bought in Enosima by Dr. Gottsche, an indication of the longitudinal
folding of the cup-wall could be recognised. A trace of the same is also to be observed
in the figure of Aphrocallistes beatrix given by Gray.

Since the thickness of the cup wall amounts to 5 mm., the mesh spaces, which are
about 1 mm. in width, have become canals, which penetrate the wall transversely in
a radial direction. The dermal membrane, which is still clearly visible in these dried
specimens, extends in the form of a delicate skin over the whole outer surface. With
a lens one can recognise a fine quadrate lattice-work formed of apposed dermalia. A
quadrate lattice-like network of this kind is indeed entirely absent on the inner side
of the partially-preserved gastral membrane, which has rather an irregular streaky
appearance.

A more accurate examination of the dictyonal framework of the septa between the
radial six-sided prismatic canals, shows that it consists of a single-layered network,
w^ith meshes predominantly three- or four-sided. The beams of the network bear rays
directed at right angles or oblic|ue]y to the dermal surface, and projecting freely

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

into the lumen of the canals. From an examination of the often vexy obviously marked
axial canals, it may be seen that the rays of adjacent dictyonalia are partly fused in the
familiar longitudinal fashion (somewhat as in Farrea), partly in a more irregular
arrangement, crossing one another arbitrarily, or connected at the intersections. Some-
times all the six rays are concerned in the formation of the network, which lies
approximately in one plane — an arrangement which is obviously only possible through
the great curvature of some rays. Usually, however, one ray is bent at right angles or
obliquely inwards, and is provided with a free point, which projects into the canals on
either side. Where the margins of three adjacent canals meet one another, the lattice-
like networks are slightly separated, and an irregular interspace is thus formed.

As in Aphrocallistes hocagei, the dermal marginal pegs of the dictyonal framework
stand at right angles to the dermal membrane, while the longer gastral marginal pegs are
in part curved inwards. The pegs on the inner surface sometimes project obhquely
towards the dermal surface into the lumen of the canals, and are sometimes applied quite
close to the surface of the wall, but the free tuberculated end is always directed outwards
towards the dermal membrane.

The beams of the meshwork often appear almost entirely smooth, while in other cases
they are more or less richly beset with small tubercles. The freely projecting pegs all
exhibit a rough or tubercled surface.

The dermal skeleton is distinguished by the strong development of the distal fir-tree-
like ray of the dermal hexacts. This is richly pronged and more bushy than in the other
species of the Aphrocallistes. The numerous scopulse, which are present in the dermal
skeleton, exhibit a shaft which runs to a point beneath, and forks externally into two,
more rarely into three branches, after forming a simple expansion or an annular thickening.
The branches are rough on their outer extremities, and terminate either in a simple
rounded manner (PL LXXXV. fig. 7), or in a very slight knob-like thickening.

In the gastral membrane, as in Aphrocallistes hocagei, the hexacts are replaced by
simple, straight, rough or pronged diacts of variable length, with rounded extremities and
central knots (PI. LXXXV. fig. 6) ; pin-like monacts are also scattered here and there
(PI. LXXXV. fig. 10). Whether the scopuljB, which are entirely absent in the gastral
skeleton of Aphrocallistes hocagei, occur in the present instance remains doubtful. It is
true that in the dried specimen and on the inner side of the cup-wall scopulae occur
which, like the dermal, consist of a terminally pointed shaft, and of two, more rarely
three knobbed and externally roughened branches, but I am not sure that these are not
subsequent extrinsic intrusions.

The uncinates are distinguished by their length and also by the fact that their
greatest dilatation usually lies much nearer the dermal than the gastral extremity.
This latter appears much more slender, and usually runs out into a smooth (PI. LXXXV.
fig. 2) point, which is less frequently beset with lateral prongs. It is important to note

REPORT ON THE HEXACTINELLIDA. 319

that the irregularly scattered loose parenchymal hexasters, which are present in large
numbers, all bear terminal rays the ends of which are knobbed or provided with small
thick transverse discs (PI. LXXXV. figs. 3, 5, 8, 9). The diameter of these disco-
hexasters varies from 0-08 to O'OS mm. The principal rays remain, as a rule, uni-
formly short and crowded, but the two rays of one axis are often greatly prolonged in
comparison with the others, and it is just in such cases that the latter usually remain
simple (PI. LXXXV. fig. 5), while the former become divided into two to four ter-
minal rays.

4. Aphrocallistes ramosus, n. sp. (PI. LXXXVI.).

Both among the sponges of the Challenger Expedition and among the Hexactinellida
dredged by Dr. Doderlein in the Sagami Bay, there are dichotomously branched round
tubes from 5 to 10 cm. in height, which are only from 3 to 4 mm. broad at the
liase, but become gradually wider in the upper branches, and finally open out by
cup-shaped lateral and terminal branches from 8 to 10 mm. in width. The specimen re-
presented from a photograph in PL LXXXVI. fig. 1, in its natural size, was obtained
from the Philippines (Station 210 of the Challenger Expedition, lat. 9° 26' N., long.
123° 45' E.), from a depth of 375 fathoms and a blue mud ground.

The tube wall consists of the same honeycomb-like framework of six-sided meshes or
prismatic radial tubes, as in the cup-wall of the other species of Aphrocallistes; and
the microscopic examination shows that the minute structure of the meshes or prismatic
septa does not difi"er essentially from that abeady described.

The tolerably smooth network of beams, which is only here and there provided with
delicate tubercles consists of irregularly fused hexacts, and exhibits predominantly
triangular narrow meshes. While the conical pegs on the dermal side run out to simple
points, and are directed at right angles to the bounding surface, the terminal pegs on the
gastral surface are elongated, provided with a rough pear-shaped end-swelling, and are
frec^uently obliquely directed or somewhat curved round. The conical pegs projecting
from the surface of the network of beams into the lumen of the radial tubes are directed
obliquely outwards, that is to say, towards the dermal surface.

In the dermal skeleton, hexacts occur with a very variously developed distal ray,
whicb is sometimes quite fir-tree-like (PI. LXXXVI. fig. 8), sometimes club-like with
lateral prongs (PI. LXXXVI. fig. 4), aometimes simply rod-like or pear-shaped, or even
quite rudimentary and knob-like (PI. LXXXVI. fig. 3). The proximal ray equals or
usually exceeds the four cruciately disposed transverse rays in length. More rarely it is
shorter than the others, which often exhil)it an externally convex curvature (PI.
LXXXVI. fig. 2) Besides the dermal hexacts, numerous dermal scopulre occur. These
present a smooth, pointed, or terminally rounded stalk, and the outer expansion bears

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

slender diverging branches, each provided with a pear-shaped barbed terminal swelling.
The branch stalks are smooth or rough, and either straight or uniformly curved towards
the exterior, or occasionally slightly flexuous (PI. LXXXVI. figs. 5, 9). This seems to
vary according to habitat, but also according to the individual.

The gastral skeleton consists exclusively of long rod-like diacts, which are rough
throughout or terminally, and are provided with a central node of intersection. Their
extremities, which are embedded in the gastral membrane, are simply rounded or slightly
swollen. Of gastral hexacts or scopulse I have found no trace.

The uncinates of the parenchyma vary greatly in length and form. Sometimes the
greatest breadth occurs just about the middle, sometimes nearer the outer extremity ;
sometimes the barbs are densely crowded, sometimes more widely disposed, and so on.

The numerous irregularly scattered hexasters are, on the one hand, oxyhexasters with
a variable number of terminal rays, which are not unfrequeutly curved, similar in fact to
forms already described in the other species of Aphrocallistes (PI. LXXXVI. figs. 6, 11),
and, on the other hand, regular or irregular discohexasters in which the terminal rays
are also curved, and provided with rounded terminal knobs (PL LXXXVI. fig. 10). In
addition to these, simple regular hexacts occur, in some specimens very abundantly. In
these the rays are slender and tolerably long, smooth or rough, and always ending in
fine points (PI. LXXXVI. fig. 7).

The soft parts, which I was able to examine on some well-preserved spirit specimens,
do not differ in disposition or minute structure from what has been already described in
Aphrocallistes hocagei.

Family IIL Coscinoporid^, Zittel (Pis. LXXXVII.-XCL).

The plate-like wall of the cup-, goblet-, or plate-shaped, firmly attached body is trans-
versely penetrated by more or less elongated funnel-shaped straight canals, which open
alternately on one or other surface (covered only by the sieve-like bounding skin), but are
pointed and blend at the other end. Their length thus always corresponds to the thick-
ness of the sponge body-wall.

Genus Chonelasma, n. gen. (Pis. LXXXVII.-XCL).

The dictyonal framework of the beaker or almost plate-hke specimens is traversed
by two systems of oppositely directed funnel-shaped meshes or passages, which appear so '
arranged that each blind funnel extremity of the one system always occurs between the
circumjacent funnel openings of the adjoining passages of the other and oppositely
directed system. Those oppositely directed passages represent the incurrent and ex-

REPORT ON THE HEXACTINELLIDA. 321

current canal spaces, and in every transverse section they may be observed to alternate
with tolerable regularity.

The dermal skeleton consists of pentacts or hexacts with pronged terminal rays, and
in addition to these numerous scopula3 of different forms occur.

The gastral skeleton consists of pentacts or hexacts, which exactly resemble the cor-
responding dermalia. It also contains in most cases scopulse like those which are found
in the outer skin.

In addition to the parenchymal uncinates, numerous discohexasters occur, more
rarely oxyhexasters and sometimes also simple regular hexacts in variable numbers.

1. Chonelasma lamella, n. sp. (Pis. LXXXVIL, LXXXVIIL).

In the neighbourhood of the Kermadec Islands (Station 170a, lat. 29° 45' S., long.
178° 11' W., 630 fathoms), the trawl brought up some fragments of a plate about the
size of a little hand, from 5 to 10 mm. in thickness, and provided with attenuated, smooth,
irregularly undulating margins. They are in some places somewhat bent and irregularly
thickened, but on the whole they appear tolerably flat. The soft parts are well preserved.
The two lateral surfaces exhibit to the naked eye no noteworthy difl'erences. Both sides
are covered by a fine delicate porous skin, through which are seen the round openings of
the passages, which are about 1 mm. in width, and traverse the plate in alternately
opposite directions. The distribution of these transverse canals is not, indeed, quite
regular, yet the general arrangement of the rectangularly crossed longitudinal and
transverse rows cannot be mistaken.

The two sides of the other macerated specimen are represented, from a photograph,
in their natural size, on PI. LXXXVIL figs. 1 and 2. They exhibit a slightly bent, but
otherwise smooth macerated plate of 3 "5 mm. in thickness, which was obtained at Station
148a (lat. 46° 53' S., long. 51° 52' E.), from a depth of 550 fathoms on hard ground,
while the skeletal fragment figured on PI. XC. figs. 9, 10, and 11, in natural size,
which was dredged at Station 56 (lat. 32° 8' 45" N., long. 64° 59' 35" W.), from a depth
of 1075 fathoms on Coral mud, appears to belong to the same species.

The macerated dictyonal framework exhibits notable variations in the difi'erent
regions of the plate. While the beams in the neighbourhood of the two surfaces
form a tolerably narrow-meshed, and somewhat irregular framework, which surrounds
the wide round openings of the funnel-shaped transverse canals, the middle portion of
the plate exhibits a regular system of perfectly square or rectangular meshes (PI.
LXXXVIIL fig. 1). Since these rectangular meshes in the middle layer are much wider
than the meshes in the neighbourhood of the two bounding surfaces, and since the
middle framework consists of beams, which are not only longer but thinner than those on
the surfaces, it is easy to understand how this middle layer may readily break, and

(ZOOL. CHALL. EXP. — PART LIU. 1887.) Ggg 41

322 THE VOYAGE OF H.M.S. CHALLENGER,

then become spread over the entire plate in two lamellse. The beams of the framework
are more or less richly beset with pointed tubercles. The freely -projecting prongs or
conical pegs are further especially rough and tubercled, while in the interior of the
skeleton portions occur in which the beams appear almost or entirely smooth.

The dermal skeleton consists of strong peutacts of variable size, and in these the
distal ray is entirely absent. The four cruciately disposed transverse rays are slightly
and uniformly bent inwards, and each terminates in a blunt point. While the outer
surface of these transverse rays is thickly beset with strong conical prickles, which
gradually decrease in height towards the lateral margin, the inner surface is quite smooth
(PI. LXXXVIII. fig. 3). The simple conical attenuated proximal ray varies in length,
and is, on the other hand, uniformly beset all round with a few simple conical prongs,
which stand out at right angles.^ In addition to the proximal radial ray of the pentacts,
scopulse extend towards the dermal membrane. These are to be reckoned among the
smaller types. The stalk always ends in a somewhat rough point, and exhibits close
beneath the forking an often sharply defined annular thickening. The four (more rarely
three or five) teeth are quite distinct from one another, and somewhat divergent. They
are either simply blunted or provided with an insignificant knob-like swelling.

The gastral skeleton lying on the other side of the flat section exhibits the same
structure. Here also we find the same pentacts with a pronged upper surface, curved
transverse rays, and a conical radial ray. The scopulee are also similar to those of the
dermal surface (PL LXXXVIII. fig. 1).'

Among the parenchymalia, in addition to the normal uncinates of variable length and
thickness, numerous small discohexasters occur with somewhat rough straight rays, which
may be slightly bent here and there, and bear terminally a small somewhat convex
marginally fringed transverse disc. By di\dsion of one, or a few, or all of the rays into
two or more (seldom more than four) terminal rays, discohexasters of various forms arise,
in which the terminal rays are about three times as long as the shaft from which they
spring (PI. LXXXVIII. figs. 8, 9). Between these discohexacts and discohexasters,
oxyhexacts and oxyhexasters of the same size, and on the whole of similar form, occur.
They also exhibit a similar roughened surface.

The structure of the soft parts presents a general resemblance to that which we have
already seen in the Euretidaj, except that, in relation to the much greater thickness of
the wall, the aff"erent and eS"erent passages are longer and sometimes slightly ramified.
From the subdermal lacunae straight canals extend to near the subgastral spaces, and
between these afi"erent passages lie the efi"erent canals, which are also straight, and extend
from below the subdermal lacunae to the subgastral cavities into which they open widely.

1 111 the figure (PI. LXXXVIII. fig. 3) the prickles of the radial ray are by mistake provided witli incurved,
instead of perpendicularly disposed prongs.

^ By mistake, the gastral scopulte are omitted in the figure.

REPORT ON THE HEXACTINELLIDA. 323

The efferent Ccanals are surrounded by medium-sized, sharply defined thimble-like
chambers, with their blind ends all turned towards the afferent canals and the sub-
dermal space. They are, indeed, surrounded by the loose trabecular framework, which
extends between them and the dermal membrane, and by the subdermal spaces, as well
as by the afferent canals which traverse the framework without being always sharply
defined from it. It is noteworthy that the chamber membrane is often so pene-
trated by the small parenchymal oxyhexasters that some rays of the latter project for
a greater or less distance into the lumen of the chamber. On the outer wall of the
chambers one frequently observes those groups of small cells which have been already
several times discussed.

2. Chonelasma hamatum, n. sp. (PI. XCI.),

In the locality in which the large plates of Chonelasma lamella were trawled (Station
170a, lat. 29° 45' S., long. 178° 11' W., in a depth of 630 fathoms, on volcanic mud),
there was also found the smaller plate figured in its natural size on PI. XCI. fig. 1.
This is bounded by an irregularly undulating surftice, and is only about 3 mm. in
thickness, while it is provided with a thin rounded margin. Apart from some slight
curvatures this plate also in general forms a flat expansion, and appears to have been
firmly attached by the narrow extremity.

The structure of the tolerably regular dictyonal framework agrees essentially with
that of Chonelasma lamella. The meshes are, however, narrower and are not markedly
wider in the middle layer of the plate than near the surfaces. The tuberculation on the
surface of the tolerably strong network of beams is not uniform throughout. In some
parts it is very pronounced, in others it is almost entirely absent, with the exception of
the projecting pegs. Small tubercled hexacts occur very regularly and are transversely
or obliquely soldered to the regularly formed framework of beams.

The alternating funnel-shaped canals, which traverse the plate, opening on the one side
and terminating blindly on the other, have a maximum width of 0"5 mm.

The dermal skeleton consists of hexacts, each of which has a short distal ray, provided
with a slightly knob-shaped thickening and densely beset with tubercles, while their
four cruciate, long, straight, transverse rays and the still larger proximal ray are smooth at
the base, but rough at the pointed extremities (PI. XCI. figs. 2, 5). In addition to these
dermal hexacts, scopulae of various forms occur, in which the stalk is provided Avith a
rough pointed or rounded internal terminal portion, while the expanded outer extremity
exhibits, close beneath the teeth, an annular thickening or four cruciately disposed
archings, or else gradually passes into the bases of the teeth. I have never been
able to determine more clearly than in this thickening beneath the teeth an intersection
of the principal canal by two short transverse canals disposed at right angles to it.
These transverse canals do not, however, pass into the teeth, but remain straight and

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

terminate in tlie cii-cular swelliug or iu one of the four tubercles, while the prolongation
of the principal canal is likewise straight, and terminates close beneath the teeth.
Hence I am of opinion that I can with certainty exclude the view that the teeth are to
be rewarded as curved principal rays, and maintain the opinion that they are terminal
rays of a greatly shortened radial principal. These terminal rays are as a rule rough and
straio-ht, and provided with terminal knobs. There are usually four, but three may also
occur. Exceptionally an S-shaped curvature occurs, as on PI. XCI. fig. 4.

The o-astralia are so essentially similar in form and position to the dermalia that no
special description is required. I may simply refer to PI. XCI. fig. 2.

In the parenchyma tolerably regular uncinates of various dimensions occur in great
numbers, but extend usually only through half the thickness of the plate. They are all
disposed at right angles to the surface. In the uncinates occurring near the dermal
surface aU the barbs have their points towards the gastral side, while in the uncinates
which lie closer to the gastral side, the points of the barbs are turned round and
directed towards the dermal surface. On both surfaces therefore the uncinates would
on gentle motion project outwards from the sponge body. The numerous small hexacts,
which are for the most part fused to the dictyonal framework, but partly also occur free
in the soft tissue, are provided with pointed tubercles, and each of the rays is obliquely
pointed at the extremity. Moreover many elegant discohexasters also occur in the
parenchyma, and exhibit various forms. Very frequently small rosettes, such as are
figured on PL XCI. fig. 6, occur, and each of the moderately short basal rays bears four,
five, or six terminals, which are curved in an S-like manner like petals, and knobbed
or provided with terminal discs. More rarely forms occur in which the basal rays are
somewhat longer and more slender, and are united in a spherical central knot. These
bear on their outer extremities from three to four very fine long terminal rays with
terminal discs. "Whether the rosettes with long rough basal rays, which are represented
on PI. XCI. fig. 7, and which I found here and there in preparations of Chonelasma
hamatum really belong to this species and are not merely intrusions, I have lately seen
reason to doubt. In any case, however, should they really belong to this sponge, they
are of very rare occurrence.

The structure of the soft parts difi'ers from that described in Chonelasma lamella at
most in the greater simplicity of the straight afierent and efferent canals which traverse
the plate at right angles (PI. XCI. fig. 2).

3. Chonelasma dcederleinii, n. sp. (PI. XC. figs. 1-7).

Among the HexactineUida which were collected by Dr. Doderlein in the Sagami Bay
(Japan) there are some irregularly bent plates from 2 to 3 mm. in thickness which are pre-
served in the dry state. These are provided with a thin undulating margin, are from 2 to
3 cm. in breadth, from 4 to 5 cm. in height, and without doubt were attached to a solid

REPORT ON THE HEXACTINELLIDA. 325

substratum by means of the narrow inferior surface. There is also a fragment of similar
structure preserved in spirit, and this has been especially serviceable for microscopic
examination. Although the macroscopic and microscopic structure of this fragment
essentially agree with those oi Chonelasma hamatum, there are several points of difference,
especially in regard to the form of the isolated spicules, and these differences appear to
me to be sufficient for the erection of a distinct species. I have named this species in
honour of Dr. Doderlein who gave it over to me for examination.

In the dictyonal framework the two funnel-shaped systems of canals, which open
alternately on the two lateral surfaces, are narrower than in Chonelasmcc hamatum, but
this might be associated with the slight thickness of the entire plate. The beams of the
network exhibit the more or less tubercled surfaces noted in the above species. There is
further great regularity in the square or rectangular meshes of the middle layer, while the
meshes near the two surfaces appear more irregular and rounded.

The strands of beams which run parallel to both sides, and cross each other
approximately at right angles, are for the most part perpendicular and parallel to the
upper boundary. When this therefore is somewhat curved, as in the specimen represented
in PI. XC. fig. 1, a more circular and radial disposition of the beam-strands results.

The hexacts of the dermal and gastral skeleton are provided with a short, broad, almost
oval, brush or fir-tree-like freely projecting ray, while the four moderately long trans-
verse rays and the strong conical proximal radial (which varies in length and penetrates
into the parenchyma), are smooth at their base and rough towards the extremities (PL XC.
figs. 2, 3). The scopulae which occur both in the dermal and gastral skeleton all have a
shaft which runs out to a fine point, but vary greatly in the form of the teeth and of that
expanded portion of the shaft which bears them. The latter sometimes exhibits a simple
thickening, or in some cases an annular swelling or four cruciately disposed knobs, and
into these the short transverse rudiment of the axial canal extends.

The teeth are almost always in fours and appear either as rods thickly beset with
barbs provided with slender terminal knobs, and running almost parallel to one another ;
or while remaining similar in form they may diverge slightly outwards (PL XC. fig.
6) ; or they are slender and show a tendency to become somewhat curved in an S-bke
manner (PL XC fig. 5).

In addition to the uncinates of variable strength which I have found only in the
middle and dermal portions of the parenchyma, the latter contains discohexasters of
variable strength, and these bear on the extremity of every short basal ray three or four
long, thin, diverging, somewhat curved terminal rays with pronged terminal discs
(PL XC. fig. 7). Sometimes the number of terminal rays on one or other principal ray is
reduced, and I even observed some discohexasters with irregular undulating rays.

The structure of the soft parts does not exhibit any essential deviation from that
already described in Chonelasma lamella and Chonelasma hamatum (PL XC. fig. 2).

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

4. Chonelasma calyx, n. sp. (PI. LXXXIX.).

In contrast to the above described simple plate-like forms, the species now to be
noted, which I have established from three dried and partly macerated specimens
obtained by Dr. Doderlein in Enoshima, has the form of a cup several decimeters high,
with nipple or glove-finger-like radial protuberances from the wall, which is 5 mm. in
thickness. The majority of these parietal protuberances, which have the thickness of one's
finger or thumb, exhibit a terminal or subterminal circular aperture from 3 to 5 mm. in
diameter, and are continued as thinner tubular or half channeled hollow irregular
cyhndrical processes, which either extend obliquely downwards and reach the firm sub-
stratum, or unite with the adjoining processes of neighbouring diverticula (PI. LXXXIX.
figs. 1, 2). The whole wall is traversed by canals which open alternately on the inner and
outer side, but the dried specimens at my disposal do not show the arrangement with
equal clearness throughout.

The dictyonal framework, which consists partly of perfectly smooth beams and partly
of beams sparsely covered with tubercles, exhibits a thicker and less regular meshwork in
the neighbourhood of the outer and inner bounding surfaces than in the middle layer,
where the meshes are large and more uniformly rectangular. The dermal skeleton consists
of hexacts, each of which has a terminally corroded, knobbed, and fir-cone-like distal ray
which is beset with pointed prongs, while the long conical proximal ray, and the four
moderately long transversals, are either quite smooth or beset on their outer extremities
with small tubercles. In addition to the hexacts, the very numerous radially projecting
scopulge usually bear six, more rarely five or four, barbed, more or less markedly diverging,
knobbed terminal rays, or a cap-like swelling on the shaft which terminates in a rough
point (PI. LXXXIX. fig. 6). The spicules of the gastral skeleton exhibit a similar
structure, but the freely projecting (proximal) ray of the hexacts is often somewhat more
slender than in the dermal hexacts. The scopulse, moreover, vary considerably in the
number of their barbed and usually markedly knobbed terminal rays.

The parenchymal uncinates are remarkable for their strength. The greatest swelling
always occurs in the anterior third dermal portion. Among the smaller loose parenchy-
malia numerous discohexasters and oxyhexasters must be noted. The number of the
moderately/ long terminal rays varies, as has already been described and figured in
Chonelasma lamella. The same small discohexacts appear as in the above.

Chonelasma sp. (PL XC. figs. 8-11).

Besides the largest specimens figured in PI. LXXXVII. figs. 1 and 2, fragments of
macerated dictyonal framework belonging to a Chonelasma form were found at various
stations on the Challenger Expedition. I have not been aljle to refer these with any

REPORT ON THE HEXACTINELLIDA. 327

certainty to one or other of the above described forms, nor are the characteristics
sufficiently distinctive to permit of the erection of a definite species. In some of
the fragments, however, several features are very distinctly developed, to some of which
it seems worth to refer. In PL XC. figs. 9, 10, and 11, I have given figures from
photographs of a plate-like specimen dredged near the Bermuda Islands (Station 56,
lat. 32° 8' 45" N., long. 64° 59' 35" W.), from a depth of 1075 fathoms and a coral
mud ground. One surface, probably the dermal, has unfortunately been very much rubbed
(fig. 11), so that on surface view only the deeper layers of the framework are to be
seen. The surface represented in fig. 10 seems to me to be gastral, so that the very
regularly formed layer of the framework, displayed with its distinctly rectangular meshes
in fig. 11, lies close beneath the gastral surface. A section at right angles through
the best preserved portion of the plate (fig. 9), displays very distinctly the alternation of
the straight funnel-shaped spaces of the afierent and efferent canals.

Among numerous small macerated fragments dredged near the Penguin Islands
(Station 148a, lat. 46° 53' S., long. 51° 52' W.), from a depth of 550 fathoms and a bottom
of hard ground, gravel, and shells, besides the larger plates represented in PI. LXXXVII.
figs. 1 and 2, there was another which exhibited a very characteristic thickness. In the
cross section (PI. XC. fig. 8), as in the above instance (PL XC. fig. 9), the alternation
of funnel-shaped canal spaces can be detected. I am inclined to believe that the
Scleroplcgma herculeum (from Santa Cruz), referred to by Oscar Schmidt in his Spongien
des Meerbusens von Mexico (p. 57), is a similar form.

In the neighbourhood of St. Thomas (Station 23b, lat. 18° 28' N., long. 63° 35' W.),
from a depth of 590 fathoms, and a Pteropod ooze bottom, other small Chonela^ma frag-
ments were obtained, as also on the coast of Portugal at localities not recorded.

Family IV. T r e t o d i c t y i d ^, F. E. Schulze (Pis. XCII.-XCVIIL).

With irregularly arranged afferent and efi"erent canals, which penetrate the body-wall,
and especially the more or less thick dictyonal framework, not transversely but obliquely,
or in a longitudinal direction, or even in a coiled course.

Genus 1. Hexactinella, Carter (Pis. XCIII.-XCVL).

1885. H. T. Carter, Ann. and Mag. Nat. Hist, ser. 5, vol. xv. p. 387.

History. — After I had completed my investigation of the Challenger Hexactinellida,
and had given names to the new forms which I had determined, I discovered that the form
designated by Carter Hexactinella ventilabrum} was identical with the species to which

'.1(H!. and Mag. Nat. Hist., ser. 5. vol. xv. p. 387, 1885.

H.

328 THE VOYAGE OF H.M.S. CHALLENGED.

I had given the title Tretodiotyum cyathus. I do not hesitate, therefore, to abandon my
generic and specific designation for this form, and to adopt Carter's title, which has the
priority. Thus the generic title Tretodictyum is wholly given up, and the two other
species which belong to the same genus are no longer called Tretodictyum tuhidosum and
Tretodictyum latum, but Hexactinella tuhulosa and Hexactinella lata.

The specimen figured by Carter, unfortunately not in its entirety, was obtained from
Misaki, Japan, at the extrance of the Bay of Tokio. According to his description, it
closely resembles the Phakellia ventildbrum, Bwk., figured by Bowerbank in his Mono-
graph of British Sponges.^ " The surface is on both sides even and uniform ; uniformly
scattered over internally with circular apertures about -^ inch in diameter, and about the
same distance apart, and externall)' with a dermal, quadrilateral, spicular reticulation."
" Wall about \ of an inch thick, composed of two layers, viz., one on each side of
an irregular central plane of condensed tissue, each layer consisting of plumose fibre
curving upwards and outwards florally from the centi'al plane of condensed tissue,
strengthened by transverse fibres in their course."

Isolated spicules are represented, according to Carter, by the following types : — (1)
dermal pentacts ; (2) uncinates, called barbulte by Carter ; (3) thin oxydiacts ; (4)
scopulse with two, three, or four slightly divergent teeth ; (5) simple or spinose oxy-
hexacts ; (6) oxyhexasters and discohexasters of diff"erent kinds.

Character of the Genus. — The wall of each of the cup-shaped or tubular specimens
is traversed by canals, which are not exclusively disposed at right angles to the bounding
surface.

The dictyonal framework is principally composed of radial, longitudinal, straight
or slightly bent fibrous reticulate 2:)lates, about 1"5 mm. in breadth. These are separated
from one another by spaces of similar form and breadth, but are at the same time bound
together laterally by numerous transverse beams. A more irregularly developed fibrous
network with round openings extends in some species over the outer, in others over the
inner (gastral) surface of the dictyonal framework, and thus conceals either on the
outside or inside the above mentioned longitudinally directed radial plates and their
cleft-like interspaces.

The dermal and gastral skeleton consists of pentacts or hexacts, and numerous scopulae
of various forms. In addition to delicate uncinates, numerous oxyhexasters, disco-
hexasters, and more rarely small oxyhexacts and discohexacts appear in the parenchyma.

1. Heasactinella tuhdosa, n. sp. (PL XCIIL).

Simple or dichotomously branched, sometimes also laterally anastomosing tubes about
1 cm. in diameter, with a wall varying from 2 to 3 mm. in thickness, rise from a flat

' Vol. iii. pi. xxii.

REPORT ON THE HEXACTINELLIDA. 329

base, which may either have grown upon a solid body, or may have formed a flat ex[tan-
sion on somewhat loose clayey ground.

The largest of the specimens, bought by Dr. Doderlein in Enoshima in a dried con-
dition, attains a height of 10 cm. On the outer extremity of the tubes a simple smootli-
margined oscular opening occurs. Through the dermal and gastral sieve-like skin the
canals, which are from 1 to 2 mm. in width, may be recognised.

In the dictyonal framework, which has been isolated by maceration, the gastral
margins of the longitudinally disposed dictyonal plates project freely on the inner surface,
while an irregoilar network of beams with round openings from 1 to 2 mm. in diameter
extends over the exterior.

The dermal skeleton is formed of strong pentacts, with numerous inequalities over
the entire surface. A round knob or peg represents the atrophied distal (sixth) ray.
The long proximal ray and the four tangentials generally terminate in sharp points.
Scopulse also occur, each provided with a moderately long, terminally rough and pointed
stalk, and Ijearing on the outer simple conical expanded extremity of the latter four
rough, slightly capitate or entirely unknobbed, parallel or slightly diverging, terminal
rays (PL XCIII. fig. 4). In some specimens numerous fine spicules, which run out to a
point at both ends, occur in groups close to the proximal ray of the pentacts, and
project somewhat above the skin.

In the gastral skeleton, curiously enough, I have found no pentacts or hexacts,
although numerous scopulse similar to the dermal forms were present in the familiar
radial position (PL XCTII. figs. 2, 5).

The loose parenchymalia consist of very thin and short uncinates (PL XCIII. fig. 3) of
simple, slender, or somewhat rough hexacts (PL XCIII.) and oxyhexasters, each of which
bears two diverging, short, and strong straight terminal rays on each of the moderately
long basal principals.

2. Hexactinella lata, n. sp. (Pis. XCIV., XCV.).

Dichotomously branched tubes, which tend to anastomose, varying from 1 to 4 cm. in
diameter, expanding superiorly in funnel-like form. They are also frequently crowded
together laterally, and appear to be pitted or swollen out here and there in an irregular
manner. The wall of the tube has a thickness of 2 to 2 "5 mm. Through it one can
clearly recognise the radial and also the generally longitudinal, frequently undulating or
meandering plates, with the intervening spaces, while on the inner sitrface only ii-re-
gular round spaces are to be observed through the gastral skin (PL XCIV. fig. 1). In
completely macerated specimens it may be seen that, in the parallel longitudinal radial
plates of the dictyonal framework, the longitudinal fibrous bands exhibit an external

(ZOOL. CHALL. EXP. — PART LIII. — 1887.) Ggg 42

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

curvature towards tlie dermal surface. The lateral union of these longitudinal radial
plates, which are separated by spaces of uniform breadth, is effected by means of trans-
verse bridges of the same dictyonal framework. These extend uniformly on the gastral
side, where they are perforated only here and there by long oval openings, while similar
bridges arching over the interradial spaces on the dermal side have a more limited longi-
tudinal extension. In this manner a canal system is formed, which is, for the most ])art,
hmoitudinally disposed, and is here and there dichotomously branched. It opens towards
the outside by somewhat long longitudinal clefts, towards the inside by round or oval pores ;
and since the inner openings are frequently arched over by the external connecting
bridges of the adjoining radial plates, it often results that a transverse section of a com-
plete tube exhibits the appearance of a folded plate (PI. XCIV. fig. 3 ; PI. XCV. fig. 1).
The beams of the fibrous framework bear small superficial knobs, which are either
irregularly scattered or arranged in incomplete transverse rows (PI. XCV. fig. 2). The
nodes of intersection are somewhat thickened and knobbed here and there, especially in
the outer layer of the framework.

On the outer (dermal) surface the dermal membrane extends quite uninterruptedly
over all the clefts and openings, while on the inner surface the gastral membrane passes
into the excurrent openings of the larger efferent canals, which in most cases traverse the
wall vevy obliquely.

The strong dermal pentacts almost invariably exhibit a button-, peg-, or knob-like
rudiment of a sixth distal ray. The proximal ray, which varies in length, and the four
tolerably long, and in most cases slightly curved tangential rays, terminate in points
and are beset with small conical knobs all over in the case of the larger spicules, but
only on the extremities of the, rays in the smaller (younger) forms. Strong scopulaj'
also occur, in which the four parallel or slightly diverging, thick or slender, rough
knobbed terminal rays usually arise just above an annular quadri-tuberculate thickening
of the stalk (PI. XCIV. fig. 8).

Finely pointed slim oxydiacts extend here and there in bundles, close to the proximal
ray of many dermal pentacts, in the dermal membrane, and even above the latter.

The gastral skeleton almost completely resembles the dermal. A distinction could
only be found in the fact that the pentacts are less strong, and their projecting rudiment
of the sixth ray is somewhat longer, so as almost to warrant the term hexact. The
gastral scopulae agree completely with the dermal.

The parenchjnoia contains delicate uncinates, and numerous isolated, thin-pointed
spicules, which are united in somewhat loose bundles. Besides small hexacts of
various dimensions and provided with small terminal knobs (PI. XCV. figs. 3, 4),
sphjerohexasters occur. These have a variable number of curved terminal rays, which
are in most cases only of medium size, and are convex externally (PI. XCIV. figs. 6, 7,

1 The scopulse figured on PI. XCIV. figs. 5, 9, do not belong to this form. They are casual intruders.

REPORT ON THE HEXACTINELLIDA. 331

i:^). Oxyhexasters, with long pointed terminal rays, rarely occur' (PI. XCIV. fig. 10).
The structure of the .soft parts does not essentially diifer from that observed in the Euretidae.
The few specimens of this form were trawled near the Little Ki island (Station 192,
lat. 5° 49' 15" S., long 132° 14' 15" E.), from a depth of 140 fathoms, and a blue
mud ground.

3. Hexactinella ventilabrum, Carter (PI. XCVI.).

The elegant simple cup represented on PI. XCVI. figs. 1, 2 (measuring 8 cm. in
height and 12'5 in greatest breadth), belongs to the collection of Hexactinellida which
was made by Dr. Doderlein in Enoshima, Japan. From a base about the size of a half-
crown, a strong, laterally-compressed stalk, 3 cm. in breadth and 1 cm. in thickness arises,
and this is continued into the somewhat bent body, which is compressed in the same
direction. This somewhat resembles a boat, and its cavity exhibits the openings of a
row of twelve branching canals of approximately ecjual width. The circular openings
measure from 2 to 4 mm. in diameter (PI. XCVI. fig. 2).

Since the dermal and gastral skin in this dried specimen are only preserved here and
there in small remnants, the coarser structure of the dictyonal framework can be
recognised without further trouble. As in Hexactinella lata, the free outer margins of
the radially and longitudinally disposed, reticulate, fibrous plates (1 mm. in breadth),
distinctly project on the external dermal surface, while on the gastral surface of the cup
they are covered by the evenly expanded fibrous network. Instead, therefore, of clefts
I mm. in breadth, only round pores of equal breadth occur on the inner surface. The
meshes of the dictyonal network of fibres appear in many places to be almost regularly
square, although the longitudinal fibrous strands frequently run not quite parallel to
the bounding surface, but in arched bundles from the middle layer of the wall, and
diverge slightly towards the free outer and inner surfaces, so that their extremities
project freely on the surfaces. The beams of the fibrous network are irregularly, aud
more or less abundantly beset with low knobs. The dermal skeleton includes strong
pentacts like those in the skin of Hexactinella tuhulosa and Hexactinella lata, also
strong, four-toothed scopulse, and numerous fine spicules which are pointed at both ends.
Constituent elements of exactly the same character are to be found in the gastral skeleton.

The parenchyma contains minute, rough rods, with pointed extremities, which corre-
spond somewhat to the uncinates. There are also oxyhexasters, with tolerably long,
diverging, terminal rays (PI. XCVI. fig. 8), further, discohexasters with long basal
])rincipals, and numerous short (externally slightly convex) terminal rays (PI. XCVI.
fig. 5), and finally, discohexasters with short basal princij)als, and a few (three to
five) straight or S-shaped, tolerably long terminal rays (PI. XCVI. fig. 4).

1 The spicules figured on PL XCIV. figs. 11, 12, seem to have come in accidentally, and were not found on careful

re-examin:ition.

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

Geuus 3. Cyrtavlon, F. E. Schulze (PI. XCIL).

Volvulina,^ 0. Schmidt.

1880. O. Schmidt, Spongien des Meerbusens von Mexico, p. 58.

History. — Among the sponges of the Gulf of Mexico, Oscar Schmidt^ found some
He.xactinellida of cylindrical, globular, or beaker-like form, which were procured from
Morrolight, Barbados, and St. Vincent, from a depth varying from 100 to 300 fathoms.
These he has embraced under the specific name of Volvulina sigsheei. Their
siliceous networks are united in an irregular lattice-work by strands, varying from ^ to 3
mm. in thickness. Between these strands there are pores and passages, the outer
openings of which appear to be covered by a membrane. A true gastral cavity was
not observed, but, on the other hand, a pitting here and there was regarded as
pseudogaMral.

The dictyonal framework exhibits knobbed and also smooth beams. These unite in
some species in round tubercled nodes of intersection, while thickenings of this kind
are entirely absent at the intersections in other species.

In the outer skin hexacts or pentacts occur, and these are arranged in a regular
network. Moreover, among free spicules there are (l) slender thorny spicules (uncinates),
which, in the specimens procured from St. Vincent, always exhibited a central
circular swelling; (2) club-shaped brooms (scopula3) ; (3) hexacts with large finely-
toothed hooks on the ends of the rays ; (4) umbel-like rosettes (discohexasters), the
single ray in each of which is prolonged without forking, and terminates in a point, but with
a little thickening before the extremity.

In the fragment figured in PL XCIL fig. 9, which was very kindly intrusted to me by
Professor Oscar Schmidt for comparative examination, I was able to corroborate his
important remarks on the striking want of uniformity in the dictyonal framework. I
find on the outer surface greatly thickened beams, which are richly beset with large wart-
like knobs ; in the interior of the framework the beams in some places are ornamented
merely with small pointed knobs, while in others, on the contrary, they are quite smooth.
While in some species densely tuberculate nodes of intersection are present, these are
completely absent in others.

I have indicated above the general characters which may be inferred from the
specific features noted 1)y Oscar Schmidt, but the generic diagnosis of Volvulina, in
contrast to that of some allied genera, may be formulated in the following manner : —

Cyrtaulon, F. E. Schulze (= Volvulina, 0. Schmidt).

Cylindrical, conical, or beaker-shaped forms, in which the dictyonal framework con-

' The name " Volvulina^' has been applied to a Gasteropod since 1865.
'^ O. Scliniidt, Spongien des Meerbusens von Mexico, p. 58, 1880.

REPORT ON THE HEXACTINELLIDA. 333

sists of irregularly united fibres, reticulated beams, with interposed irregular spaces and
passages, the outer orifices of which are covered by a dermal membrane. In addition
to the regularly disposed dermal jjentacts or hexacts there are scopul* with knobbed
branches. The parenchyma contains, besides uncinates, scopula-like spicules, each of
which is provided with one prolonged principal ray, which runs out to a fine point.

1. Cyrtaulon sigsbeei, 0. Schmidt (PI. XCII. fig. 9).

Since I am inclined to believe that the description which (J. Schmidt {loc. cit., pp.
58, 59) has given of his Volvulina sigsbeei was based on specimens of difi"erent species,
I must, in regard to the minute structure, restrict myself in the first instance to that
specimen of which only a fragment is available, as represented from a photograph in
PI. XCII. fig. 9. For an account of the more macroscopic characters I must simply refer
to 0. Schmidt's results.

We have here to deal with very variously shaped, often goblet-like specimens, whose
wtill consists of an irregular feltwork, with anastomosing cavities both on the external
and on the internal bounding surface, covered over by a porous skin.

The dictyonal framework consists of tuberculate beams, in which the nodes of inter-
section are here and there, and especially near the surface of the body, thickened and
beset with wart-like elevations.

The parenchyma contains uncinates with central thickened nodes, and scopula-like
spicules with several thin prongs radiating out from the terminal knob of the stalk, and
bearing marginally-toothed terminal discs. The dermal skeleton consists of pentacts
and hexacts, and also of scopulae with knobbed prongs.

As to localities, 0. Schmidt mentions (1) Barbados, 100 fathoms; (2) lat. 32° 9' N.,
long. 82° 23' W., 158 fathoms ; (3) Morrolight, 292 fathoms ; (4) St. Vincent, 124 fathoms.

2. Cyrtaulon solutus, n. sp. (PI. XCII. figs. 1-8).

Among the numerous Hexactinellida which were obtained by the Challenger at Station
192 (lat. 5° 49' 15" S., long. 132° 14' 15" E.), near Little Ki Island, from a depth of 140
fathoms, on blue mud ground, there is one dried specimen which grew in the tubular cavity
of a piece of limestone, and which had the form of a cylinder 5 cm. in length and 2 cm.
in breadth. As may be observed in the sketch given on PI. XCII. fig. 1, the dictyonal
framework passing through the body consists of irregularly united flat or rounded beams
from 1 to 2 mm. in thickness, which surround interspaces and passages from 3 to 4 mm.
in breadth. As may be clearly recognised in some parts of the surface which have been
specially well protected by the surrounding stone, the whole was covered by a cylindrical
mantle-like veil, which exhibits in the dried state a square-meshed network with large

334 THE VOYAGE OF H.M.S. CHALLENGER. •

and siiKiU spaces, and extended continuously over the spaces aud openings. No wide
gastral cavity or oscular opening is to be recognised.

The microscopic examination of the dictyonal framework reveals a somewhat irregular
network of l^eams which are beset with moderately large smooth tubercles of irregular
number and arrangement. The beams never exhibit a spherical thickening of the nodes
of intersection. In dead portions the axial canals of the dictyonal hexacts are here and
there greatly enlarged, and accordingly well marked. The dermal skeleton consists
chiefly of moderately large, smooth, or only terminally somewhat tubercled pentacts,
which form by their apposition a beautiful square meshed lattice-work. In addition to
this, bundles of fine spicules pointed at both ends jut out, and scopulse of different
forms occur, but especially forms provided with rough pointed shafts and four strong,
almost or absolutely parallel teeth, densely beset with barbs. These teeth spring
from a simple conical expansion of the shaft, and possess no knob-like terminal swellings
(PL XCII. fig. 5).

On the surface of the large inner strands and plates of the dictyonal fibrous frame-
work there are no pentacts, but only scopulse, and unpointed fine spicules which also
appear to constitute the skeleton of the gastral membrane.

In the parenchyma, and between the reticulated beams of the dictyonal framework,
there are uncinates, which usually exhibit a slight curvature, and further, those spicules
which are characteristic of the genus Cyrtaulon and were first described by Oscar
Schmidt, who regarded them as discohexasters with a greatly prolonged ray. These
forms exhibit a simple shaft, which runs out to a fine point, and is provided on the
pointed terminal portion with rough knobs ; on the other extremity with a knob- or
ball-like thickening, wliich bears a somewhat large number of thin terminal rays with
terminal discs. The thickening which occurs close beneath the pointed extremity of the
arrow-like shaft in Cyytaulon ( Volvulina) sigsbeei is here entirely wanting. This peculiar
form of spicule may be best regarded as a modification of a hexaster, but I have not found
any indications of the basal parts of principal rays, except the long shaft. All the thin
terminal rays originate directly from the knobbed terminal thickening of that shaft.
In addition to the scopulte with four thick rough teeth, already mentioned in the dermal
skeleton, other forms with four thin smooth teeth and simple terminal knobs apparently
occur in the parenchyma, in addition to those which bear six thin smooth similar ter-
minal rays with terminal knobs or discs (PL XCII. fig. 6). If one imagines the numlier of
these teeth to be increased and their position more irregularly radial, one can also under-
stand the unusual form of parenchymalia. I have not found any regular hexacts in
Cyrtaulon sohdus.

REPORT ON THE HEXACTINELLIDA. 335

Genus 4. Fieldingia, Sav. Kent (PI. XCVIL).

1870. Saville Kent, Ann. and Mag. Nat Hist., ser. 4, vol. vi. p. 219.
1876. Marshall, Zeitschr. f. wi.'^s. Zool., Bd. xxvii. \\ 124.

History. — On a specimen of Lopfiohelia prolifoxi, vnr. antJiophyllites, which was
dredged ten miles off the sliore of Cezimbra, Portugal, from a depth of 500 fathoms,
Saville Kent found' in 1870 a Hexactinellid, distinguished by a pustulate rind of fine
recticulatcd laminiB and numerous spherical internal bodies, varying in diameter from
"SIT to j^ of an inch. He named it Fieldingia in honour of Mr. Edward Fielding,
adding the specific name of lagettoides, in reference to the delicate reticulate laminae
associated with the cortex, which are so strongly suggestive of the internal lace-like
layers in the bark of the lace-bark tree. His brief diagnosis of the new form was : —
" Sponge adherent, consisting of a cortex of irregular reticulated spicula, having on its
interior surface numerous reticulated laminae of extremely delicate consistence. Common
cavity of the sponge containing numerous spherical aggregations of spicular reticulations ;
these invested and brought into relation with the cortex by loose reticulated fibres of
coarser structure, having a general hexradiate arrangement ; these fibres cylindrical, and
to a considerable extent minutely and erectly spined ; frequently attached to them very
diminutive spicula of the ' rectangulated hexradiate ' type, these also minutely and
erectly spined. Nutritive and exhalent functions most probably performed through the
general reticulations of the cortex."

Marshall, who in 1876 ranked' Fieldingia along with Aphrocallistes as "aberrant
forms " among his Pleionacida?. suggested that the form was probably a young
A'plirocallistes.

Fieldingia lagettoides, Sav. Kent (PI. XCVIL).

The drawing on PI. XCVII. fig. 1 represents in its natural size one of the two
dried specimens which were trawled by the Challenger oft" Little Ki Island (Station 192,
lat. 5° 49' 15" S., long 132' 14' 15" E.), in 140 fathoms, on blue mud ground. The
surface of the knob-Hke specimens, which are as large as a walnut, consists in the
perfectly intact regions of a strong external rind, associated with adherent granules of
sand and other foreign l^odies. The rind appears here and there somewhat blistered and
almost spongy, and consists of several lamellae of a very narrow meshed and irregular
network. An irregular framework of strong siliceous beams with partly square meshes
(from 1 to 3 mm. in width) extends from the rind across the lumen. This framework of
beams exhibits numerous thick spherical knots about 1 mm. in diameter, and with tolcr,il)ly

■ Ann. and Mag. Nat. Hift., ser. 4, vol. vi. p. 219.
' Zcitschr.f. xoiss. Zool., Bil. x.wii. p. 124.

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

uniform distribution. These occur at intervals of 2 to 3 mm., and are not otherwise
remarkable, except as local condensations of the network of beams. Both on the beams of
the wide meshed part of the framework, and on the outer surface of the knots, small
simple hexacts occur fused together in variable numbers. All the beams of the frame-
work and the small fused hexacts are more or less richly beset with small tubercles.

From what has been said there can be no doubt that the specimens in question
belong to the species Ficldingia lagettoides, Sav. Kent, which Saville Kent found on a
specimen of Loj'thoheUa at Cezimbra, Portugal, and which he has figured and described.
But it is to be regretted that in many parts of the present specimens the remarkable
blistered rind, which is of several layers, is insufficiently preserved for accurate
examination. From the fragments, however, it may be seen that we have to deal with
lattice-like lamellfe, which exhibit the crossed tangential rays of numerous pentacts, in
which the fifth ray projects inwards. Between these pentact rays, which are remarkable
for their distinct axial canals, there extends a more or less well-developed, in part very
narrow-meshed, network of connecting beams without central canals, and provided with
rounded meshes. This is similar to what is found in the basal plate of many fixed
Hexactinellida, and in the regions of contact between these and sofid foreign bodies.
Since the free outer surface is covered with sand and other extrinsic elements, the
supposition is confirmed that we have here to deal with a dermal skeleton altered by an
admixture of foreion bodies.

Among isolated spicules several apparently typical forms occur. It is not possible to
determine with certainty their normal arrangement or stratification. In addition to
various strong uncinates (PI. XCVII. fig. 9), simple straight smooth diacts frequently
occur. These exhibit a central swelling or four cruciately disposed knobs, and their
two extremities form a point which arises by a process of jjitting (PI. XCVII fig. 8).
In addition to simple smooth and delicate hexacts (PI. XCVII. fig. 7), oxyhexasters
occur, which have several or all of their rays forked. The principal ray usually remains
very short (PI. XCVII. fig. 6), and sometimes, indeed, it is so much abbreviated that the
terminal rays appear to rise almost directly from the node of intersection, and in this
way a simple star with ten or more rays results (PL XCVII. fig. 3). Discohexasters
with two or more fine long terminals on each of the short principal rays, and with
marginally toothed somewhat convex terminal discs, are tolerably frequent (PL XCVII.
fig. 4).

The scopulse, which probably belong to the dermal skeleton, exhibit four strong
almost parallel teeth, which are beset with barbs, and are slightly knobbed, or terminate
without any swelling. The teeth spring from the conically thickened extremity of the
rough and pointed shaft. Less frequently the teeth diverge in the manner represented
on PL XCVII. fig. 5.

REPORT ON THE HEXACTINELLIDA. 337

Genus 5. Sclerothamnus, Marshall (PI. XCVIIL).

1875. Marshall, Zeitscbr. f. wiss. ZooL, Bd. x.vv., Suppl., p. 171.

1876. J. Murie, Trans. Linn. Soc. Lond. (ZooL), ser. 2, vol. i.
187G. Carter, Ibid., Appendix.

History. — The term Sclerothamnus clausii was applied by Marshall ' to a bushy-
branched Hexactinellid obtained from an unknown source. It measured 50 cm. in
height, while the cylindrical branches, which were repeatedly forked, had a diameter of
3 "5 cm. towards the extremities, and of 13 cm. at the basal ends. In the tolerably
uniform framework of siliceous beams, which consists of fused hexradiate sj)icules with
tubercled external surfaces, Marshall thought he perceived a continuous connection
between the axial canals, which exactly met one another in the fusion of adjacent hex-
radiate spicules, and thus remained in open communication. The broad (5 mm. in
diameter) and narrow (1 mm. in diameter) canals, which traverse the branches of the
sponge, often anastomose with each other, and open outwards by orifices of variable size.
Many of these excurrent passages, moreover, possess a fringe in the form of a freely pro-
jecting cuff, but on the whole the canal system seems to be indistinctly and irregularly
developed in the dense tissue.

Among free spicules Marshall found (1) large-knobbed hexradiate forms, which fuse to
form the continuous framework ; (2) very small fine regular hexradiate spicules ; (3)
broom-forks with five clubs, beset with delicate warts on each of the expanded portions,
while their stalk, which is covered with similar warts down to a slisfht swellinsf, ter-
minates in a short point; (4) longer broom-forks (Besengabeln) with four long clubs, in
which the stalk, like the spicular shaft, is beset with fine recurved booklets.

In the following year, 1876, there appeared in the Transactions of the Linnsean
Society (Zoology), ser. 2, vol. i., a detailed paper illustrated by numerous good figures,
On Steere's Sponge, a New Genus of the Hexactinelbd Group of the Spongidse, by James
Murie. This bushy branched Hexactinellid was procured by natives from a depth
varpng from 80 to 100 fathoms in the neighbourhood of the Philippines, " between the
strip-like and parallel islands of Negros and Cebu." In its dried state it measured about
80 cm. in height, and exhibited branches as thick as one's finger. The form is accurately
described under the name of Dcndrospongia steerei. Professor Steere had bought this
beautiful specimen during his stay in the Philippines from fishermen " on the opposite
eastern side of the island of Cebu, where no sponges are said hitherto to have been got
from the sound or channel above-mentioned." Murie reports ^ — " One day, while in the
forest, Professor Steere was suddenly apprised of the arrival of the Challenger Expedi-
tion by a note from Professor AVyville Thomson, who had heard of him and his where-
abouts. He hastened on board, passing, as he assures me, a most agreeable day in the

' Zeitschr.f. vnss. ZooL, Bd. xxv., Suppl., p. 171, 1875.

' Loc. cit., p. 221.

(zOOL. CHALL. EXP. — PAKT LHI. — 1887.) -Ggg 43

338 THE VOYAGE OF H.M.S. CHALLENGED.

company of our countrymen. He doubts if the English naturalists obtained any other
specimen of this sponge [Dendrospongia) , though he told them of his one, and where
and how acquired." In a note Murie made the following addition : — " Since reading this
communication, Sir Wyville Thomson's return has enabled me to show him the accom-
panying plates ; the species he does not recognise as being among their collections,
although it is possible a less perfect specimen or fragmentary portions may be found
among the stores of material as yet imperfectly worked out."

According to Murie's description and figures, " The general surface exhibits several
broad but very shallow impressions or concavities which traverse the stem obliquely.
The distinguishing feature of the branches is a series of tufts or rosettes, so continuous
and interwoven as in the main to present a whorl running successively round from base
to apex, composed of a bunch of long parallel placed spicules, which issue from the axis
of the branch at an oblique angle, and slightly spread out at their free ends. The spiral
hollow between the frill shows a delicate gossamer lacework composed of minute spicules,
forming a rectangular chequer."

The principal fibrous bands of the network of beams, which is composed of fused
hexradiate spicules, run, according to Murie's representation, in a longitudinal direction
in the axial part of the branches, but bend laterally in a bow-shaped curve towards the
exterior, and terminate freely in the more radially directed obliquely projecting fibres of
the spiral frill. All were covered with short mucronate spines.

Among the spicules which lie isolated Murie described (l) long acerate, fusiform,
inequilateral sj)icules of two sizes, large and small, both covered with spines, all sloping in
the same direction ; (2) two forms of scopuline spicules, the larger with a straight shaft,
and with microspined and indistinctly capitate terminal rays, from two to four in number,
the smaller with two to four rays opposite to one another, and expanded laterally like
petals.

In regard to the position of these scopuline spicules, Murie' records Carter's conjecture
— illustrated by a woodcut — that they may have lain parallel to the surface in the dermal
layer, and may have been crossed in such a way that square meshes were formed. He
found (3) two forms of rosettes, of which the smaller bore six straight, smooth rays rising
at right angles from the centre, and each terminating in a little discoid swelling bearing
four to eight rays spreading in what Carter terms Vijieur-dc-Us. Each ray terminates in a
swelling which is expanded into a circular convex head, bordered by four opposite and
recurved spines. The larger rosettes are very rare, and the globular still more so. Each
of the six short, stout rays terminates in a quadrangular swelling, which bears several long,
straight rays, ending in a quadrangular, or, more frequently, pentangular cap or head,
with free convex surface, but jjrovided with recurved spines at the angles of the opposite
surface. He also notes (4) single hexradiate dermal spicules, which form by the over-

' Loc. cit, p. 227.

EEPORT ON THE HEXACTINELLIDA. 339

lapping of their horizontal rays the squares of the skin. The outermost tip of the
exterior ray frequently bears the small rosette above described.

The most important peculiarities of the new form have been summarised by Murie in
the following brief diagnosis; — " Hexaetinellid sponge characterised by its dendritic or
shrubby contour, occasionally attaining a height of 3 feet or possibly more. Branches
forking or dichotomous, with continuous whorled series of spicular tufts from base to
apices. Skeleton only known ; basework composed of relatively stoutish glassy fibres
of coalesced sexradiate and spinomucronate spicula, disposed in tolerably compact tra-
beculse. Main direction of fibre longitudinal to axis in parallel, straightish, or slightly
bent lines, where continued into exterior whorls ; in crossing fibres more irregular, as are
the very numerous excretory canals. Oscula and pores of moderate size distributed all
over the free surfaces. Flesh-spicula abundant, and of scopuline, acerate, and rosette
shapes. A dermal veil of slender interwoven Hexaetinellid spicula probably clothes the
major portion, or possibly the entire sponge."

In a postscript Murie finally calls attention to the fact that his Dendrospongia steerei
may be identical with the Sclerotliamnus clausii, Marshall, described a year before by
Marshall,' so that the latter name must be accepted as the earlier.

To the memoir by Murie, Carter has added an appendix in which he expresses the
opinion that the fragment from the basal tuft of Euplectella cucumer, which was at first
referred by Bowerbank to Farrea occa, and later by Carter to Farrea densa as distinct
from Farrea occa, and which was found to be rich in " scopuline spicules," is most pro-
bably to be referred to Sclerotliamnus claiisii.

Sclerothamnus clausii, Marshall (PL XCVIIL).

Of this remarkable bushy branched species, which difters essentially in external
appearance from all known Hexactinellida, only two skeletal fragments belonging to a
dead specimen were trawled by the Challenger Expedition in the neighbourhood of
Timor (Station 194a, lat. 4° 31' 0" S., long. 129° 57' 20" E.), from a depth of 360 fathoms
on volcanic mud. These are small, irregularly rounded, somewhat bent, and slightly
knee-shaped branches of the thickness of one's finger, and from 12 to 15 cm. in length.
The outer portion has been destroyed by abrasion or otherwise. The tolerably compact
fibrous framework exhibits, in longitudinal section, strands of fibres for the most part
longitudinally directed, or arched towards the exterior, and laterally curved. They
terminate freely on the outer surface, while between them other fibres extend approxi-
mately at right angles. The entii-e disposition of the dictyonal framework, and especi-
ally the above-mentioned direction of the fibres, corresponds exactly wdth the structure of
the framework described by Murie, and represented in a figure which has been copied here

^Zeitschr.f. wiss. ZooL, BJ. xxv., Suppl.

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

on PI. XOVIII. fig. 3. Moreover, the results of microscopic examination of the beams of
the framework are in complete accordance with the description by Marshall and ]\Iurie.
All the beams of the network, the meshes of which are not always quite regular, though
generally square, are more or less richly beset with prongs which are either low and
pointed, or larger, broader, and more acute, as shown in PL XCVIII. fig. 11. The whole
framework is traversed by round canals, from 2 to 3 mm. in width, which open out
laterally.

In the comparatively well-preserved large specimen studied by Murie, whose original
figure I have copied on PI. XCVIII. figs. 1, 2, and 4, ring-like or spiral wreaths were
found to originate laterally on the branches of the stock, while between these bands
of approximately uniform breadth occur. It is only in these wreaths that the ex-
tremities of the laterally bent longitudinal fibres of the dictyonal framework project
freely (PL XCVIII. figs. 2, 4). Although, moreover, in the case of the specimen examined
by Murie, only the deepened furrows were covered with a dermal lattice-like network of
delicate hexacts, forming square meshes, he still regarded it as possible that the entire
surface of the whole sponge was covered with such a dermal network.

Since I was able, in the British Museum, to examine some microscopic prepara-
tions, which were probably made from Murie's original specimen, and to compare the
loose spicules preserved in great abundance, and partly in their natural position, with
those which could be discovered here and there in the fragments from the Challenger
Expedition, I was able to demonstrate the most complete agreement between the two
forms. The identity of the species is therefore indubitable.

The slender hexacts of the dermal skeleton are covered with small pointed tubercles,
especially on the extremities of the straight rays which run out to simple points. The
distal ray is shorter, the proximal longer than the four equal intersecting tangentials.
The outer extremity of every distal ray seems to be associated with a regular floricome-
like hexaster. In the best preserved portions I found these on almost all hexacts of the
dermal network (PL XCVIII. fig. 5).

Each of the six strong round and moderately short basal rays bears six terminals,
which have the form of a liliaceous perianth, being slightly curved. They become
gradually thicker towards the outer extremities, and finally terminate in a spherical
knob. It seems to me noteworthy that these freely projecting spicules differ in the
formation of the outer extremities of their terminal rays from the true floricomes of
the Euplectellidse, &c., though they agree in position and general form (PL XCVIII.
fig- 6).

In addition to the dermal hexacts with radial axial ray, strong dermal scopulte occur
with rough shaft and four strong almost parallel teeth, which are likewise rough, and
pass into a small terminal thickening (PL XCVIII. fig. 9).

Between the beams of the lattice-like dictyonal framework, strong uncinates occur

REPORT ON THE HEXACTINELLIDA. 341

(PI. XCVIII. fig. 10), and also the singular spicules already described by Marshall
as aberrant broom-forks (Besengabcln), which indeed resemble the scopulse in many
respects, though not agreeing with them completely. The somewhat rough straight
shaft, which runs out at one extremity into an abrupt point, is more or less richly
beset with rough finger-like or conical teeth of variable size, and perpendicularly inserted.
On the other somewhat thickened extremity it bears four, more rarely five or three,
rough finger-like terminal rays, which diverge more or less markedly, and are, as a rule,
slightly curved {PI. XCVIII. figs. 7, 8).

It is improbable that the floricome-like hexacts with spherical terminal knobs, which
are to be found in the dried specimens here and there between the beams of the network
of the dictyonal framework, are to be regarded as parenchymal spicules. I am rather of
opinion that they are accidental additions to the parenchyma, and derived from the skin.

Tribe II. INERMIA, F. E. Schulze (Pis. XCIX.-CI. ; Pis. CIII., CIV.).
Dictyonina without uncinates and scopulte.

Family M^ANDROSPONGiDJi:, Zittel (Pis. XCIX.-CI. ; Pis. CIII., CIV.).

The body consists of a system of rounded, anastomosing tubes of tolerably uniform
calibre, with a system of cavities between. Through the latter the water reaches the
interior, passing through the wall of the tubes, and is passed out through their lumen
either into the gastral cavity or directly to the exterior.

Genus 1. Dactylocalyx, Stutchbury (Pis. XCIX., C).

1780. Rozier, Journal de Physique, October, t. xvi. p. 315, pi. i.

1781. Lichtenberg, Magaziu fiir das Neueste aus dor Physik und Naturgeschichte, Gotha, 1781,

Bd. i. pp. 68, 69.
1841. Stutchbury, Proc. Zool. Sec. Lond., vol. ix. p. 86; Ann. and Mag. Nat. Hist., ser. 1,
vol. ix. p. 504 {DadylocaliiK pnmiceus).

1858. Bowerbank, Phil. Trans., vol. cxlviii. p. 279.

1859. Gray, Proc. Zool. Soc. Lond., vol. xxvii. p. 437 ; Ann. and Mag. Nat. Hist., ser. 3,

vol. V. p. 495 (Macandrewia).

1860. Max Schultze, Hyalonema.

1861. Ehrenberg, Monatsber. d. k. preuss. Akad. d. Wiss. Berlin, vol. i. p. 448.

1862. Bowerbank, PhO. Trans., vol. clii. p. 747.

1863. Johnson, Proc. Zool. Soc. Lend., p. 257; Ann. and Mag. Nat. Hist., ser. 3, vol. xiii. p. 257.

1864. Bowerbank, Monograph of the British Spongiadoe, vol. i.
1864. KulLiker, Icones hist., Protoz.

1867. Gray, Proc. Zool. Soc. Lond., p. 492.

1868. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. i. p. 161.

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

1868. Clrtus, Ueber Euplectella aspergillum.

1869. Wyvillc Thomson, Phil. Trans., vol. cli.K. p. 701.
1869. Bowerbank, Proc. ZooL Soe. Loud., p. 66.

1869. 0. Schmidt. Mittheil. Naturw. Vereines Steierm., p. 89.

1870. 0. Schmidt, Grundziige einer Spongienfauna des Atlant. Gebiet.

1871. Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. vii. pp. 112-141, 268.

1872. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. ix. p. 442.

1873. Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. xii. p. 349.
1875. Carter, Ann. and Mag. Nat. Hi.st., ser. 4, vol. xvi. p. 1.
1875. Marshall, Zeitsohr. f. wiss. Zool., Bd. xxv., Suppl.

1875. Bowerbank, Proc. Zool. Soc. Lond., p. 272.

1876. Marshall. Zeitschr. f. wiss. Zool., Bd. xxvii. p. 113.

1877. Sollas, Ann. and Mag. Nat. Hist., ser. 4, voL xis. p. 1.
1877. Sollas, Ann. and Mag. Nat. Hist., ser. 4, vol. xx. p. 285.

1877. Zittel, Abhandl. d. k. Baier. Akad. d. Wiss., Bd. xiii. p. 1.

1878. Zittel, Abhandl. d. k. Baier. Akad. d. Wiss., Bd. xiv. pp. 21-48, 65.

1879. Sollas, Journ. Roy. Micr. Soc, vol. ii. p. 122.

1879. Agassiz, Bull. Mus. Comp. Zool., vol. v. p. 289.

1880. 0. Schmidt, Die Spongien des Meerbusens von Mexico.

History. — In 1841 a report was given by Stutchbury^ on a sjionge which was procured
at Barbados, one of the Antilles, and which exhibited a siliceous skeleton composed of
fine tubes with netted walls. The following diagnosis was proposed : — " Sponge fixed,
rigid, siliceous ; incurrent canals uniform in size ; excurrent canals large, forming deep
sinuosities on the outer surface, radiating from the root to the outer circumference."
Stutchbury named it Dactylocalyx pumiceus.

Probably an illustrated notice which has passed from Rozier's Journal de Physique -
into Lichtenberg's Magazine ' also refers to the same genus. For a knowledge of this
publication, which is chiefly of interest on account of the date of its publication, I have
to thank Professor Marshall of Leipzig, who has sent me a copy of the text and of the
figures. The title in Lichtenberg's Magazine * is in the following words : — " Nachricht
von einem seltenen Polypengebiiude." " The structure of the polype in the first
copperplate shows a peculiar character and form not hitherto met with in any collection
of the products of the sea. It is dazzling white, resembles a net, and the magnifying
glass reveals a number of scattered chambers which are perforated in all directions ; it is
therefore unusually light, and is very readily pulverised. Its elegant form approaches
very near to tbat of the Coupe de Neptune ; but the resemblance is not confirmed when
the two polype forms are placed side by side. It is difficult to refer this structure to
its proper class. When kept under water it does not become larger, nor does it acquire
any flexibility. With nitric acid it does not effervesce. When a piece of it is calcined

■ Proc. Zool. Soc. Lond., vol. ix. p. 86 ; Ann. and Mag. Nat. Hist., ser. 1, vol. is. p. 504.
"- October 1780.

^ Magazinfiir das Neueste aiis der Physik mid Naturgeschichfe, Bd. i. pp. 68, 69, Gotha, 1781.
* Loc. cit., p. 68.

REPORT ON THE HEXACTINELLIDA. 343

it gives off, indeed, a smell like that of a burning animal body, but it does not expand
like other lithophites. If it is heated in a crucible over a strong fire it smells like
burned gore, assumes a darkish colour, and becomes finally dazzling white, so that it
cannot, like the ordinary Madrepores, become transformed into lime. Nor docs the
residue vitrify by the application of strong heat. A portion of this form gave a
yellowish colour to white oil of vitriol, which was due to the action of this acid on the
animal part. The stone on which this structure is fixed, and which Mr. Badiez brought
from Martinique, is a dense green lava mixed with crystals and dark prismatic schorl."^

In 1858, Bowerbank noted in the Philosophical Transactions (ph xxvi. p. 312) some
stellate spicules oi Dactylocalyx pumiceus, Stutchbury, as examples of his " trifurcato-
hexradiate stellate " and " spinulo-trifurcated hexradiate stellate spicules," and gave '^ a
good figure of one of these.

The Macandrewia azorica, which Gray first described in 1859,^ and which he united
in a group along with Dactylocalyx and Myliusia callocyathes, does not belong, as
later accurate descriptions show, to this category, being in fact a Lithistid. The same
is true of Dactylocalyx prattii, Bowerbank, of which some typical skeletal spicules
are figured,* as weU as of Dactylocalyx hoiverhanJcii, Johnson, which Johnson described,*
and of which Bowerbank figured an isolated quadriradiate spicule in 1864 in his British
Spongiadse, vol. i. pi. ii. fig. 53.

In his great Sponge System Gray formed in 1867" a special family — the Dactylo-
calycidse — within his order Corallispongia, and this family he characterised as " Sponges
with massive, expanded, or flabellate, the network with angular meshes." The members
of this family possess a "network irregular, not symmetrical," and consist of Macan-
dreioia and Myliusia and of the genus Dactylocalyx, which was partly identified with
Tphiteo7i, Valenciennes.

The diagnosis of Dactylocalyx is as follows : — " Sponge expanded, with large sunken
groves and oscules on the upper and lower surface. Spicules of skeleton tuberculated ;
spicular network, rugose, tubercular. Sarcode with scattered, radiated, or stellate spicules,
divided into branches near the base, and with knobs at the tip of the rays. Sarcode
studded with many-rayed stellate spicules, the six principal rays diverging on aU sides,
and divided near the base into several elongated cylindrical linear rays, which diverge
from each other and are tipped wii\\ a small apical knob like the head of a pin."

Besides the Dactylocalyx prattii, Bowerbank, which is a Lithistid, and does not there-
fore belong to this group, Gray established two other species of Dactylocalyx, namely,

1 Eozier, Journal de Physique, October 1780, t. xvi. p. 315.
^ Loc. cit., fig. 1.

2 Proc. Zool. Soc. Land., vol. xxvii. p. 432 ; Ann. and Mag. Nat. Hist, ser. 3, vol. v. p. 495.
* Phil. Trans., 1862, p. 747, pi. xxvii. fig. 8.

» Proc. Zool. Soc. Lond., 1863, p. 257 ; Ann. and Mag. Xat. Hist., ser. 3, vol. xiii. p. 257.
« Proc. Zool. Soc. Lond., 1867, p. 505.

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

(l) Dacti/localyx piimiceus, Stutchbury, named Dactylocalyx jjumicea by Gray, and
identified with Iphiteon 2)cinicea, Valenciennes (Paris Museum), as a " sponge broad
expanded, upper surface rather concave. Hab. West Indies, Barbados, St. Vincent ; " '
and (2) Dactylocalyx subglobosus, Gray, a " sponge subglobose, with a deep central
concavity above ; the outer surface with irregular anastomosing oscules. Hab.
Malacca (?)" (pi. xxvii. fig. 1).

A short account of the nature of the continuous siliceous framework of Dactylocalyx
was given by Glaus in 1868 in his treatise on EuplecteUa aspergillum. He says on page
23 : — " We have thus to deal here not with united siliceous spicules, but with a fibrous
tress-work of siliceous substance like the ceratose fibrous networks which occur in the
Ceratospongia." On the other hand, Wyville Thomson in the same year remarks,- — " I
believe that it would be safe to accept the generalisation that the continuous siliceous
network, wherever it occurs in the vitreous sponges, is produced by the fusion of spicules
of the hexradiate type."

A detailed description of all these sponges which Bowerbank ascribed to the genus
Dactylocalyx, Stutchbury, was given by the same author in 1869.^ After first separating
oflF the genus Iphiteon, Valenciennes, from Dactylocalyx, he characterised the genus
Dactylocalyx, Stutchbury, in the following manner: — Skeleton siliceo-fibrous, fibres solid,
cylindrical." Iphiteon, on the other hand, thus : — " Skeleton siliceo-fibrous, fibres solid,
cylindrical. Eeticulations symmetrical. Areas rotulate, confluent."

To Dactylocalyx he ascribed the following, in part, newly described species : —
(1) Dactylocalyx pumiceus, Stutchbury ; (2) Dactylocalyx heteroformis, Bow^erbank ;
(3) Dactylocalyx macandrewii, Bowerbank ; (4) Dactylocalyx prattii, Bowerbank ;
(5) Dactylocalyx masonis, Bowerbank ; (6) Dactylocalyx howerhanhi, Johnson ; (7)
Dactylocalyx polydisctis, Bowerbank. Of all these, however, only the first, according
to the figures and descriptions, namely, Dactylocalyx pumiceus, Stutchbury, is truly a
Hexactinellid, all the others being Lithistida or Tetractinellida.

From Bowerbank's long description of Dactylocalyx pumiceus, Stutchbury, I shall
quote only the diagnosis : — " Sponge cyathiform, slightly pedicelled, surface even.
Oscula and pores unknown. Expansile dermal system, connecting spicula furcated
attenuato-patento-ternate, and dichotomo-patento-ternate. Dermal membrane — tension
spicula small, acerate and subequiangular triradiate spicula ; retentive and defensive
spicula acerate or cylindrical verticil! ately spinous, whorls of spines numerous and very
large ; and also attenuato - stellate, very minute and numerous. Skeleton : — rete
irregular ; fibre stout, irregularly and abundantly tuberculated, apices of the tubercles
mainly papillous. Auxiliary skeleton-fibres more or less rectangular, hexradiate,
profusely spinous, distal terminations clavate, large and numerous. Tension spicula

' Figured he. cit, pL ixvii-. fig. 2. - Ann and Mag. Nat. Hist., ser. 4, vol. i. p. 122.

2 Proc. Zool. Soc. Loud., 1869, p. 76.

REPORT ON THE HEXACTINELLIDA. 345

rectangular hexradiate, smooth, long and slender, radii subclavate. Retentive spicula
trifurcated attenuato-hexradiate stellate, minute and very numerous. Gemmules mem-
branous, aspicular." " Colour unknown in the living state. Hab. — Barbados (Dr.
(Cutting) Martinique, M. Plec, 1829. Examined in the skeleton state."

Of the genus Iphiteon — a name which Bowerbank has ascribed to Valenciennes,' —
Bowerbank accurately described five species in 1869,' namely, I'phiteon ^xt/ricect (Val-
enciennes), Bowerbank, Iphiteon heatrix, Gray, Iphiteon suhglohosa, Gray, Iphiteon
ingalli, Bowerbank, and Iphiteon callocyathes, Gray.

The essential characters of the first species, Ip>hiteon panicea, are established by
I'owerbank from a specimen obtained at Porto Rico, and preserved in the Museum of
the Jardin des Plantes, Paris. The diagnosis reads as follows :" — " Sponge cyathiform,
slightly pedicelled. Surface of rigid skeleton even. Oscula, pores and dermal mem-
l>rane unknown. Skeleton symmetrically radial, radii short and stout, areas of the rete
mostly six-sided spaces, within triangular ; fibre cylindrical, incipiently spinous. Tension
spicula simple, hexradiate, slender, abundantly spinous, radii terminally more or less
clavate. Rententive spicula spinulo-pentafurcated hexradiate stellate, few in number.
Gemmules simple, membranous, subspherical, irregularly disposed, very numerous."

The second species, Iphiteon heatrix, Bowerbank, resembles Aphrocallistcs heatrix.
Gray, and has already been referred to in detail under the genus ApthrocaUistes.

The third species, Iphiteon suhglohosa, Bowerbank, agrees with Dactylocalyx
suhglobosa, Gray. Bowerbank diagnoses this form* as " a massive sponge, somewhat
cyathiform, sessile ; surface uneven. Oscula and pores unknown. Dermal membrane-
retentive spicula spinulated biternate minute, very numerous. Skeleton symmetrical
radial ; areas confluent, somewhat irregular, mostly six-sided, spaces within triangular.
Skeleton fibre at the external surface coarsely and irregularly tuberculatcd ; disposition
of the tubercles sublinear. AuxiHary fibres rectangulated hexradiate, abundantly
spinous ; radii spinulated. External defensive spicula fusiform acerate, very large and
long, distal terminations occasionally incipiently spinous. Interstitial spicula rectangu-
lated hexradiate, very slender, radii subclavate, basal ray very long. Spicula of the
membranes retentive spicula spinulo-quadrifurcate and pentafurcate hexradiate stellate,
numerous ; margins of the spinulate terminations crenulate."

The fourth species, Iphiteon ingalli, Bowerbank, agrees, according to Bowerbank
himself, with Dactylocalyx pumicea, Gray, which is said to be distinct from Dactylocalyx
jmmiceus, Stutchbury. Bowerbank gives the following diagnosis*: — " Sponge cup-shaped.
Rigid skeleton ; upper or exhalent surface with large intermarginal excurrent canals

» Valenciennes never really diagnosed a genus Iphiteon. The only foundation for this supposition is a preparation
iu the Museum of the Jardin des Plantes in Paris bearing a label as above. He cannot be referred to as the founder
of the genus.

2 Proc. Zool. Soc Land., 1869, pp. 323-357. ^ Loc. cit., p. 324.

* Loc cit., p. 329. ^ Loc cit., p. 321.

(ZOOL. CHALL. £Xr.— PART LIII.— 1887.) Crgg 44

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

radiating irregularly from the centre towards the circumference. Surface even. Oscula,
pores and expansile dermal system unknown. Skeleton fibre stout, more or less furnished
with scattered warty tubercles. Auxiliary fibres abundantly tuberculated, terminating
spinulately. Insterstitial spicula rectangulated, hexradiate, large ; radii nearly equid,
attenuated and acutely terminated. Retentive spicula spinulo-quadrifurcate hexradiate
stellate ; terminal radii long."

The fifth form, Iphiteon callocyathes = Myliusia callocyathes, Gray, belongs to
another genus, Myliusia, Gray, which will be described more closely below.

Oscar Schmidt ^ first called attention to the fact that the greater number of the forms
described by Bowerbank as Dactylocalyx are not Hexactinellids at all, but belong to the
Lithistida. According to Oscar Schmidt, the genus Dactylocalyx strictly embraces those
sponges which are characterised by a triaxial type of spicule, " their siliceous network
resembles neither the wide tubes of Farrea nor the prismatic tissues of A2?hrocallistes,
but a more dense irregular tresswork. The habit of the body may accordingly be very
variable."

In addition to the meshed forms, Dactylocalyx pumiceus and Dactylocalyx suhglo-
bosus, Oscar Schmidt described and figured a third species, of very different appearance,
under the designation Dactylocalyx crisiyus. From a short compact solid base, there
rises a simple or divided funnel-shaped tube with a thin and somewhat folded wall, while
the upper terminal opening appears to be either irregularly round or else sinuous and
fissure like. Short tube-like lateral protuberances or branches which are here and there
arranged in distinct longitudinal rows open to the exterior by a rounded aperture.

In an article on the classification of the sponges, published in 1872,- Gray distinctly
curtailed his family of the Dactylocalycidae as established in 1867. He removed the
genus Macandreiuia, which was regarded as representative of a special family, the
Macandrewiadse, retained the genera Daxitylocalyx, Myliusia, Kaliapsis, and Disco-
dermia, and characterised the restricted family as follows : — " Sponge massive or ex-
panded or cup-shaped. Skeleton more or less regularly reticulated, with angular
openings diverging from the centre."

In 1873 Carter^ united Dactylocalyx pumiceus, Stutchbury, Daxitylocalyx piumicea.
Gray, and Iphiteon. p>umicea, Valenciennes, in his division of the Vitreohexactinellida, in
the subdivision of " species massive, excavated, shallow," and in a family for which
" rosettes or flesh-spicules, many rayed, rays of equal length, straight, capitate, some-
times only pointed," were said to be characteristic, while he formed a special family
for Dactylocalyx subglohosa, Gray, with the following peculiarities : — " Rosette many
rayed, rays of equal length, straight capitate, or with multitudinous rays of unequal

1 O. Schmidt, Spongienfauua des Atlant. Gebietes, p. 18, 1870.
^ Ann. and Mag, Nat. Hist., ser. 4, vol. ix. pp. 453-457.

2 Ann. and Mag. Nat. Hist, ser. 4, vol. xii. p. 357.

REPORT ON THE HEXACTINELLIDA. 'M7

Ifiigtii, straight and capitate, pappiform, or occasionally of the first kind once branched,
or occasionally with echiuated heads." The difference between Dactylocahjx pumiceus,
8tutchbury, and Dactylocalyx pumicea, Gray = IphiteoJi ingalli, Bowerbank, Carter found
to consist in the fact that " the latter is charged, especially towards the surface, Avith long
linear spicules (slender, fusiform, slightly inflated, and spined for some distance at each
ond), while these are not to be seen in Dactylocalyx pumiceus, Stutchbury.

In 1876 Marshall' placed the genus Dactylocalyx among his Pleionacidse, and
chax'acterised it in the following words : — " Fused latticed tissue of little regularity.
Free spicules are represented by rosettes and irregular hexradiate forms of unknown
significance." As species, he noted (l) Dactylocalyx pumiceus, Stutchbury, and (2)
Dactylocalyx crispus, 0. Schmidt.

In 1877, in his paper on Stauronema,^ Sollas referred Dactylocalyx and Aphrocallistes
to his family of the Aphrocallistidse, distinguished by " sex-radiate skeleton spicules with
rays making any angle with each other," and in an article On the Action of Caustic
Potash on the Siliceous System of Sponges,^ he pointed out the irregular arrangement
of the spicules in the skeletal framework of Dactylocalyx pumiceus and Dactylocalyx
sxilxjlohosus.

In 1877, in his studies on fossil sponges,* Zittel referred Dactylocalyx with Pei-i-
p)hragella, Marshall, and Mylitisia, Gray, to his Mseandrospongidae, which consist of
meandering, entangled, and simple anastomising tubes or plates.

In the Journal of the Microscopical Society,'* Sollas described a simple plain cup-
shaped variety of Dactylocalyx pumiceus, Stutchbury, which he named Dactylocalyx
pumiceus, var. stutchhuryi, or simply Dactylocalyx stutcliburyi. He called attention
to the typical alternation of the outer and inner, or upper and under radial furrows or
furrow-like depressions, described the continuous skeletal framework composed of united
hexradiate spicules, and further some six-rayed, five-rayed, and simple s^^indle-shaped
or thread-like isolated " flesh spicules." At the knots of these originally isolated six-
rayed spicules, which are here united to the skeletal framework without recognisable
order, Sollas frequently saw the beginning of a hollow octahedron, such as occurs in the
Ventriculites. This was seen to become so completely covered by siliceous lamellae that
the knots in the older portions of the framework appeared to be solid throughout.

Among the Hexactinellida procured by the American Expeditions from the West
Indies, and examined by Oscar Schmidt in 1880, several species of Dactylocalyx
were found, which corroborated on the whole the opinions of Sollas. The form which
had previously been described by Schmidt as Dactylocalyx crispus was now separated
from the genus, and regarded as a young form of Schmidt's Syringidium zittelii.

• Zcilschr.f. wiss. ZooL, Bd. xxvii. \). 122. ^ Ann and May. Nat. Hist., ser. 4, vol. .\i.\. \k 1.

^Ann. and Mag. Nat. Hist, ser. 4, vol. x.v. pp. 285-300. * Abhandl. d. k. Baier. Akad., vol. -xiii. p. 38.

!■• Vol. ii. pp 122-133.

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

Dactylocalyx j^umicetis, Stutchbury, was found in the neighbourhood of Barl)ados
at a depth of 103 fathoms; Dactylocalyx suhglobosus, Gray, near St. Lucia, in 11 G
fathoms, and also to the north-west of Havanna in 190 fathoms.

As a new species Schmidt described a drinking-horn -shaped Dactylocalyx {Dactylo-
calyx potatorum) dredged from a depth of 190 fathoms near St. Lucia. Its wall measured
1-5 to 2-5 cm. in thickness, and consisted of a very light and fragile lattice-work of fine
tubes. It also exhibited external as well as internal deep pittings, the openings of
which alternate on the surfaces of the walls, and appear on the inner side to be arrangccl
in lou'i-itudinal rows. Further, the form described by Gray as Myliusia callocyathes was
referred by Oscar Schmidt to the genus Dactylocalyx.

Generic Characters. — The cup-shaped, thick-walled body exhibits both on the outer
and on the inner surface irregular, but generally radial ridges, and interjacent furrows.
Since internal furrows correspond to external ridges, the whole wall seems to be
folded. It consists of a system of anastomosing, but mainly radial tubes, which open
internally, and perhaps also externally. The system of spaces lying between these
anastomosing tubes is closed on the internal gastral surface. It is probable that in the
living form both the .external dermal and the internal gastral surface of the entire sponge
are covered by a flatly expanded dermal or gastral membrane. In the dermal membrane
pentacts occur with unpaired hypodermal ray.

1. Dactylocalyx pumiceus, Stutchbury.

From the original description of Stutchbury, afterwards corroborated by Bowerbank
and by Sollas in reference to two specimens from the Antilles Island, Barbados, and pre-
served in the Bristol Museum, Dactylocalyx p^imiceus has a firm, flat, cup-shaped body,
borne on a short, thick, massive stalk. It consists of a flat, plate-shaped roundish
mass, 30 cm. in breadth, as thick as a thumb, and bordered by a slightly involute,
slightly sinuous, rounded margin.

The radially-disposed furrows, which are frequently interrupted, alternate on the
lower dermal and the upper gastral surface, and here and there exhibit dichotomous
division. The very fine tubular network which forms the entire body is stone-like
and very narrow in its meshes. The wall of the narrow tubes is composed of finely
tuberculate beams, without thickening at the nodes of intersection.

Sollas found loose spicules in the parenchyma in the form of small hexacts, in which
the slender rays were for the most part terminally thickened, while individual rays were
frequently considerably elongated and somewhat curved." According to Bowerbank ' the
parenchyma includes oxyhexasters, with three long, slightly curved terminals on each of
the six short principal rays, and also discohexasters with somewhat long terminals. The

' Proc. Zool. Soc. Lond., 1869, p. 77, pi. iii.

REPORT ON THE HEXACTINELLIDA. 349

flormal skeleton consists, according to Sollas, of numerous rough oxypentacts. No
uncinates or scopulse were found.

2. Dactylocahjx suhglobosus, Gray (PI. XCIX.).

Through the kindness of Professor 0. Schmidt I obtained the half of an unfortunately
much injured and partially macerated dried specimen of that sponge which Gray
described in 1867 under the title Dactylocalyx siibglobosus, and figured in the
Proceedings of the Zoological Society of London, 18G7, pi. xxvii. fig. 1. I was thus
able to study the skeletal structure of this form. While I can, without further comment,
simply corroborate the results of Gray and 0. Schmidt as to the general form and
macroscopic features, I have been led to some divergent conclusions as to the minuter
structure. To these I must therefore briefly refer. Like Gray's specimen, and that
figured by 0. Schmidt in his Spongien des Meerbusens von Mexico (Taf. iv. fig. 8),
the specimen lent me for examination, and represented from a photograph in the outer
view in PI. XCIX. fig. I, exhibited a deep, thick- walled, pear-shaped cup, with a compara-
tively broad basis. The wall consists of a much folded plate. The clefts and furrows of
the internal surface are narrower and more lonoitudinal than the more irregular cavities
and furrows on the exterior.

While former investigators have seen in the folded plate, which forms the entire wall
of the goblet, only a compact mass traversed by small round canals, I find it to be
composed of a fine network of anastomosing tubes, 0"3 to 0-5 mm. in transverse
diameter. The main direction of the tubes forming the framework is indeed radial to
the surface of the entire sponge, but so many lateral branches of equal width are given
off in every direction, anastomosing with adjacent tubes, that the whole a^ipears as a
thick and irregular felt-work. The cavities and interjacent spaces which occur between
these anastomosing tubes, and, of course, communicate with one another in all directions,
have the same diameter as the tubes. While I found this intercanalicular system of cavities
closed on the inner gastral surface by a reticulated plate, I saw their free apertures here
and there on the outer dermal surfiice. Here too, however, they are doubtless covered
over by the delicate porous dermal membrane. The water streaming in from outside
must find its way first into these interstitial spaces, then through the walls of the tubes
into their lumen, and thence into the large gastral space, and perhaps also to the
exterior by the dermal terminal apertures which are probably present during the life of
the sponge. In fig. 2 of PI. XCIX. I have given a diagrammatic representation
illustrating my conception of the structure.

The dictyonal framework supporting the waU of these narrow tubes exhibits
moderately strong irregularly tuberculate beams, with square, triangular, or round
meshes without any definite regularity. The parenchyma includes, besides simple

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

tuberculate discohcxacts, numerous discoliexasters, in which each of the short simple
principal rays bears four long terminals. These become gradually thicker towards
the exterior, and finally end in a small compact transverse disc. The size of these
discoliexasters varies from 0"05 to 0"1 mm. in diameter.

The dermal membrane contains pentacts with rudiments of a sixtli distal ray. The
four tangentials end in a rough terminal knob, while the longer proximal radial is
usually pointed (PI. XCIX. figs. 3, 4).

Gray's original specimen is said to have been brought from Malacca (?). 0. Schmidt
obtained his forms from St. Lucia, West Indies, from a depth of 116 fathoms, and
from lat. 23° 1' N., long. 83° 14' W., from a depth of 190 fathoms.

3. Dactylocalyx{?) f>atella, n. sp. {PL C).

The completely macerated dictyonal framework displayed in PI. C. fig. 1, probably
belongs to a cup- or platter-shaped sponge, with walls at least a finger's-breadth in thick-
ness. The tissue consists of an anastomosing network of tubes of very different calibre,
varying from 2 to 4 mm. in width, and running in a generally radial or slightly curved
course. As to the structure of the natural bounding surface nothing certain cau
be said.

The beams of the dictyonal framework, which form a network with approximately
square or more irregular meshes, are moderately strong and quite smooth. Only the
conical bosses jn-ojecting radially into the lumen of the canals or beyond its free bounding
surface are beset with small pointed tubercles.

Since no loose spicules were found, it must remain somewhat doubtful whether the
generic designation is correct.

Some specimens of this species were obtained by the Challenger Expedition off the
coast of Portugal, others to the south-west of Bermudas, from a depth of 1075 fathoms,
and coral mud ground.

Genus 2. Sclerojjlegma, 0. Schmidt (PI. CI. figs. 1, 2).

From the diagnosis which the author of this genus has given in his Spougien des
Meerbusens von Mexico, p. 56, it unquestionably follows that the genus should be
ranked in the family Maeandrospongidte, Zittel. It seems to me doubtful, however,
whether all the forms referred by 0. Schmidt to this genus were really ScJeroiolegmata.
On the contrary, from his own description, I am forced to conclude that his Sclcrojilegma
lanterna and Scleroplegma herculeum belong to another set of forms. But with this I
have not much to do, since I have not been able to examine the original specimens of these
two species. I have, however, a half specimen of Scleroplegma conicum, Schmidt, which

REPORT ON THE HEXACTINELLIDA. 851

though macerated has the dictyonal framework well preserved. This specimen, preserved
by 0. Schmidt himself, 1 have been able to use as a basis for an independent investiga-
tion of the genus and species.

I accept Schmidt's generic characters except in their last clause, which is based on the
two doubtful species above referred to. In abbreviated form it reads as follows' : —
" Sponges of cylindrical or truncated conical form, with a similarly shaped thick-walled
body cavity. The peripheral felt-work is wide-meshed, but strong and firm, or brittle.
It forms round or prismatic tubes, which generally run obliquely from the exterior to the
interior, either isolated or united with one another, and opening into the gastral cavity.
Between them irregular intercanals occur."

1. Scleroplegma conicum, 0. Schmidt (PL CT. figs. 1, 2).

Of this species I could unfortunately procure only the single specimen figured from a
photograph in PL CI. fig. 1. It was completely macerated, so that of the isolated skeletal
elements no trace was left. The tubes, 2 to 3 mm. in width, which form the thick wall
(5 to 8 mm.) of the cup-shaped body, are generally radial, and directed somewhat
obliquely upwards and outwards. They open both on the internal gastral and on the
external dermal surface by round apertures, and here and there exhibit anastomoses. The
intercanals occurring between these anastomose abundantly, and have a uniform width ;
they open, however, only on the outer dermal surface, where the water enters.

The dictyonal framework consists of strong beams, superficially beset with small
pointed tubercles, usually disposed in transverse rows. The nodes of intersection are
spherically thickened and beset with pointed tuberculated warts, which are especially well
developed in the bounding surfaces of the whole framework (PL CI. fig. 2). There is
a remarkable absence of the conical bosses which, on the bounding surfaces of the
dictyonal framework of other Dictyonalia, project for a variable distance at right angles
to the nodes of intersection.

Scleroplegma conicum was found near Morrolight, West Indies, at a depth of 292
fathoms.

Genus 3. Margaritella, 0. Schmidt.
Margaritella cceloptychioides, 0. Schmidt (PL CI. figs 3-8).

From the somewhat indefinite figure and description which 0. Schmidt gives of this
form in his Spongien des Meerbusens von Mexico,^ the body seems to exhibit a flat, cup-
like form, resulting from a connected system of tubes. In the fragment which I received
for comparison from Professor 0. Schmidt, some free spicules were fortunately retained.

1 hoc. cil., p. 5G. ' P. 54, pi. v-ii. fig. 7.

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

The slightly irregular dictyonal framework consists of moderately thick beams,
somewhat uniformly beset with fine tubercles. The nodes of intersection exhibit no
thickeninsj, and on the boundina; surface slender tubercle-like bosses of moderate length
project freely (PI. CI. fig. 3). The dermal membrane contains rough pentacts disposed
I'egularly so as to form square meshes. In these sj)icules, the four uniformly long tangential
rays end in a knot-like swelling, while the much longer jiroximal radial, which becomes
narrower inferiorly, is simply rounded off or pointed. The parenchyma includes delicate
oxyhexacts with somewhat undulating rays, which sometimes divide, and may also
exhibit thickened points of intersection (PI. CI. figs. 5, G). More frequently sphsero-
hexasters occur of variable size, and with a variable number of terminal rays. The
principal rays are either moderately short, bearing six to eight knobbed terminals, about
three times as long, somewhat markedly divergent and slightly curved (PI. CI. fig. 8), or
they are so shortened that they appear rather as a spherical thickening of the node of
intersection. In the latter form the numerous, long knobbed terminal rays look like
simple radii from a s^Dherical centre (PL CI. fig. 7).

0. Schmidt notes that the form was dredged at Havanna, West Indies, from a depth
of 158 fathoms.

Genus 4. Myliusia, Gray (PI. CIIL).

1859. Gray, Pioc. Zool. Soc. Lond., vol. xxvii. p. 437 ; Ann. and Mag. Nat. Hist., ser. 3, vol. v.

p. 495 (Mi/Iiusia calloajcdhes).
1867. Graj', Proc. Zool. Soc. Lond., p. 492.

1869. Bowerbank, Proc. Zool. Soc. Lond., pp. 66, 334 {Myliusia grayi).

1870. 0. Schmidt, Grundziige einer Spongienfauna des Atlant. Gebiet.

1872. Gray, Ann. and. Mag. Nat. Hist., ser. 4, vol. i.x. p. 442.

1873. Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. xii. p. 349.
1875. Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. xvi. p. 1.

1875. Marshall, Zeitschr. f. wiss. Zool., Ed. xxv., Suppl.

1876. Marshall, Zeitschr. f. wise. Zool., Bd. xxvii.

1877. Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. xix. p. 121.

1877. Sollas, Ann. and Mag. Nat. Hist., ser. 4, vol. xix. p. 43 {Mijliusia gyayi) ; and voh xx.
p. 285.

1877. Zittel, Abhandl. d. k. Baier. Akad.

1878. Marshall and Meyer, Mittheil. d. Zool. Mus. zu Dresden, I'.d. ii. p. 261.

1879. 0. Schmidt, Spoogien der Meerbusens von Mexico, vol. i.

1880. O. Schmidt, Spongieu der Meerbusens von ]\Iexico, vol. ii.
1882. Weltner, Beitrage zur Kenntniss der Spongien.

History. — In 1859 Gray' described and figured a new sponge procured from the West
Indies, which he regarded as representative of a new genus, and named Mj/Iiusia, with
the specific name callocyathes.- The generic title was given in honour of Christlob
Mylius, who was the first] to descrilie the singular Unibellularia groenlandica. The

' Proc. Zool. Soc. Lond, p. 439 ; A iin. and. Mag. Nat. Hist., ser. 3, vol. v. p. 497.
- Properly, callocyathus, I'roiii y.iix6o:, a cui).

REPORT ON THE HEXACTINELIJDA. 353

description ran : — " Sponge, siliceous, funnel-shaped, fixed by the base ; the upper sur-
face smooth, marked with numerous minute perforations placed in nearly parallel
grooves, radiating from the centre to the circumference, and with numerous large, olilong,
rather unequal-sized perforations, which are fringed on the lower side with a high wall of
a similar structure to the rest of the sponge ; these edges of the cavities causing the
under surface to be covered with unequal irregular-shaped tubes of nearly the same
length, and more or less confluent together : some of these tubes are simple and sub-
cylindrical, others are expanded and more or less crumpled on the edge around the
cavity, so as to end in two, three, or even four more or less circular mouths." A magni-
ficent delineation of the elegant sponge completed this accurate diagnosis.

In the systematic review of the sponges, which Gray gave in the Proceedings of
the Zoological Society London, in 1867, he placed Myliusia in the family of the Dactylo-
calycidae (" sponge massive, expanded or flabellate, the network with angular meshes,")
close to Dactylocalyx, and briefly characterised it on page 506 in the following words : —
" The sponge conical, cup-shaped, pierced with numerous short truncated tubes, forming
raised folded anastomising laminse on the lower surface."

A diagnosis of the genus Myliusia, Gray, founded for the most part on the micro-
scopic structure of the skeletal framework, was published by Bowerbank in 1869 in the
Proceedings of the Zoological Society of London, part i. p. 66, and ran as follows : —
" Skeleton siliceo-fibrous. Fibres solid, cylindrical. Rete symmetrical, disposed in a
series of crypt-like layers parallel with the external surface, with intervening plades of
perforated siliceous tissue."

Wyville Thomson,^ in his scheme of Porifera vitrea, rejected Gray's genus Myliusia,
and referred the form in question simply as Dactylocalyx callocyathus to the genus
Dactylocalyx. Moreover, Bowerbank" subsequently found, "on microscopic examina-
tion of the structure of the type specimen of Myliusia, Gray," that it was identical
with that of his genus Iphiteon. He therefore referred Gray's Myliusia callocyathus, as
Iphiteon callocyathus to the genus Iphiteon. Thus, then, the generic name Myliusia
became as it were free, and Bowerbank availing himself of the fact, designated another
recently studied remarkable Hexactinellid from St. Vincent, West Indies, with the
generic title Myliusia, and added to it the specific name grayi. The accurate descrip-
tion of the new species Myliusico grayi, Bowerbank, which was illustrated with two good
figures of the microscopic structure of the skeleton, undoubtedly showed the essential
difference between this species and Gray's iH7//i"«siacaifoc?/a^/iW5, and certainly established
the generic distinctness of the two. It is, however, very much to be regretted that
Bowerbank did not invent a new name for his new genus, as many errors and ani-
V)i(i;uities would thus have been avoided.

Later authors have, however, by no means always attended to the essential difference

' Phil. Trans., p. 713, 1869. -Pm. Zool. Soc. Load., \\ .335, 1869.

(ZOOL. CHAIX. EXP. —PART LIII. — 1887.) Ggg 45

354 THE VOYAGE OF H.M.S. CHALLEN(}ER.

between the two species Mylinsia {Iplikcon, Bowerbank) callonjatlnis, Gray, and
Myliusia grayi, Bowerbank, but without further inquiry have assigned both to the
same genus Myliusia — a generic name which, after what has been said, ought to be
applied to Myliusia callocyathus, and not to Myliusia grayi, Bowerbank, in case the
latter of the two species does not belong to the same genus.

In 1873 Carter' established a characteristic difference between Myliusia callocyathus,
(!ray, and Myliusia grayi, with respect to the rosettes. In Myliusia callocyathus, the
" rosettes are many rayed; rays of equal length, capitate, flexed, and grouped en fleur-
de-lis, or occasionally with straight and capitate rays ; " in Myliusia grayi " the
rosettes are many rayed, rays of equal length, straight, capitate."

In the memoir which Marshall published in 1876 on the affinities of the Hexactin-
ellida- he seemed inclined to identify Myliusia callocyathus, Gray, with Dactylocalyx
crispus, 0. Schmidt.

In the description of Myliusia grayi, Bowerbank, which Carter gave^ in 1877, he
notes that^ " although Myliusia grayi presents the convoluted cerebriform appearance of
Myliusia callocyathus, yet its minute structure is totally diff'erent, inasmuch as the knots
or junctions of the fibre in the latter are solid and round, not hollow and lantern-shaped
as in Myliusia grayi."

In his studies on fossil sponges Zittel'^ places Myliusia grayi — 2:)robably with exclusion
oi Myliusia callocyathus, Gray — along with Dactylocalyx, Stutchbury, AnAPeriphragella,
Mar^all, in his family of the Mfeandrospongidaj, " in which the sponge body consists of
meandering, intertwining, and anastomosing thin-walled tubes or laminse, the canal
system absent or scarcely developed, the intercanal system, on the other hand, present. A
covering layer is wanting, or forms a coherent siliceous skin on the surface."

In a report from the Dresden Zoological Museum (1878), Marshall and Meyer
accurately described a new sponge from the Philippines as Myliusia zittelli, Marshall and
Meyer. But since this form is doubtless closely related not to Myliusia callo-
cycdhus, but to Myliusia grayi, Bowerbank, we shall not at this stage take it into
consideration (see Aulocystis). The species in question was found by 0. Schmidt among
the West Indian sponges of the American Expeditions.

1. Myliusia callocyathus. Gray (PI. CIIL).

Specimens of this elegant species were procured from three stations by the Chal-
lenger Expedition, but of these none attains the size of the example described by J. E.
Gray, and figured in the Proceedings of the Zoological Society of London, 1859, pt. xvi.

1 Ann. and Maij. Nat. Hist., ser. 4, vol. xii. p. 358. - Zeitschr.f. wiss. Zool, Bd. xxvii. pp. 113-1.36.

3 Ann. and Mag. Nat. Hist., .ser. 4, vol. xix. pp. 120-131. ■• Loc cit., p. 128.

° Abhandl. d. k. Baier. Acad., p. 38, 1877.

REPORT ON THE HEXACTINELLIDA. 355

They are, however, in part very well preserved iii alcohol, so one cau learn someLhing of
the structure of the soft parts. The two specimens represented in natural size on
PI. cm. figs. 1 and 2 (of which fig. 1 represents two individuals united into one)
were obtained along with some fragments of other specimens, on blue mud ground, in the
vicinity of Little Ki Island (Station 192, lat. 5° 49' 15" S., long. 132° 14' 15" E.) in
140 fathoms. A somewhat larger cup was procured in the neighbourhood of Timor
(Station 194, lat. 4° 34' 0" S., long. 129° 57' 30" E.) on volcanic mud, at a depth of
200 fathoms ; and a small dried specimen, only about 1 cm. in height, from St. Thomas,
West Indies (Station 24, lat. 18° 38' 30" N., long. 65° 5' 30" W.) on Pteropod ooze in
390 fathoms. All these specimens have the form of a thick-walled cup which expands
upwards from a narrow basal plate.

The lateral wall consists of a plate from 1 to 3 mm. thick, which is simply folded and
drawn out into radial open tubes. While in the small dried specimen from Station 24
these radial excurrent tubes of the gastral cavity are only 3 mm. broad, and are
almost all isolated towards the exterior, in the larger specimens from Stations 192 and
194, they have a tolerable breadth (as much as 5 mm. or more), and are fused to one
another laterally, so that they project externally in meandering lines or interwoven reticu-
late plates (PI. cm. figs. 1, 2). The free upper margin of the entire cup exhibits a sinuous
external curvature. This is also true of the regular funnel-shaped large specimen (10 cm.
broad, and 7 cm. high) figured by Gray (loc. cit.). The figure suggests that the radial
tubes have originated through closure of the parietal folds by growth of the entire body.
One may regard the entire upper aperture of the cup or funnel as the chief osculum,
and the lateral orifices of the tubes as secondary or incidental oscula. At all events the
water stream passes through the coherent plate in such a manner that it enters from the
outside into the depressions, furrows, &c., between the radial tubes, permeates the
continuous plate, and then reaches the lumen of the tubes which open out laterally, or
the central space which is in open communication with these, and thus reaches the
exterior, whether it be through the outer openings of the radial tubes, or through the
large terminal aperture of the central (gastral) cavity.

The finer structure of the body is best studied in sections at right angles to the
wall plate ; one of these I have combined diagrammatically on PI. CIII. fig. 3 from
numerous preparations.

The dictyoual framework consists of moderately strong beams. These are beset with
numerous small conical hooks, disposed in great part in regular cross rows. The beams are
united with tolerable regularity, and form for the most part distinctly square, more rarely
three-sided or five-sided meshes. The somewhat thickened nodes of intersection are
provided with rough flat papillary rounded swellings. It seems to me noteworthy that on
the (outer) dennal as well as on the gastral surface of the dictyonal framework the
nodes of intersection which are here very richly provided with protuberant papilhv;, are

356 THE VOYAGE OF H.M.S. CHALLENGElt.

to be contrasted with the corresponding nodes of most other Dictyonina in possessing, as
a rule, no freely projecting radial ray, but simple slightly arched extremities (PL CIII.

fig- 3).

The dermal skeleton contains h5q5odermal hexaets sometimes mixed with pentacts,
and provided with smooth rays which are occasionally roughened at the extremities. Of
these the distal ray, which is absent in the pentacts, is shortest, often simply rounded or
even terminating in a small swelling, while the four tangentials which run out to a
point, and the much longer proximal ray always taper to a point. On the outer
extremities the hexact hypodermalia support a floricome-like discohexaster in which
each of the short principal rays bears seven (more rarely fewer) S-shaped terminals,
which run out externally into a thickened hemispherical transverse disc with a notched
margin (PI. CIII. figs. 7, 8, 9). The spicules which constitute the gastral skeleton have
an exactly similar form and disposition (PI. CIII. fig. 3).

Scattered throughout the dictyonal framework of beams I have also to note small
simple hexaets with somewhat rough uniformly tapering rays (PL CIII. fig. 4), and
oxyhexasters with short principals and long S-shaped pointed terminal rays (PL CIII.
figs. 5, 6). Since there are two or three terminals on each principal ray, and since each
of the six principals may either remain simple and undivided, or may be divided into
terminals, all transitions occur between a hexact with five undivided somewhat rough
rays and only the sixth ray forked, and an oxyhexaster in which each of the principal
rays is divided into three terminals, giving a total of eighteen (PL CIII. fig. 5).

Genus 5. Aulocystis, n. gen. (PL CIV.).

History. — In his sponge-system Gray^ added to the description of Myliusia callo-
cyathus, the following brief notice : — " There are two smaller specimens in the British
Museum which probably belong to the same species. The smaller one was collected by
the Rev. L. Guilding at St. Vincent in 1840, and the other was received from the West
Indies by Mr. Scrivener in 1842."

These sjDecimens, which were not further studied b}' Gray, were subsequently
examined by Bowerbank, who, in 1869,^ gave the following account of them. While he
maintained that the specimen procured by Mr. Scrivener in 1842 agreed in structure with
Myliusia {Iphiteon, Bowerbank) callocyathus, Gray, " he found that the specimen collected
by the Rev. L. Guilding at St. Vincent difi'ered widely in the structure of its skeleton
from either Ipliiteon, Bowerbank, or Dactylocalyx. He proposed, therefore, to apply
Dr. Gray's generic title Myliusia to this species, instead of to Myliusia {Iphiteoit, Bower-
bank) callocyathus, Gray. He accordingly named the form Myliusia grayi, and briefly
diagnosed it as follows : — " Sponge sessile, massive. Dermal surface unkno'RTi. Surface

' Proc. Zool. .S'oc. Loiul, 1867, p. OdC. ^ p^oc. Zool. Soc. Luwl, 1S69, p. 335.

REPORT ON THE HEXACTINELLIDA. 357

of rigid skeleton uneven and excavated. Oscula, pores and expansile dermal system
unknown. Skeleton stratified, forming a series of expanded crypt-like spaces. Fibre
cylindrical, incipiently or minutely spinous. Interstitial spicula numerous, acerate, large
and long, variable in size, disposed in lines at right angles to the stratification in loose
fascicula of two to four or five together. Retentive spicula spinulo-multifurcate,
hexradiate, stellate."

The surface of the rigid skeleton exhibited a series of square or irregularly angular
areas, the angles of which were occupied by thin 'perforated angle plates with their inner
margins curved, so that, when combined, they left a large circular or oval orifice in the
middle of each space. The mass of the oval sponge itself he found to consist of a series
of thin sinuous skeletal plates, not more than one-third of a line in thickness. In sec-
tions of the rigid skeleton at right angles to the surface of the sponge, he found that
" it was formed of a series of crypt-like layers of skeletal fibre, each layer forming as it
were a distinct and extensive crypt-like space, with short, stout, cylindrical pillars with
gradually expanded bases and capitals, the intervening portions of the shafts of the
columns being u'regularly studded with acutely conical incipient spines."

The two excellent figures which Bowerbank has appended to his description {Joe. cit.,
pi. xxiii. fig. 8. and pi. xxv. fig. 1), show more clearly than the description itself that
the knots of the rectangular meshes of the beam network are perforated in a manner
quite similar to the fossil Ventriculites. They exhibit in the interior a delicate hexradiate
axial cross, but consist externally of siliceous strands, which correspond to the margins
of a regular octahedron. This, too, was definitely stated by Carter in his description
of the Hexactihellida in 1873.' " Myliusia grayi, hoth.\vi generic form and structure,
has many distinguishing characters, especially that which consists in an octahedral
lantern-like appearance of the joints produced by the vitreous fibre stopping short of the
joint, and proceeding thence directly on from one arm of the hexradiate spicule to the
other, so as to leave the centre intact and visible through the interstices."

An exhaustive and accurate description of this Myliusia grayi, Bowerbank, was given
by Carter in 1877.'^ Here he describes it as follows :—" General form hemispherical;
general appearance enteromorphous or cerebriform, sessile, consisting of tortuous anasto-
mosing tubular canals or passages separated by equally tortuous labyrinthic intervals.
Tubular canals or passages now terminating on the surface in round patulous or long-
tortuous gutter-like openings. Surface of tubular passages, both externally and inter-
nally covered with a dermal layer of small sexradiate spicules, whose horizontal arms,
overlapping each other, form a continuous quadrilateral meshwork. Internal or body
structure of the wall of the tubular passages composed of lozenge-shaped or lantern-like
knots of vitreous fibre applied end to end, three or more layers deep, thus forming a
laminate mass of trapezoids united to each other at their angles in successive rows, with

1 Ann. and Mag. Nat. Hist., ser. 4, vol. xii. p. 365. , ' Aim. and Nat. May. Hist., ser. 4, vol. xix. pp. 106-131.

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

cylindrical intervals between them covering each other more or less rectangularly.
Skeleton spicules of three forms, viz., (1) small sexradiate, arms not inflated at their
junction, alternately pointed, and thickly spined throughout ; (2) much larger se.x-
radiate, the same but with the arms slighted inflated at the extremity; (3) still much
larger, unispined, but uneven on the surface, and here and there microspined. Flesh -
.spicules of two forms, viz., (1) rosette globular, consisting of six short arms, each of
wliich is surmounted by five long capitate rays expanded in a vasiform manner ; (2)
bundles of minute hair-like undulating acerates like the tricurvate or low spicule."

" The small sexradiate," Carter continues, " become the centres respectively of the
trapezoids, which are thus formed by the extension of a thread of vitreous sarcode, from
one end of each of the arms of the sexradiate spicules to the other, strengthened at eacli
attachment by subsiding threads and the arm of each end of the latter ; finally increasing
in thickness throughout till the trapezoid is fully formed and presents four sides,
with eight lantern-like holes in them, one in each triangular face, through which the sex-
radiate form of the original spicule may be seen in the centre intact. Spicules 2 and
3 form the fringe round the apertures which interknits with the body structure of the
lamina internally, the latter or the supposed acerate form extending beyond the former,
])otli distally and proximally, while the flesh-spicules are scattered throughout the
structure unequally, that is much more numerously towards the surface."

Although Carter clearly recognised the essential difi"erence between Myliusia callo-
cyathus, Gray, and Myliusia grayi, Bowerbank, their generic separation appeared to
him unnecessary, perhaps because of the agreement in the " convoluted cerebriform
appearance." He included both in the same genus Myliusia. Carter, however, called atten-
tion to the fact that although the skeletal structure of Myliusia grayi agrees very closely
with that of many fossil Hexactinellids, such as Scyphia, Ventriculites, Cceloptychium,
&c., yet " the general structure of Myliusia grayi, although convoluted, is massive and
labyrinthic throughout, not cup-shaped or lioUow in the axis, as that of Ventriculites,
while Cceloptychium consists of radiating tubes, more or less branches round a hollow
axis or stem, which in the horizontal section resembles Ventrictdites."

On the other hand, SoUas described in the same volume of the Annals in 1877 a new
fossil genus Stauronema, and assigned Myliusia grayi to the Ventriculitidse, which are
characterised by a " skeletal network having the nodes complicated by the presence of
an octahedral lantern about each one."

In 1878 Marshall and Meyer' described a Hexactinellid from the Philippines closely
related to, yet specifically distinct from, Myliusia grayi, Bowerbank. This they named
Myliusia zittelii.

Of the three specimens examined, the best preserved exhibits a system of meandering
folded funnels which have fused with one another, besides wider tubes with walls sup-

' Mittheil. d. Zool. Museum, Dresden.

REPORT ON THE HEXACTINELLIDA. 3r)9

])()rted by three to five layers of regularly coalescent hcxradiate forms enclosing culjical
meshes. Here, too, as in Myliuda grayi and in many fossil sponges, each ray of the
liexradiate spicules is united at a certain distance from the node of intersection with one
of the four adjoining rays by means of a siliceous bridge passing off at an angle of 45°,
so that by means of these the appearance of an octahedral lantern with an internal axinl
cross is produced. Marshall, therefore, proposed the expressive name "lantern spicules."
As to the oblique bridges between the arms of the central sexradiate spicules, Marshall
was able to demonstrate their origin from the fusion of lateral hooks and prickles, which
had become apposed to one another.

Among free spicules Marshall and Meyer found in the interior of the adult tissue
scattered rod-like forms slightly bent at one end, and very small hexradiate rosettes,
which bear on the extremity of each ray from four to six teeth, arranged like tulip petals.
(Jn the other hand, the covering of the soft parts contains on both sides of the tube-walls
numerous four-, five-, and six-rayed spicules. On the upper slender margin of the walls of
th(^ tubes there was in some places a dense palisade row of smooth uniaxials, which are
probably to be explained as a peristomial wreath. The relationship between Myliusia
and Cceloptychium was particularly emphasised by JMarshall and Meyer, but this
difference was noted, that Cceloptychium is monozoic, Myliusia, on the other hand,
polyzoic.

Oscar Schmidt also found the same species in the West Indian and Mexican region.'
He adds to Marshall's description the observation that many nodes of intersection in the
dictyonal framework of the specimens examined by him are unperforated. He found,
however, a central cavity with which a markedly irregular plexus of externally opening-
tubes communicated. The meandering course of these tubes often makes it difticult to
distinguish the canals and intercanals. Among the free skeletal elements lie noted
" prism rosettes."

Schmidt regarded Myliusia as a Cystisj)ongia without a covering layer, and provided
with conspicuous lantern nodes in the lattice framework.

To the genus Cystispongia, Roemer, Oscar Schmidt has assigned ;i new living form—
Cystisjyongia siiperstes. He has given a careful account of the characteristics of the
genus Cystispongia, established by Roemer, and more accurately defined by Zittel, and of
the single known species Cystispongia bursa, from the chalk, as also of the living
Cystispongia superstes added by him.

1. Aulocystis grayi (Bowerhank), (PL CIV. fig. 7).

There is in the British Museum a dried round specimen about 2 cm. in breadth,
and 1^ cm. in height, which Gray first described (1859 and 1867) as Myliusia

' Spongien dea Meerbusens von Mexico, p. 52.

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

callocyathus, but which Bowerbank afterwards (1869) distinguished as Myliusia (jrayi
According to the label the form was found at St. Vincent, West Indies. Bowerl)arik
briefly characterised it^ in the following diagnosis: — " Sponge sessile, massive. Dermal
surface unknown. Surface of rigid skeleton uneven and excavated. Oscula, jDores, and
e.xpansile dermal system unknown. Skeleton stratified, forming a series of expanded
• lypt-like spaces. Fibre cylindrical, incipiently or minutely spinous. Interstitial spicula
numerous, accratc, large and long, variable in size ; disj)osed in lines at right angles to
the stratification in loose fasciculi of two to four or five together. Eetentive spicula
spiculo-multifurcate hexradiate stellate." Bowerbank also gave^ good figures of several
specimens. From these two figures and from preparations which I was able to make
during my stay in the British Museum at South Kensington, it can be distinctly seen that
the dictyonal framework of Aulocystis grayi differs in several particulars from Aulocy.stin
zittelii, Marshall. The most important difference consists in this, that on each of the
nodes of intersection the octahedral edges are formed not of twelve simple, cylindrical,
oblique buttresses, as in Aulocystis zittelii, but of twelve plates lying in the plane of the
two connected axes, and perforated by several round or oval smooth margined holes of
variable size. These plates expand to some extent in their different planes, so that in
some places, especially on the free bounding surfaces of tlie entire dictyonal framework,
conspicuous perforated siliceous membranes may be formed, as represented in one of
Bowerbank's figures.'' The small conical tubercles which occur all round the beams in
Aulocystis zittelii, are here present in abundance on the freely projecting conical bosses,
but elsewhere only on the edges of the perforated plates, and less abundantly, or not
at all on the lateral surfaces of the same. The " long acerate interstitial spicules " oi
Bowerbank are slender, smooth, cylindrical needles which are disposed at right angles to
the surface, and are in my opinion not free oxydiact parenchymalia, but the very much
elongated proximal radial rays of the pentact dermalia.

Isolated pai-enchjTnalia are represented by numerous discohexasters with short
principal rays, varying in size and in the number, strength, and length of the terminals.
One of these is represented in PI. CIV. fig. 7. Besides these I frequently observed
bundles of those long and extremely fine terminal rays which characterise the graphio-
hexasters of Aulocystis zittelii. I found only a quite isolated occurrence of small simple
oxyhexasters. I did not observe that special form of discohexaster with medium-sized
principals which is so abundant in Aulocystis zittelii (PL CIV. fig. 6).

Consequently we are led to regard Aulocystis grayi, Bowerbank, as a form nearly
related to Aulocystis zittelii, but yet different enough to be referred to a distinct species.

1 Proc. Zool. Soc. Land., 1869, p. 335.

2 Loc. cit., pi. xxiii. fig. 8, and pi. xxv. fig. 1.
^ Loc. cit., pi. .xxiii. fig. 8.

REPORT ON THE HEXACTINELLIDA. 361

2. Auloci/stis zittelii (Marshall) (PI. CIV. figs. 1-6).

Two specimens of this remarkable Mseandrospongid were included in the rich collec-
tion of Hexactinellids trawled by the Challenger at Station 192, Little Ki Island, from
a depth of 140 fathoms and a l)lae mud ground. They are fairly well preserved in spirit,
have an oval form, and are about as large as a little man's fist. A smaller third specimen,
dead but intact in form, was trawled off Banda Islands (Station 194a, lat. 4° 31' S., long.
129° 57' 20" E.), from a depth of 360 fathoms and a volcanic mud ground. In form
and size it resembles a hen's egg (PI. CIV. fig. 2). The better preserved of the two
other specimens is figured on PI. CIV. fig. 1, in its natural size. It forms an oval, some-
what bulbous mass, 11 cm. in length, and 6 '5 in greatest breadth. The outer surface
consists of a thin, smooth, soft covering plate. In certain regions this plate exhibits
irregular stellate clefts, while at others thin, translucent, and fine sieve-like perforated
regions are seen, with an interjacent connected network of narrow, not translucent zones.
Since not all the portions of this covering are preserved, it cannot be determined whether
a round oscular aperture, which is suggested on the other specimens, is here really absent.
Nor is it possible to discover the mode by which the sponge was fixed to its solid sub-
stratum, or its original position.

Traversing the entire internal space of this capsule is a well-preserved framework of
irregular round tubes, a finger's breath in diameter, with firm walls 1 to 2 mm. in thick-
ness, and exhibiting on both surfaces numerous small groove-like depressions of variable
depth and breadth. From a median central space or short main passage, tubes which
sometimes anastomose pass out radially to the external capsular plate, on which they are
directly inserted transversely. These zones of insertion correspond to the non-trans-
lucent zones of the capsular (covering plate, while the lumina of the tubes usually
correspond to one of the round regions bearing a stellate cleft. Between this connected
system of tubes there is a system of similarly anastomosing intercanals, which are covered
over externally by the thin finely-perforated regions of the covering capsular plate
(PI. CIV. fig. 1). Through the latter the water passes from the exterior into the inter-
canalicular system, and through the tube-walls into the lumen of the anastomosing
system of canal tubes, whence it regains the exterior through the clefts of the covering
plate or through the oscular opening.

The connected dictyonal framework which supports the wall of the anastomosing tubes,
is characterised by the great regularity of the lattice of beams which surrounds the
strictly square meshes and corresponding cubical spaces. The simple nodes of inter-
section are furnished with twelve beams representing the edges of a regular octahedron.
These beams extend as oblique buttresses between each two adjacent beams of the
framework, at an angle of 45", and at equal distances from the node of intersection. Both
these buttresses and the freely projecting portions of the beams of the general framework

(ZOOL. CHALL. KXP. — PART LIII. — 1887.) Ggg 46

362 THE VOYAGE OF H.M.S. (JHALLENGEl^.

;ire furnished with numerous small tubercles and pointed elevations. The portions of the
main beams, on the other hand, which lie within the space enclosed by the buttresses are
perfectly smooth, and are not inconsiderably thinner than the outer portions of the
beams (PI. CIV. fig. 3). As Marshall and Meyer have noted in detail in the Mittheil.
d. Zool. Museums, Dresden (1877), the oblique buttresses round about the nodes of
intersection are really spine-like processes of the main beams, growing out in a given
direction from the external surface of two adjacent beams, until they finally meet and
fuse with one another. That they are in fact simple spines is further demonstrated
by the fact, which Marshall and Meyer have emphasised, that they have no axial canals,
which can be distinctly detected, on the other hand, in all the beams of the main frame-
work, and even on those thin portions which are enclosed by the lantern -like arrangement
of the buttresses.

It need hardly be noted that here also the entire quadrate lattice-work is built up of
single hexacts, in which the corresponding parallel rays are enclosed in a common
siliceous sheath, and thus united.

Here and there, especially near the bounding surfaces, there are also simple, solid,
slightly thickened nodes of intersection in the dictyonal framework.

The covering plate which encloses the sponge as in a capsule, apparently
arises from the outward bending of the tube-wall at right angles, or in a trumpet-like
curve. This is distinctly seen in the macerated specimen represented in PI. CIV. fig. 2.
In the two larger specimens it contains no connected skeleton, while in that trawled off"
Banda Islands (PI. CIV. fig. 2), it is supported over a large extent by a dictyonal frame-
work, which corresponds exactly to that of the tubular network, and is in fact a direct
continuation of the latter.

In this form, better than in those obtained off" Little Ki Island, a double principal
canal can be seen, which seems to have opened with a free margin at the narrower end
(PL CIV. fig. 2).

Sections through the wall of the tube show distinctly in many cases the structure
of the soft parts. The deeply folded chamber layer is connected to the dermal
membrane by a somewhat uniformly developed external trabecular framework, which
extends also into the afferent clefts and passages. The internal trabecular, on the other
hand, only extend between the sieve-like gastral membrane and the internally pro-
jecting folds of the chamber-layer (PL CIV. fig. 3), without being to any marked extent
continued into the eff'erent canals. The chambers themselves are comparatively small
and simply thimble-shaped.

The capsule has an average thickness of 0'3 mm., and exhibits between the outer
and the inner porous limiting membrane numerous passages and vesicular cavities, but
in the Little Ki specimens, at least, no chambers.

As to loose spicules, I found in the parenchyma between the chambers in the tra-

REPORT ON THE HEXACTINELLIDA. 363

becular network : — (1) sim^ile oxyhexasters which are always beset with pointed spines ;
(2) discohexasters of various sorts, including isolated forms with short principal rays
and six to ten long markedly divergent terminals, in some cases strongly developed
(PI. CIV. fig. 7), and in others quite thin and delicate (PI. CIV. fig. 5). There is, on
the other hand, a very abundant occurrence of discohexasters in which each of the long-
principals bears a tuft of six to ten, slightly divergent, fine, moderately long terminals
(PI. CIV. fig. 6). Here and there (3) graphiohexasters occur, with short principal rays
and very long, extremely delicate, slightly undulating terminals, which form a loose
divergent tuft on each principal (PI. CIV. fig. 4). In some places as a support to the
dermal membrane of the tubes, moderately strong oxypentacts occur with a hypodermal
radial ray, usually somewhat elongated. While the internal portion of their rays is
usually smooth, the outer ends are more or less abundantly beset witli pointed tubercles.
A rounded-oft' rudiment of a sixth (distal radial) ray is of very frequent occurrence, and of
variable size. In structure and size resembling these hypodermalia, hypogastral pentacts
occur with their tangential rays in the gastral membrane. On the external limiting
membrane of the capsule there is a layer of hypodermal pentacts- which entirely resemble
those above described. On the inner surface of the capsule there is in many places no
special skeleton, while other regions exhibit a layer of large oxypentacts with toothed
ends to the rays, and arranged in contrast to the former pentacts with the radial
unpaired ray directed outwards. The strands of tissue between the two parallel limiting
lamellse of the capsule contain either no spicules at all, or in certain positions somewhat
long straight rough diacts, wdth central swelling, and pointed or slightly rounded toothed
extremities.

BRIEF SYSTEMATIC SURVEY OF THE HEXACTINELLIDA
DESCRIBED IN THIS REPORT.

Order HEXACTINELLIDA or TRIAXONIA.

Sponges with very loose soft parts, and with siliceous spicules, which are either
isolated or united into a connected framework, and lielona; or are reducible to the triaxial
system.

Suborder I. LYSSACINA, Zittel.

Hexactinellida in which the spicules either remain altogether isolated, or are in part
subsequently and irregularly united by siliceous masses or transverse synapticula.

Tribe I. HEXASTEEOPHOEA, F. E. S.

Hexasters are always present in the parenchyma. The ciliated chambers are sharply
separated from one another, and are thimble-shaped.

Family I. Euplectellid/E, Gray.
The dermal skeleton contains sword-shaped oxyhexacts with long proximal ra5^

Subfamily 1. Euplectellin.e, F. E. S.

Tubular forms with transverse terminal sieve-plate. The lateral wall exhibits
circular parietal apertures. The projecting distal ray of the dagger-shaped hexact
dermalia bears a floricome.

Genus 1. Euplectella, Owen.
With basal root-tuft. Parenchymal oxyhexasters.

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

Species 1. Euplectella asi^evgillum, Owen.

Someivhat curved tube, slightly widened towards the iqyper end, bearing external
ridges and cuff-like fringe. Circular parietal gaps alternate with closed meshes, and
are disposed in spiral rows, ascending at an angle of 4.5°. The principal supporting
spicules of the square-meshed framework are oxytetracts. The annular membrane round
the parietal gaps contains strongly developed compressed oxypentacts. Zebu (Philippines) ;
95 fathoms.

Species 2. Euplectella suherea, Wyville Thomson.

Straight, sonieivhat bulging tube, without ridges or cuff. Circular parietal gaps
alternate with closed arched meshes, and are spirally arranged, ascending at an angle of
45° The principal supporting spicules of the loosely united meshed framework are
strongly developed oxypentacts with distal radial ray. The annular memhrane of the
parietal gaps contains straight knotted diacts and sceptres. West of Gibraltar, 600 to
1090 fathoms ; north-east of Bahia, 1600 fathoms; off the Berlingues, 1600 fathoms.

Species 3. Eupleetella cucumer, R. Owen.

Straight, somewhat bulging tube, without ridges or cuff. Circular parietal gaps
alternate with closed arched projecting meshes, and are disposed in spiral rows, ascending
at an angle of 45°. The summit of each protruding arched mesh bears the projecting
distal ray of a strongly develop)ed oxyhexaster. Seychelle Islands.

Species 4. Euplectella jovis, 0. Schmidt.

Straight tube, somewhat widened towards the upper end, without external ridges,
but with a terminal marginal cuff. Round parietal gaps occur in approximately regular
.spiral rows. The principal supporting spicules of the square-meshed framework are
strongly developed oxypentacts, with long projecting distal ray. The annular mem-
brane of the parietal gaps contains sceptres and S-shaped clasps. Antille Islands, 42.3
fathoms.

Species 5. Euplectella oweni, Marshall.

Straight phallus-shaped tube, with oval cross section, without ridges or cuff. Cir-
cular parietal gaps arranged in longitudincd and transverse rows. The principal
supporting spicules of the square-meshed framework are strong oxytetracts. The annular
membrane of the parietal gaps contains compass-shaped oxydiacts disposed tangentially.
Japan.

REPORT ON THE HEXACTINELLIDA. 367

Species 6. Euplectella crassistellata, u. sp.

The priucipal parenchymal spicules (priucipalia) arc long smooth oxydiacts with
knotted thickening in the centre. The parenchymal oxyhexasters have strikingly thick
and somewhat long principal rays, each with four moderately short terminals. Middle
of Pacific, 2750 fathoms.

Species 7. Euplectella (?*) nodosa, n. sp.

Straight slightly bulging tube (with irregular parietal gaps), without external ridges.
The principalia are strong oxypeutacts, in which the distal ray does not project Ijeyoud
the external surface. In some dermal oxypentacts the floricome is replaced l^y a tuft of
oxydiacts with central iiodes. Instead of the ordinary oxyhexasters, lophohexasters
occur. Possibly belonging to the genus Holascus. Bermuda Islands.

Genus 2. Regadrella, 0. Schmidt.
A tubular form firmly attached by means of a knobbed basal portion.

Species 1. Regadrella phcenix, 0. Schmidt.

The principalia are large oxy|3entacts, which have their four tangential rays variously
disposed in difi"erent directions. Parenchymal discohexasters with short principal and
long terminal rays. Antille Islands, 221 to 280 fathoms.

Subfamily 2. Holascin^.

Tubular. Without parietal yaps. Without superficially situated floricomes.
Parenchymal oxyhexasters.

Genus 1. Holascus, n. gen.

With root-tuft and transverse sharply defined terminal sieve-plate. Tlie internal
gastral surface bears projecting longitudinal and transverse lattice-work ridges. Hypo-
dermalia, sword-shaped hexacts with short rough distal ray, l)esides which ox5'diacts an;
radially apposed.

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

Species 1. Holascus stcllatus, n. sp.

Priucipalia arc oxypeiitacts with distal radial ray, or rarely oxyhexacts. The paren-
(;hyma includes oxyhexasters, some with straight, and others with long curved terminals.
Graphiohexasters and probably discohexasters also occur. The hypogastralia resemljle
the hypodermalia. Buenos Ayres, 2650 fathoms.

Species 2. Holascus Jibulatus, n. sp.

The principalia are oxytefcracts. Small oxyhexacts more externally. The paren-
chyma includes, besides' the oxyhexasters, numerous markedly curved two- (or frequently
three-) rayed fibulae. The gastralia are oxypentacts without inwardly projecting radial
ray. South of Australia, 2600 fathoms ; west of Crozet Islands, 1375 to 1600 fathoms.

Species 3. Holascus polejacvii, n. sp.

The principalia are oxyhexacts with long, sometimes curved, tangential rays. The
parenchyma contains, besides the simple oxyhexasters, other forms in which the principal
rays are externally thickened, and bear tufts of strongly developed terminals. The
gastralia are somewhat large oxypentacts without freely projecting radial ray. South
of Australia, 1950 fathoms.

Species. 4. Holascus ridleyi, u. sp.

The principalia are substantial oxytetracts. The parenchyma includes, besides the
ordinaiy oxyhexasters, other forms in which the principal rays are much thickened
externally, and bear tufts of strong terminals. Simple oxyhexasters with curved terminal
rays have also an occasional occurrence. The gastralia are slender oxyhexacts. Philip-
pines, 2225 fiithoms.

Genus 2. Malcosaccus, n. gen.

Sack-like or tubular forms with loose walls, externally somewhat smooth, and internally
honeycoml^ed. The principalia are oxyhexacts with very long, thin, pliable tangential
rays. The hypodermalia and hypogastralia are sword-shaped hexacts with rough pro-
jecting ray, which perhaps bears a floricome.

Species 1. Malacosaccus vastus, n. sp.

The parenchyma contains oxyhexasters with very fine long terminal rays, and small
discohexasters with many terminals on the transverse disc-shaped terminal expansion
of each principal ray. West of Crozet Islands, 1375 fathoms.

REPORT OX THE HEXACTINELLIDA. 369

Species 2. Malacosaccus unguiculatus, n. sp.

Besides numerous oxyhexasters with fine straight or curved terminal rays, the
parenchyma contains discohexasters with three or four long and slightly curved terminal
rays on each principal. The terminal rays of the floricomes only exhibit two to three
strongly developed terminal claws. South of Sierra Leone, 2450 fathoms.

Subfamily 3. T.egerin.e, F. E. S.

The thin wall of the sack-like or tubular body is penetrated by parietal gaps of
irregular shape and distribution. The skeletal lattice-work of the wall usually forms an
irregular meshwork. Each projecting distal ray of the sword-shaped hypodermalia bears
u fioricome.

Genus 1. Tsegeria, n. gen.

With the single species, Taigeria pulchra, n. sp.

A rigid sack-shaped liody firmly attached by means of a knotted base. Irregular
roundish parietal gaps. The superior round terminal aperture exhibits an external
wreath of short straight, and an inner crown of long arched spicules. The principalia
are oxytetracts, oxytriacts, and oxydiacts. The parenchyma contains discohexasters and
discohexacts. Graphiohexasters occur near the outer skin. The dermal skeleton contains
slender hypodermal oxyhexacts. The gastral skeleton consists of strongly developed
pentacts. Fiji Islands, 610 fathoms.

Genus 2. Walteria, n. gen.

With the single species, Walteria fiemmingii, n. sp.

The wall of the sack-shaped body consists of a lattice-work with irregularly angular
meshes of various size and form. It is continued at one end into a funnel with square
meshes. The distal ray of the hypodermalia is thickened and rounded off. The
parenchyma contains discohexasters with many, and others with few terminal rays. The
Horicomes have numerous (fifteen) terminal rays on each principal. North of Kermadec
Islands, 630 fathoms.

(ZOOL. CH.^LL. EXP. — PART LIII. — 1887.) Ggg 47

'370 THE VOYAGE OF H.M.S. CHALLENGER.

To the Euplectellidse are also to be referred the following seven genera, which have
not as yet been sufficiently investigated.

Genus 1. Hahroclictyum, Wyville Thomson.

With the single species, Hahroclictyum speciosum (Quoy and Gaimard).

Tubular forms with irregular parietal network, firmly attached by a knotted basal
portion. The lateral wall is directly continued without definite margin or cuff into the
uniform gently arched terminal cupola. The parenchyma contains oxyhexasters with
medium-sized principal rays and short terminals. Molucca Islands.

Genus 2. Eudictyum, Marshall.

With the single species, Eudictyum elegans, Marshall.

The terminal sieve-plate is distinct indeed from the lattice-work of the tubular parietal
skeleton, but is not markedly different. The parenchyma contains small discohexacts.

Genus 3. Dictyocalyx, n. gen.

With the single species, Dictyocalyx gracilis, n. sp.

A funnel-shaped skeletal network with altogether irregular meshes, firmly attached
to a substratum by means of a solid compact stalk. The parenchyma includes several
kinds of discohexasters, some of which bear on their terminal rays marginally toothed
convex discs, and others lacerate campanulate terminal umbels. South Pacific, 2385
fathoms.

Genus 4. Rhahdodictyum, 0. Schmidt.

With the single species, Rhahdodictyum delicatum, 0. Schmidt.

A very lank almost tubular cup, borne on a small compact basal plate. The parietal
skeleton is traversed by round smooth holes, and consists of irregularly disposed long-
rayed hexacts which are cemented together. Bequia, 1591 fathoms.

Genus 5. Rlmhdopectella, 0. Schmidt.

With the single species, Rhahdopectella tintinnus, 0. Schmidt.

A firmly attached cup with skeletal beams cemented together below, but consisting
superiorly of free and isolated spicules. The parenchyma contains the usual oxyhexasters,
and also other forms in which the strong terminal rays appear to spring from a central

REPORT ON THE HEXACTINELLIDA. 371

node as the result of the disappearance of the principals. Discohexasters also occur in
which the long lank curved terminals bear on their thickened external end a marginally-
toothed convex disc. Other discohexasters occur with five terminal rays on each
moderately short principal. Of the five lacerate campanulate terminal discs of these
terminal rays, the four outer are always markedly larger than the central. Rough, some-
what spirally curved diacts also occur. Antille Islands, 994 fathoms.

Genus 6. Hertwigia, 0. Schmidt.

With the single species, Hertivigia falcifera, 0. Schmidt.

An irregular lattice-work labyrinth is borne on a firmly attached knotted base. Some
of the parenchymal hexasters bear on each of the short principal rays four sickle-shaped
hooks representing terminals. Antille Islands, 611 fathoms.

Genus 7. Hyalostylus, u. gen.

With the single species, Hyalostylus dives, n. sp.

A long slender stalk bears a soft folded cup of a somewhat flattened, or triangular
bilaterally symmetrical form. The two larger smooth lateral surfaces pass by a rounded
edge into one another, while the third narrower side is folded. Besides thread-like diacts
with swollen roughened ends, the parenchyma includes four diff"erent forms of rosettes,
in which the terminal rays either terminate in long cylindrical clubs, or in terminal
claws with convex terminal discs. Besides these, rough slightly spiral diacts occur.
South Pacific, 2550 fathoms.

Family II. A s c o N e m a T i D ^, (Gray).

The dermal and gastral skeletons contain pentact or hexact pinuli. The hypoder-
malia and hypogastralia are pentacts. AVith parenchymal discohexasters.

Subfamily 1. Asconematin>e, F. E. S.

Cup-, funnel-, or tube-shaped forms borne on a stalk. The wall has the form of a
thin loose plate.

Genus 1. Asconema, Sav. Kent.

With the single species, Asconema setuhalense, Sav. Kent.

Funnel-shaped form. The principalia are long diacts. Between these in the paren-

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

chyma small oxyhcxacts, oxyhexasters, and discohexasters occur. The pinuli are for the
most part pentaets. West of Portugal, Morocco aud Scotland, 200 to 400 fathoms.

Genus 2. Aulascus, n. gen.

With tlie single species, Aulascus joh^istoni, ii. sp.

Tubular form. The principalia are oxyhexacts and diacts. Between these in the
parenchyma discohexasters and single plumicomes occur. On the dermal and gastral
pinuli the ray turned towards the parenchyma is more or less developed. Prince Edward
Islands, 310 fathoms.

Subfamily 2. Sympagellin^, F. E. S.

Ovoid, thick- walled, usually (or always?) stalked goblets with smooth, thin superior
margin. The parenchyma between the principal hexacts and diacts contains small
discohexasters.

Genus 1. Sympagella, 0. Schmidt.

With the single species, Sympagella nux, 0. Schmidt.

On the terminal branches of a ramified stock ellipsoidal goblets are borne. The
parenchyma contains, besides isolated plumicomes, numerous discohexasters with two to
four delicate terminal rays on each principal, and also small elongated stars. Florida,
98 to 123 fathoms; coasts of Spain and Portugal; Cape Verde Islands, 100 to 128
fathoms.

Genus 2. Polyrliahdus, n. gen.

With the single species, Polyrhahdus oviformis, u. sp.

Unstalked (?). Egg-shaped goblet slightly narrowed superiorly. The dermal pinuli
are hexacts with thick scaly distal ray. The parenchyma contains discohexasters with
numerous terminal rays. South Indian Ocean, 1975 fathoms.

Genus 3. Balanites, n. gen.

With the single species, Balanites pipetta, n. sp.

Obliquely stalked. The parenchyma contains numerous small discohexacts. The
dermal and gastral pinuli are hexacts with compressed, scaly, freely projecting ray. On
the superior oscular margin they are directly continuous. Besides the small discohexacts,
the parenchyma also includes discohexasters with long principal rays, which bear a
tuft of short terminals. South Indian Ocean, 1950 fathoms.

REPORT ON THE HEXACTINELLIDA. 373

Subfamily 3. Caulophacin.e, F. E. S.
Mushroom-like, with long, cylindrical, hollow stalk.

Genus 1. Caulophacus, n. gen.

The dermal surface of the body, which is flattened from above downwards, is either
slightly convex or has become concave by the downward folding of the marginal portion
of the flat body, The projecting ray in the dermal pinuli is short and broad, but long
and weakly developed in the gastral.

Species 1. Caulophacus latus, n. sp.

The body consists of a flat circular disc with a downward-folded sharp-edged
margin. The long cylindrical stalk passes by a trumpet-shaped expansion into the
middle portion of the disc. The parenchyma contains discohexacts, oxyhexasters, and
numerous discohexasters. The short dermal pinuli are almost all hexacts, the long
gastral pinuli are, on the other hand, pentacts. West of Crozet Islands, 1600
fathoms.

Species 2. Caulophacus elegans, n. sp.

The disc is either biconvex, or with down-turned marginal portion convex-concave.
The parenchyma contains, besides numerous discohexacts, also four-rayed discohexasters.
The dermal and gastral pinuli are hexacts. The radial rays of the hypodermal pentacts
are toothed. East of Japan, 2300 fathoms.

Genus 2. Trachycaulus, n. gen.

With the single species Trachycaulus gurlittii, n. sp.

The long stalk is covered with long pointed dermal pentact pinuli, and contains as
principalia long diacts which lie parallel in longitudinal series, and are firmly united by
means of synapticula. Subdermally large sickle rosettes occur, with four terminal rays on
each principal. South Pacific, 2550 fathoms.

374 THE VOYAGE OF H.iM.S. CHALLENGER.

Family III. E o s s E L l i d ^, F. E. S.
The dermalia are always without a distal radial ray.

Genus 1. Lanuginella, O. Schmidt.

With the single species, Lanuginella pupa, 0. Schmidt.

The body forms a cocoon-shaped, thick-walled sack with superior circular aperture,
and is directly attached by its blunt lower end. The parenchyma contains delicate
discohexasters, plumicomes, and small discohexasters with numerous terminal rays on
the broad terminal discs of the principals. Cape Verde Islands ; Little Ki Island, 140
fathoms.

Genus 2. Polylophus, n. gen.

With the single species, Polylaphus philipinnensis (Gray).

The cup-shaped, thick-walled body, provided with a wide circular oscular aperture, is
rooted in the mud by means of a basal tuft. The external surface bears spherical
papillae, from the rounded summit of which a tuft of long pleuralia projects. From the
papillae buds are frequently developed. The parenchyma contains numerous oxy-
hexasters with long rough principal rays, each with three long markedly diverging
terminals. Single plumicomes also occur. In the skin, above the medium-sized hypo-
dermal oxypentacts, there lie small, rough, somewhat incurved, cruciate autodermal
tetracts. The pleuralia and basalia pass inferiorly into anchors with four slightly curved
transverse teeth. Little Ki Island, 140 fathoms.

Genus 3. Rossdla, Carter.

Thick-walled goblets of an egg- or barrel-like form, with circular oscular aperture
and deep gastral cavity. From regularly distributed small boss-like elevations of the
external surface a group of diact and pentact pleuralia project radially outwards, and by
the association of their tangential rays form a kind of veil. By a deep but regular and
simple folding of the chamber layer, alternating afferent and efferent, narrow, funnel-
shaped, radial canals are formed. The parenchyma contains oxyhexasters with very
short principal rays and various discohexasters. In the dermal membrane rough pentacts
almost exclusively occur.

Species 1. Rossella antarctica, Carter.

Elongated egg- or barrel-shaped forms, firmly fixed or rooted among small stones by
means of short processes. The tangential rays of the pleural pentacts are so displaced to

REPORT ON THE HEXACTINELLIDA. 375

one side that the outermost do not include much more than a right angle. The paren-
chyma contains oxyhexasters with long terminal rays disposed in a perianth-like
fashion. South-east of Prince Edward Island, 140 fathoms; south of Kerguelen, 150
fathoms; east of Buenos Ayres, 600 fathoms.

Species 2. Rossella velata, Wyville Thomson.

At the base tufts of long basalia project like the pleural tufts, and unite into a root
anchoring the sponge in the mud. The four tangential rays of the pleuralia are crossed
at risht angles. The basalia end in anchors, the four teeth of which are somewhat
markedly recurved. The parenchyma contains plumicomes. West of Gibraltar,
651 fathoms.

Genus 4. Accmthascus, n. gen.

Thick-walled goblets firmly attached by the blind lower end. The gastral cavitj-
opens superiorly by a simple smooth-margined round oscular aperture. Diact pleuralia
project radially on the sides. The parenchyma contains oxyhexasters and discohexasters
in varied form. The dermalia are small rough tetracts and pentacts.

Species 1. Accmthascus grossularia, n. sp.

The pleuralia project as isolated spicules, so that the egg-shaped body comes to
resemble a gooseberry. The parenchyma contains discohexasters in which numerous
terminals are borne on the broad terminal plate of the principal rays, while others have
the long terminals arranged in perianth-like fashion. The dermalia are for the most
part tetracts, though pentacts also occur. The gastralia are spinose oxyhexaets. Posses-
sion Island, 210 fathoms.

Species 2. Accmthascus duhius, u. sp.

The parenchyma contains oxyhexasters in which the long terminal rays seem to
spring directly from the central node, as the result of a marked shortening of the prin-
cipals. Besides these, discohexasters occur with short principal rays. The dermalia are
pentacts, the gastralia are rough oxyhexaets. South of Puerto Bueno, in Patagonia, 400
fathoms.

Species 3. Accmthascus cactus, n. sp.

From numerous gentle elevations on the external surface, tufts of radially disposed
oxy diact pleuralia project. The parenchyma contains discohexasters with eight or more

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

rays, which break up at their extremities into a brush-like tuft of terminals, and also
small discohexasters, in which numerous terminals arise from the disc-shaped expansion
of the principal rays. The dermal and gastral skeletons contain rough tetracts and
pentacts. Japan.

Genus 5. Bathydorus, n. gen.

Sack- or wine-skin-like forms with loose walls, and thin (always V) oscular margin pro-
vided with a cuff-like marginal fringe of spicules. The parenchyma contains, besides large
hexacts and diacts, also oxyhexasters, and in some species discohexasters. In the dermal
skeleton, besides the smooth hypodermal oxypentacts, numerous rough autodermal oxyte-
tracts occur, and in some also autodermal diacts and even monacts. The gastral skeleton
contains exclusively rough oxyhexacts.

Species 1. Bathydorus Jimhriahis, u. sp.

A large smooth wine-skin-like form, in which the thin superior oscular margin is
equipped with a wreath of oxydiact marginalia. The parenchyma contains, besides
the principal diacts, oxyhexasters with long frequently S-shaped terminal rays. The
autodermalia are rough cruciate oxytetracts ; the gastralia are rough or spinose ()xy-
hexacts. North Pacific, 2300 to 2900 fathoms.

Species 2. Bathydorus stellatus, n. sp.

From the external surface of the sack-shaped body, isolated oxydiacts project radially
or obliquely outwards. The parenchyma contains numerous stars. The dermal and
gastral skeletons resemble those of Bathydorus Jimbriatus. Messier Channel, Patagonia,
140 fathoms.

Species 3. Bathydorus spinosus, n. sp.

Externally like Bathydorus stellatus. The oscular margin bears a wreath of spicules.
The parenchyma contains oxyhexasters with long, somewhat wavy terminal rays. The
autodermalia are rough oxytetracts, the gastralia oxyhexacts of varying size, sometimes
rough, sometimes smooth. Penguin Island, 1600 fathoms.

Species 4. Bathydorus hacxdifer, n. sp.

Externally like Bathydorus spinosus. The parenchyma contains oxyhexasters and
discohexasters. The dermal skeleton includes numerous rough autodermal diacts and
monacts. In the middle of the South Pacific, 2335 fathoms.

REPORT ON THE HEXACTINELLIDA. :J77

Genus 6. Rhahdocalyptus, n. gen.

Moderately thin-walled cups with smooth external surface,, and sharp-edged smooth
margin to the wide oscular aperture. The inner surface exhibits holes of variable size —
the apertures of the eflferent canals. The principal parenchymalia are diacts of variable
length. The parenchyma also contains discohexasters and oxyhexasters, and further
eight-rayed rosettes with disc-bearing terminal rays on each of the medium-sized
principals. The dermal membrane contains rough diacts either exclusively, or plus
rough pentacts, tetracts, and monacts. The gastralia are rough oxyhexacts.

Species 1. RhabdocalyjJtus mollis, n. sp.

A funnel-shaped, somewhat compressed cup with a short round stalk. The paren-
chyma contains small oxyhexacts in which the rays near the point of intersection are
beset with inward directed spines. The rays are frequently curved and may be also re-
duced to four or two. The dermal membrane contains only rough diacts, and the gastral
membrane only rough hexacts. Japan, 100-200 fathoms.

Species 2. Rhahdocalyptus roeperi, n. sp.

A sack-shaped cup, which is directly attaiched by its thick blind end. The paren-
chyma contains oxyhexasters with short principal rays and long undulating terminals.
The dermal membrane contains rough pentacts, tetracts, diacts, and monacts. In the
gastral and canalicular membrane there are weakly developed hexacts with tubercled or
uniformly rough rays. South of Puerto Bueuo, Patagonia, 400 fathoms.

Genus 7. Crateromorpha, Gray.

Stalked goblets with smooth external surface. On the gastral internal surface the
L'tiereut canals open directly into the gastral space, with round apertures of variable
width. In the loose parenchyma, besides medium-sized hexacts and diacts oxyhexasters
and discohexasters occur of variable form and size. The dermal membrane contains
rough tetracts, pentacts, and sometimes also diacts, all three with club-shaped thickened
ends. The gastral membrane contains rough pentacts.

Species 1. Crateromoipha meyeri, Gray.

A tulip-shaped cup, varying in size from a hen's to a goose's egg, and seated on a
Ik of the length and thickness of one's little finger. On the oscular margin of the wide

(ZOOL. CHALL. EXP. I'AUT LIII. 1887.) Ggg 48

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

gastral cavity there is a thin skin-fringe, 4 mm. in height. The parenchyma contains
oxyhexasters with very short principals, and small discohexasters with many terminal
rays. The dermal membrane contains, besides many pentacts, some tetracts. The inner
surface of the efferent canals bears small rough oxyhexacts. The stalk is penetrated
by canals. Zebu (Philippines), 95 fathoms ; Japan.

Species 2. Crateromorpha thier/elderi, n. sp.

The hollow stalk, somewhat thicker than a goose-quill, passes gradually into a spindle-
shaped cup, about the size of a hen's egg. The sharp oscular margin bears a thin
dermal fringe. The outer surface of body and stalk is smooth ; the inner surface exhibits
the round variously-sized apertures of the efferent canals. The parenchyma contains
oxyhexasters with somewhat externally curved terminal rays and small discohexasters.
The dermal membrane contains rough tetracts and pentacts. Little Ki Island, 140
fathoms.

Species 3. Crateromorpha murrayi, n. sp.

A broad, superiorly campanulate cup larger than a man's fist, with a strong hollow
stalk as thick as a thumb, and a sharp smooth oscular margin. The external surface is
smooth ; the internal surface exhibits round, variously sized apertures of the efferent
canals. The parenchyma contains oxyhexasters with terminal rays somewhat curved at
their ends. The dermal membrane contains small rough pentacts, tetracts, and less
frequently diacts. The gastral membrane contains rough pentacts and tetracts. Little
Ki Island, 140 fathoms.

Species 4. Crateromorpha tumida, n. sp.

The superiorly expanded stalk bears an irregularly ridged cup, somewhat larger than
a fist. The supei'ior, irregularly frilled, smooth, and somewhat sharp oscular margin pro-
jects straight upwards. The parenchyma contains oxyhexasters in which the long
terminal rays are frequently curved at their extremities, and also discohexasters with
rather numerous Ions; terminals. The dermal membrane contains tetracts. and less
frequently diacts, or even triacts. Banda Islands, 360 fathoms.

Genus 8. Aidochone, n. gen.

Cylindrical or spherical cup-like forms, borne on long tubular stalks. The oscular
margin is folded backwards in such a way that a portion of the originally internal gastral

REPORT ON THE HEXACTINELLIDA. 379

wall has become the external wall of the body, while the other internal portion of the
gastral wall, which exhibits the roundish excurrent apertures of the efferent canals, is
continued directly into the internal surface of the stalk lumen. The fine quadratic
lattice-work of the dermal membrane lies on the inferior surface of the body. The
parenchyma contains between long diacts numerous oxyhexasters, and, though less
abundantly, also discohexasters. The dermal and gastral membranes exhibit a pre-
dominant or exclusive occurrence of small rough pentacts.

Species 1. Aulochone cylindrica, n. sp.

The cylindrical body, about the size of a child's fist, exhibits a funnel-shaped gastral
cavity, and on the superior everted margin of the gastral membrane a somewhat sharp,
undulating edge, while the lower or true oscular margin is more uniformly sharp and
smooth. Among the numerous oxyhexasters of the parenchyma, much dwarfed forms
occur. The discohexasters have numerous terminal rays arranged in a tuft on each
principal. Between the rough pentacts of the dermal membrane there are sometimes
also similarly formed tetracts. The gastral membrane contains exclusively rough
pentacts. North-east of the Kermadec Islands, 600 fathoms.

Species 2. Aulochone lilium, n. sp.

The lateral surface of the hemispherical, rather than cylindrical body is some-
what puffed out, and passes not so much by a superior edge as by an arched marginal
region into the funnel-shaped wall of the gastral cavity, while towards the dermal
inferior surface it is very distinctly defined by a sharp-edged oscular margin. The
parenchyma contains, besides numerous oxyhexasters with long rays, also discohexasters
with S-shaped terminals arranged in calyx-like fashion. Both dermal and gastral
membranes contain rough pentacts. Meangis Islands, north-east of Celebes, 500
fathoms.

Genus 9. Caulocalyx, n. gen.

With the single species, Caulocalyx tener, n. sp.

The simple cup-shaped body is gradually contracted inferiorly into a solid narrow stalk.
From the smooth lateral wall long isolated oxydiacts project radially. Between the long
diacts of the parenchyma discohexasters of variable strength project, in which the long
terminal rays are gradually thickened towards the outer end, and terminate in a margin-
ally toothed hemispherical transverse disc. The dermal skeleton contains oxypentacts
with spines on the four tangential rays. West of Tristan da Cuuha, 2025 ftithoms.

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

Genus 10. Atdocalyx, n. gen.

With the single .species Aulocalyx irregularis, n. sp.

A thiu-walled cup, much folded, extended into lateral diverticula, and also continued
into short laterally projecting tubes. The cup is fixed by a firm irregular base. The
connected framework of beams consists of much curved hexacts, partly united by synap-
ticula, partly soldered together. The parenchyma contains loose discohexasters wdth
short or with medium-sized principal rays, bearing S-shaped terminals disposed in
perianth-like fashion. Under the skin there are large hexasters in which each of the short
principal rays bears six long diverging terminals, which gradually increase in thickness
towards the round outer end, and are beset all round with backward bent pointed hooks.
The dermalia and gastralia are rough medium-sized oxypentacts. Marion Islands, south-
east of the Cape of Good Hope, 310 fathoms; between Marion and the Crozet Islands,
600 fathoms.

Genus 11. Euryplegma, n. gen.

With the single species Euryplegma auriculare, n. sp.

Cup- or ear-shaped plate, 3 to 5 mm. in thickness, with smooth convex outer surface,
penetrated, however, by numerous longitudinally disposed, oval incurrent apertures, and
a concave internal surface exhibiting numerous longitudinal ridges and papilla-like bosses
3 to 4 mm. in breadth. The excurrent apertures, 1 to 2 mm. in width, are found
between the longitudinal ridges and on the summit of each protuberance. The beams
of the very irregular dictyonal framework, which exhibits numerous ladder-like structures
formed by transverse synapticula, are beset with small scattered pointed tubercles, and
exhibit no thickening of the nodes of intersection. The loose parenchymal spicules are
small simple rough oxyhexacts and discohexasters of various kinds. Some of the latter
bear three to five somewhat markedly divergent terminals, while in others the moderately
short principals exhibit three to ten long S-shaped rays, united in a slender calyciue
arrangement.

The dermal and gastral skeletons consist exclusively of oxy^jentacts, which bear small
tubercles at the ends of the sharpened rays, but are otherwise smooth. Ofi' Eaoul Islands,
north-east of New Zealand, 630 fathoms.

Tribe II. AMPHIDISCOPHOEA, F. E. S.

Amphidiscs are always present in the limiting membranes. The parenchyma contains
no hexasters whatever. A basal tuft of filjres is always present, anchoring the cup- or

REPORT ON THE HEXACTINELLIDA. 381

club-shaped sponge in the mud. The chambers are not exactly thim1)lc-shapcd, nor
sharply marked off from one another, but form irregular diverticula of the membrana
reticularis.

Family H Y a L o N E M A T i D .E, Gray.
Both dermal and gastral membranes contain numerous pentact pinuli.

Subfamily 1. Hyalonematin^, F. E. S.

The compressed usually cup- or goblet-shaped body bears on the upper surface a
more or less sharply contoured round excurrent (oscular) region, and is only excep-
tionally split laterally.

Genus 1. Hyalonema, Gray.

At the lower pole of the funnel-shaped or more spherical body, there is a long, narrow,
sharply defined root-tuft, in which the spicules are at their lower end equipped with a four-
toothed anchor structure. No uncinates. The slender marginalia are superiorly pointed
diacts with toothed distal ray.

Subgenus 1. Hyalonema, s. str.

The superior aperture of the gastral cavity is covered Ijy a sieve network which
extends from the annular rim with its cuff-like fringe of fine marginalia, either flatly
over the whole gastral aperture, or sunk into a funnel-shaped depression.

Species 1. Hyalonema sieboldii, Gray.

The almost cylindrical, inferiorly rounded, superiorly truncate body shows at its
infeiior pole a twisted tuft of long spicules, almost as thick as a little finger. The
terminal sieve-plate exhibits a cruciate zone of imperforate skin, Ij'ing above the four
crossed radial septa, and bears in the middle a central cone.

The upper portion of the root-tuft is surrounded by an encrustation of Palythoa
fatua, Max Schultze, and other commensal Anthozoa are seated on the external skin. The
parenchyma contains small oxyhexacts with straight, and others with curved toothed
rays. The larger amphidiscs have broad arched umbels with eight rather broad paddle-
shaped rays. Japan.

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

Species 2. Hyalonema gracile, n. sp.

The inverted pear-shaped, superiorly truncate, inferiorly pointed body, is hardly as
large as a plum, and forms a slender untwisted basal tuft without a Palyihoa encrusta-
tion. The delicate terminal sieve-plate extends freely over the gastral cavity, which
exhibits internally a central columella and four cruciately disposed radial septa. The
parenchyma contains small oxyhexacts with straight, or curved smooth rays. The large
amphidiscs have moderately long arched terminal umbels with eight rather broad lancet-
shaped rays. Philippines, 2225 fathoms.

Species 3. Hyalonema divergens, n. sp.

The body has the form of an inverted, not belly-shaped, bell with the maximum breadth
at the superior, transversely truncated end. Over the flat gastral cavity with its free
central columellar cone, a delicate latticed network extends transversely. The basal,
somewhat narrowed end gives off a bundle of rather straight and thick spicules, forming a
basal tuft without a Palythoa encrustation. The parenchyma contains small oxyhexacts,
some with straight, and others with curved smooth rays. The large amphidiscs have
terminal umbels with somewhat transversely truncated extremity. The moderately long,
but not very broad uml^el-rays, with parallel margins, are almost straight and diverge
rather markedly. Mid-Pacific, east of Maldon Island, 2425 fathoms.

Species 4. Hyalonema toxeres, Wyville Thomson.

The spherical body, which is puffed out in the young forms, bears on its pointed
inferior end a brush-like diverging basal tuft with encrusting Palythoa on its upper
portion. Free terminal sieve-plate. The parenchyma contains small oxyhexacts, some
with straight, and others with curved rays, of which the latter are beset vdth barbs.
Large, thick, somewhat curved diacts also occur. Large amphidiscs like those of
Hyalonema sieboldii. Near St. Thomas, West Indies, 390 fathoms.

Species 5. Hyalonema kentii, 0. Schmidt.

Funnel-shaped body, with the sieve-plate intruding into the gastral cavity. The
parenchyma contains small weakly developed oxyhexacts, some with straight, others
with curved, almost smooth rays. Large amphidiscs with broad hemispherical terminal
umbels, in which the eight rays are broad lancet-shaped plates. West Indies, 300 to
1500 fathoms.

REPORT ON THE HEXACTINELLIDA. 383

Species 6. Ilyalonema poculum, n. sp.

Funnel-shaped. Sieve-plate bent inwards. The small parenchymal oxyhexacts have
all straight and somewhat rough rays. The large amphidiscs have hemispherical terminal
umbels and narrow pointed rays. Near the Island of Juan Fernandez, west of Valparaiso,
1375 fathoms.

Species 7. Hyalonema conus, n. sp.

Conical form. The sieve-plate somewhat flatly stretched. Parenchymal oxyhexacts
with straight smooth rays. Large amphidiscs with flat, broad, terminal umbels with
eight broad paddle-shaped rays. South of Australia, 1800 fathoms.

Subgenus. 2. Stylocalyx, n. subgeu.

The aperture of the gastral cavity is not covered by a common sieve-plate, but is quite
open. The gastral space is eruciately divided into four chambers by radial septa.

Species 1. Stylocalyx thomsonii, Marshall.

The body is approximately spindle-shaped. From the open aperture of the gastral
space the columellar cone projects for some distance as a long and narrow smooth conical
elevation, on which the four broad radial septa of the gastral space are extended upwards.
The somewhat twisted tuft of basal spicules is superiorly covered by a Palythoa
encrustation. The parenchyma contains small oxyhexacts with straight smooth rays.
North of Shetland Islands, 550 fathoms.

Species 2. Stylocalyx apertus, u. sp.

From the lower pole of the spherical or bulging body a slightly-twisted basal tuft
(5 mm. thick) projects, covered on its superior portion by a Palythoa crust. The apex
of the central cone does not quite reach the level of the oscular margin. The rays of the
small parenchymal oxyhexacts are curved and beset with barbs. The large amphidiscs
have hemispherically arched -terminal umbels with six broad, lancet-shaped radiating
rays. The dermal and gastral pinuli have short toothed basal rays, and a slightly
spinose radial. Sagami Bay, Japan, 345 fathoms.

Species 3. Stylocalyx depressus, n. sp.

The body is of a flat cake-shape or approximately spherical, and attains the size of a
man's fist. The oscular aperture is comparatively narrow, and surrounded by a smooth-

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

margined membrauous collar. The root-tuft bears uo Palythoa encrustation, aud consists
of tolerably straight spicules. The central cone rises to the opening of the gastral
cavity, which is traversed by four cruciate radial septa. The small oxyhexacts of the
parenchyma are almost wholly provided with curved rays, which Ijear small barljs.
Between these there are a few oxyhexacts with small barbs on the straight rays. The
laro-er amphidiscs bear somewhat long, lacerate, campanulate terminal umbels, with eight
slender pointed rays. The dermal pinuli have short, thick, rough basal rays, and a radial
beset with long lateral spines. The gastral and canalicular pinuli have long rough basal
rays, and a weakly developed radial with short thin lateral spines. North of Mellish
Islands, Mid-North Pacific, 2050 fathoms.

Species 4. Stylocalyx claviger, n. sp.

The oscular margin of the oval body is sharp. The small parenchymal hexacts have
delicate, smooth, usually quite straight rays. The large amphidiscs have broad, or some-
what highly arched terminal umbels, with eight broad paddle-shaped rays. The dermal
pinuli are compressed, with bushy superiorly somewhat truncated radial ray. The gastral
pinuli, on the other hand, have a long spindle-shaped solid radial ray, which is beset only
with short teeth and scales. Near the Penguin Islands, 1600 fathoms.

Species 5. Stylocalyx globus, n. sp.

The spherical body bears at the basal pole a slender root-tuft of straight needles,
without any Palythoa encrustation, while the superior pole exhibits a comparatively
small, circular, sharp-edged oscular aperture without a fringe of spicules. From
the free central cone four cruciate sharp-edged radial septa extend. The small paren-
chymal oxyhexacts have smooth, slender, straight rays. The larger amphidiscs have
broad terminally truncate, tolerably short, terminal umbels, in which eight straight
paddle-shaped rays diverge markedly outwards. The dermal pinuli have long, firm,
roughened basal rays, and a moderately short bushy radial ray, in which the strongly
developed axial portion bears long bent lateral spines, and ends in a conical swelling.
The gastral pinuli have also moderately long and rough basal rays, but the radial
which runs to a thin point bears short curved lateral spines. Near the Banda Islands,
360 fathoms.

Species 6. Stylocalyx elegans, n. sp.

The approximately spherical, Ijut inferiorly conical body bears in the middle of the
obliquely truncated upper surface an irregularly contoured oscular aperture, above which

REPORT ON THE HEXACTINELLIDA. 385

a slender pointed central cone distinctly projects. Tlie inferior pole bears a slender basal
tuft of a few long gently curved spicules. The parenchyma contains many small
oxyhexacts with smooth curved rays. Tiie larger amphidiscs have lacerate and bell-
shaped terminal umbels, with eight somewhat long rays, the pointed extremities of which
extend almost to the middle of the amphidiscs. The dermal pinuli have somewhat short
rough basal rays, and a long slender radial, running out to a thin point. South-east of
Christmas Island, Mid-Pacific, 2425 fathoms.

Species 7. Stylocahjx tener, n. sp.

The oval or pear-shaped body exhibits a flat conical narrowed basis, giving off a
narrow much twisted basal tuft of long thin spicules. The parenchyma contains numerous
small oxyhexacts with smooth curved rays. The large amphidiscs are approximately
spherical, with hemispherical terminal umbels of eight to twelve paddle-shaped rays,
alternating with those of the other side and almost reaching them. The dermal pinuli
have a long, slender, pointed and short-toothed terminal ray. South Pacific, 2550
fathoms.

To these there have to be added some species of Hijalonema which, owing to the
insufficient preservation of the upper portion, or to inadequate description, cannot yet
be referred to either of the above subgenera : —

Species 1. Hyalonema lusitanicum, Barboza du Bocage.

The long and strong spirally-twisted root-tvil't is surrounded l;)y a large Pahjthoa
crust. The parenchyma contains small oxyhexacts with sti'aight, or with curved smooth
rays. The dermal pinuli exhibit four short basal rays and a somewhat long distal with
moderately long lateral spines. The large amphidiscs are elongated and slender. Their
deep bell-shaped terminal umbels have eight slender blunt rays. Near Setubal, Portugal.

Species 2. Hyalonema cebuense, Higgin.

The body has the form of a sculptor's mallet with the broad portion upwards. The
basal tuft is thicker than a goose-quill, and is spirally twisted. The small paren-
chymal oxyhexacts have thin straight rays, which are curved outwards at their outer
ends and bear lateral spines. The dermal pinuli are somewhat small and slender. The
large amphidiscs are narrow, with deep bell-shaped umbels, in which the eight arms are
slender. The parenchyma contains slender oxydiacts with teeth which arc directed
towards the centre. Zebu, Philippines.

(ZOOL. CHALL. EXP. — PART LHI. — 1887.) Ggg 49

386 THE VOYAGE OF H.ISLS CHALLENGEE.

Species 3. Hyalonema tenuc, n. sp.

With narrow oval oscular aperture. The parenchyma contains small weakly
developed oxyhexacts wth smooth curved rays, and also somewhat larger forms with
thin straight smooth vays. The large amphidiscs have campanulate terminal umbels
with twelve to thirteen lancet-shaped rays. The dermal pinuli have somewhat long
smooth liasal rays, and a long, very thin radial. East of the mouth of the Rio de la
Plata, 1900 fathoms.

Species 4. Hyalonema robustum, n. sp.

The parenchyma contains small oxyhexacts with smooth or slightly roughened curved
rays. Besides these somewhat larger oxyhexacts occur with spinose straight rays. There
is a characteristic occurrence of unusually large, approximately spherical amphidiscs, in
which the hemispherical terminal umbels almost, or actually meet one another in the
middle. They consist of eight, or less frequently as many as twelve broad paddle-shaped
rays. The dermal pinuli have short, thick, rough, basal rays, and a long club-shaped,
but at the same time pointed radial ray with moderately short lateral spines. West of
the North Pacific, 2300 fathoms.

Genus 2. P/ieronema, Leidy.

The basal terminal surface of the round, goblet- or cup-shaped, thick-walled body bears
a broad root-tuft, not always sharply marked off. The lateral surface of the body also bears
projecting tufts of spicules. The circular oscular margin is surrounded (with the probable
exception of one species) by a cuflp-like freely projecting collar. The parenchyma of the
body contains uncinates. Each spicule of the root-tuft bears on its lower end a bidentate
anchor. The marginalia and pleuralia end externally in a club-shaped swelling.

Species 1. Pheronema annw, Leidy.

An egg-shaped body which is somewhat truncated both at its superior narrower and
inferior broader pole. The gastral cavity is cylindrical, about half the length of the
body, and with a flat floor. The broad basal tuft consists of numerous individual tufts
about half the length of the body. Slender tufts of spicules project radially in irregular
distribution from the lateral surface of the body. At the somewhat rounded oscular
margin there seems to be no special wreath of spicules. The parenchyma contains small
oxyliexacts sparsely beset with lateral spines. The large amphidiscs have campanulate
terminal umbels with eight broad lancet-shaped rays, and a thick knolibed axial lieani.

REPORT OX THE HEXACTINELLIDA. 387

The dermal piiiuli have four tolerably loug, smooth and gradually pointed basal rays,
and a moderately long distal which bears strongly developed upward bent spines. Santa
Cruz, West Indies, 180 to 248 fathoms.

Species 2. Pheronema carpeiiteri (Wyville Thomson).

Tiic body has a somewhat broad ellipsoidal form, slightly truncated at the upper and
lower pole, of which the latter bears the broad basal root-tuft as long as the body, while
the superior pole is surrounded by the cuff-like fringe of spicules round the osculum.
The radial tufts projecting in irregular distribution from the lateral surface of the body
form an annular zone for a finger's Ijreadth below the oscular margin. The parenchyma
contains long and short uncinates. The amphidiscs have hemispherical terminal umbels,
each with eight moderately Inroad paddle-shaped rays. The dermal pinuli have lonw,
smooth, pointed basal rays, and a moderately long, somewhat bushy radial. North of
Scotland, 530 to 1600 fathoms.

Species 3. Pheronema graiji, Sav. Kent.

The body has the forari of a Ijullfinch nest, with broad gastral cavity, occupying
aliout two-thirds of a sphere, and bearing a loug beard-like basal tuft. The prostalia
lateralia are almost uniformly distributed over the whole outer surface, but form
besides, as in Pheronema carpenteri, a thick zone measuring a finger's breadth across
below the oscular cuff. Otherwise the forms of spicules agree essentially with the above
species. Portugal.

Species 4. Pheronema hemisphairic^un (Gray).

Body urn-shaped, with a concave indentation (annular groove) both above and below
a gentle median swelling beset with tufts of spicules. The spicules of the broad
basal tuft project downwards as far as the body is long, and form a broad annular zone
round a central free portion. The parenchyma contains small oxyhexacts, in which the
rays bear raised lateral spines. The dermal pinuli have rough basal rays directed some-
what obliquely downwards, and a short slightly bushy radial. The gastral pinuli, on the
other hand, have a long, .slender radial ray. Zel )U, Philippines.

Species 5. Pheronema ylohosum, n. sp.

The body occupies three-fourths of a sphere the size of a man's fist, and bears on its
superior truncated extremity the wide oscular aperture of the flat gastral cavity. The

388 THE VOYAGE Of H.M.S. CHALLENGER.

somewhat sharp-edged oscidar margin bears a wreath of pointed spicules. Numerous
tufts of radially disposed spicules project from the lateral surface of the body. The basal
tuft is formed of numerous separate tufts of spicules, which are interlaced at the outer
ends. In the external skin a thick network with irregular stellate structures is recognis-
able. The parenchyma contains, besides medium-sized uncinates, very numerous smaller
fomis with curved lateral spines. The large amphidiscs have a rather thick axial rod
and hemisphcrically arched terminal umbels wit\\ eight broad paddle-shaped rays. The
dermal pinuli have four rough, substantial, somewhat downward bent liasal rays, and a
moderately long l)usliy radial. Near Little Ki Island, 140 fathoms.

Species 6. Phcronema giganteum, n. sp.

An ellipsoidal body as big as a man's head, with a round oscular opening measuring
a hand's breadth across, and surrounded by cuff-like wreath of spicules. The smooth
external surface bears small scattered tubercle-like elevations, from the apex of each of
which a tuft of long radial pleuralia projects. These increase in length towards the lower
end, and measure more than a hand's length in the root-tuft at the base. The paren-
chyma includes, besides strong oxypentacts probably originating from the dermal skeleton,
small spinous oxyhexacts and uncinates of various length and with short barbs. The
large amphidiscs have a slender axial rod and hemispherical umbels with eight narrow
rays. The dermal pinuli have rough basal rays of medium length, and a somewhat long-
uniformly broad tufted radial. Near Little Ki Island, 140 fathoms.

Genus 3. Polio2)ogon, Wyville Thomson.

The body forms either a plump goblet or a thin ear-shaped involute plate. The
low^er end forms a broad basal tuft, while the upper bears a delicate fringe of spicules
round the sharp-edged oscular margin. The smooth external lateral surface hears
no pleural prostalia. The parenchyma includes, besides rough or spinose small
oxyhexacts, uncinates, and in one species also small smooth oxydiacts in variable al>un(l-
ance and size. The two anchor teeth of the basalia stand approximately at right angles
to the long almost smooth shaft. The marginalia are externally club-shaped.

Species 1. Poliopogon amadou, Wyville Thomson.

An ear-shaped semi-involute thick plate with sharpened oscular margin, bearing a
short and slender fringe of spicules. A fine square-meshed latticework is borne both on
the convex outer dermal, and concave inner gastral surface. The parenchyma contains
spinose oxyhexacts and numerous medium-sized uncinates. The larger amphidiscs have

REPORT ON THE HEXACTINELLIDA. 389

somewhat conical, superiorly truncate terminal umbels with eight somewhat In-oad paddle-
shaped rays. The dermal and gastral pinuli have moderately long toothed basal rays,
bent in figure 8 fashion, and an elongated somewhat bushy but pointed distal. South-
west of Canary Islands, 1525 fathoms.

Species 2. Poliopogon gigas, n. sp.

A very large, thick-walled, plump goblet, with a spacious gastral cavity more than a
span wide, and opening by a circular osculum. Besides the parenchymal spicules
mentioned in Poliopogon amadou, there are here small smooth spindle-shaped oxydiacts.
The large amphidiscs resemble in form those of Poliopogon amadou, but have somewhat
longer terminal umbels. In the pinuli the radial ray is shorter and less thickly spinose
than in the otherwise very similar pinuli of Poliopogon amadou. Between the Raoul and
Macaulay Islands, north of New Zealand, 630 fathoms.

Subfamily 2. Semperellin.e.

With the single genus Semperella, Gray.

With the single species Semperella schultzei, Graj^

The elongated club-shaped body bears at its base a brush-like root-tuft, but neither a
simple round oscular aperture at its superior, gently conical end, nor a simple internal
gastral cavity within. It is traversed by a connected system of thin-walled tubes, as
thick as a little finger, associated with an axial main tube. Between these an approxi-
mately equal set of connected interspaces are left. While the lumen of the connected
(efferent) tubes opens on the four to six, rounded off and irregular, longitudinal sides, and
on the superior conical extremity of the club-shaped body by a sieve-shaped covering
with comparatively wide meshes, the interjacent (afferent) canal system is covered on the
flat sides of the body by a fine-meshed quadratic framework. The skeletal spicules
resemble, for the most part, those of Poliopogon amadou. The long uncinates have,
however, more strongly developed and somewhat curved spines. Besides the long
spinose oxyhexasters, numerous reduced forms of the same occur, down to long spiny
oxydiacts, and more rarely small uncinates with short spines and conically pointed
extremities. The dermal pinuli have tolerably short, for the most part 8-shaped basal
rays, and a short, strongly developed, and pointed distal with strong prominent lateral
spines. The pinuli which occur on the dermal sieve-work of the excurrent region are,
on the other hand, long and slender, w4th somewhat short, upward bent lateral spines.

The aljundant and large amphidiscs, which occur especially in the dermal sieve-work
of the excurrent region, have a strongly developed and knotted axial rod, with eight broad
paddle-shaped terminals, somewhat conical, terminally truncated, short principal rays.

;59() THE VOYAGE OF H.M.S, CHALLENGEE.

The boundaxy between excurrent and incurrent regions is marked by projecting toothed
oxydiacts which run to a point or are slightly club-shaped. Zebu, Philippines ; Little
Ki Island, 140 fathoms.

Suborder II. DICTYONINA, Zittel.

Hexactinellida in which the large parenchymal hexacts are from the first more or less
reo-ularly united as dictyonalia in a firmly connected framework.

Tribe I. UNCINATAEIA, F. E. S.
With uncinates.

Subtribe I. Clavularia, F. E. S.

Besides the pentact hypodermalia and hypogastralia there are groups of radially-
disposed clavulfe.

Family I. F a R R E i D .■£, F. E. S.

In the youngest (most external) portions of the tubes the dictyonal framework
consists solely of a single-layered network with square meshes, in which the nodes of
intersection bear on either side a conical boss projecting at right angles.

Genus 1. Farrea, Bowerbauk.

The stock, which is firmly attached by an expanded base, consists of a dichotomously
branched tube tending to form anastomoses, or of a simple funnel-shaped cup. Beside
each of the medium-sized hypodermal and hypogastral pentacts, there is a group of
radially disposed clavulse.

Species 1. Farrea occa (Bowerbank), Carter.

Richly ramified dichotomous and anastomosing tubes, which are gradually widened
superiorly till they become as thick as a glove-finger. The parenchyma contains
numerous slender oxyhexasters with long cylindrical principal rays, each with three to four
short, markedly divergent terminals. The dermal clavute exhibit on their distal end,
above a terminal swelling, an externally convex marginally toothed transverse disc, or
else form transitions towards a simple club type, with ovoid or citron-shaped terminal

REPORT ON THE HEXACTINELLIDA. 391

knob, frequently with a transverse ring of teeth. The gastral clavula; are ahnost always
well-developed, verticillate forms, with convex marginally toothed terminal disc and a
very variable number of teeth. Sagami Bay, Japan, 100 to 200 fathoms; west of
Manila, 700 fathoms.

Species 2. Farrea sollasii, n. sp.

The parenchyma contains discohexasters with short principal rays and long delicate
gently diverging terminals. They otherwise closely resemble Farrea occa. Sagami Bay,
Japan, 100 to 200 fathoms.

Species 3, Farrea vosmaeri, n. sp.

The parenchyma contains strongly developed oxyhexasters with long and strong
principal rays, and four short, markedly divergent, usually somewhat outward bent terminals.
Discohexasters also occur with short principal rays and weakly developed long terminals,
as also discohexasters with somewhat strongly developed, long principals, each with four
short S-shaped terminals disposed in a cup-like form, and bearing terminal knobs.
Sagami Bay, Japan, 100 to 200 fathoms.

Species 4. Farrea clavigera, n. sp.

A straight main tube rising at right angles gives off transverse and terminal,
branched and anastomosing secondary tubes. The parenchyma contains oxyhex-
asters with long principal rays, each bearing four markedly diverging terminals. The
dermal clavulse always exhibit a club-shaped distal end, which is sometimes elongjited
and quite smooth, in other cases compressed and weakly developed, and provided with
a simple or double transverse circle of hooks. The gastral clavulas all exhibit four
cruciately disposed, much recurved anchor teeth. Near the Banda Islands, 200 to 360
fathoms.

Subtribe II. Scopularia, F. E. S.
Besides the hypodermal and hypogastral pentacts radially disposed scopulse occur.

Family I. E u R E T i d .^, F. E. S.

Branched and manifoldly anastomosing tubes, whicli either form an irregular frame-
work or the lateral wall of a cup. The dictyonal framework in the tubular wall always

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

consists of several layers, so that the outermost ends of the tulies never exhiljit (as in
Farrea) a single-layered network of beams.

Genus 1. Eurete (Semper), Carter.

A richly anastomosing framework of approximately uniform tubes, attached at several
places to the substratum, and exhibiting numerous oscular apertures.

Species 1. Eurete semperi, n. sp.

The oscular terminal apertures of the tubes, which are almost as thick as a little
finger, always occur on the free, transversely truncated ends of the numerous stumps.
The somewhat irregular dictyonal framework consists of strong beams irregularly beset
with few little spines. The beams are united in spherically thickened nodes of
intersection, which are richly beset with small tubercles. The parenchyma contains,
besides the uncinates and simple oxyhexacts, numerous small discohexasters with
short principal rays and moderately long terminals. The dermal scopulse bear on the
end of a small conical expansion, four (less frequently five or six) cruciately disposed
smooth, spindle-shaped prongs. The gastral scopulae have four cruciately disposed
prongs, in which the thin outward bent stalk ends in a club-shaped swelling, which
is covered on to the smooth apex with small barbs. Near the Little Ki Island, 140
fathoms.

Species 2. Eurete schmidtii, n. sp.

A somewhat narrow-meshed system of anastomosing tubes which measure about
5 mm. in width. The meshes of the dictyonal framework are predominantly square, and
consist of smooth or slightly spinose beams with slightly thickened moderately spinose
nodes of intersection. The parenchyma contains, besides the variously disposed
uncinates, small rough oxyhexacts and oxyhexasters with short principal rays, each with
two terminals. Most of the dermal scopulse bear on the distal end of the inferiorly
pointed smooth stalk, a knotted swelling, with two to four smooth, pointed terminal
prongs, while the others exhibit a thickening with four prongs with thin stalks and pear-
shaped terminal knobs with lateral barbs. The gastral scopulee belong exclusively to the
latter type. Philippines, 102 fathoms; Japan.

Species 3. Eurete farreopsis, Carter.

The irregularly anastomosing system exhibits tubes 4 to 5 mm. in width, and some-
what wider intermediate spaces. The nodes of intersection of the predominantly square-

REPORT ON THE HEXACTINELLIDA. 303

meshed dictyonal framework are somewhat thickened and slightly spinosc. The paren-
chymalia contain, besides the uncinates, disposed at right angles to the bounding surfaces,
and small rough oxyhexacts, small disco- and sphsero-hexastcrs, with four moderatelj'
long terminals on each of the short principal rays. Both the dermal and the gastral
scopulie exhibit a marked and sharp dislocation on their four to six terminal prongs,
which are equipped with pear-shaped terminal knobs bearing lateral barbs. Near the
Little Ki Island, 140 fathoms.

Species 4. Eurete carteri, n. sj).

Somewhat narrow -meshed anastomosing systems of tubes 3 to 10 mm. in width,
which open externally by round or oval apertures not only terminally but also laterally.
The somewhat irregular dictyonal framework exhibits spinose beams with but slight!}-
thickened nodes of intersection, varying in diifferent regions, but tolerably constant in
the development of spines. The parenchyma contains, besides the radially disposed
uncinates and numerous small oxyhexacts, irregularly scattered small discohexasters with
short principal rays, and moderately long, irregularly imdidating terminals, three of
which usually occur on each principal. Both the dermal and gastral scopulae have four
to six straight or slightly curved, but never dislocated terminal prongs with barbed
terminal knobs. Near the Little Ki Island, 140 fathoms; Sagami Bay, Japan, 150
fathoms.

Species 5. Eurete marshalli, n. sp.

A somewhat wide-meshed framework of tubes of varying width (3 to 8 mm.), with
round terminal excurrent apertures. The dictyonal framework with its predominantly
square meshes exhibits, as in Eurete farreopsis, Carter, somewhat thickened and spinose
nodes of intersection. The parenchyma contains, besides the I'adial uncinates and small
oxyhexacts, also oxyhexasters with short principal rays, each with three to five long
terminals. The dermal and gastral scopulse exhibit no dislocation of the four terminal
prongs which end in club-shaped barb-beset extremities. Near the Little Ki Island, 140
fathoms.

Species 6. Eurete howerhankii, n. sp.

The single fragment as yet known consists of a funnel-shaped expanded tubide
3 to 10 mm. in diameter, provided with lateral branches 5 mm. in breadth. The
somewhat irregular dictyonal framework consists of slightly spinose beams ^-ithouL
markedly thickened nodes of intersection. The parench}Tna contains, besides the radial

(ZOOL. CHALL. EXP. — PART LIII. 1887.) Ggg 50

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

uncinates and small oxyhexacts, numerous oxyhexasters ivith strongly developed long
principal rays, each with four short, markedly diverging terminals. The dermal and
gastral scopute exhibit straight, or slightly curved, but never dislocated prongs with
barbed terminal knobs. Japan, 80 to 200 fathoms. Sagami Bay, Japan, 100 to 200
fathoms.

Genus 2. Periphragella, Marshall.

With the single known species Periphragella elis^, Marshall.

A cup- or funnel-shaped form, about a hand's length in height, with a lateral wall
formed of an irregularly anastomosing system of tubes. A somewhat solid basal portion
about as thick as a finger forms an attaching expansion, and passes gradually into
the parietal tubes, which measure at first 3 to 4 mm. in width, but become gradually
wider upwards. These tubes arise from the folding of the primary plate-like funnel-wall,
and are thus on the one hand in open communication with the large internal gastral
cavity, and on the other iDy the round terminal aperture with the exterior. The beams
of the tolerably regular, square or rectangularly meshed dictyonal framework are smooth
or slightly spinose, and exhibit no markedly thickened nodes of intersection. The
parenchyma contains, besides the usually, but not exclusively radially disposed uncinates
and small oxyhexacts, oxyhexasters with rather short principal rays, each with two to
four moderately long, externally curved terminals, and also discohexasters, with five or
more medium-sized, tliiu, somewhat externally convex terminals on each of the strongly
develojied medium-sized principals. The dermal and gastral scopulse exhibit a stalk
with club-shaped swollen ends, while on the other end, turned towards the free bounding
surface, there are four prongs with pear-shaped or spherical, l^arb-beset, terminal knobs.
Many of the dermal scopulse exhibit no thickening below the origin of their four gently
outward-bent knobbed prongs. Moluccas ; Japan, 80 to 200 fathoms.

Genus 3. Lefroyella, Wyville Thomson.

With the single species Lefroyella decora, W5rvrille Thomson.

In the compact skeletal framework of the (1 cm. thick) wall of the syringe-shaped body,
there are radially disposed longitudinal plates 2 to 4 mm. in Ijreadth, which project inwards,
enclosing longitudinal furrows between them, while the external surface exhibits in the
firm cortical layer numerous spirally or circularly disposed round apertures, 3 to 4 mm.
in width, with projecting annular fringe. The beams forming the usually distinctly
square meshes of the dictyonal framework, are in the neighbourhood of the free
bounding surfaces of the tubes slightly spinose, but are otherwise smooth, only here and
there exhibiting somewhat thickened nodes of intersection. Bermudas, 1075 fathoms.

EEPORT ON THE HEXACTINELLIDA. 39.>

Family 11. M e l i t t i o N i D ^, Zittel.

Sponges in the form of a ramified tube or of a cup witli lateral blind diverticula.
The dictyoual framework forms very irregular meshes. The parietal skeleton is honey-
comb-like, with tolerably regular hexagonal, radially disposed canals, arranged at right
ansfles to the bounding surfaces. In each of these little canals the membrana rcticu-
laris, is continued in a funnel-shaped extension of the laterally disposed, glove-
finger-hke chambers, across the canalicidar lumen. The free outer surface is covered
with the dermal membrane, and the inner surface with the gastral. The gastral
skeleton is without scopula3.

Single Genus Aphrocallistes, Gray.

On the hexact dermalia the distal radial ray is more or less distinctly fir-tree-like
with a lateral coating of spines. Besides these there are dermal scopula;, in which the
prongs are usually knobbed, more rarely pointed. The parenchyma contains irregular
oxyhexasters, and in many cases also discohexasters.

Species 1. Aphrocallistes bocagei, Perceval Wright.

An elongated, almost tubular cup, a hand's length or more in height, with numerous
glove-finger-likc radial, and somewhat downward directed diverticula from the wall.
The lumen of the cup is traversed at considerable intervals by several loose lattice-w'ork
septa, which coiTespond in their insertion on the sides to the lateral diverticula, and
exhibit semicircular sections. The loose parenchymalia are represented by uncinates
directed at right angles to the bounding surface, and also by hexasters of different kinds.
Besides oxyhexasters with approximately equal principals and terminals, there are others
in which the axis is considerably elongated, and in which the two strong, more or less long
principal rays are divided at their extremities into four markedly divergent moderately
long terminals, while the four rays representing the two other axes remain, as a rule,
undivided. Small discohexasters occasionally occur. Corresponding to the dermal
hexact pinuli, there are on the gastral surface autogastral diacts of variable length, which
are more or less rough, especially on their rounded extremities. The dermal scopulse
possess, for the most part, four straight somewhat diverging prongs with pear- shaped
terminal knobs bearing lateral barbs. Besides these there are dermal scopula; with four
pointed rough prongs, and others with four strikingly long, rough prongs equipped with
small spherical knobs. Cape Verde Islands; Florida, 283 fothoms ; English Channel,
700 fathoms; coasts of Spain and Portugal; Antilles Islands, 1G4 to 400 fathoms;
Atlantic coast of France.

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

Species 2. Aphrocallistes beatrix, Gray.

lu form resembling Aphrocallistes hocagei, Wright, but much smaller and more

0

ielicate. Besides the uncinates the parenchjrma contains almost exclusively much
elongated oxyhexasters, in which one (principal) axis possesses two long strong principal
rays, occasionally with a few lateral spines, and dividing into four strongly developed,
moderately long, and somewhat markedly divergent terminals, while the four other rays
exhibit uniformly short, simply pointed rays without terminals. Malacca.

Species 3. Aphrocallistes vastus, n. sp.

A large cup, 40 cm. or more in height, and 30 cm. in breadth, rising from a
compact knobbed basal portion growing on the substratum. The lateral diverticula of
the wall are flat and pouch-like, and longitudinally disposed. The parenchyma contains
besides long uncinates, most swollen near the outer end, numerous disco- or sphsero-
hexasters with moderately long terminal rays of variable form and size. On the dermal
hexact pinuli the free distal ray is broad and bushy. The dermal scopulae usually bear
only two or three straight, somewhat uniformly thick, terminally rough prongs without
distinct terminal knob. The gastral membrane contains straight rough diacts, and
occasionally monacts. Japan, 180 fathoms.

Species 4. Aphrocallistes ramosus, n. sp.

Dichotomously branched tubes, 4 to 10 mm. in diameter, attached by means of a
compact basal plate. The branches open with round terminal apertures. The loose
parenchymalia are represented not only by uncinates and by numerous small rough
oxyhexacts, but also by oxy- and sphsero-hexasters elongated in the direction of the main
axis, with principal rays of various length, and with moderately long terminals, usually
developed only on the two principals of the long axis. The dermalia have a weakly developed
distal ray, which is frequently either almost, or wholly without the lateral spines of the
pinuli. The adjacent dermal scopulse possess four straight, or slightly dislocated prongs
with spiny terminal knobs. The gastral membrane contains diacts of various length,
with rough and rounded ends. Japan, 80 to 200 fathoms ; Phillippine Islands, 375
fathoms.

Family III. Coscinoporid^, Zittel.

The smooth wall of the cup-, goblet-, or j)late-like fii-mly fixed body is traversed by
more or less elongated, funnel-shaped, straight canals, which open alternately on either

REPORT ON THE HEXACTINELLIDA. 397

surface, covered only by the sieve-like Ijounding membrane, but end in a blind point at
the other extremity. They thus always correspond approximately in length to the
parietal thickness of the sponge.

Single Genus Chonelasma, n. gen.

The body consists of an upright plain or slightly curved plate about the size of a hand,
and exhibiting an irregularly chambered smooth free margin, or of a funnel-shaped cup
with lateral glove-finger-like parietal diverticula.

Species 1. Clionelasma lamella, n. sp.

The dictyonal framework enclosing rectangular meshes exhibits in its median region
a loose structure with wider meshes than in the neighbourhood of the two bounding
surfaces. The beams of the framework are usually beset with small pointed tubercles,
and are connected without thickening of the nodes of intersection. The parenchyma
includes, besides the uncinates and simple rough minute oxyhexacts, small rough disco-
hexacts, oxyhexasters, and discohexasters with long terminal rays.

The dermalia and gastralia are strongly developed pentacts, in which the proximal or
distal radial ray, as also the sides of the four tangentials turned towards the free surface,
are richly beset with short spines. Besides these the dermal skeleton contains scopulse,
with four knobbed, straight prongs, which rise from a thick stalk-node. Kermadec
Islands, 520 fathoms; west of Kerguelen Islands, 550 fathoms; Bermuda Islands,
1705 fathoms.

Species 2. Chonelasma hamatum, n. sp.

An upright, hand-shaped plate, 3 to 5 mm. in thickness, with a slightly undulating
curvature, and with irregular roundish marginal contour. The dictyonal framework
resembles that of Chonelasma lamella. The loose parenchymalia include, besides uncinates
and numerous small rough oxyhexacts which tend to become soldered to the dictyonal
framework, small discohexasters, some of which bear short principal rays -n-ith six to eight
long, S-shaped terminals in perianth-like arrangement, while others possess long, rough,
strongly developed principals, with short, S-shaped terminals. The dermal and gastral
skeleton is formed of hexacts, in which the projecting radial ray is short and beset with
inconspicuous spines. Besides these the dermal skeleton especiaU}^ contains abundant
scopulse, some of which bear near the outer end of their stalk an annular swelling or four
cruciately disposed bosses, in which an intersection of axial canals can be distinctly seen.
Their four straight or slightly S-shaped prongs end cither in a simple rounded off
extremity, or in a minute knob-like thickening. Kermadec Islands, 520 fathoms.

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

Species 3. Clionclasma dceclerlei7m, n. sp.

A slightly undulating upright plate, with irregular rounded margin and somewhat
thickened firmly attached basis. The dictyonal framework agrees essentially with that
of Clionclasma lamella. The parenchyma contains, besides the uncinates, numerous large,
delicate discohexasters, with short principal rays on each of which three or four long,
thin, somewhat curved or undulating terminals are borne. The dermal and gastral
skeletons consist of hexaets with bushy, freely projecting radial ray, and of scopulae with
four straight or slightly curved prongs, covered with short barbs, and ending either in
simply rounded, or in small knob-like extremities. Japan, 80 to 200 fathoms.

Species 4. Chonelasma calyx, n. sp.

The body is cup-shaped, and over a hand's length in height. The dictyonal frame-
worK of the cup and of the radial glove-finger-like parietal diverticula consists of smooth
or slightly spinose beams, and is slightly more irregular than that of Chonelasma lamella.
The loose parenchymalia resemble those of the latter. In the dermal and gastral hexaets
the freely projecting ray is somewhat compressed and clubbed, and is further beset with
scale-like lateral teeth. The dermal and gastral scopiilse bear four to six straight or
S-shaped prongs, beset with barbs, and either rounded off at the ends or with a club-
shaped thickening. Japan, 80 to 200 fathoms.

Family IV. T r e t o d i c t y i d ^, F. E. S.

Uucinataria with irregular afferent and efferent canals which penetrate the body-
wall, not at right angles to the bounding surfaces, that is transversely, but obliquely or
longitudinally, or even in curved course.

Genus 1. Hexactinella, Carter.

Cup- or tube-shaped forms, with canals which traverse the body-wall in an oblique
longitudinal dii-ection. The dictj'onal framework contains reticulate plates, extending
between parallel canalicular spaces in a longitudinal or radial fashion.

Species 1. Hexactinella tuhulosa, n. sp.

Dichotomous and anastomosing tubes as thick as a little finger, rising from an
encrusting base, and opening terminally by terminal apertures. The beams of the
dictyonal framework are tubercled, and united without thickened nodes.

REPORT ON THE HEXACTINELLIDA. 399

On the internal surface of the entire tubular skeleton there arc longitudinal and radial
plates alternating with grooves of equal breadth, while on the external surface the frame-
work is flatly expanded. The loose parenchjonalia are represented not only by very
delicate uncinates and by numerous small rough oxyhexacts, but also by numerous
oxyhexasters, with rather long strongly developed principal rays, each bearing two short
moderately divergent terminals, besides simple medium-sized oxypentacts. The dermal
skeleton contains scopulae, with four approximately S-shfipcd rough prongs, which are
at their extremities only slightly, or not at all swollen. The gastral skeleton contains
numerous similar scopulae, in which, however, the prongs are almost straight. There
seem to be no pentacts in the gastral skeleton. Japan.

Species 2. Hexactinella lata, n. sp.

Ramified strong-walled tubes, in which the branches, more than a thumb's breadth
in thickness, expand superiorly in funnel- or syringe-like fashion, and open by wide
terminal apertures several centimetres in width. In the longitudinally disposed radial
plates of the dictyonal framework, which alternate with grooves or canals of equal breadth,
and are especially distinct on the external surface of the tubes, it may be seen that the
dictyonal beams radiate from the inside and from below towards the exterior and
superior surface. On the surface of the predominantly square-meshed framework there
are minute tubercles, arranged for the most part in transverse rows. The loose paren-
chymal needles are represented by variously disposed weakly developed uncinates and
numerous thin oxydiacts, also Ijy small disco- or sphsero-hexacts, with cylindrical rays
and terminal knobs, by oxyhexasters with long thin terminals, and, lastly, by sphsero-
hexasters, with two to six simple straight or S-shaped terminals of medium length.
In the moderately large dermal and gastral pentacts there is usually a more or less
conspicuous rudiment of the sixth freely projecting radial ray. The dermal and gastral
scopulse have four knobbed, slightly divergent, rough prongs. Little Ki Island,
140 fathoms.

Species 3. Hexactinella ventilabrum, Carter.

A thick-walled cup, widely open above, and laterall}' somewhat compressed ; borne
by a broad firm base fixed to a solid substratum. The radial plates which extend
lietween the predominantly longitudinal and oblique canals, traversing the wall often in
curved courses, are more distinctly seen on the outer, than on the inner surface of the
skeleton. The beams of the predominantly square-meshed framework are sparsely beset
with minute irregularly distril>uted pointed tubercles. The loose parench}Tiialia are
represented not only by a few uncinates with slightly developed barbs, l)ut also by

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

oxyhexasters with short principal and long terminal rays, and by discohexasters. Of these
the one set have moderately short principals and four to six long S-shaped terminals, while
the others have long substantial principals and numerous short, somewhat convex,
terminals. The dermal and gastral skeletons contain compact j)entacts and four-pronged
scopulas, in which the rough, slightly curved, or almost straight rough prongs are equipped
with slight terminal knobs. Besides the dermal pentacts, there are bundles of fine
radially disposed oxydiacts projecting beyond the bounding surface. Japan.

Genus 2. Cyrtaulon [Volvulina, 0. Schmidt), n. gen.

The variously shaped body is traversed by very irregular canals. The dictyonal
framework forms an irregular feltwork of plates and strands. The parenchyma contains
peculiar spicules, which may be regarded either as scopulse with tuft-like or radially-
disposed prongs bearing terminal discs, or better, as discohexasters with one much elon-
gated and strongly developed principal ray, and five others always much shortened, and
bearing tufts of terminal rays.

Species 1. Cyrtaulon sigsbeei (0. Schmidt).

A variously shaped, not unfrequently goblet-like form, in which the wall consists of
an irregular feltwork, with cavities in which the apertures, both on the external and on
the internal bounding surface, are covered with skin. The dictyonal framework itself
consists of tubercled beams, in which here and there, especially near the surface of the
sponge, the nodes of intersection are thickened and beset with wart-like prominences.
The parenchyma contains uncinates with central thickening and modified scopulse, or
discohexaster forms with numerous thin prongs radiating out from the terminal node of
the stalk, and bearing marginally toothed terminal discs. The dermal skeleton consists
of medium-sized pentacts and hexacts, with typical scopulse with knobbed prongs.
AntUle Islands, 100 to 300 fathoms.

Species 2. Cyrtaulon solutus, n. sp.

The only specimen as yet known is a cylindrical body within a ca^dty in a stone.
The supporting framework is irregular with strands and plates. The dictyonal network
exhibits small superficial tubercles, but no marked thickening of the nodes of intersection.
The genericaUy characteristic modifications of scopulse or discohexasters, which occur in
the parenchyma instead of the typical hexasters, exhibit on the knobbed and thickened
end of the (otherwise pointed) stalk nvimerous thin radiating prongs or terminal rays, with
small convex, marginally toothed terminal discs. Besides the pentacts of the dermal

REPORT ON THE HEXACTINELLIDA. 401

skeleton there are scopulse of the ordinary type, some of which bear four uniformly thick
and barb-beset prongs without marked terminal swelling, while others exhil)it six thin,
smooth, slightly S-shaped prongs, with minute marginally-toothed terminal discs. In
the gastral skeleton I found no pentacts, or indeed any hypogastralia, except simple
scopulse in their usual position, with four uniformly thick, barb-beset, unknobbed prongs.
Little Ki Island, 140 fathoms.

Genus 3. Fieldingia, Sav. Kent.

With the single species, Fieldingia lagettoides, Sav. Kent.

An irregular round, sometimes almost spherical body, which is usually ensheathed in a
thin parallel enveloping capsule, and traversed internally by an irregular framework of thin
round strands, which exhibit numerous spherical compact knots, about 1 mm. in diameter,
and occurring at distances of 2 to 3 mm. While the thin strands consist of a few long
beams, beset with small, scattered and pointed tubercles, the spherical knots are formed of
a thick framework of beams with a similar superficial appearance. Both the large, more
or less elongated, and the short reticulate beams of the spherical knots have abundant
small rough hexacts soldered on to them, usually at right angles. These doubtless serve
to enlarge or thicken the dictyonal framework. The leaf-like enveloping capsule which
surrounds the larger portion of the sponge, consists of pentacts bound together in plates
by a narrow-meshed network of syuapticula extending in the tangential direction. The
parenchyma contains, besides long uncinates and simple small hexacts, oxydiacts with
central nodes and sharpened extremities, oxyhexasters with very short, in some cases
almost undeveloped principals, and long smooth terminals, and also discohexasters with
short principals and rather long, slightly curved terminals. The dermal skeleton
contains, besides the above-mentioned pentacts, scopulse with four uniformly cylinfh-ieal,
slightly divergent prongs, which are wholly covered with barbs, but exhibit no terminal
swelling. Portugal, 500 fathoms ; Little Ki Island, 140 fathoms.

Genus 4. Sclerothamniis, W. Marshall.

With the single species Sclerothamnus clcmsii, W. Marshall.

A shrub-like, dichotomously branched stock, with long, round, and terminal Invinches
as thick as a little finger, and Ijearing spiral or annular pads a little finger's breadth
across, and alternating with somewhat narrower grooves. The somewhat thick dictj'^onal
framework supporting the body consists of beams which enclose predominantly rectangular
meshes, and exhibit superficially numerous small, irregidarly scattered, wart-like and
pointed tubercles. They are united without nodal thickening. While the main strands
of fibres in the axis of the branches lie longitudinally, ''•<'■, parallel to the axis, in the

(ZOOL. CHALL. EXP. — PART LIII. — 1887.) Ggg 51

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

external portion of the Ijranclies the strands bend out towards the lateral surface,
and end below the skin, as is readily apparent on the macerated skeletons. The
transverse beams between these principal strands are for the most part disposed at
right angles. The afferent and efferent canals which penetrate the whole body are
irregular roundish passages, running in various directions, but in general across the
branches. Since the spiral or circular grooves on the branches are covered over with a
fine square dermal lattice-work, while on the annular pads the outward bent skeletal
strands and the associated efferent canals run at right angles to the lateral surface, it
may be inferred that the afferent canals pass to the interior from the large sub-
dermal spaces under the skin of the annular grooves, while the efi'erent canals
open out on the pads. Isolated spicules in the parenchyma are represented not
only by strongly developed uncinates, Ijut by scopula-like spicules with a long,
straight rough stalk, from which a few conical or finger-shaped elevations here and
there project transversely, and with four markedly diverging, uniformly thick, rough
unknobbed prongs rising from the thick extremity. The dermal skeleton contains
hypodermal sword-shaped oxyhexacts, with a floricome-like spha^rohexaster on the
short radial ray. The principal rays of the latter are moderately short, and bear
six long terminals, disposed in perianth fashion, diverging slightly outwards, gradually
thickened, and ending in a spherical knob. Besides these hypodermal hexacts there
are scopulfe with rough stalks narrowed inferiorly, and bearing at the thick distal
end four slightly S-shaped or straight, uniformly thick prongs, which are wholly l^eset
with small barbs, and end in slight, knob-like, terminal swellings. Philippines ; Timor,
360 fathoms.

■ Tribe II. INEEMIA, F. E. S.
Dictyonina without uncinates or scopul^.

Single Family M^androspongid^, Zittel.

The body consists of a connected system of labyrinthine anastomosing tubes of
approximately uniform width, between which there is a connected interstitial system of
interspaces. The water passes by the latter into the interior, penetrates the walls of the
tubes, and passes by the tubes either into the gastral cavity or du'ectly to the
exterior.

Genus 1. Dactylocalyx, Stutchbury.

The thick wall of the plump and generally cup-shaped body exhibits both on the
outer and on the inner surface broad irregular pads, and interjacent clefts or grooves.
The external swellings correspond to the internal grooves, so that the whole appears to

REPORT ON THE HEXACTIXELLIDA. 403

be folded. The thick wall of the cup thus folded consists simply of a system of anasto-
mosing tubes, which open internally into the gastral cavity, and probal)ly also directly to
the exterior. On the other hand, the connected interstitial system of spaces between the
above tube work is closed on the internal gastral surface of the body, but probably
covered externally only by a porous skin, which admits the incurrent water.

Species 1. Dactylocalyx pumiceus, Stuchbury.

The body forms a broad flat cup which is borne on a short, somewhat meshed and
thin- walled stalk firmly fixed to the substratum. The cup expands into a broad (30 cm.)
thin-walled plate- or cup-shaped body, which is laterally provided with a somewhat back-
ward bent, gently sinuous, rounded margin. The radially disposed grooves of the
external inferior and internal superior surface here and there exhibit a dichotomous
external division. The tubular network is very narrow meshed and the whole thick
connected skeleton is firm and strong. It is composed of finely-tuliercled beams without
thickened nodes of intersection. The loose spicules of the parenchyma are represented
by small hexacts with lank, terminally thickened, in part elongated and somewhat curved
rays, and also, according to Bowerbank,' by oxyhexasters with three long, slightly curved
terminals on each of the short principals, and lastly by discohexasters with somewhat
long terminals. The oxypentacts described by Sollas are to be referred to the dermal
skeleton. There is no trace of uncinates or of scopulse. Barbados ; West Indies.

Species 2. Dactylocalyx s2ibglohosus, Graj'.

A deep, thick-walled goblet in which the grooves on the gastral or internal surface
are less broad than those on the external. The parenchyma contains, besides hexacts
with terminal knob-like thickenings (sph^rohexacts) numerous discohexasters of various
size, with long terminal rays, but no oxyhexasters. The tangential rays of the rough
dermal pentacts are terminally club-shaped, while the longer proximal radial ray runs
out to a point. West Indies.

Species 3. Dactylocalyx 2)cUella, n. sp.

The dictyonal framework of this probably flat patelliform species, of which only a
fragment of macerated skeleton was procured, consists of smooth beams forming an
approximately square-meshed network, and forms an anastomosing system of tubes
which in many respects resemble those of the other species of Dactylocalyx, but are at
least twice as broad. (Perhaps identical with loanella compressa, 0. Schmidt, loc cit.)
Bermuda, 1075 fathoms ; coast of Portugal.

' Pnc. Zool. Soc. Lonil, 1869, p. 7", pt. iii.

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

Genus 2. Margaritella, 0. Schmidt.

With the single species Margaritella cceloiotijchioides, 0. Schmidt.

From the somewhat indistinct description and figure given by 0. Schmidt/ the
body of this sponge forms a flat cup, the wall of which consists of a system of connected
tubes and equally wide intercanals. The dictyonal framework enclosing polyhedral
meshes consists of weakly developed, richly tubercled beams, which are united by slightly
thickened nodes of intersection. The latter bear tubercle-like warts where they occur
near the surface of the network. The parenchyma contains loose spicules in the form
of weakly developed oxyhexacts, delicate oxyhexasters with a few short terminals and
very short or wholly reduced principals, and finally somewhat substantial sphgero-
hexasters in which the principal rays are often so much shortened and connected by
siliceous masses into a central knot, that the numerous long knobbed terminals appear as
if radiated out directly from the centre. The dermal skeleton contains rough pentacts in
which the four tangential rays are knobbed, while the proximal radial is simply rounded
off" or somewhat narrowed at its extremity. Havana, 158 fathoms.

Genus 3. Scleroplegma, 0. Schmidt.

Thick-walled cup or cylinder in which " the brittle parietal feltwork consists of
round or prismatic tubes, which run for the most part obliquely from the outside inwards,
either isolated or united in small numbers, and open into the gastral cavity." Between
the tubes there are irregular intercanals.'

The only species really known to me is Scleroplegma conicum, 0. Schmidt.

The somewhat smooth internal wall of the spherical goblet exhibits the openings of
several wide tubes, arranged in toleral^ly distinct longitudinal rows. The external
surface is traversed by a labjrrinth of grooves. The beams of the polyhedral meshwork
are beset with transverse rows of conical tubercles, and are united in nodes of inter-
section, which are thickened especially on the surface of the lattice-work, and beset
with groups of tubercle-like warts on the free surface. The loose spicules are quite
undetermined. West Indies, 292 fathoms.

Genus 4. Myliusia, Gray.

With the single species, Myliusia callocyathus, Gray.

A thick-walled substantial cup, with an originally simple plate-shaped wall, from
which, by continued parietal folding, a system of tubes has arisen. This system com-

1 Spongien des Meerbusens von Mexico, p. 54, Taf. vii. fig. 7.
- 0. Sclimidt, loc. cit., p. 56.

REPOET ON THE HEXACTINELLIDA. 405

municates on the one hand with the funnel-shaped gastral cavity, and on the other
hand opens laterally to the exterior by numerous radially disposed and occasionally
connected tubes. Between the latter there is a toleral:)ly wide intcrcanalicular system,
which does not open, however, into the gastral space. The polyhedral dictyonal frame-
work, which penetrates and supports the much-folded plate or tube-wall, consists of fii-m
beams, which are beset with numerous transverse rows of small conical tubercles, and
united in slightly thickened nodes of intersection, beset with broad tubercled warts. On
the dermal and gastral surface of the dictyonal framework there are specially thick nodes
of intersection, which never exhibit freely projecting spherical bosses, as in most of the
other Dictyonina. The loose spicules of the parenchyma are small lank oxyhexacts,
further, oxyhexasters with short principal rays and long S-shaped terminals, disposed in
perianth-like fashion, and, lastly, discohexasters with short principals, each with a varying
number of long S-shaped externally thickened terminals, forming a tuft and l^earing
marginally toothed terminal discs. The dermal and gastral skeletons contain pentacts or
sword-shaped hexacts with rounded ends, and with a floricome-like discohexaster of the
above type, usually attached to the freely projecting radial ray. West Indies ; Little Ki
Island, 140 fathoms; Timor, 200 fathoms; St. Thomas, West Indies.

Genus 5. Aulocystis, n. gen.

The body consists of an anastomosing system of tubes. The dictyonal framework
encloses scjuare meshes, i.e., cubical spaces, and consists of tubercled beams, in which the
nodes of intersection exhibit twelve oblique rod-like buttresses, extending between the
six intersecting beams, so that the sides of a regular octahedron are represented. The
parenchyma contains loose oxyhexasters and discohexasters. Ox}^ientacts occur in the
dermal and gastral skeleton. Graphiohexasters are found iDcneath the skin.

Species 1. Aulocystis grayi, Bowerbank.

The octahedral edges of the (" perforate ") nodes of intersection in the dictyonal
framework are properly not formed of simple, cyUndrical, obliquely disposed buttresses,
but the external margins of plates which stretch between the intersecting beams, and are
transversely perforated by several round pores of various size. Over the free bounding
surface of the whole dictyonal framework these plates are so much developed, especially
in the tangential dii-ection, that they form here and there, by fusion, a perforate siliceous
membrane. The parenchyma contains numerous discohexasters with short principal rays
of different sizes, and with a variable number of terminals; also small isolated oxyhexasters,
with graphiohexasters here and there under the skin. Discohexasters, with medium-

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

sized principals and terminals of equal length, such as occur so abundantly in Aulocystis-
zittelii, are here wholly absent. On the pentact dermalia the proximal radial ray is
much drawn out. St. Vincent, West Indies.

Species 2. A idocystis zittelii, Marshall.

Pear- or egg-shaj^ed form, from the size of a hen's egg to that of a man's fist. The
system of anastomosing thin-walled tubes, as thick as a finger, exhibits a central main
passage, or two may be present. From the latter, simple or slightly branched anasto-
mosing tubes radiate outwards, and between these there is an irregular system of wide
anastomosing intercanals. The whole system of tubes is covered externally with a thin
smooth enveloping capsule, which at the end of the principal passage and lateral tubes
exhibits cleft-like or irregularly stellate apertures, while the portions of the capsule above
the intercanals consist of a more uniformly porous plate or skin, through which the water
enters the sponge. The dictyonal framework supporting the walls of the tubes seems to
be very regularly constituted, and consists of beams with pointed tubercles, enclosing
meshes usually exactly square or cubical. They are united by nodes of intersection,
which are so surrounded by strong beam-like oblique buttresses, with tubercles but
without axial canal, that the edges formed from the latter are the edges of a regular
octahedron. The direct continuations of the beams within the octahedron are weakly
developed and smooth, but provided with axial canals. The loose parenchymalia include
small, somewhat regularly formed, oxyhexasters, various discohexasters, isolated delicate
graphiohexasters, and in certain regions long oxydiacts with central swelling. The
ordinary form of discohexaster is that with medium-sized principals and almost equally
long terminals, and more rarely that form with short principals and long delicate
terminals, or that with short principals and long strongly developed terminals, with thick
terminal discs or knobs. The dermal skeleton consists, like the gastral, of oxypentacts,
with rough ends to the rays. There is almost always a rudiment of the atrophied sixth
ray represented by a small rounded tubercle. Philippines ; West Indies ; Little Ki
Island, 140 fathoms.

ItEPORT ON THE HEXACTINELLIDA.

407

Key for the Determination of Genera and Species.

This Table includes those living Hexactincllida in wliich the structure is sufficiently
well known to admit of accurate characterisation.

PAGE AND PLATE

A. Witli ampliidises and without hexasters,

B. Witliout aniphidiscs and with hexasters,

I. Ampiiidiscophoiia, including the single family Hyalonematidae.

Cup-shaped (exceptionally plate- or ear-shaped) body
■with a simple round sharply contoured terminal
oscular region, .....

Elongated club-shaped body with internal anastomosing
tubular network and narrow lateral oscular zones, .

II. Hyalonematin.e.

A slender, long, very sharply defined basal tuft of spicules

with four-toothed anchor spicules {Hyalonema, Gray),
A broad, short, less sharply separated basal tuft of spicules

with bidentatc anchor spicules,
III. Hijalonema.

The rays of the small parenchymal oxyhexacts are

altogether .straight, ....

Some of the small parenchymal oxyhexacts have straight

rays, while others are distinctly curved.
The rays of the small parenchymal oxyhexacts are

altogether curved, .....
1 The outer ends of the rays of the small parenchymal

oxyhexacts bear outwardly bent lateral spines,
The rays of the small parenchymal oxyhexacts have no

curved lateral spines, and are either perfectly smooth

or slightly rough, .....
"2 Sponge-body of a perfectly conical form with oscular

sieve-plate on the broad upper end. The small

parenchymal oxyhexacts are rough.
Sponge not exactly conical but more or less swollen out.

The parenchymal oxyhexacts are smooth, .

3 The terminal umbels of the larger amphidiscs have

slender, pointed marginal rays.

The terminal umbels of the larger amphidiscs hear broad,
shovel-shaped, marginal rays,

4 The dermal pinuli have short, somewhat closely appressed

lateral sjjines on the projecting ray, which is always
slender and usually ends in a long drawn out point.

The dermal pinuli have a broad, superiorly truncate,

' radial ray with long distant lateral spines. Tliere

is no oscular sieve-plate, ....

L
VII.

II.

SempereUa schultzei,

III.
IV.

1

6

12

Hyalonema eebuense,

. 261

LI., LII.

228

Hijalonema poculnm,

. 208

xxxm.

HyaloTienia lonus,

. 209

.\xxin.

Hyalonema thomsonii,

. 211

XXXIV.

408

THE VOYAGE OF H.M.S. CHALLENGER.

in. Hijaloucma — continued.

5 Eody egg-shaped. The gastral pinuli have a long thick,

spindle-shaped swollen radial ray, with short lateral
teeth. Terminal umbels of the large amphidiscs are
arched, ......

Body exactly spherical, with firm dermal skeleton. The
gastral pinuli have a thin radial ray, with short,
curved, lateral spines. Terminal umbels of the
large amphidiscs are conical,

6 Both kinds of parenchymal oxyhexacts, or only those

with straight rays bear lateral spines on the rays, .

All the parenchymal oxyhexacts have exclusively smooth

rays, ......

7 The lateral spines on the rays of the parenchymal oxy-

hexacts have their j)oints directed towards the
centre. Oscular sieve-plate present.

The lateral spines on the rays of the parenchymal oxy-
hexacts are directed towards the exterior. Oscular
sieve-plate present, ....

The lateral spines are inserted at right angles to the
straight rays of the parenchymal hexacts, .

8 Body of an exactly conical form. Large amphidiscs occur

with very broad terminal umbels and broad smooth
marginal rays. An oscular sieve-plate is present, .

Body more or less swollen out. There are no large
amphidiscs, .....

9 The dermal pinuli bear short, strongly developed basal

rays, ......

The dermal pinuli have long, slender, basal rays,

10 The dermal pinuli have a very slender radial ray with

apposed short lateral teeth and slender pointed
terminal portion. The straight, slender, long, lateral
teeth of the terminal truncated amphidisc-terminal
umbels diverge inwards. An oscular sieve-plate
is present, ......

^he dermal pinuli have somewhat long lateral teeth on
the distal ray. The terminal umbels of the larger
amphidiscs are deeply bell-shaped, .

11 The dermal pinuli have a long, very thin, markedly

pointed radial ray, with short closely appressed
lateral teeth. The large amphidiscs bear twelve to
thirteen lancet-shaped teeth onthemargin of the discs.

PAGE AND PL.\TE

H yalonema clavigeniiit, . 220

XLI.

Hyalonema globus,

7
8

Hi/alonema toxeres,

10
11

221

XL.

201

XXIX.

Hyalonema sieholdii, . 190
XXVII.

Hj/alonema rohustum,

229

XXXII.

Hyalonema kentii,

. 207

xsx.

Hyalonema diver gens, . 199

XXVIII.

Hyalonema lusitanicujii, . 225

XXVIII., XXXIX.

Hyalonema feniie,

228

XXX.

REPOET ON THE HEXACTINELLIDA.

409

III. Iljialonehta — continueil.

The dennnl pinuli have moderately long lateral teeth on
the long strongly developed radial ray. The large
amphidiscs have eight somewhat broad teeth on the
margin of the discs, ....

12 The small parenchymal o.xyhexacts have oblique in-

wardly directed lateral spines or barbs on their rays,
The rays of the small parenchymal oxyhexacts are quite
smooth, ......

13 The body has a cup-like or truncated oval form, with

wide oscular aperture, without distinct sieve-
membrane. The large amphidiscs have broad,
hemispherical arched terminal discs with only six
broad terminally lancet-shaped marginal rays.

The body is spherical or flat and cake-shaped, with narrow
oscular aperture, without distinct sieve-membrane.
The large amphidiscs have campanulate terminal
discs with eight slender pointed rays,

14 The larger, though still comparatively small amphidiscs,

have campanulate terminal discs with eight rather
long rays, whose pointed ends reach almost to the
middle, ......

The larger amphidiscs have an approximately spherical
form. The hemispherical terminal umbels exhibit
eight to twelve broad terminally rounded marginal
rays, which often meet, or alternate with those of
the opposite side, .....

IV. Numerous radially projecting pleuralia {Pheronema,

Leidy), ......

Without radial pleuralia. External surface smooth
{Poliopogon, Wyv. Thomson),
V. Pherune)aa, Leidy.

Body markedly longer than broad,

Body not longer than broad, ....

1 Body egg- or pear-shaped, narrowed superiorly. The

basal tuft consists of several isolated tufts of needles,

Body resembles a somewhat broad ellipsoid. The broad
basal tuft is bushy, ....

2 Parallel to the cuff-like wreath of marginalia, projecting

as a fringe round the oscular margin, there is below
the oscular margin a connected annular zone of
pleuralia, measuring about a finger's breadth,

(ZOOL. CHALL. EXP. — PART LIII. — 1887.)

PAGE AND PL.\TE

Hyalomma ijiwilr, , l 'J6

XXVII.

13
U

Hijalonema aperttmi, . 214

XXXVII., XXXVIII.

Hyalonema depressum, . 217
XXXV., x.xxvi.

Hyalonema dojaiw, 223

XXXI.

Hyahni'ina fenermn,

V.
VI.

1

3

Pheronema annse.

224

XXXI.

239

XLII.

Pheronema carpenieri, 241

XI.III.

Ggg52

410

THE VOYAGE OF H.M.S. CHALLENGER.

J. Phcronema — continued.

No connected circlet of pleuralia below the oscular
border, but somewhat uniformly distributed tufts of
pleuralia, ......

3 Body resembling a chaffinch's nest, forming about two-
thirds of a sphere, with very long basal tufts.

On the external surface of the body, which occupies
three-fourths of a sphere, there is both below the
oscular border and above the basis an annular
constriction, so that the whole has a bird's-nest form.

The body, which occupies three-fourths of a sphere, has
no oscular constriction, and is superiorly somewhat
flattened, ......

VL Poliopogon, Wyv. Thomson.

The body forms an ear-shaped plate. The parenchyma
contains no smooth oxydiacts, but only .sjsinose
uncinates, ......

PAGE AND PLATB

Pheronema (lirjanteum, . 250

XLV., XLVI.

Phevonema grayi,

246

Pheronema hemisphmcum, 246

Pheromma globosum, . 248

xuv.

Poliopogon amadou,

254

The body forms a very thick-walled, approximately
spherical goblet. The parenchyma includes small
spindle-shaped smooth oxydiacts, .

VII. The skeleton either consists exclusively of isolated

spicules, or exhibits synapticula and ladder-like
connections, resulting in an irregular secondary
union of various large spicules into a connected
compact framework (Lyssacina, Zittel, p.p. ; Hexas-
terophora, F. E. S.),

Besides the isolated spicules there is an early and regular
fusion of uniform hexacts, independent of synap-
ticula, and forming a compact connected framework,
with no ladder-like structures (Dictyonina, Zittel),
VIII. Hexasterophoha, F. E. S.

The dermal skeleton contains hypodermal dagger-shaped
hexacts (EdplectelliDjE), ....

The autodermalia and autogastraUa are pentact or hexact
pinuli, in which the parenchymal radial ray is
altogether wanting or only weakly developed

(ASCONEMATID^), .....

The autodermalia have no distal radial ray (Rossellid-e),

IX. EUPLBCTELLID^.

Tubular forms, with terminal sieve-plate and lateral
circular parietal pores. On the distal radial ray of
the dagger-shaped he.xact hypodermalia a floricome
is seated (Edplectellinjl),

Tubular forms without lateral parietal apertures, and
without external floricomes (Holascin.e.), .

Poliopogon giyan, . 257

XX VII., XLVIII.

VIII.

XLIX.

IX.

XIX.
XXX.

X.
XIII.

REPORT ON THE HEXACTINELLIDA.

411

IX . Eu PLECTELLi d;e — rnntin nod.

The lateral wall of the sack or tube-like body is traversed
by irregular parietal apertures of variable form and
size, and is partially supported by a framework of
united parenchymal spicules. The distal radial ray
of the dagger-shaped hexact hypodermalia bears a
floricome (T.egerin.e), ....

X. EuPLECTELLINiE.

With basal root-tuft. The parenchyma contains oxyhex-
asters {EuiilecMJa, Owen),

Attached by a firm basis to solid bodies. The
parenchyma contains discohexasters {Regadrella,
0. Schmidt), .....

XI. Eupledella, Owen.

In each of the meshes which are formed by the inter-
section of the longitudinal and circular principal
strands of fibres, there is a parietal pore. The
parietal pores are thus arranged in regular transverse
and longitudinal rows, ....

Perforate meshes alternate, usually with some regularity,
with imperforate, so that the parietal pores are
disposed in oblique spiral rows or irregularly,

1 Curved tube, with externally projecting ridges and

superior annular cuff, ....

Straight tubes, without externally projecting ridges,

2 In the middle of each parietal convexity, lying between

every four parietal pores, there is a strongly
developed oxyhexact, with a distal ray projecting
for some distance outwards,
Without these principal hexacts in the centre of the
convex parietal elevation, ....

3 The annular membrane of the parietal gaps contains

sceptre-like monacts, ....

No sceptres in the annular membrane of the parietal

gaps, ......

4 The annular membrane of the parietal gaps includes,

besides the sceptres, S-shaped clasps.

Without S-shaped clasps, ....

b The abundant parenchymal oxyhexasters have long and
extraordinary thick principal rays.

Parenchymal oxyhexasters, not as yet observed, and
therefore either very sparsely present or altogether
absent, ......

PACK AND PLATE

XVL

XI.

XII.

Eupledella oicenii.

78

VI.

Eupledella aspergillum, . 64

I.-IV.

76

Eujiledella rnrumer,

3

i

5

Eupledella jovis, . 77

VI.

Eupledella suberea, , 73

v., VI.

Eupledella crassistellata, . 81

XIII.

Eupledella nodosa,

82

XIV.

412

THE VOYAGE OF H.M.S. CHALLENGER.

Xn. Reijadrella, 0. Schmidt, with tlie single species,

Xin. On the internal surface of the somewhat firm and

compact tube-wall, there is a rectangular meshwork
formed from longitudinal and annular bands, and
forming pit-like depressions in the meshes
{Holascm, F. E. S.),
The internal surface of the very loose body- wall exhibits
irregularly distributed round apertures of the
efferent canals [Malacosaccus, F. E. S.), .

XIV. Holascm, F. E. S.

The principalia are oxypentacts or oxyhexacts, .
The principalia are oxytetracts, besides which oxyhexacts
also occur, .....

1 The gastralia are hexacts, ....

The gastralia are pentacts, ....

2 The gastralia are oxypentacts. The parenchyma con-

tains fibulfe, .....

The gastralia are oxyhexacts. No fibulfe in the
parenchyma, .....

XV. Malacosacxus, F. E. S.

The parenchyma contains small discohexasters with
numerous terminal rays on the transverse terminal
discs of the principal rays,

Regadvella phoenir,

PAGE AND PLTAE

84

XIII.

XIV.

XV.

1

2

Holascus stellatus,

Hulascm polejaevii,

86

XIV., XV.

89

XVII.

HoJasi'us fibulafiis,

87
XV., XVI.

Holascus rklleyi,

90

XVII.

Malacosaccus vastus

There are in the parenchyma no small discohexasters
with numerous terminal rays on the transverse
terminal portions of the principal rays. Floricomes
with two or five terminal claws on each terminal ray, Malacosaccus unguiculatus

XVI. TiEGERIN^, F. E. S.

Roundi-sh, usually circular parietal gaps. Over the
oscular aperture there is an arched cupola of radial
spicules {Tsecieria, F. E. S.),
Parietal gaps irregular, angular (Walteria, F. E. S.),
XVII. Tsegefia, F. E. S., with the single species,

91

XVIII.

93

XIX.

XVIII. Walteria, F. E. S., with the single species,

XIX. ASCONEMATID.^;, F. E. S.

The wall of the cup-, funnel-, or tube-shaped body consists
of a thin loose plate (AscoNEMATiNiE),

The wall of the (always 1) stalked goblet-shaped body is
somewhat thick and firm (Sympagellin.*),

The body has a mushroom-like shape, with long hollow
stalk (Caulophacin^), ....

XVII.
XVIII.

Txgeria pulchra, 94

VII., VIII., XI.

Walteria flemmingii, . 96
IX., X., XI.

XX.
XXIII.

XXVII.

REPORT ON THE HEXACTINELLIDA.

41:5

XX. Cup- or fiinncl-shaped. The principalia are long diacts

(Asccmema, Sav. Kent.),
Tubular. The principalia include not only long diacts,
but also o.xyhcxacts (Aulascu.s, F. E. S.), ..

XXI. Asconema, Sav. Kent, with the single species,

XXII. Aulascus, F. E. S., with the single species,

XXIII. SYMPAGELLINiE.

On each of the terminal branches of a slightly ramified
stock, there is a terminal oval goblet. The dermal
pinuli are pentacts. The parenchyma contains
small " roller stars " (Sympagella, O. Schmidt),

The oval body, which is narrow superiorly, contains
dermal hexact pinuli {PoJijrhahiluit, F. E. S.),

The clay-pipe-like body contains in its parenchyma
discohexasters, with long principal rays and short
terminals {Balanites, F. E. S.), .

XXIV. Sijmpagella, O. Schmidt, with the single species,

XXV. Polyrliahdus, F. E. S., with the single species, .
XXVI. Balanites, F. E. S., with the single species,

XXVII. Caulophacin.«, F. E. S.

The body, which is flattened from above downwards, is
either biconvex, or convex above and concave below
by the folding over of the marginal portion. With-
out sickle-rosettes in the stalk ( Cauloph acus, F. E. iS . ),

"With sickle-rosettes in the long, hollow, rough stalk
(Trarhyranlus, F. E. S.), ....
XXVIII. Caulopha<:us, F. E. S.

The margin of the body is folded over downwards. On
the convex gastral surface there are long pentact pinuli

The disc of the body is biconvex or with a downward
folded border. The gastral pinuli are all hexacts, .

XXIX. Traehycatdm, F. E. S., with the single species,

XXX. RoSSELLIDiE.

Unstalked (or quite shortly stalked), cup- or sack-sliaped
forms, ....

Seated on a distinct long stalk, .
XXXI. Principalia do not tend to become soldered together, nor

do they even in oM specimens form a connected
framework, .....

The principalia, which show a tendency to be united and
soldered together by synapticula, form in the older
parts of the body a connected framework, .

PAGE AND PLATE
XXI.
XXII.

Asconema setubalense, . 116

XXI.

A ulascus jolinntoni, . 118

XXIV.
XXV.

XXVI.

Sympagella nux, .

Polyrhahdus ovifornii.^.
Balanites pipetta.

120

XXII.

121

XXIII.

122

XXIIl.

XXVIII.
XXIX.

Caulophacus latus, . 124

XXIV.

Caulophacus eleganx, . 126

XXV., XXVI.

Trachycaulus gurlitfii, . 1 28

XXVI.

XXXI.
XLIV

XXXII.

XLIII.

414

THE VOYAGE OF H.M.S. CHALLENGER.

PAGE AND PLATE

XXXIL External surface surrounded by a veil of tangential rays

belonging to pentact pleuralia. The autodermalia
are predominantly pentacts {RosseJla, Carter),

External surface without a veil of tangential rays,
XXXIII. Rossella, Carter.

The four tangential rays of the pentact pleuralia intersect
at right angles, .....

The four tangential rays of the pentact pleuralia do not
intersect at right angles but form acute angles with
one another, .....

XXXIII.
XXXIV.

Rossella velata.

143

XXXIV.

XXXV.
XXXVL

The external surface of the body bears radially projecting
papillfe, from each of which a tuft of long pleuralia
projects. The root tuft is fixed in the mud
{Pohjlophus, F. E. S.), .

The external surface of the body bears no papillae or
basal tuft, ......

Pohjlophus, F. E. S., with the single species, .

Body-wall thin and loose. Oscular margin (always?)
with a cuff-like fringe of marginalia (Bathydorus,
F. E. S.),

Body-wall somewhat thick and firm. Oscular border
smooth, without marginalia,

XXXVII. Bcdhijdoms, F. E. S.

The autodermalia consist solely of rough oxytetracts,
The autodermalia consist solely of diacts and monacts.
The j^arenchyma contains besides oxyhexasters also
discohexasters, .....

1 In many of the parenchymal oxyhexasters the jsrincipal

rays are very much shortened, and the terminals thus
appear to spring directly from a central node,

The terminal rays of the parenchymal oxyhexasters all
spring from developed principals, .

2 The terminal rays of many parenchymal oxyhexasters are

curved in an S-shaped fashion,

The terminal rays of many parenchymal oxyhexasters
are somewhat undulating in their curvature,

XXXVIII. From the external surface of the body pointed pleuralia

project, either isolated or united in tufts. The
autodermalia are pentacts and tetraots {Acaidliascus,

F. E. S.),

The external surface of the body is smooth, without
pleuralia, ... ...

Rossella antarctica, . 139

LV.

XXXV.

XXXVI.

Polylophtis pkilippinensis, 133

LIII., LIV.

XXXVII.

XXXVIII.

1

Bathydorus hacuUfer,

Batliydorus stellatus,

154

LIX.

152

LIX.

Bathydonis fimbriatus, . 151

LVIII.

Bathydorus spinosus, , 153

LIX.

XXXIX.

XL.

REPORT ON THE HEXACTINELLIDA.

415

PAGE AND PF.Al K

XXXIX.

XL.

XLI.
XLII.

XLIII.

XLIV.

XLV.
XLVI.

Acanthascus, F. E. S.

The dermal skeleton contains exclusively pentacts,

The dermal skeleton contains pentacts and tetracts,
I Isolated pleuralia, .....

Pleuralia proj(>cting in groups from small elevations on
the external surface, ....

The autodermalia of the cocoon-shaped body are all

tetracts {Lanngiwlla, O. Schmidt),
Besides the tetract autodermalia many diacts occur

{Rhahdoralyptiis, F. E. S.),

Lanuginella, O. Schmidt, with the single species,

Rhabdocalyptus, F. E. S.

Funnel-.shapod. Autodermalia exclusively diacts.

Sack-shaped. Autodermalia are pentacts, tetracts,
diacts and monacts, ....

With great parenchymal hexasters, in which each of the
short principal rays bears six long diverging
terminals, which gradually increase in thickness
towards the round outer end, and are beset all
round with backwardly bent pointed hooks
(Aulocalij.r, F. E. S.), with the single species.

Without .such great hexasters, whose long terminal rays
increase in thickness towards the round outer end,
and are beset all round with backwardly bent
pointed hooks {Euryplegma, F. E. S.), witli the
single species, .....

From the lateral surface of the body long pleural diacts

project (Caulocah/.i; F. E. S.),
The external surface of the body is smootli,

Caulocalyx, F. E. S., with tlie single species, .

The oscular margin is not folded back, but remains up-
right, surrounding with a sharjj fringe the superior
gastral cavity opening of the cup-sbajied body
{Crateromorpha, Gray), ....

The oscular margin is folded back for some distance so
that a portion of the gastral surface forms tlie
external wall of the body (AuIocJione, F. E. S.),

XLVII. Crateromorplta, Gray.

Body (not including stalk), higher than broad, .
Body (not including stalk), broader than high, .

Acunthasem ditbius,

1

Acanthascus grossularta,

Acanthascus cactus,

XLI.

XLVII.

XLVIIL

1
•1

147

LVII.

145

LVI.

148
r.vii.

XLII.

Lanuginella pupa,

130

LIII.

Rhahdocalyptm mollis, .

1.55

LXIV.

Rhabdocalyptus roeperi, .

158

LXV.

Atdoralyx irregularis, . 174

LX.

Euryplegma auriculare.

176

CII.

XLV.

XLVI.

Caulocaly.e tener.

172

LXIX.

416

THE VOYAGE OF H.M.S. CHALLENGER.

XLVII. Gi-aieroiiwrpha — continued.

1 The form of the body is tulip-like. The oscular fringe

forms a thin annular plate almost 1 cm. in height.
The terminal rays of the parenchymal oxyhexasters
are straight, .....

The form of the body is oval. The oscular marginal
fringe is inconspicuous, about 2 mm. The terminal
rays of the parenchymal oxyhexasters are distinctly
S-shaped, ......

2 The oscular marginal fringe is sharply angular, and forms

the broadest portion of the body. The terminal
rays of the parenchymal oxyhexasters are bent
round at their extremities,

The oscular fringe is less extended than the median
portion of the puffed out body. The terminal rays
of the parenchymal oxyhexasters are straight,

XLVin. Aulochone, F. E. S.

The parenchyma contains, besides other spicules, ttori-
comes, ......

The tloricomes are altogether absent,

XLIX. DicTYONiNA, Zittel.

With distinctly developed uncinates (Unoinataria,

F. E. S.),

Without distinct uncinates (Inermia), .
L. Uncinataria.

The dermal and gastral skeletons contain clavulie
(Clavularia), .....

Without clavulse, but with scopulae (Scopularia),
LI. Clavularia, with the single family Farreidse, and the single
genus Farrea, Bwk.
Branched tubular stock with a straight principal stem,
the dermal clavulse have mostly smooth edges,

Branched anastomosing tubular stocks, without a straight
principal stem, .....

1 The parenchyma contains no discohexasters,

Discohexasters occur in the parenchyma,

2 The hexasters occurring in the parenchyma are ex-

clusively discohexasters, with long terminal rays, .

Of the parenchymal hexasters, some are oxyhexasters,
with short terminal rays, and others discohexasters,
with Ions terminals, ....

PAGE AND PLATE

Craieromarpfia meyeri, . 161

LXI.

Crateromorplia thierfelderi, 164

LXII.

Crateromorpha murrayi, 164

LXIIL

Crateromorpha fumida, . 166

LXVn., LXVIII.

Aulochone lilium,

171

Lxvin.

Aulochone cylindrica,

168

LXVI.

LXVIII.

L.

LXX.

LI.

LII.

Farrea davigera,

1

287

LXXV.

Farrea occa, . .277

LXXI.-LXXIIL, LXXVL

Farrea sollasii, .

. 286

LXXIV.

Fairea oosmaeri,

. 286

LXXIV.

REPORT OX THE HEXACTIXELLIDA

4i:

LII. SCOPULARIA.

Tlie dictyonal framework exhibits a honeycomb-like
structure ; the wall of the cup or tube-like skeleton
is regularly traversed by hexagonal parietal apertures

(il£LITTIOSID;E), .....

The dictyonal skeleton exhibits no such honeycombed
structure, ......

LIII. ^lELiTTioNn).E, with the single genus AphrocalHsles, Gray

The body consists of a dichotomously branched tube,

The body has the form of a cup with lateral diverticula,

1 The traverse lateral diverticula of the cup-shaped body

are flat and pouch-like, and disposed perpendicu-
larly. The parenchyma contains no oxy-, but only
disco- and sphaero-hexasters,

The lateral diverticula are glove-finger-shaped. The
parenchyma contains oxj'hexasters,

2 The oxyhexasters of the parenchyma are all much

elongated in a longitudinal direction. The four
principal rays at right angles to the latter, remain
simple and short, .....

The oxyhexasters of the parenchyma are, for the most
part, uniformly stellate. Some sphaerohexasters

PAGE A!n) PLATE

occur, ......

LIV. Body plate or cup-shaped. The dictyonal framework

exhibits regularly alternating, straight, or funnel-
shaped passages, belonging to the afferent and
efferent canals, which penetrate the body wall
transversely (CosciNOPOBiDiE),

The dictyonal framework does not exhibit regularly

alternating, straight, afferent and efferent passages,

traversing the wall, ....

LV. CosciNOPOEiD^, Zittel, with the single genus Chonelasma, F.E.S.

The body has a flat tabular form.

The body has the form of a cup with lateral thimble-
like diverticula, .....

LYI.

The dermal skeleton contains strongly developed
pentacts, without rudiment of a sixth distal ray,
and with tangential rays beset with spines on the
outer surface, .....

LIII.
LIV.

Aphrocallistes ramosus, . 317

LXXXVI.

1

Aphrocalligtes vagtus, . 315

LXXXV.

Aphrocalligtes heatrix, . 309

LXXXIV.

Aphrocalligtes hocagei, . 311

LXXXni., LXXXIV.

LV.

LVII.

LVI.

Chonelasma falyj-, . 324

LXXXIX.

Clionelagma lamella, . 319
Lxxxvu., Lxxxvni.

Hexacts occur in the dermal skeleton, with a short
distal laterally spinose ray, ... 1

1 The parenchyma contains discohexasters, with long

rough principal rays, and short S-sbaped terminals, Chonelatma hamatum.

321

SCI.

(ZOOL. CHALL. EXP. — PART LIII. — 1887.)

Ggg53

418

THE VOYAGE OP H.M.S. CHALLENGER.

LVI. — continued.

Tho parenchymal discohexasters have only sborfc
principal rays, .....

LVII.

The body consists of a system of branched anastomosing
tubes, which form either an irregular feltwork, or
the wall of a cup (EuREHDiE, F. E. S.),

The more or less thick wall of the usually cup-shaped
body is traversed by oblique or curved afferent and
efferent canals (TRETODiCTYiDiE),

LVIII

PAGE AND PLATE

CJicmelasma doederleinii, 322

xc.

LVIII.

LXIII.
LIX.

LX.

The tubular feltwork does not enclose a central funnel-
shaped cavity. Eurete {Semper), Carter, .

The anastomosing tubular work encloses a central
funnel-shaped cavity, and thus forms the wall
of a cup, ......

LIX. Eurete, Carter.

The dermal scopulre have four smooth prongs, with
spindle or club-shaped extremities,

The dermal scopuhc bear smooth pointed prongs,

The dermal scopulas bear prongs, with club-shaped
barbed extremities, . . . .1

1 The parenchyma contains oxyhexasters whose principal

rays are about twice as long as the short terminals, Eurete hmeerhanldi ,

The parenchyma contains exclusively hexasters whose
principal rays are much shorter than the terminals, 2

2 The hexasters of tho parenchyma all bear long pointed

terminal rays, .....

Eurete semperi, .

292

LXSVII.

Eurete schmidtii.

. 293

LXXVIII.

Eurete marshalli.

LX.

LXI.

The hexasters of the parenchyma are provided with

knobs or transverse discs at the end of their terminal

rays, and are therefore sphsro- or discohexasters, .

3 Both the dermal and gastral scopulre exhibit a distinct

break-like bend on the pedicel of the prongs.

The pedicels of the prongs are not thus bent either in
the dermal or gastral scopulfe, but are straight or
gently curved.

The internal surface of the cup-shaped body exhibits no
longitudinal ridges on the dictyonal skeleton. The
whole body-wall consists of a loose feltwork of
thin-walled tubes {Periphragella, Marshall),

On the internal surface of the somewhat firm dictyonal
framework of the cup-shaped body there are distinct
fu-m longitudinal ridges {Lefroyella, Wy v. Thomson),
Periphragella, Marshall, with the single species.

Eurete farreopsis.

Eurete carteri.

LXI.

LXIL
Periphragella

297

LXXIX.

297

LXXIX.

. 295

LXXIX.

. 296

LXXVIII.

. 299

LXXX., LXXXI.

KEPORT ON THE HEXACTINELLIDA.

419

LXII. Lefroyella, Wyv. Thomson, with the single species,

LXIII. TretodictyidjE, F. E. S.,

Bush-like ramified stock with somewhat solid cylindrical
branches, on which annular or spiral zones of in-
current and excurrent regions alternate {Sclero-
thamnus, Marshall), ....

Body cup-, tube-, or ear-shaped, ....
LXIV. Sderothanmus, Marshall, with the single species,

LXV. Body surrounded by a leaf-like enveloping capsule

{Fieldimjia, Sav. Kent), ....
Body witliout leaf-like enveloping capsule,
LXVI. Fieldingia, Sav. Kent, mth the single species,

LXVII. The parenchyma contains scopula-like spicules bearing

numerous prongs with transverse terminal discs
springing from a terminal knob on the long pedicel
{Cijrf avion, F. E. S.), . . . . LXVIII.

Body in the form of a ramified, slightly anastomosing
tubular feltwork, or of a cup, in which the thick
wall is traversed by oblique, or twisted afferent and
efferent canals, without scopula-like parenchymal
spicules with radial prongs (Hexactinella, Carter), . LXIX.
LXVIII. Cyrtaulon, F. E. S.

The body-wall is supported by a feltwork of plates of
the dictyonal framework, which enclose round
cavities. The nodes of intersection are here and
there thickened, and near the surface of the body
beset with warts, ....

PAGE AND PLATE

Lefroyella decora, . 301

LXXXII.

LXIV.

LXV.

Sderothamnus claudi,

. 337

XCVIII.

LXVI.

LXVII.

Fieldingia lageitoides,

. 333

XCVII.

Cyrtaulon sigsbeei,

The dictyonal framework forms an irregular network
of beams and strands which traverse the cylindrical
body. Nodes of intersection without thickening or
warts, ......

LXIX. Hexactinella, Carter.

The body consists of a dichotomously ramified feltwork
of uniformly wide, here and there anastomosing tubes.

The body consists of a dichotomously branched tube,
the terminal twigs of which exhiljit funnel-shaped
expansions, .....

The body forms a cap seated on a firm basis, becoming
markedly wide, but somewhat laterally compressed,

LXX. The body consists of a system of thin- walled tubes of

approximately uniform calibre, with an intermediate
system of cavities (JLEANDRosPONGiDiE), .

331

XCII.

Cyrtaulon solutus,

. 331

XCII.

HexadineUa tuhulosa,

. 326

XCIII.

Hexactinella lata.

. 327

XCIV., xcv.

Hexactindla ventilabrum, 329

XCTI.

LXXI.

420

THE VOYAGE OF H.M.S. CHALLENGER.

PAGE AND PLATE

LXXL MiEANDBOSPONGID^, Zittel.

The very regular dictyonal framework enclosing cubical
meshes has perforated nodes of intersection with
octahedral edges {Aulocystis, F. E. S.),

The nodes of intersection of the dictyonal framework
are simple and imperforate,
LXXII. Aulocystis, F. E. S.

The octahedral edges of the perforated nodes of inter-
section of the dictyonal framework are formed of
simple oblique round buttresses. The parenchyma
contains abundant discoliexasters in which the
terminal rays are as long as the prtucipals, .

The octahedral edges of the perforate nodes of the
dictyonal framework are formed of the sometimes
repeatedly perforate plates, which extend between
the intersecting beams. The principal rays of the
jiiarenchymal discohexasters are much shorter than
the termiuals, .....

LXXIIL The nodes of intersection of the dictyonal framework

are thickened and beset with broad tuberculate
warts (Mj/Uusia, Gray), ....
The nodes of intersection of the dictyonal framework
possess no broad tuberculate warts (Dactylocalyx,
Stutchbury), . . , , .

LXXIV. Myliusia, Gray, with the single species,

LXXV. Dactylocalyx, Stutchbury.

The body forms a flatly expanded thick-walled cup, which
consists of a system of narrow tubes, only 1 to 2
mm. in width. The parenchyma contains, besides
other isolated spicules, oxyhexasters,

The body forms a dee^j thick-walled goblet, which
consists of a system of narrow (only 1 to 2 mm.
in width) tubes. No oxyhexasters in the paren-
chyma, ......

The flat body consists of tubes 3 to 5 mm. in width.

LXXII.
LXXIIL

Aulocystis zittelii,

359

CIV.

Aulocystis grayi, .

LXXIV.

LXXV.

Myliusia calloeyathus,

357

CIV.

352
cm.

Dactylocalyx pumieeus, . 346

Dactylocalyx suhglobosus, 347

XCIX.

Dactylocalyx patella, . 348

GEOGRAPHICAL DISTHIBUTION.

HISTORY.

A careful synopsis of the habitats of all the Hexaetinellida definitely known in 1873
is given by Carter in his memoir On the Hexactinellidas and Lithistidce.* The results
of Carter's summary are tabulated here with several additions, and with the use of the
specific designations which I have adopted.

Species,

Locality.

1. Euplectella aspergillum, Owen.

2. Euplectella cucumer, Owen.

3. Hahrodidijum speciosum, Wyville Thomson.

4. Asconema setubalense, Sav. Kent.

5. Stjmpagella mix, 0. Schmidt.

6 Lamiginella pupa, O. Schmidt.

7. Pohjlopihus phiUppiinensis, Gray.

8. Rossella aniwdira, Carter.

9. Rossella velata, Wyville Thomson.

10. Crateromorpha meyeri. Gray.

11. Hyalonema sieholdli, Gray.

12. Hyalonema lusitanicum, I3ocage.

13. Pheronema annx, Lcidy.

14. Pheronema carpenteri, Wyville Thomson.

15. Pheronema grayi, Sav. Kent.

16. Pheronema liemisphxrira, Gray.

17. Semperella sdtuUzei, Semper.

18. Farrea occa. Carter.

19. Farrea sp. {woodicanli, Sav. Kent).

20. Farrea sp. {facunda, O. Schmidt).

21. Farrea (?) infundibularis. Carter.

22. Eurde sp. {simpHeissinia, Semper).

23. Aplirocallistes hocagei, Wright.

24. Aphrocallistes heatrix, Gray.

25. Fieldingia lageiloides, Sav. Kent.

26. Dadylocalyx pumiceus, Stntchbury.

27. Dadylocalyx suhglohosns. Gray.

28. Mylhuda calloryathus, Gray.

29. Aidocystis grayi, Bowerliank.

Philippines.

Seychelles.

Moluccas.

Portugal.

Florida.

Cape Verde Islands, St. lago.

Philippines.

Antarctic Sea (South Pacific).

South of Fffiroe Islands.

Philippines, Zebu.

Japan.

Portugal.

Santa Cruz, West Indies.

South of Fairtie Islands.

Portugal.

Philippines, Zelm.

Philippines, Zebu.

Seychelles.

Portugal.

Between Florida and Cuba.

Caribbean Sea, West Indies.

Philippines, Zebu.

Portugal.

Malacca.

Portugal.

Barbados, West Indies.

Jlalacca.

West Indies.

St. Vincent, West Indies.

1 Ann. and Mag. Nat. Hist, ser. 4, vol. xii. p. 349.

•422 THE VOYAGE OF H.M.S. CHALLENGER.

From this it appears that in 1873 six specimens were known from the neighbourhood

of Portugal, viz

1. Asconema setuhalense, Sav. Kent.

2. Hijaloiiema ludtanicum, Bocage.

3. Pheronema ijraiji, Sav. Kent.

4. Fan-ea sp.

5. ApTirocallistes hocagei, "Wright.

6. Fieldinrjia lagettoides, Sav. Kent.

from the PhiHppiues six species were also known, viz.:

1. Eupledella aspergiUum, Owen.

2. PohjlopliMS pliilipjnnensis, Gray.

3. Crateromorpha meyeri, Gray.

and from the West Indies five species :-

4. Pheronema Tiemisplixrieum, Gray.

5. SempereUa sclmltzei, Semper.

6. Eurete sp. {simpKcissima, Semper).

1. Phcronetna annx, Tueidj. I 3. Dadylocalyx pumiceus, Stutchhuij.

2. {Farrea ?) infundibularis, Carter. | 4. MyUusia callocyatMis, Gray.

5. Aulocystis grayi, Bowerbank.

and lastly, one or two species from widely separated localities.

From a list published a few years ago by Marshall,^ in which several additional forms
were included, and a few bathymetrical data were also furnished, the author drew the
following conclusions (oj:>. cit., p. 151) : — "From this survey it is seen that the Hexac-
tinelhda have a horizontal distribution from about 65° north latitude to about 50° south,
and a vertical distribution from about 500 to over 4000 fathoms. It may seem striking
that several localities have yielded a relatively large number of Hexactinellid species, e.g.,
Portugal, Florida (and the West Indies), and the Philippines, but it must be noted that
the two first regions have been specially explored by deep-sea expeditions, while the
Philippines are the home of an exceedingly keen fishing folk, who are wide awake to the
fact that any new form means gold to the finder."

Without entering on any detailed review of the several reports which have been made
on the distribution of the HexactineUida, I shall tabulate all the trustworthy and definite
notices of locality with which I am acquainted, so far as they refer to sufiiciently defined
species. The first of the two synoptic tables is arranged zoologically according to my
system, the second according to the localities. I have further marked all the localities on
the accompanying map of the world. In this way the state of knowledge independent
of the results of the Challenger Expedition is clearly indicated.

1 Zdtschr. f. wiss. Zool., 1875, Bd. xxv. Suppl., p. 150.

REPORT ON THE HEXACTINELLIDA.

423

TABLE I. — Synopsis of the Hexactinellida hitheeto knowxV, with their
Habitats, as ascertained apart from the Challenger Collection.

A. Clearly defined Forms.

Species.

Habitat.

Depth in
Fathoms.

Authority.

1. Suplectellaaspergillum, Owen,

1. Philippines, . . . .

Owen.

2. I'hilippines (Bohol, Zebu),

Bowerbank.

3. Philippines, ....

Semper.

4. Philipi)ines (near Zebu),

Gray.

5. Philippines (Zebu),

Moore.

6. Philippines (Zebu),

Chimnio and others.

2. Eupleetella suberea, ^\yr. Thomson,

1. North of Scotland,

AVyv. Thomson.

2. AVidely distributed in the N. Atlantic,

450-ilOO

Filhol.

3. EupkctcUa cucumer, Owen,

Seychelles Islands, ....

Owen.

4. Eupleetella jovis, 0. Schmidt,

Gulf of Mexico, St. Lucia, Grenada,

416 and 423

O. Schmidt.

5. EuplcctcUa oweni, Herkl. and Marshall, .

Japan, ......

Marshall.

Eupleetella oweni, Herkl. and MarshaU, .

Japan (.Sagami Bay),

iod--200

Dciderlein.

6. Regadrclla phcenix, 0. Schmidt, .

Giilf of Mexico, Barbados, Santa Cruz, .

221, 248, and

288

O. Schmidt.

7. Habrodictpum speciosum, Quoy and

Moluccas, Mauritius,

Quoy and Gaimard.

Gaiinard,

Valenciennes.

8. Dictjioealyx gmeilis, F. E. Schulze,

Gulf of Mexico (Bequia), .

1591

O. Schmidt.

9. Rhahdudictimm delicntum, O. Schmidt, .

Gulf of Mexico (Bequia),

1591

0. Schmidt.

10. Bliabdopeetella tuitinnus, O. Schmidt,

Gulf of Mexico (Grenada), .

291, 994

O. Schmidt.

11. Hertwiffia faleifera, 0. Schmidt, .

Gulf of Mexico (Dominica),

611

O. Schmidt.

12. Ascoiiciiia setubalcnse, Sav. Kent, .

1. Portugal (Cape St. Vincent),

SaviUe Kent, Filhol.

2. North-west of Scotland,

327-430

John Murray.

13. Siimpagella mix, 0. Schmidt,

Florida,

98-123

O. Schmidt.

14. Lanugmclla pupa, 0. Schmidt,

Cape Verde Islands (St. lago).

0. Schmidt.

15. Polf/fapkus pliilippuuusis, Gray,

Philippines (Zebu), .

A. B. Meyer.

16. Hosseflc irut'ii'dica. Carter, .

Antarctic Ocean, lat. 74° S.,long. -

75 "W".

J. Ross, Carter.

17. Jtossillii nlata, Wyv. Thomson, .

"West of Strait of Gibraltar,

651

Wyv. Thomson.

IS. AeanthaS'.us eactus, F. E. S.,

Jai>an (Sagami Bay),

100-200

DUderlein.

19. Rhabdocalyptus mollis, F. E. S., .

Japan (Sagami Bay),

100-200

Doderlein.

20. Crateromarpha meyeri, Gray,

1. Philippines (Zebu),

Gray.

2. Japan (Sagami Bay),

100-200

Doderlein.

21. Hyalonema sicboldii. Gray, .

Japan (Sagami Bay),

100-200

Gray.

Semper.

Max Schultze, Hilgen-

dorf, Doderlein, and

others.

( Grenada,

338, 576

)

22. Hyalonema kentii, 0. Schmidt,

Westlndies^ '^^^
(. Bequia,

565

583
1507

VO. Schmidt.

23. HyaloneTna (Stylocalyx) thomson ii, MarshaU,

North of Shetland Islands, .

550

W. Marshaa

24. ZfWnHcma sp., Wyv. Thomson, .

North of Hebrides, .

Wyv. Thomson.

25. Hyalonaua lusitanicum, Bocagc, .

Portugal,
Portugal,

Bocage.
Sarnie Kent.

26. Hyalonema cebuense. Biggin,

Philippines (Zebu), .

Higgin.

27. Pheroruma annce, Leidy,

West Indies (Santa Cruz), .

180, 248

Leidy.

O. Schmidt.

28. Pheronema grayi, Sav. Kent,

Portugal (Setubal)

Sav. Kent.

29. Pheronema carpcnteri, AVyv. Thomson, .

From Butt of Lewis (Hebrides) to
Gibraltar,

( 500-1000 )
1(300-800) i

Wyv. Thomson.

30. Pheronema hemispheericum, Gray,

1. Singapore, . . . . .

Gray.

2. Philippines (Zebu),

JIarshall and Meyer.

31. Sempcrella schultsei, Semper,

1. Philippines,

2. Philippines (Zebu),

Semper.

Marshall and Meyer.

32. Farrea sp. (occa, Bowerbank),

SeycheUe Island,

Owen.

Farrea sp. (oeca, Bowerbank),

Portugal,

Sav. Kent.

33. Farrea occa (Bowerbank), Carter,

Japan (Misaki).

Carter, Doderlein.

34. Farrea sollasii, F. E. S. ,

Japan (Sagami Bay),

100^200

Doderlein.

35. Farrea vvsmticri, F. E. S., .

Japan (Sagami Bay),

100-200

Doderlein.

36. Farrea sp. (woodwardi, Sav. Kent),

Portugal,

Sav. Kent.

37. Eurele sp. (aimplicisaima, Semper),

Philippines,

Semper.

38. Euretes^t.,

Philippines (Zebu), .

Marshall and Meyer.

39. Eurete schmidtii, . . . ,

Japan (Sagami Bay),

10d-i200

Doderlein.

40. Eurete farreopm. Carter, .

Philippines,

Carter.

41. Eurete sp., .....

Japan (Sagami Bay),

100^200

Doderlein.

424

THE VOYAGE OF H.M.S. CHALLENGER.
Table I. — continued.

Species.

Habitat.

Depth in
Fathoms.

Authority.

42. Pcriphraijella clisce, Marshall,

43. Aphrocallistes bocagei, Perceval Wright,

44. AphrocaUistes ramosiis, F. E. S.,
4.5. AphroriilUstes vastus, F. E. S.,

46. C/iniuIiisma dadcrhinii, F. E. S.,

47. i^linuchismii nihi.r, F. J'^. S.,

48. Ht.tadinella riulilnhrum, Carter,

49. Hcxactindla tubulosa, F. E. S.,

.'iO. Cyrtaulon sigsheei, O. Schmitlt,

51. Fieldingia lagettoides, Sav. Kent,

52. Sclerothamnus clausii, MarshaD,

53. Dactylor.alyx pumicms, Stutchbury,

54. Dacti/localyx subglobosus, Gray,

55. SeJcropJegnui conicnm, O. Schmidt,

56. Mi/fiusia callocyathus, Gray,

57. Aulocystis graiii^'Bovtexh^vtk,

58. Aulocystis zittelii, Marshall,

1. Moluccas, ....

2. Japan (Sagami Bay),
Cape Verde Islands, Japan,
South-east of Florida, Gulf of Mexico,
Coast of Spain and Portugal,

Japan (Sagami Bay),
Japan fSagami Bay),
Japan (Sagami Bay),
Japan (Sagami Bay),
Japan (Sagami Bay),
Japan (Sagami Bay),

Gulf of Mexico { <^"^f ^I'.fe"™''^^''
Portugal, . . . . .

Philippines (Zebu), .

West Indies (Bai'bados, St. Vincent),

Malacca, ....

West Indies (St. Lucia),

West Indies (North-west of Havanna),

West Indies (Morrolight), .

West Indies, ....

West Indies (St. Vincent), .

Barbados, ....

Philippines (Zebu), .

West Indies (Guadeloupe),

100-200

100-200
100-200
100-200
100-200
100-200
100-200

100-292

116
190
292

100
150

Jlarshall.

Doderlein.

Perceval Wright.

O. Schmidt.

Sav. Kent, Milne-
Edwards, Filhol, and
others.

Dtiderlein.

Doderlein, Gottsche.

Doderlein.

Doderlein.

Carter, Doderlein.

Doderlein.

O. Schmidt.

Sav. Kent.

Marshall.

Murie.

Stutchbuiy, O. Schmidt.

Gray.

O. Schmidt.

O. Schmidt.

O. Schmidt.

Gray.

Bowerbanlc.

O. Schmidt.

Meyer and Marshall.

O. Schmidt.

B. Insufficiently defined Forms.

1. E II pi r, -till a sp.,

2. EliplrrlilhlS^.,

3. EuphctUla sp. ,

4. IloUenia pourtcdesii, O. Schmidt,

5. HoHenia saccus, O. Schmidt,

6. Hiinhiiifma sieboldi? .

7. Hiiidniuiiia sp.,

8. Lciubolidiam, O. Schmidt, .

9. Lyssacine,

10. Fiinrafacunda, 0. Schmidt,

11. Farrea infundibuUforniis, Carter,

12. Farrea gassioU, Bowerbank,

13. Farrea pociUum, Bowerbank,

14. Fan'ea parasitica, Bowerbank,

15. Farrea spinulenta, Bowerbank,

16. Farrea perarviata, Bowerlmnk,

17. Fa7i-ea irregularis, Bowerbank,

18. Diarctula cornu, O. Schmidt,

19. Diaretula muretta, O. Schmidt,
'■iO. Cyathella lutca, O. Schmidt,

21. Syringidium zittelii, O. Schmidt,

22. Cystispongia superstes, 0. Schmidt,

23. Diirfiiloiiihjx potatorum, O. Schmidt,

24. Jufindht rompressa, 0. Schmidt,

25. Sdcropkgma lanterna, 0. Schmidt,

26. Scleroplegma seriatum, 0. Schmidt,

27. Scleroplegma herculcum, O. Schmidt,

28. Diplacodium mixtum, O. Schmidt,

29. Pachaulidium sp. ?, O. Schmidt,

30. Bhabdostauridium retortula, O. Schmidt,

North Sea, ....

Between Ireland and the Faeroe Islands,

Mediterranean,

Florida (Sand Key), .

Florida,

Between England and Ireland,

Gulf of Mexico (Grenada), .

Gulf of Mexico (Bequia),

West of Norway,

Florida, Cuba, Gulf of Mexico,

Caribbean Sea,

West Indies, ,

West Indies, .

West Indies, .

Tripoli, . _ .

West Indies, .

Algiers,

Gulf of Mexico (Morrolight),

Gulf of Mexico (Morrolight),

Gulf of Mexico (Bequia),

( St Lucia,

I Martinique,
Gulf of Mexico -j Santa Cruz,

I Guadeloupe,

(^Morrolight,

("Yukatan,
Gulf of Mexico < Morrolight,

( Martinique,
Gulf of Mexico, St. Lucia, .
Gulf of Mexico, 23° 2' N., 83° 13' W.

Gulf of Mexico! 2^jJ'^N.^^-j.j_

Gulf of Mexico (Morrolight),

Gulf of Mexico (Santa Cruz),

2.5°.33'N., S3°21' W.,

Gulf of Mexico (Morrolight),

Gulf of Mexico (Santa Cruz),

GuH of Mexico, 23'' 54' N., 88° 55' W,

134-324

416
1507
1081

805
805
1591
116
136
218, 245
138, 150, 87S
200-450
20(?)
292
136
151
287
320
292
200-300
580
101
292
680
804

Bowerbank.

Carter.

]\Iarsball.

O. Schmidt.

O. Schmidt.

Carter.

O. Schmidt.

O. Schmidt.

Armauer Hansen.

O. Schmidt.

Carter.

Bowerbank.

Bowerbank.

Bowerbank.

Bowerbank.

Bowerbank.

Bowerbank.

O. Schmidt.

O. Schmidt.

O. Schmidt.

O. Schmidt.

O. Schmidt.
O. Schmidt.

[■ 0. Schmidt.

O. Schmidt.

i 0. Schmidt.

O. Schmidt.
0. Schmidt.
0. Schmidt.

J

REPORT ON THE HEXACTINELLIDA.

425

TABLE II. — Synopsis of the Habitats of Hexactinellida, as known apart
FROM THE Challenger Collection, arranged Geographically.

I.— ATLANTIC OCEAN.

Locality

A. Mediterranean.

1. Algiers, . . . .

2. Tripoli, . . . .

3. Naples, . . . .

4. Mediterranean,

B. North Sea.

5. North Sea, . . . .

C. Irish Channel.

G. Irish Channel,

D. North Atlantic.

7. West of Norway,

8. North of .Shetland Islands,

9. North of Scotland, .

10. North of Hebrides, .

11. From Butt of Lewis to Gibraltar, .

12. North West of .Scotland, .

13. Between Ireland and Fcei'cie Island

14. "West from English Channel,

15. Coast of Spain and Portugal,

16. Off Portugal,

Depth in
Fathoms.

17. Cape Verde Islands,

18. Florida," .

19. South-east of Florida, .

20. Cuba,

E. GnLF OF Mexico and Caribbean Sea.

21. Lat. 23° 2' N., long. 83° 13' W.,

22. Lat. 23° 54' N., long. 88° 55' \V.,

23. Lat. 25° 33' N., long. 85° 21' W.,

24. Yukatan,

25. North West of Havanna, .

F. Antilles.

26. Santa Cruz,

27. Bequia,

>»

28. Guadeloupe,

29. Dominica,

lOSl
550

500-1000
327-430

98-123

287
804

101

20(?)

190

288

218, 245

580

580

1591

1591

1507

1591

583

138, 150,

878

150

611

Species.

Dictyonine {Farrea in't^pjifarh, Bowerbank),
Dictyonine (Fanra xpiiiiUciUttj Bowerbank),
Dictyonine, ....

Lyssacine (Euplectella sp., Marshall), .

Lyssacine {Euplectella sp., Bowerbank),
Lyssacine (Hyalonema sicholdii (.'), Carter),

Lyssacine, .....

Hijiiloniiiui {Sf '/local ifx) thmumml, Mar.shall,
Ei(plertrllii <iiiUn'r, WyviUe Thomson, ,
Miialoncma sji., Wyv. Thomson,
PJieroi\^ina [Holtenia] carpenteri, Wyv.

Thomson, and Hjialoneina {lusitanicum?)

Bocage, AVyv. Thomson,
Asconeina aetubahiisc^ .Sav. Kent,
JjyfiSAciue {Euplcctclta sp. , Vniter),
Aplirncallistes boccu/ei, Wright, .
AphrocalHstes bocaf/ei^ Wright, .
Euplectella {fispergillum ?), Owen,

Asconema selnhalctise, Sav. Kent,
Hyalonema lusitanicum, Bocage,
Pkeronema grayi, Sav. Kent,
Farrea sp. , and Farrea [looodwardi?), sp.,
Fieldiwjia lagcttoides, Sav. Kent,
Lanuf/inella pupa, O. Hchniidt, .
AphrocalHstes bocagei, Wright, .
Si/mpaffella nux, O. Schmidt,
Lyssacine [Holtema pourtalcsii, O. .Schmidt), .
Lyssacine (Holtenia saccus, O. Schniiilt),
Dictyonine (Farrea facuivJa, O. Schmidt),
Aphnjcallistes boc'agci, Wright, .
Dictyonine (Farrea facunda, O. Schmidt,)

Dictyonine (JoaneUa comprcssa, O. Schmidt), .
Dictyonine (Rhabdoatauridium, rctortula,

O. Schmidt),

Dictyonine (Sclernplerima herculcum, O- Schmidt),
Dictyonine (C'/stixpoit{/ia supcrstcs, O. .Schmidt)
Dacti/localyx subyloboaus, Cray, .

Rcgadrella plMnix, O. Schmidt, .
Pheroncma anna;, Leidy, .

Dictyonine (Siirinyidium zittclii, O. Schmidt),
Dictyonine (Pachaulidium sp., O. Schmidt),
Dictyonine (Sclcropkgma herculcum, O. Schmidt),
Dictiiocal ;ix gracilis, F. E. Schiilze,
Jilialxhdietyum (ttiiratum, O. Schmidt, .
Lyssacine {Lirobntiillitm sp., O. .Schmidt),
Dictyonine ('>/^A>//i( liil/'i, O. Sclimidt),
Hi/aloncnia kentii, O. Schmidt, . •

Dictyonine (Si/ringidum Mtelii, O. Schmidt),

A ulociistis zittelii, Marshall,
Hertwigia falcifera, O. Sclimidt,

Authority.

Bowerbank.
Bowerbank.
F. E. Schulze.
Marshall.

Bowerbank.

Carter.

Armauer Hansen.

Marshall.
Wyv. Thomson.
Wyv. Thomson.
Wyv. Thomson.

John Murray.
Carter.

Wyv. Thomson.
Wyv. Thomson.
Higgin and Wyv. Thom-
son.
S.iville Kent, Filhol.
Socage, Sav. Kent.
Sav. Kent.
Sav. Kent.
Sav. Kent.
O. Schmidt.
O. Schmidt.
O. Schmidt.
O. Schmidt.
O. Schmidt.
O. Schmidt.
O. Schmidt.
O. Schmidt.

O. Schmidt.
O. Schmidt.

O. Schmidt.
O. Schmidt.
O. Schmidt.

O. Schmidt.
Leidy.
O. Schmidt.
O. Schmidt.
O. Sclimidt.
O. Schmidt.
O. Schmidt.
O. ScRmidt.
O. Schmidt.
O. Schmidt.
O. Schmidt.
O. Schmidt.

O. Schmidt.
O. Schmidt.

(ZOOL. CHALL. EXP. — PART LIII. 1887.

Ggg54

426

THE VOYAGE OF H.M.S. CHALLENGER.

Table II. — continued.

Locality.

Depth in
Fathoms.

.Species.

Authority.

30. Martinique, .

31. St. Lucia, '. '.

32. St. Vincent,

33. Barbados, .

34. Grenada,

35. Mor'roUght, .

36. Gulf of Mexico, !

37. Caribbean Sea,

38. West Indies,

»»

565
136
416
116
116
151

221

100

423

291, 994

338, 576

416

292
805
805

Btialonema Jcenlii, O. Schmidt, .
Dictyonine {St/riugidium zitteJii, 0. Schmidt), .
Euptectclla jovis, O. Schmidt,
Dacti/locahix suhglobosus, Gray, .
Dictyonine (Siiriwjiilimn zittcHi, O. Schmidt), .
Dictyonine {Diwt^ihiraftix potatorum, O. Schmidt),
Ciirtaiijon siti.^lxri, O. Scliniidt, .
Dacti/localiixpumiccus. Stutchbury,
Aulocystis grayi, Bowerbank,
RegmlreUn ph<jenix, O. Schmidt,
Ctirtaulnii i^ifinhi'ci, 0. Schmidt, .
Aiiliiriisth zitti'liU'Mavsha.W.
Eiipliiiilld juris. 0. Hchmidt,
RhaliiloprrtcUa tintinnus, 0. Schmidt, .
Hiiahjnciiiii kintii, O. Schmidt, .
Lyssacine {Hifi/onema sp., 0. Schmidt),
Ciirtatthni siiishcci, O. Sclimidt, .
Scleropiff/rna ronicum, O. Schmidt.
Dictyonine {Diaretula cornu, O. Schmidt),
Dictyonine {Diaretula cornu, 0. Schmidt),
Dictyonine {Farrea factauia, 0. Schmidt),
Dictyonine (ScleropUgma lanterna, O. Schmidt),
Dictyonine (Farrea infundibul iformis, Carter),
Dictyonine {Farrea gassioti, Bowerbank),
Dictyonine {Farrea podUuin, Bowerbank),
Dictyonine {Farrea parasitica, Bowerbank),
Dictyonine {Fatrea perarmata, Bowerbank), .

0. Schmidt.
O. Schmidt.
0. Schmidt.
O. Schmidt.
O. Schmidt.
O. Schmidt.
O. Schmidt.
O. Schmidt.
Bowerbank.
O. Schmidt.
O. Schmidt.
O. Schmidt.
O. Schmidt.
O. Schmidt.
O. Schmidt.
0. Schmidt.
O. Schmidt.
O. Schmidt.
O. Schmidt.
0. Schmidt.
O. Schmidt.
O. Schmidt.
Carter.
Bowerbank.
Bowerbank.
Bowerbank.
Bowerbank.

II.— INDIAN OCEAN.

39. Seychelle Islands, . . .

40. Mauritius,

EupIccteUa cucumcr, Owen,
Dictyonine {Farrea occa, Bowerbank), .
Dictyonine {Farrea sp., Bowerbank),
Eahrodictyum speciosuiri', Quoy and Gaimard, .

Owen.

Bowerbank.
Bowerbank.
Valenciennes.

III. PACIFIC OCEAN.

41. Japan,

42. Japan, .Sagami Bay,

S» J,
St »>

>) JJ
)> J,
,» I,
J» ),
»» J.
)) »
It »
i» ;»
»j »

43. Philippines, .

tt

44. Zebu (Philippines),

" "
» >>
it ,,

5» 1,

100-200

Hyalonema sieboldii, Gray,

Euplectella oweni, Herkl. and Marshall,

Euphctella oxoeni, „ ,,
Acanthascus cactus, F. E. Schulze,
Rliabdoeahiptus mollis, F. E. Schulze, .
CratC7'oniorpha racyeri, F. E. Schulze, .
Hyalonema sieboldii, Gray

Farrea occa (Bowerbank), Carter,
Farrea sollasii, F. E. Schulze, .
Eurete schmidiii, F. E. Schulze. .
Eurcte sp.,

Pcriphrofjclhi: clisce, Marshall,
Aphrocailistes bocagei, Perceval Wright,
Aphrocallistes ramosus, F. E. Schulze, .
AiyhrocaUistcs vastus, F. E. Schulze,
Chonelasma d(£d€rleinii, F. E. Schulze, .
Ghonelasma calyx, F. E. Schulze,
Sexactinella ventilabrum, Carter,

Euplectella aspcrgilhim, Owen, .
SempereUa sell ult:ii. Semper,
Eurete sp. {simjyUcissima, Semper),
Eurete farreopsis. Carter,
EupIccteUa aspergillum, Owen, .

Semperella schultzci, Semper,
Polylophus philippinensis. Gray,
Crateromorpha meyeri. Gray,
Hyalonema ccbuense, Higgin,

Gray, Max Schultze,
Hilgendorf.

Herkl. and Marshall.

Doderlein.

Doderlein.

Doderlein.

Doderlein.

Carter, Doderlein, and
others.

Carter, Doderlem.

Doderlein.

Doderlein.

Doderlein.

Carter, Doderlein.

Doderlein.

Doderlein.

Doderlein.

Doderlein.

Doderlein.

Carter, Gottsche, Doder-
lein.

Owen, Semper.

Semper.

Semper.

Carter.

Bowerbank, Gray, Moore,
Chimmo and others.

Marshall and Meyer.

Gray, Marshall & Meyer.

Gray.

Higgin.

J

REPORT ON THE HEXACTINELLIDA.

427

Table II. — continued.

Locality.

Depth in
Fathoms.

Species.

Authority.

44. Zebu (Philippines),

45. Malacca,

46. Singapore, .

47. Moluccas,

Pheronenia luinisphwricum, Gray,
Sclerothamntm rlmisii. Marshall,
Aulocystvs ziUiiii, IMarshall,
Eurete sp., ....
Dactylocalyx mbgloboims, Gray, .
Phcronnna lii'misphrvrinim. Gray,
Halirinlirii/iiin n/uri<:siim. (^luoy and Gaimard,
Pcrijilirnffcl/ii. cli.vr, MAvtihdili, .

Marshall and Meyer.

Jluiie (iMarshall).

Meyer and Marshall.

Meyer and Marshall.

Gray.

Gray.

Quoy and Gaimard.

Marshall.

IV.— ANTAECTIC OCEAN.

48. Lat. 74° S., long. 175° W., .

Rossella antarctica, Carter,

Ross, Carter.

A glance at the map shows that there were four princij)al regions from which the
considerable collection of Hexactinellida gathered before the Challenger Expedition had
been obtained, viz.: —

1. West of Europe from the Shetland Islands to the Straits of Gibraltar.

2. Near the West India Islands, from Florida to Trinidad.

3. Near Japan, and especially in the Bays of Tokio and Sagami.

4. Near the Philippines, especially off the island of Zebu.

Besides these there are a few isolated localities in the North Sea and the Mediterranean,
off the Cape Verde Islands, the Seychelles, Isle de Bourbon, Malacca, Moluccas, and in
the Aiitai'ctic Ocean.

From these results it is oljviously impossible to gain anything but an approximately
plausible conception of the true distribution of the HexactineUida. At the outset the
fact has to be noted, that only in the four localities specially noted as the origin of the
majority of the known forms and individvials, was the sea-bed satisfactorily examined
below a depth of 100 fathoms. This was done on the Western European coasts by the
dredgiugs of English and French expeditions, off the West India Islands by the explora-
tions of Agassiz and Count Pourtales, off the Philippines by the enterprising and
successful activities of the fisher folk, and off Japan by a few naturalists and especially
by Dr. Doderlein.

It is of course impossible to decide whether other localities noted in the Mediter-
ranean, Indian Ocean, &c., represent the occurrence of Hexactinellida in altogether
sporadic distribution, or the existence of a richly developed and perhaps widely distributed
fauna, of which a few localities have been by chance discovered.

In regard to the distribution of the several suborders and families it is impossible
for the same reason to glean many certain results. It is, however, worth noting in this
connection that none of the localities are tenanted exclusively by any one family, but
that aU include representatives of different groups of the system.

428

THE VOYAGE OF H.M.S. CHALLENGER.

GEOGRAPHICAL DISTRIBUTION OF THE CHALLENGER
HEXACTINELLIDA.

After the above brief summary of the most important facts in regard to distribution
known before, and independently of the Challenger Expedition, I may proceed to an
account of the results which are now available.

I shall first give a tabular survey of aU those stations of the Challenger Expedition
at which Hexactinellida were obtained, including notes of the distribution in depth, the
nature of the ground, and the number of specimens. The stations are arranged
according to the route of the expedition, and I have also noted the distribution on the
map accompanying this Report, in such a way that the marks indicate at least
the great division, and as a rule the family to which the species in question belong.

TABLE III. — List of Hexactinellida collected during the Challenger
Expedition, arranged according to Stations.

station.

Locality.

Depth

in
Fathoms

Nature
of Bottom.

Bottom
Tem-
perature
F°.

Species.

No. of
Speci-
mens.

IV.

South-east of Cape St. Vincent,

600

Blue mud, .

1.

2.

Eupkctdla sulerca, AVyv. Thomson, .
Htndomma lusitanicum, Barboza du
Bocage,

1
1

V.

West of Gibraltar, .

1090

Globigerina ooze,

3S-5

1.

Jluplcciella sulcrea, Wyv. Thomson, .

1

3

South-west of Canary Islands,

1525

Hard ground.

37-0

1.

PoUopogon amadou, Wyv. Thomson, .

2

23

Off St. Thomas, West Indies, .

4.50

Pteropod ooze, .

1.
2.
3.

Aphrocallistes hocagci, Perc. Wright, .

ChonelasriMsp

Farrea sp. ,

1
1

1

24

North of St. Thomas. West
Indies,

390

Pteropod ooze, .

1.
2.
3.

4.

Hyalonema toxeres, Wyv. Thomson, .
Farrea occa, Carter, ....
Aphrocallistes bocagH, Perc. Wright, .
Mifliusia callociiatlmSj'Gi&y,

3

few
1
1

33

East of Bermudas, .

435

Coral mud.

1.

Lefroyella decora, Wyv. Thomson,

2

S6

South-west of Bermudas,

1075

Coral mud,

3S-2

1.
2.
3.
4.
5.
6.
7.

Eitplectella nodosa, F. E. S.,
RJiahdodictyuvi delicatum, F. E. S., .
Aiilocaliix irregularis, F. E. S., .
Farrea sp., .
Aphrocallistes bocagei, Perc. Wright,

Chonelasma sp. ,

Dactylocalyx patella, F. E. S. , .

1
2
1
1
several
1
2

73

West of Azores,

1000

Pteropod ooze, .

39-4

1.

Hpaloncnia (Stvlocalux) thomsoni,
F. E. S.,

1

94

South of Cape Verde Islands,

1150

Volcanic mud, .

1.

Sympagclla nux, 0. Schmidt,

1

124

East of Kio San Francisco,
BrazU.

1600

Bed mud, .

I.

2.

Eupleclella suberea, Wyv. Thomson, .
Pheronenia carpenteri, Wyv. Thomson,

1
1

KEPORT ON THE HEXACTINELLIDA.

429

Table III. — continued.

station.

Locality.

Depth
in

Fathoms

Nature
of Bottom.

Bottom
Tem-
perature
F°.

Species.

No. of
Speci-
mens.

135E

Off Tristan da Cunha, .

1000

Hard ground,
shells, gravel,

1.

Farrea sp.,

1

145

South-east of Prince Edward
Island,

140

Volcanic sand, .

1.

Rossella antarctica, Carter,

1

145a

South-east of Prince Edward

310

Volcanic sand, .

1.

Aidaftcus johnstoni, F. E. S.,

2

Island,

2.

Auluealf/x imyularisy F. E. S,, .

1

146

West of the Crozet Islands, .

1375

Globigerina ooze,

35'6

1.
2.

Holascus fbulatus, F. E. S.,
Malacomccus vastus, F. E. S.,

1
1

147

West of the Crozet Islands, .

1600

Diatom ooze.

34-2

1.

2

a

4.
5.

6.

Holascus fbulalus, F. E. S.,
Coulophacus lalus, F. E. S.,
Bathydorua spiiwsus, F. E. S., .
Aulocalyx irregularis, F. E. S., .
Btialonema (Sti/localux) claiir/eruin,

F. E. S.,
Farrea sp.,

1
2
1
1
1

1

148a

South of the Crozet Islands, .

550

Hard ground,

1.

Acan,thascusgrossularia,F. E. S.,

1

gi-avel, shells.

2.
3.

Chomlasvm lamella, F. E. S.,
Hexactinella sp..

2
1

149h

Off Kerguelen Island,

127

Volcanic mud, .

...

1.

Rossella antarctica, Carter,

several

150

South of Kerguelen Island, .

150

Coarse gravel, .

35-2

1.

Rossella antarctica, Carter, .

many

156

Lat. 62° 26' S., long. 95' 44' E.,
near Antarctic Circle,

1975

Diatom ooze,

1.

Polyrhahdus oviformis, F. E. S.,

1

157

Lat. 53° 55' S., long. 108° 35' E.,

1950

Diatom ooze,

32-1

1.

Holascus polejnem, F. E. S.,

1

Antarctic Ocean,

2.

Balanites pipelta, F. E. S., .

1

158

Lat. 50° rs., long. 123° 4' E.,

1800

Globigerina ooze,

33-5

1.

Hyalomma (Hyalonema) conus, F.E.S.,

1

160

South-west of Melbourne,

2600

Ked clay, .

33-9

1.

Holascus fihulatus, F. E. S.,

1

169

East of New Zealand,

700

Blue mud, .

40-0

1.

Hexactinellid (needles).

170a

North-east of Kermadec

G30

Volcanic mud, .

39-5

1.

Walteria fiemminfjii, F. E. S., .

1
J.

Islands,

2.
3.
4.
5.
6.

Euryplegma auricularc, F. E. S.,

Poliopogongigai,F. E. S

Farrea occa. Carter. , .
Chonelasma lamella, F. E. S.,
Chonelasma uncinatum, F. E. S.,

2
1

1
2

1

171

North of Raoul Islands,

600

Hard ground.

39-5

1.

Aulochone C!/lindrica,F. E.S., .

1

174c

Off Fiji Islands, .

610

Coral mud,

39-0

1.

Ta'gcriapulc?ira,?.'E.P,.,.

1

184

South-east of Cape York,
Torres Strait,

1400

Globigerina ooze.

36 0

1.

Hyaloncma sp., .....

1

192

Off Little Ki Island, .

140

Blue mud, .

1.
2.
3.
4.

h.

6.

7.

8.

9.
10.
11.
12.
13.
14.
15.
10.
17.
18.

Laiutginella pupa, O. Schmidt.,
Foli/luph us ph Uippimnms, Carter,
Crate roMvrplia IJiiir/i Itleri, F. E. S., .
Craltroinorplta murrayi, F. E. S.,
Fherotiema gigantcum, F. E. S.,
Phcroncma globosum, F. E. S., .
ScmpcrcUa schuHzci, Semper,
Eurete scmperi, F. E. S.,
Furete marshalli, F. E. S., .
Eurete cartcri, F. E. S.,
Eurcte farrcopsis. Carter, .
Euretide, ......

HcxaclincUa lata, V. E. S., .
Cyrtauloit suhitus, F. E. S.,
Fiiklingia lagettoidts, Sav. Kent,
Myliusia caliurt/ut/ius, Gray,
Aulocystis zittelii, Marshall,
Dictyonine, .....

few
several
1
1
1
5
2
1
1
1
1
2
2
1
1
3
2
1

430

THE VOYAGE OF H.M.S. CHALLENGEE.
Table III. — continued.

station.

Locality.

Depth
in

Fathoms

Nature
of Bottom.

Bottom
Tem-
perature
F°.

Species.

No. of

Speci-
mens.

194a

Off Banda Islands,

360

Volcanic mud, .

1. Cfatryomorpha fujnida,

2. Ht/alnncjiK! {Sttil'irdbix) globus, F. £. S.,

3. Farrca rfavi;/ffa, F. E. S.,

4. Sclcrothamuns clausii, Marshall,
.5. MyHusia calloci/athns, Gray,

6. Aulocystis dtlelii, Marshall,

3

1
1

1
1
2

201

Off Samboangan, Mindanao,
Philippines,

82

Stones, gi'avel, .

1. Eurete uhmidtii, F. E. S.,

2

205

■West of Luzon,

1050

Blue mud, .

37-0

li Hi/cdoncma sp.,

1

207

Philippines, ....

700

Blue mud, .

61-6

1. Farrea occa, Carter, ....

1

209

Philippines, Zebu, .

95-100

Blue mud, .

71-0

1. FuplecteUa aspcrgillum, Owen, .

2. Polylophus p)dllippinensis. Carter,

3. Ci'atercmorpha meyeri^ Gray,

4. Semperella schuUzei, Semper,

many

several

few

2

210

Philippines, ....

375

Blue mud, .

64-1

1. Aphrocallistes ramosus, F. E. S.,

1

211

West of Mindanao, Philppines,

2225

Blue mud, .

50-5

1. ffolascus ridleyi, F. E. S. , .

2. Hyalonema gracik, F. E. S.,

1

1

214
2.32

South-east of Mindanao,

Philippines,
Sagami Bay, Japan,

500
345

Blue mud,
Green mud.

41-8
411

1. Aulochonc lUiuni, F. E. S.,

1 . Hyalonema (Stylocalyx) apertum, F. E. S.,

1
several

236

South of .Japan,

775

Green mud.

37-6

1. Dictyonine,

1

■ 241

East of Nipon, Japan, .

2300

Bed clay .

35-1

1. Caulopliacus elegans, F. E. S., .

2. Bathydorusfimbriatus, F. E. S., .

3. Hyalonema robustum, F. E. S., .

4. Hyalonema sp.,

3
1

1
1

246

Ut. 36° 10" N., long. 178" 0' E.,

2050

Globigerinsk ooze.

35-1

1. Caulopliacus sp.,

2. Hyalonema (Stylocalyx) depressum,

F. E. S.,

1
several

248

Lat. 37° 41' N. , long. 177° 4' "W. ,

2900

Red clay,

35-1

1. Bathydorus Jimbriatus, F. E. S.,

1

271

Lat. 0°33'S.,long.]5r34'-\V.,

2425

Globigerina ooze,

35 0

1. Hyalonema divere/ens, F. E. S., .

2. Hyalonema (Stylocalux) depressum,

F. E. S.

3. Hyalonema [Stylocalyx) elegams, F. E. S.,

4. Hyalonema sp., .

1

1

1

1

272

Lat. 3° 48' S., long. 152° 50' W.,

2600

Kadiolarian ooze,

35-1

1. Farreid {?),

1

274

North-west of Caroline Islands,

2750

Kadiolarian ooze,

35-1

1. Euplectella crassistellata, F. E. S.,

1

2S1

South of Tahiti,

2385

Red clay

34-9

1. Dictyocalyx gracilis, Y.'E.S.,

2. Dictyonine, .....

1
1

286

Lat. 33° 29' S., long. 133° 22' W.,

2335

Red clay, .

34-8

1. Bathydorus baculi/er, F. E. S., .

2. Dictyonine, .....

3
1

289

Lat. 39° 41' S., long. 131° 23' W.,

2550

Red clay, .

34-8

1. Hyalostyhis dives, F.E. S.,

2. Trachycaul.us gurlittii, F. E. S., .

3. Hyalonema (Stylocalyx) elegans, F. E. S.,

1
1
1

293

Lat. 39° 4' S., long. 105° 5' W.,

2025

Globigerina ooze.

34-4

1. Dictyonine,

1

300

Off Juan Fernandez,

1375

Globigerina ooze.

35-5

1. Hyalontma pocidum, F. E. S., .

1

307
310

Channel leading to the

Magellan Strait,
Ch.innel leading to the

Magellan Strait,

140

400

Blue mud,
Blue mud.

46-5

1. Bathydorus stdlatus, F. E. S., .

1. Acanthascus duhius, F. E. S.,

2. Bhabduealyptus rcepcri, F. E. S.,

1

1
1

320

South-east of IMonte Video, .

600

Green sand,

37 '2

1. Rossella aniarctica, Carter,

2

REPORT ON THE HEXACTINELLIDA.

431

Table III. — contiuned.

Static

Locality.

Depth

in

Fathoms

Nature of
Bottom.

Bottom
Tem-
perature
F°.

Species.

No. of
Speci-
mens.

323
325
333

343

344
348

E^sl of Monte Video,
East of Jlonte Video,
West of Tristan da Cvmha,

South of Ascension,

Off Ascension,

Lat. 3° Iff N., long. 14° 51' W.

1900
2650
2025

425

420
2450

Blue mud, .
Blue mud, .
Globigerina ooze

Volcanic sand.

Volcanic sand,
Globigerina ooze

33-1
32-7
35-3

40-3

1. Hi/alontma tciiuc, F. E. S.,

1. Holascus stellatus, F. E. S.,

1. Caulocalyz tener, F. E. S., .

2. Hyalonema sp., .

1. Aphrocallistes bocagci, Perc. Wright,

2. Dictyonine, ....

1. Aphrocallistes bocagci, Perc. Wright,
1. Malacosaccus unguiculatus, F. E. S.,

A glaijce at the map shows that the Hexactiuellida are by no means confined
to the few localities hitlierto chronicled. On the contrary they are widely distributed in
all the oceans, and the majority of the Challenger specimens alone have not been found
in localities where Hexactinellida were previously discovered, but in localities which arc
new for the group.

As the table shows, it was only at 58 of the 275 stations {i.e., 21'1 per cent.) which
were explored with dredge and trawl that Hexactinellida were obtained. And although
these localities occur very uniformly over the whole route, nowhere is there any great
interval between two successive localities. The greatest distance occurs between Stations
94 and 124, but it has to be noted that this portion of the return route, which
included twelve stations where no Hexactinellida were found, was crossed at one point
(Stations 102 and 348, lat. 3° 10' N., long. 14° 51' W.), where (Station 348) a Hexac-
tinellid was discovered.

If we follow the course of the expedition, we see that in the first place to the west of
the English channel (Stations IV., V.) some Hexactinellids were captured, and that after-
wards, to the south-west of the Canary Islands (Station 3), the stately PolioiMgon
amadou was fished up. The booty became somewhat richer off" the West India Island of
St. Thomas, where, at the two Stations 23 and 24, six difierent species were procured.
The next locality, ofi" the Bermudas Islands (Stations 33 and 5G), yielded in all eight
species. West of the Azores, and afterwards ofi" Cape Verde, two species were found.
After a long pause two forms were captured ofi" the coast of Brazil (east of the Eio San
Francisco). On the tour from Baliia to the Cape of Good Hope, a fragment of a
Hexactinellid was obtained near Tristan da Cuuha, while near Prince Edward Island
three difi'erent species rewarded the search. This last region seems indeed to have been
tolerably rich in Hexactinellida, since some degrees further east first two, and soon
afterwards six, difi'erent species, and ofi' Crozets Islands, again three were obtained. Near

432 THE VOYAGE OF H.M.S, CHALLENGER.

Kerguelen Island and 5° south of it there was an abundant occurrence of the
large Rossella antarctica, while still further southwards, at Station 156, another
Hexactinellid was found. At Station 157 (lat. 53° 55' S., long. 108° 35' E.) two
forms were found, and 24° further east (Station 160) another, and again a fourth
south-west of Melbourne. Neither on the south-east coast of Australia, nor on the
voyage from Sydney to New Zealand, was there any sponge booty captured ; but to the
east of the North Island of New Zealand some Hexactinellid spicules at least were
obtained, and near the Kermadec Islands as many as six different species. Near the
Raoul Island was the habitat of the delicate Aulochone cyliiidrica, and off the Fiji
Islands the beautiful Tiegeria was trawled. After a break some Hyalonema fragments
were again obtained to the south-east of the Torres Strait. The richest haul of
Hexactinellida was obtained in the neighbourhood of the Little Ki Island, where no
fewer than eighteen different species were found, some of them of large size and repre-
sented by several specimens ; near the Banda Islands also six species w^ere obtained, but
after that it was not till the Philippines — long known as rich Hexactinellid localities —
were reached that several forms rewarded the search. Of the seven different stations in
the Philippine region, the locality between the islands of Zebu and Bohol, famous as
the seat of the almost mercantile capture of Eiqytectella aspergilhim, yielded not only
numerous specimens of Euplectella aspergillum, but several of the well-known typically
Philippine species, such as Pohjloplms philippinensis, Crateromorpha meyeri, and
Semperella schultzei, while at the remaining six stations, only one, or at most two species
were obtained. On the way from New Guinea to Japan no Hexactinellids were obtained,
l)ut in Japanese waters, in the Sagami Bay so industriously explored by Dr. Doderlein,
one species of Hyalonema at least was found in great abundance. A Dictyonal form was
also found to the south of the bay. Eight degrees east of Yokohama, at considerable
depths, four remarkable forms were found, and again as many degrees further east some
closely allied or identical species. On the long voyage from the Sandwdch Islands to
the equator no traces of Hexactinellids were discovered. On this side of the equator,
however, at some adjacent localities near the line, as many as four different species of
Hyalonema were trawled. After that two species were found some degrees south of
Tahiti, and 8° to 10° to the south-east first two and then three forms. Again in the
south-east portion of the Atlantic (Station 293) a fragment of a Dictyonal sponge was
discovered ; while near the island Juan Fernandez a Hyalonema, and at two stations in
the Magellan Strait, first one and then two species were captured. On the return
voyage through the Atlantic, the southern portion yielded no Hexactinellids, until
several degrees to the east of Monte Video, at three distinct stations, the search was
again successful. After that to the west of Tristan da Cunha and near Ascension Island
single species were obtained, and lastly 3° north of the equator, at considerable depths,
a single form belonging to the remarkable genus Malacosaccus was procured.

REPORT ON THE HEXACTINELLIDA. 438

In regard to the number of species found at the different localities this general fact
may be noted, that at most of the stations only one or two species were found, while
only a few yielded several or many forms. The statistics on which I based this result
in my preliminary communication' have been slightly altered by the additions of some
localities, and by changes in the determination of a few species, but these modifications
are on the whole unessential, and the results stand thus. Of the fifty-eight soundings on
the occasion of which Hexactinellida were found —

34 yielded only 1 species.

12

2

3

3

4

4

3

6

1

7

1

18

In his work La vie au fond des Mers, Filhol cites the statistics in question from my
preliminary notice, and adds the following remark : — " II resulte de ce tableau que dans
plus de la moitie des cas les especes d'Hexactinellides ont ete trouvees isolees. Cette
observation ne concorde pas avec celles que nous avons pu faire dans la parti e de
I'Atlantique nord, parcourue avec le Talisman, oil les chaluts n'ont rapporte qu'
exceptionellement une seule espece. Les Askonema ont et6 trouvees avec les Aphro-
callistes ; les Hyalonema, les Euplectelles etaient preque toujours associees." A possible
solution of this apparent divergence in the results of two deep-sea expeditions, will be
alluded to below when the general results as to the geographical distribution of the
Hexactinellida are summed up.

The number of individuals of the same species found at one place is but rarely
considerable. Generally only one or two specimens of each species were obtained at the
same locality. Sometimes, however, a considerable number of specimens were found at
once, as was the case with Farrea occa near St. Thomas, West Indies, Aphrocallistes
hocagei ofi" Bermudas, Rossella antarctica near the Kerguelen Islands, Polylaphus
2)hilippinensis and Pheronema glohosum at Little Ki Island, Hyalonema depression in
the middle of the North Pacific, and finally, Crateromorplia meyeri, Polylophus
philippinensis, and especially Euplectella aspergillum at the Philippine Island, Zebu.
Of the latter indeed, which was eagerly sought after, in a well-known locality, nearly a
hundred specimens were obtained.

As to the richness of the difi'erent seas and regions in Hexactinellida, it is, in the first
place, of interest to notice the absolute number of species found in the three great oceans.
The following tables (IV. and V.) exhibit the state of the case in regard to this and
similar points : —

1 Narr. Chall. Exp., vol. i. part i. p. 449.
(ZOOL. CHALL. EXP. — PART LIIL — 1887.) Ggg 55

484

THE VOYAGE OF H.M.S. CHALLENGER.

TABLE IV. — The Hexactinellida of the Challenger Collection,
LYSSACINA.

ATLANTIC OCEAN.

SOUTH IIJDIAN OCEAN.

PACIFIC OCEAN.

I. EUPLECTELLIDX.

,1.

EuplectcJla siiberea.

1.

Holascus fibulatus.

1.

Euplectella aspergillum.

(15 Species).

2.

EuphctcUa nodosa.

2.

Holascus polejaevii.

2.

Euplectella crassistellata.

3.

Holascus stellatus.

3.

MaZacosaccus vastus.

3.

Holascus ridleyi.

i.

Mfilacosaccus unguiculatus.

4.

Tageria pulchra.

5.

Rhabdodictyum delicatum.

5.
6.

7.

Waltci'ia femmingii.
ZHctyocalyx gracilis.
Hyalostylus dives.

II. AsCONEMATrD.E.

1.

Sympagella nux.

1.

Aida-scus johnstoni.

1.

Caulophacus elegans.

(8 Species).

2.

Pobjrhahdus oriformis.

2.

Caulophacus sp.

3.

Balanites pipctta.

3.

Trachycaulus gurlitti.

4.

Caulophacus lotus.

III. EOSSELLIDi.

1.

JRossella antarctica.

I.

Eossclla antarctica.

I.

LanugincUa pupa.

(19 Species).

2

Canlocabjx Uner.

2.

Acanthascus grossularia.

2.

Polyloph us ph ilippinensis.

3'.

Aulocalyx irregularis.

a

Bathyilorus spinosus.

3.

Acanthascus dubius.

4.

Aulocalyx irregularis.

4.
5.
6.

7.

8.

9.
10.
11.
12.
13.
14.

Bathydorus Jimbriatus.
Bathydorus stellatus.
Bathydorus baculifer.
Phabdocalyptus rceperi.
Crateromoipha mcyeri.
Crateromorpha thierfelderi.
Crateromo7-plux murrayi.
Crateromorpha tumida.
Aulochont cylindrica.
Aulochone lilium.
Euryplcgma auriculare.

IV. HTAL0NE3r.iTrD.E.

1.

Hiialonema toxcns.

I.

Hyalonema conus.

1-

Hyalonema gracile.

(21 Species).

2.

Hyalonona (Stylocalyx) thomsoni.

2.

Hyalonema (Stylocalyx) clavi-

2.

Hyalonema divergens.

3.

Hyalonema lusHanicum.

gerum.

3.

Hyalonema pocuhtm.

4.

Hyalonema tenue.

4.

Hyaloneiim(Stylocalyx) apertum.

5.

HyaJorwina sp.

5.

Hyalonema (Stylocalyx) depres-

6.

Phcronema carpenteri.

sum.

7.

Poliopogon amadou.

6.

7.

8.

9.
10.
11.
12.
13.

Hyalonema (Stylocalyx) globus.
Hyalonema (Stylocalyx) elegans.
Hyalonema robustum.
Hyalonema sp.
Pheronema globosum.
Pheronema giganteum.
Poliopogon gigas.
Semperella schuUzei.

Species of Lyssacina,

16

13

37

REPORT ON THE HEXACTINELLIDA.

435

ARRANGED ACCORDING TO THE OCEANS WHENCE THEY WERE OBTAINED.

DICTYONINA.

ATLANTIC OCEAN.

SOUTH INDIAN OCEAN.

PACIFIC OCEAN.

V. FAEREIDiE.
(4 Species).

VI. EnKETII).E.
(6 Species),

VII. Melittionidj:.
(2 Siiecies).

VIII. COSCINOPORIPJE.

(3 Species).

IX. TrETODICTTID/E.
(5 Species).

X. MiEANDRO-

SPONGIDiE.
(3 Species).

1. Farreaocca,

2. Farrea sp.

1. Lcfroyella decora.

1. ApUrocallistcs bocagci.
1. Chondaema sp.

1. DacUjlocaliix patella.

2. Myliusia eallocyatkus.

1. Dictyonine.

1. Farrea sp.

1. 0wnelasma lamella.
1. Hexactinella sp.

1. Farrea occa.

2. Farrea clavigera,

3. Farreid.

1. Eurete semperi.

2. Eurete schmidtii.

3. Eurete farreopsis.
i, Eurete carteri.

5. Eurete marshalli.

6. Euretid.

1. Aphrocallistes ramosus.

1. Chonelasma lamella.

2. Chonelasma hamatum

1. Hexactinella lata.

2. Cyrta-ulon solutus.

3. Fieldingia lagettoideg.

4. Selerothamnus clausii.

1. Myliusia calhcyathus.

2. Aulocystis zittdii.

1. Dictyonine.

Species of Dictyonina,

8

3

19

I. HexactineUid.

Species of Hexacti-
nellida,

24

16

57

436

THE VOYAGE OF H.M.S. CHALLENGER.

TABLE V. — The Hexactinellida or the Challenger Collection, arranged

List of all the Species in the CnALLENGBR
Collection.

Species.

I.

Bupleo-

TELLIDjE.

1. Enphctetta asperr/iUum, Owen, .

2. EuplecteUa subcrea, Wyv. , Thorn- (

son (

3. Euplcctdla crcKsistcUata, F. E.

SchiUze,

4. EuplecteUa (?) nodosa, F. E.

Schulze,

5. Holascus stellatus, F. E. Schulze,

6. Holascus fihulatus, F. E. Schulze <

7. Holascus polejaevii, F. E. Schulze,

8. Holascus ridleyi, F. E. Schulze,

9. j\Ialacosa<;cus vastus, F. E. Schulze,

10. Malmosaccus unijuiculatus, F. E.

Schulze,

11. Tcufjcria jmlchra, F. E. Schulze, .

12. Waltcria flemmingii, F. E.

Schulze,

13. Dictyocali/x gracilis, F.E. Schulze,

14. Rhahdodictyum delicatum, O.

Schmidt, .....

15. Hyalostylus dives, F, E. Schulze,

209
IV., v.,

124

274

56
325
146, 147
160
167
211
146

348
174c

170a
281

56

289

Station.

NORTH ATLANTIC OCEAN.

Species.

1. EuplecteUa subcrea.

2. EuplecteUa nodosa,

3. Malacosaccus iin-
guiculatus, .

4. Rhahdodictyum de-
licatum.

Station.

IV., V.

56

348

66

SOUTH ATLANTIC OCEAN.

Species.

1. EuplecteUa suberea,

2. Holascus stcllatus, .

Station.

124

325

IL

ASCONE-
UATIDM.

("16. Avlnsens jiilinstoni, F. E. Schulze,

17. SymptKiillit nux, O. Schmidt,

18. Pohirhabdus ovi/ormis, F. E.

Schulze, .....

19. Balanites pipetta, F. E. Schulze,

20. Caulophacus latus, E. F. Schulze,

21. Caulophacus clcgans, F. E. Schulze,

22. Caulaphaciis sp., ....

23. Trachycaulus gurlittii, F. E.

Schulze

146a
94

156
157
147
241
246

289

5. Sympagella nux.

94

'24. Lanuginclla pupa, 0. Schmidt,

25. Polylophus philippincnsis. Gray,

26. Rossella antarctica. Carter, .

27. Acanthascus grossularia, F. E.
Schulze, ....

28. Acantlmscus dubius, F. E. Schulze,

29. Bathydorus jimbriatus, F. E,
Schulze, ....

30. BiUhpdorus stcllatus, F. E. Schulze,

31. Bathydorus spinosus, F. E. Schulze,

32. Bathydorus baculifer, F. E.
Schulze,

III. 33. Rhabdocalyptiis rocpcri, F. E.

RossEL- -I Schulze, . . ' .

LID.E. 34. Crateromoipha meyeri. Gray,

35. Cratcroirwrpha thicrfcldcri, F. E.
Schulze,

36. Crateromoipha murrayi, F. E.
Schulze, . . . .

37. Cratermnorjiha tumida, F. E.
Schulze,

38. Aulochone cylindrica, F. E.
Schulze,

39. A ulochone lilium, F. E. Schulze,

40. Caulocalyx teiwr, F. E. Schulze, .

41. Aulocalyx irregularis, F. E.
Schulze,

42. Euruplegma auriculare, F. E.
Schulze,

192

209

145, 150.

1C9h,

320

148
310

241,248
307
147

286

310

209

192

192
194

171

214
333

56, 145a

170a

3. Rossella antarctica,

320

6. Aulocalyx inegu
laiis, .

i. Caulocalyx tener, .

333

66

_J

REPORT OX THE HEXACTINELLIDA.

43/

ACCORDING TO THE FiVE PraXCIPAL MaEINE DISTRIBUTIONAL ArEAS.

SOUTH INDIAN OCEAN.

NORTH PACmC OCEAN.

SOUTH PACIFIC OCEAN.

Species.

Station.

Species.

Station.

Species.

Station.

1. Eupleetella aspergillum,

209

1. Ewplectella crassistellata,

274

1. ffolascus fibuUttus, . . |

146, 147,
ICO

2. Holascus polejdevii, .

3. Malacosaccm vastus, ,

157
146

2. Solascui ridleyi.

211

2. TcEgcriM pulchra, .

3. Walteria Jlemmingii, .

4. Dictyocalyx gracilis, .

5. Hyahstylus dives,

174o
170a

281

289

4. Auluscus johnstoni, .

145a

5. Polyrhahdus ovi/ormis,

156

6. Balanites pipetta,

7. Caulophacus latus.

157
147

3. Caulophacus elegans, .

4. Cauloplutcv^ sp.,

241
246

6. Trachycaulus gurlittii.

289

8. RosseUa antarctiea, . . <

9. Acanthascus grossularia .

145,

149h,

150

148

5. Polylophus philippinensis, .

209

7. Lanvginclla pupa,

8. Polylophus philippinensis,

9. Acanthascus dulius, .

192
192

310

10. Bathydorus spinosus,

147

6. Bathydtyrusfimhriatus,

241

10. Bathydorus stellatus, .

11. BatJiydorus bacidifcr, .

12. Ithabdocalyptus rceperi,

307
286
310

7. Crateromorpha meyeri.

209

13. Crateromorpha thierfelderi,

14. Crateromorpha murrayi,

15. Crateronwrpha tumida,

16. v4uto(;Aon« cylindrica, .

192

192
194a
171

11. Aulocalyx irregularis,

145.4

8. AvJechoneliliuni,

214

17. Euryplegma auriculare.

170a

438

THE VOYAGE OF H.M.S. CHALLENGER.

Table V. — continued.

List op all the Species in the Challenger
Collection.

NORTH ATLANTIC OCEAN.

SOUTH ATLANTIC OCEAN.

Species.

IV.

Hyalone-
matidj:.

Hiialnnema gracUe, F. E. Scliulze,
Hmlonenxa divcrgens, F. E.

Scliulze,

Hijalonema toxeres, Wyv. Thom-
son,

Hiialonema poculum, F. E. Schulze,
Hjialoncma conns, F. E. Schuke,
Sti/locali/x thomsoni, Jlarshall,
Sti/locali/x apcrtus, F. E. Schulze,
Sti/lora/i/xdeprcssus, F. E. Schiilze,
Sti/locahi.c ckinr/cr, F. E. Schulze,
Sliilocahix rilohus, F. E. Schulze, .
Stulocali/x elegans, F. E. Schulze,
Stiilocnlyx teller, F. E. Schulze,
Hiialonema lusitanicum, Barb. cL

Station.

Species.

211

271

Hualoncma tenue, F. E. Schulze,
Hyalonema robustmn, F. E,
Schulze, .

Hifalonenut sp. ,

Pheronema carpente:

Wyv.

globoaum,
giganleum.

F. K

Thomson
Pheronema

Schulze,
Pheronema

Schulze,
Poliopogon amadou, Wyr. Thorn

son, .....
Poliopogon gigas, F. E. Schulze,
Semperella schuUzei, Semper,

24
300

158
73
232
246
147
194a
271
2S9

IV.

323

241
184, 205,
271, 333

124

192
192

3

170a

209

7. Hyalonema toxeres.

8. Stylocalyx thomsoni,

9. Hyalonema lusitani-
cum, .

10. Poliopogon amadou,

Station.

Species.

24

73

IV.

5. Hyalonema tenue,

6. Hyalonema sp. ,

7. Plieronema car

j>cntcri.

Station

323

333
124

V.

F.VB-

BEID.S;.

' 65. Farrca occa. Carter,
J 66. Farrea clavigcra, F. E. Schulze,

67. Farrea sp., ....
L68. Farreid, ....

24, 217.
292
194a,

23, 56,

135k, 147

272

11. Farrea occa,

12. Farrea sp..

24
23,56

8. Farrca sp.,

135e

69. Etiretc scmperi, F. E. Schulze,

70. Eurcte schmidtii, F. E. Schulze,
VI. 71. Eurete farreopsis, Carter,

EUKETID.E.') 72. Eurcte carteri, F. E. Schulze,

73. Eurete marshaUi, F. E. Schulze,
L74. Lefroyella decora, Wyv. Thomsou

192
201
192
192
192
33, 56

1.3. Lefroyella decora.

33

VII.

Melitti-

ONID.E.

f75. Euretid, ....

I 76. AphrornlHstes bocagei, Perceval)

•i Wright, ....

I 77. ApkrocalUstes ramosus, F. E.

l^ Schulze, ....

192
343, 23,
24, 56,

344

210

14. Aphrocallistes

■ cagei, .

bo-

23, 24, 50

9. Aphrocallistes bo-
cagei, .

343

\TII.

COSCINO-
POKID.E.

Clwnelasma lamella, F. E. Schulze, <
Vhonelasma hamatum, F. E.

Schulze,
80. Chonclasma sp.,

148a,
170a

170a

23, 56,

143a

15. Chonclasma sp.,

23,56

f 81. Hcocaclinella lata, F. E. Schulze,

IX. I 82. Hexactiiiella sp., . . . .

Treto- -! 83. Cyrtaulnn solutus, F. E. Schulze,

DICTYIDJ;. I 84. Ficldingia lagcttoides, Sav. Kent,

L85. Sclerothamnus clausii, Marshall, .

192

148a
192
192
144a

Y (86. Dactyloralyx

M«an-
DROSPON.

imtella, F. E.
Schulze,

87. Myliusia callocyathus, Gray,

88. Aulocystis zittelii, MmrshaM, .

89. Dictyonine,

90. Hexactinellid, ....

56
24, 192,

194a
192,194a

236, 343

169

16. Vactylocalyxpatella,

17. Myliusia callocya-

thus, .

56

24

10. Dictyonine,

343

REPORT ON THE HEXACTINELLIDA.

439

Table V. — continued.

SOUTH INDIAN OCEAN.

NORTH PACIFIC OCEAN.

SOUTH PACIFIC OCEAN.

Species.

Station

Species.

Station

Species.

Station.

9. Syalonema gracile, .

211

IS. Hyalonema divergens, .

271

12. Hyalonema conus,

158

19. Hyalonema poculum, .

300

13. Stylocalyx claviger, .

147

10. Stylocalyx apertus,

11. Stylocalyx dcpreisus, .

12. HyaUmcma robustum,

232
246

241

20. Stylocalyx depressus, .

21. Stylocalyx globus, ,

22. Stylocalyx elegans.

271

194a
271

13. Byalonema sp., .

205, 241

23. Hyalonema sp., .

24. Pheronema globosum, .

25. Pheronema giganteum,

184, 271

192
192

14. Scmperclla sclmltzei, .

209

26. Poliopogon gigas, .

27. Semperella schultzei, .

170a
325

15. Farrea occa,

207

28. Farrea occa,

29. Fan-ea clavigera, .

293
194a

14. Farrea sp.,

147

30. Farreid, ....

272

192

192
192
192

16. Eurcte schmidtii,

201

31. Eurete semperi, .

32. Eurete farreopsis,

33. Eurete carteri,

34. Eurete marshaUi,

17. Aphrocallistes ramosus.

210

35. Euretid, . . . .

192

15. Chonelasma lamella, .

148a

36. Chonelasma lamella, .

37. Chonelasma hamatum..

170a
170a

16. ffexactinclla sp.,

14Sa

38. Hcxactinclla lata,

39. Cyrtatilon solutus,

40. Ficldingia lagettoides, .

192

192
192

41. Myliusia callocyath-us,

192

42. Aulocpstis ziitdii,

192
293, 286,
192, 281,

169

IS. Dictyonine,

236

43. Dictyonine, . . • . •!

44. HexactineUid,

440

THE VOYAGE OF H.M.S. CHAT.T.F.NGER.

Of the ninety different species of Hexactinellids obtained on the Challenger
Expedition —

24 species were found in the Atlantic.

16 „ ,, Indian Ocean.

57 „ „ Pacific

But it is obvious that these figures taken by themselves, without reference to the
number of dredgings and trawlings, do not give any correct representation of the
abundance of Hexactinellida in the different regions. It is necessary to take into
account for each ocean the total number of dredgings or trawlings, and the proportion of
these which yield Hexactinellida. The percentage figures for the separate regions must
then be compared with the statistics of the whole expedition-
Out of the total 276 dredgings or trawling only 58 yielded Hexactinellida, which
gives a percentage of 21 '1 over the whole. Of the 125 dredgings or trawUngs in the
Atlantic 19 yielded Hexactinellida, i.e., a percentage of 15"2; of the 32 searches in the
Indian Ocean 11 were successful, i.e., 34'4 per cent.; of the 119 in the Pacific 28, i.e.,
23 '5 per cent., were rewarded with HexactineUid booty.

Thus we see that of the three oceans the Atlantic appears to be by far the poorest in
Hexactinellids, and in its percentage of 15 '2 is not a little below the general average
of 21 '1 ; while in contrast to this the Indian Ocean is richest, and in its percentage
of 34*4 is considerable above the average, which the Pacific with its 23 "5 per cent, only
slightly exceeds.

If in the case of the two largest oceans the northern be estimated apart from the
southern haK, then the proportionate statistics stands as follows : — •

North Atlantic South Atlantic. ^q^^

Xorth Pacific.

South Pacific.

1

Total number of dredgings or

79 40

32
11

34-4

47 72

trawlings.

Xuml>er of dredgings or trawl-
ings yielding Hexactinellids.

Percentages of the dredgings
or trawlings yielding Hexac-
tinellids compared with the
total numbers of dredgings
or trawlings.
1

125

11 8

119

11 17

19

13-9 17-4

28
23-4 23-6

15-2

23-5

From this it is apparent that the northern portion of the Atlantic is not so rich in
Hexactinellids as the southern, and that in the Pacific also the northern half is somewhat
surpassed by the southern.

REPORT ON THE HEXACTINELLIDA.

441

Furthermore, it must be carefully noted that this computation takes into account
only the localities, but neither the number of the discovered species nor the abund-
ance of specimens. When the number of species are also computed the proportions stand
as follows : —

North Atlantic.

South Atlantic.

Indian Ocean.

North Pacific.

South Pacific.

Xumber of dredgings or trawl-
ings — in aU 276.

Xumber of species of Hexac-
tinellida- — in all 89.

Percentages of species in rela-
tion to the number of
dredgings or trawlings— on
an average 3 2 '4.

79 46

32
16

50

47 72

125
17 10

V

119
18 44

*

24
21-5 21-7

56
38-3 61-1

19-1

47-1

The total number of Hexactiucllid species found at the 276 Stations explored with
dredge or trawl is 89, i.e., 32"4 per cent. At the 125 localities explored in the Atlantic,
24 different Hexactinellid species ■were found, i.e., 19"1 per cent, — and of these
17 were got from the 79 dredgings or trawlings in the North Atlantic, i.e., 2r5
per cent., while in the South Atlantic 46 searches yielded 10 species, i.e., 217 per
cent. The 32 dredgings or trawlings in the Indian Ocean yielded 16 different
species = 50 per cent. In the Pacific the 119 localities explored by dredge and trawl
yielded a total of 56 species, or 47 '1 per cent., and of these 18 species or 38 "3 per
cent, were found at the 47 dredgings or trawlings in the northern, and 44 species or
61"1 per cent, in the 72 searches in the southern region.

It is evident that the percentage expression of the abundance of Hexactinellids is
quite altered when we consider the number of species as well as the localities of capture,
though the relative proportion for the different oceans, or northern and southern regions,
is but slightly changed. On this computation the Atlantic, and especially the North
Atlantic, is seen to be poorest in Hexactinellids, while the Indian Ocean is richest. Only
the southern half of the Pacific appears considerably richer than in the j)rcvious table, in
fact even richer than the Indian Ocean. The proportionate richness thus stands as
follows : —

North Atlantic. South Atlantic. North Pacific. Indian Ocean. South Pacific.

21-5 per cent. 21-7 per cent. 38-3 per cent. 50 per cent. 61'1 per cent.

The richest localities on the Challenger Expedition were in the Atlantic at the West
Indian Island of St. Thomas and at the Bermudas, in the South Indian Ocean in the region

(ZOOL. CHALL. EXP. — PAUT LUI. — 1887.)

Ggg56

442 THE VOYAGE OF H.M.S CHALLENGER.

between tlie Prince Edward and Crozet Islands, in the Pacific ofi" the Kermadec, Little
Ki, and Philippine Islands.

In the middle of the oceans, as well as near the continents, Hexactinellids were indeed
found, but genei'ally speaking the abundance of species was less at a distance from the
mainland, and in the middle of the great ocean basins, than in the neighbourhood of the
continents or island groups.

It is necessary now to pass to the more detailed distribution of the indi\ddual groups
in the various seas.

In the Atlantic 16 species of Lyssacina were found, in the Indian Ocean 11, and
37 in the Pacific.

If we consider again the number of localities at which dredging and trawUng
explorations were undertaken, we have the following results : —

For the 125 dredgings or trawlings in the Atlantic, 16 species of Lyssacina, = 12-9 per cent.

„ 32 ,, „ Indian Ocean, 11 „ „ 34--i „

„ 119 „ „ Pacific, 37 „ „ 31-0

As to Dictyonina, the Atlantic j-ielded 8, the Indian Ocean 3, and the Pacific
18 species. And in reference to the number of dredgings and trawlings in the same
sea, these figures represent for the Atlantic 6"4 per cent., for the Indian Ocean 9-4
per cent., and 15'1 per cent, for the Pacific.

In regard to the separate families of Lyssacina, the Atlantic yielded 5 species of
Euplectellidse, the Indian Ocean 3, and the Pacific 5, that is to say, in reference
to the number of dredgings and trawlings, 4 per cent, for the Atlantic, 9 "4 per cent, for
the Indian Ocean, and 6 per cent, for the Pacific. And here it should be noted further
that the three species of EuplecteUidte found in the Indian Ocean belonged to the
genera Holascus and Malacosaccus, i.e., to the subfamily Holascinse, whUe in the
Atlantic and Pacific representatives of all the three subfamilies were discovered. Of
Asconematid?e the Atlantic yielded only 1 species, while in the Indian Ocean 4, and
in the Pacific 3 were found. A reference to the relatively small number of
dredgings and trawlings undertaken in the Indian Ocean, proves a special abundance of
Asconematidse in this last region.

The Atlantic yielded 3 species of Rossellidse, the Indian Ocean 4, and the Pacific
1 4 species, which mostly belonged to the Crateromorphinae.

In the Atlantic 7 Hyalonematidie w^ere found, in the Indian Ocean 2, and in the
Pacific 13 species.

In reference to the Dictyonina, if we take all the Uncinataria together in considera-
tion of the small number of representatives of the separate families, w^e find that
the Atlantic yielded 5, the Indian Ocean 3, and the Pacific 15 species, that is, in
reference to the number of dredgings and trawlings, 4 per cent, for the Atlantic, 9 "4
for the Indian Ocean, and 12-G for the Pacific. In a comparative survey of the separate

REPORT ON THE HEXACTINELLIDA. 443

genera witliin this group it is striking that in the Pacific alone 5 species of Eurete
were found, while this genus did not occur once in the other two oceans.

Mseaudrospongidse were represented only in the Atlantic and Pacific, in each by 2 species.

Further details as to the separate genera and species yield the following results.
The genus Euplectella was found in the Atlantic and Pacific, but not in the Indian
Ocean. The beautiful Euplectella aspergillum was only obtained at the one familiar
locality, namely, at the Philippine Island, Zebu. Euplectella crassistellata was also
found in the Pacific, but at a great distance from the mainland. Eu2:>lectella suberea
and Euplectella nodosa appear to be restricted to the Atlantic.

In each of the three oceans one or two representatives of the genus Holascus were
discovered. Malacosaccus was found at one locality in the Atlantic, and in the Indian
Ocean. The two Tasgerinse occur in the Pacific ; Twgeria pulchra near the Fiji Islands,
and Walteria flemmingii ofi" Kermadec.

Of Asconematidas, only one species, namely, Sympagella nux, was found in the
Atlantic, near the Cape Verde Islands ; four species in the Indian Ocean, viz., Poly-
rhahdus oviformis, Aulascus johnstoni, Balanites pipetta, and Caulophacus latus, and
two in the Pacific, namely, Caulophacus elegans to the east of Yokohama, and the
peculiar Trachycaulus gurUttii in the middle of the South Pacific.

While the Rossellidse were represented in the Atlantic region only lay the small
Lanuginella pupa, found ofi" Little Ki Island, and by Polylophus 2^hil{pp>inensis, near the
Philippine Island, Zebu, Rossella antarctica was found both in the South Pacific and in
several regions in the South Indian Ocean, and in especial abundance near Kerguelen.
Of the genus Acanthascus, one species, Acanthascus grossularia, was found in the Indian
Ocean, the other, Acanthascus dubius, in the Pacific portion of the Magellan Strait.
The species of Bathydorus are similarly distributed in the same two oceans, Bathydoi-vs
spinosus occurred in the Indian Ocean, near the Crozet Islands ; Bathydorus fimhriatiis,
Bathydorus baculifer, and Bathydorus stellatus, in the Pacific, the first in the North
Pacific at two localities, Bathydorus bacidifer in the middle of the South Pacific, and
Bathydorus stellatus in the Magellan Strait, where Rhabdocalyptus ra-peri also occurred.
All the four Challenger species of Crateromorpha were found in the Pacific, Cratero-
morpha meyeri from the neighbourhood of the Philippine Island, Zebu, Crateromorpha
thier/elderi and Crateromorpha murrayi ofi" Little Ki Island, and Crateromorpha
tumida ofi" one of the Banda Islands. The species of Atdochone were also obtained
in the Pacific, Atdochone cylindrica from the Kermadec Islands, and Atdochone
lilium from the Philippines. On the other hand, Caulocalyx tener was found in
the middle of the South Atlantic, and Atdocalyx irregularis both in the Atlantic
near the Bermudas, and in two localities in the South Indian Ocean, namely, near Prince
Edward Island and ofi" the Crozets. Finally, the Dictyonal-like Euryplegma auricxdare
was found in the Pacific near the Kermadec Islands.

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

Of the numerous species of Hyalonema no one form was represented in two oceans.
In the Atlantic Hyalonema toxeres was found near St. Thomas, Hyalonema {Stylocalyx)
thomsoni west of the Azores, Hyalonema lusitanicum to the south-west of Portugal
and west of the Straits of Gibraltar, as also at an undefined locality in the middle of the
South Atlantic. The two s^jecies of Hyalonema found in the South Indian Ocean were
Hyalonema conus to the south-west of Australia, and Hyalonema (Stylocalyx) clavigerum
to the west of the Crozets Islands.

Of the species of Pheronema the Challenger found Pheronema carpenteri in the
Atlantic to the north-east of Bahia, and Pheronema glohosum and Pheronema giganteum
in the Pacific, oft' Little Ki Island in the Malay Archipelago. Polioj)ogon amadou occurred
in the Atlantic, south-west of the Canary Islands, and PoUopogon gigas in the Pacific,
between the Eaoul and Kermadec Islands. Semperella schultzei finally was obtained
both at the well-known locality near the Philippine Island, Zebu, and also in the Malayan
Archipelago off" Little Ki Island.

As to the Farreidfe, Farrea occa occurred both in the Atlantic off" the island of St.
Thomas, and also in the Pacific at the Philippine and off" the Kermadec Islands, while
Farrea clavigera was only found in the Pacific near the Bauda Islands. Undefined
species of this remarkable, and apparently widely distributed genus, were found in all
the three oceans at various localities.

Five species of Eurete were found in the Pacific, four of them off Little Ki Island
and one at the Philippines, but in the Indian Ocean no Euretid was found, and in the
Atlantic only the single species Lefroyella decora off" the Bermudas. Aphroccdlistes
hocagei occurred at several localities in the Atlantic, Ap>hrocallistcs ramosus only in the
Pacific at the Philippines. Chonelasma was represented in all the three oceans, by Chone-
lasma lamella both in the South Indian Ocean (near the Crozets Islands) and in the Pacific
(off Kermadec), by Chonelasma hamatum only in the Pacific at the last named locality.
Some undefined frao-ments of Chonelasma occurred at different localities in the Atlantic.

Of Tretodictyidaj the Challenger found in the Pacific Hexactinella lata, Cyrtatdon
solutus, Fieldingia lagettoidcs, and Sclerothamnus clausii, all near the Little Ki Island,
and in the South Indian Ocean an undefined Hexactinella fragment near the Crozets.

The Mseandrospongidse found in the Atlantic were Dactylocalyx patella off the
Bermudas, and near the Spanish Coast, and Myliusia callocyathus off" the West India
Island of St. Thomas. The Pacific yielded Mylkisia callocyathus off Little Ki Island
and near the Bandas, and likewise Aidocystis zittelii.

In regard to the proportionate richness of the two temperate and the tropical zones,
it is necessary first of all to note the absolute number of localities yielding HexactineUida
in these three regions. The number of species in each zone must then be noted as in
the following tables (VI., VII.), and these statistics collated with the total number of
dredgings and trawlings in the respective regions.

REPORT ON THE HEXACTINELLIDA.

445

TABLE VI. — Distribution of the Hexactinellida of the Challenger
Collection in the different Zones.

NORTH TEMPERATE ZONE.

TROPICAL ZONE.

SOUTH TEMPERATE ZONE.

LYSS!\.CINA—

Station.

Station.

Station.

I. EUPLECTKLLID^.

1. Eiiplrctella subcrca, .

IV.,V.

1. Euplcctcllaaspergillum,

209

1.

Holascus stcllatus.

325

(15 Species).

2. ICiijilirt, Hii nvdosa,

3. Ithahdudictyum

56

2. Eiqileetdla subcrea,

3. Euplcctella

124

2.

Holascus fibulatus.

( 204a,
I 160

ddicatum,

56

crassistcllata, .

274

3.

Halaseus polejaevii, .

157

i. Solascus ridleyi,

211

4.

Malucosuccus vastus, .

146

5. Malacoitarrus

5.

Walteria flcmmingii.

170a

ujujuiraltitns,

348

6.

Hyalostylus dives.

289

6. Taf/eri't pulrltra^

174c

7. Diclyocalyx gracilis, .

281

II. AsC0NE5rATID.E.

1. Oaulophacus clcgans, ■

246

1. Sympagella nux,

94

1.

Aulaaeus johnstoni,

145a

(8 Species).

2. Caulophacus sp.,

246

2.
3.
4.
5.

Polyrhabdus oviformis,
Balanites pipetta,
Caulophacus latus,
Trachycaulus gnrlitti,

156
157
146
286

III. KOSSELLIDJE.

1. Bathydorus fimbriatus,

248, 241

1. Larmginella pupa.

192

( 145,

(19 Species).

2. Aulocalyx irregularis,

56

2. Polylophtts

philippincnsis.

192

1.

Bossella antarclica, .

< 149h,
(.150,320

3. Cratcromi,rpha meyeri.

209

2.

Aeanthasciis

4. Cratcromorpka

grossularia, .

14 8a

thierfelderi, .

192

3.

Aeanthascus dubius, .

310

5. Crateromorpha murrayi,

192

4.

Bathydorus stcllatus, .

307

6. Crateromorpha tumida,

194a

5.

Bathydorus spiyiosus, .

147

7. Aulochoiie lilium,

214

6.

7.
8.

9.

10.
11.

Bathydorus baculifer,
Rhabdoealyptus rteperi,
Aulochone cylindrica,

Aulocalyx irregularis,

Caulocalyx tener,
Eurypkgma auricutare,

286
318
171
( 145a,
i 147
133
170a

IV. Htalonematid^.

1. Hyalonema {Stylocalyx)

1. Hyalonema gracile, .

211

1.

Hyalonema pocidum, .

300

(21 Species).

npertum.

232

2. Hyiiluiii mn dirergens,

271

2.

Hyalonema (Stylocalyx)

2. Hiiidotuma (Stylocalyx)

3. Hyalimcma toxeres, .

24

con us, .

158

thomaoni.

73

i. Hyalonema (Stylocalyx)

3.

Hyalonema (Stylocalyx)

147

3. Hyalonema (Stylocalyx)

depressum,

271

elavigerum,

2t9

depresmm,

246

5. Hyalonema (Stylocalyx)

4.

Hyalonema (Stylocalyx)

4. Ifyalonemalusitaivicuin,

IV.

globus, .

194a

elegans,

323

5. Hyalonema robustum,

241

6. Hyalonema (Stylocalyx)

5.

Hyalonema tenue.

6. Hyalonema sp.,

241

elegans,

271

G.

Hyalonema sp..

303

7. Poliopogon amadou, .

3

7. Hyalonema sp., <

8. Pheroncma carpenten,

9. Pheronema globosum, .

10. Pheronema giganieum,

11. Semperella schultzei, .

284,
271, 205
124
192
192
192

7.

Poliopogon gigas,

170a

Species of Lyssacina,

14

26

29

DICTYONINA—

V. Farreidj;.

1. Fan-ea sp..

56

1. Farrca occa,

24, 207

1.

Farrea occa,

170a

(4 Species).

2. Farrea clavigera,

3. Farrea sp.,

4. Faneid,

194a

23

272

2.

Farrea sp.,

135e,147

4-iG

THE VOYAGE OF H.M.S. CHALLENGER.

Table VI. — continued.

NORTH TEMPERATE ZONE.

TROPICAL ZONE.

SOUTH TEMPERATE ZONE.

VI. EnEETID.E.
(0 Species).

Station.
1. Lefroyella decora, ■ 33

Station.

1. Eurcte semperi, . 192

2. Eurcte schmultii, . 192

3. Eurcte farreopsis, . 192

4. Eurcte carteri, . . 192

5. Eurete ■nmrskalli, . 192

VII. Mklittionidj:.
(2 Species).

1. Aphrocallistes bocayci, 50

1. Aphrocallistes bocagei, \ g^_j '

2. Aphrocallistes ramosus, 210

VIII. COSCINOPORID.E.

(o Species).

1. Choiulusma sp. , . 56

1. Chonclasma sp. , . 23

1. Chonclasma lamclht, . 14S.A

2. Chonclasma uncinatum, 170a

IX. Tretodicttid.e.
(5 Species).

1. llccactimUalata, . 192

2. Cm-hiulDn siilutiLS, . 192

3. Ficldiiigia lagettoides, 192

4. Sclerotliammis clausii, 194a

1. Hexactirulla sp., . 14Sa

X. M.EANDROSPONGIDiE.

(3 Species).

1. Dactylocalyx patella, . 5G

f ''4 192

1. Myliusia caUoci/athus, < "194^^ '

2. Aulocystis zittelii, . 192,194a

1. Dictyonine, . . 236

1. Dictyonine, . | '^^%ii^'

1. Dictyonine, . . 2S6

Species of Dictyonina,

6

19

6

Species of HexactineUida,

20

45

1. Heiactinellid, . . 169
36

TABLE VII — Distribution of the Hexactinellida of the Challengek Collec-
tion IN THE Northern and Southern Halves of the Tropical Zone.

NORTH TROPICAL ZONE.

SOUTH TROPICAL ZONE.

Station.

Station.

1. Euplectella aspergillum, .... 209

1. Euplectella suberea, ..... 124

2. Holascus ridleyi,

211

2. Euplectella crassistellata,

274

3. Malacosaccus unguiculatus.

348

3. Tcegeria pulchra.

174c

4. Sympagdla nnx,

94

4. Dictyocalyx gracilis, .

281

5. Pulylnjihits philippinetisis.

209

5. Lanuginella pupa.

192

6. Crateroniarpha tneyeri, .

209

6. Polylophus 2^hilippincnsis,

192

7. A ulochone lilimn.

214

7. Crateromorpha thierfehleri.

192

8. Byaloiuma gracik.

211

8. Cratero^rwrpha murrayi.

192

9. Hyaloneina toxeres,

24

9. Crateromorpha tumiila.

194a

10. Hi/aloneiim sp., .

205

10. Hiialmiema divergeiis, .

271

11. Semperella schultzei,

209

11. Hyalonema (Stylocalyx) depressum.

271

12 Farrea occa,

24, 207

12. Hyaloneitm (Stylocalyx) globus,

13. Hyalonema {Stylocalyx) elegans.

194a

13. Farrea sp.,

23

271

14. Enrete schmidtii.

201

14. Hyalonema sp.,

184, 271

15. Aphrocallistes bocagei.

23, 24

15. Pheronema carpenieri,

124

16. Ajihrocallistes ravwsus.

210

16. Pheronema globosum, .

192

17. Cho?i€las7)UC sp., .

23

17. Pheronema giganteum.

192

18. Myliusia callocyathus,

24

18. Semperella schultzei, .

192

19. Farrea clamgera.

194a

20. Farreid, .

272

21. Eurete semperi, .

192

22. Eurete marshalli.

192

23. Eurete carteri, .

192

24. Eurete farreopsis.

192

25. Aphrocallistes bocagei, .

343,344

26, Hexactinella lata.

192

27. Ci/rtaidon solutus.

192

•

28. Fieldingia lagettoides, .

29. St'lcrofhammis clausii, .

30. Myliusia callocyathus,

31. Aulocystis zittelii.

192
194a
192, 194a
192, 194a

32. Dictyonine, 281, 343

REPOET ON THE HEXACTINELLIDA, 447

It may be seen from the above that in the north temperate zone, out of 76 localities
which were explored with dredge or trawl, only 11 yielded Hexactinellida. In the
tropics, out of 99 dredgings or trawlings 22 were successful in the discovery of Hex-
actinellida. In the south temperate zone, out of a total of 1 0 1 dredgings or trawlings, 2 5
localities yielded Hexactinellids. Expressed in percentages the results are as follows : —

In the nortli temperate zone 14-4 per cent, of the explored localities yielded Hexactinellida.
„ tropical „ 22-2 „ „ „

„ south temperate ,, 24-7 ,, „ „ ,,

It is thus evident that the north temperate zone is by far the poorest in localities
yielding Hexactinellids, and further that not the tropical, but the south temperate zone
furnished both absolutely and relatively the greatest number of Hexactinellida.

The statistics are somewhat different when we consider not the number of localities,
but the number of Hexactinellid species. In this connection we see that

76 dredged and trawled localities in the north temperate zone yielded 20 species = 26'3 per cent.
99 „ „ tropical „ 45 ,, =45'4 ,,

101 „ „ south temperate ,, 36 ,, =35'6 „

The tropics are thus richer in species than even the south temperate zone.

In regard to the proportion of Hexactinellid localities in the entire northern and
southern hemispheres, considered in reference to the total number of dredgings and
trawlings, the statistics show that in the northern hemisphere out of 126 dredged and
trawled localities 22 yielded Hexactinellids, and in the southern hemisphere out of 150
dredged and trawled localities 36 yielded Hexactinellids. The percentage proportion for
the northern hemisphere is thus 17'4 per cent., and 24'0 for the southern hemisphere, so
that the latter contained many more Hexactinellid localities than the former.

The total number of species found in the north hemisphere was thirty-four, while
sixty-six were obtained in the south. If these numbers be collated with the number of
dredged and trawled locahties in each hemisphere, we have out of 126 searches in the
north 27 per cent., as against 44 per cent, out of the 150 in the south. Thus the south
hemisphere is seen to be about twice as rich in Hexactinellid species as the north.

Of the 50 dredged and trawled localities in the north tropical zone, 11 yielded
Hexactinellida, and the number is the same for the 49 localities in the south. This,
therefore, gives for the north tropical zone a percentage proportion of 22 per cent., as
against 22 "3 per cent, for the south tropics. The above noted important diflferenee
in localities pelding Hexactinellids in the north and south hemispheres does not therefore
hold true in regard to the tropics.

The proportions are different when the abundance of species in the north and south
tropics is taken into consideration.

The number of species found in the north tropical zone amounts to 18, as against 32

448

THE VOYAGE OF H.M.S. CHALLENGER.

in tlie south. And since in the northern region 50 localities were explored with dredge
and trawl, and in the south 49 localities, the percentage proportion for the former is
36 per cent., as compared with 65-3 per cent, for the latter. The south tropical region
thus considerably exceeds the north as regards the abundance of species.

And if in the same way we compare the two temperate zones in relation to the
abundance of Hexactinellid localities and species, the northern region shows, when com-
pared with the southern in regard to localities, 14 '4 per cent, as against 247 per cent.,
and 26"3 per cent, as against 35'6 per cent, in regard to the number of species.

There is thus a much greater abundance of Hexactinellids in the south temperate
than in the north temperate zone, which is doubtless in part referable to the much
greater extent of ocean in the former.

In reo-ard further to the distribution of the subdivisions in the different zones, a
review of Table VI. yields the following results : —

i

In the north temperate zone,
(. In the south temperate zone.

a. Lyssacina < In the tropics, .

f In the north temperate zone,
In the tropics, .
In the south temperate zone,

6. Dictyonina -j In the tropics, .

14 species.

26

29

6

19

6

In relation to the total number of dredgings and trawlings the percentage proportion
stands as follows : —

f In the north temperate zone,
a. Lyssacina -J In the tropics,

L In the south temperate zone,

r In the north temperate zone,
6. Dictyonina < In the tropics,

L In the south temjjerate zone,

184 per cent.
26-3

28-7 „

7-9 „
19-1 „

5-9 „

It is thus evident that throughout the Lyssacina preponderate, especially in the
south temperate zone, where they are almost five times as numerous as the Dictyonina,
while they are more than twice as abundant in the north temperate zone, and
preponderate by at least 7 per cent, in the tropics.

In regard to the abundance of Lyssacina in the different zones, it is evident that
the north temperate zone is considerably poorer than the tropical, and that the latter is
excelled by the south temperate zone.

The relation is quite different with the Dictyonina, of which the south temperate
zone contains the fewest, while the northern region a few more, and the tropics a relative
abundance.

If the Lyssacina of the northern hemisphere be contrasted with those of the south.

REPORT ON THE HEXACTINELLIDA. 449

it is seen that 25 species are found north of the equator, as against 47 to the south,
while of Dictyoniua 13 l^elong to the north, and 20 to the south hemisphere.

And since in the north liemisphere there were 126 dredgings or trawlings, as against
150 in the south, the percentage proportions for 100 dredgings or trawlings are as
follows : — The Lyssacina occurred in the north hemisphere in the proportion of
19 "8 per cent., as compared with 3r3 per cent, in the south ; the Dictyonina are
represented in the north hemisphere by 10 "3 per cent., as against 13 '3 in the south.

The absolute number of Lyssacina found in the north tropics was 11, as compared
with 18 in the south. And if the number of dredgings or trawlings in the north and
south tropics are taken into consideration, the percentage proportions stand as follows —
22 per cent, in the north tropical zone, and 367 in the south.

Of Dictyonina 7 species were found in the north tropical zone, as against 14 species
in the south. And this expressed in percentages of dredgings and trawlings means
14 per cent, for the north tropics and 28 "6 for the south.

If we now pass to the consideration of the several families, it will be seen that most
of them are represented in all the three zones, though generally in very varying
abundance. It is noteworthy that the EuretidiB and Mceandrospongidas are wdiolly
absent from the south temperate zone, while the northern region exhibits a similar
absence of Tretodictyidse, and only one representative of all the other families of
Dictyonina. The tropics on the other hand include representatives of almost all the
families, and one at least from among the Asconematidae and Cosciuoporidse.

WhUe the Euplectellidfe are represented by several forms in all the three zones, the
north temperate zone, which contains the genus Ewplectella itself, yielded no Holascidse,
which occur both in the tropics and in the south temperate zone. It is a striking fact that
the Asconematidae, which occur somewhat abundantly in the south temperate zone and
likewise in the north, are represented in the tropics only by a single species. The
Rossellidse occur in considerable abundance in the tropics, and in especial abundance in
the south temperate zone, but are represented only by one species in the northern region.

The Hyalonematidse exhibit a tolerably uniform distribution through all the three
zones, with a slight preponderance in the tropics.

The tropics are especially rich in Farreidse and Euretidse, to which have to be added
two species of Melittionidse, several Tretodictyidaj, and two species of Mseandrospongidaj,
a distribution in striking contrast to that of the two temperate zones, in which these
families either do not occur, or are represented only by a single species.

A more detailed survey of the genera and species shows that the species of Euplectdla
occur only in the north temperate zone and in the tropics, being apparently absent in the
southern region. Euplectella aspergillum and Ewplectella crassistellata appear to be
tropical, while Euplectella nodosa was confined to the north temperate zone, and
Euplectella suberea occurred in both these regions.

^OOL. CHALL. EXP. — PART LIII. — 1887.) Ggg 57

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

The genus Holascus is especially represented in the south temperate zone, where three
species occurred, while the tropics only yielded one species, and the north temperate zone
none at all. Of the two known species of Malacosaccus one was found in the tropics, the
other in the south temperate zone. In the same way of the two Tsegerinae, the one, Tsegeria
pulchra, occurred in the south tropics, and the other, Walteria fiemmingii, in the south
temperate zone.

In regard to Asconematidse, Aulascus jolmstoni was captured in the south temperate
zone* Sympagella mix in the north tropics, and Polyrhabdus oviformis and Balanites
pipetta in the south temperate zone. Of Caulophacinae, Caulojjhacus latus was found in
the south temperate zone, the other species, Caulophacus elegans, far removed in the
northern region, and finally Trachycaulus gurlitti in the south temperate zone.

Of the genus Rossella, the only species obtained on the Challenger Expedition, Rossella
antarctica, was entirely confined to the south temperate zone, where it was found at four
different localities. In regard to the two species of Acanthascus and Batliydorus, two
at least belong to the fauna of the south temperate zone, while Bathydorus fimhriatus
occurred at two localities in the northern region. Ehahdocalyptus was represented by
one tropical species and by a second in the south temperate zone. It seems a note-
worthy fact that all known species of Crateromorpha were found in the tropics, while
Aidoclione was represented by one species in the tropics and by another in the south
temperate zone. Aidocalyx irregularis appears to be widely distributed, since it was
found both in the north and south temperate zones. The peculiar forms Caulocalyx
tener and Euryplegma aurictdare belong to the south temperate zone.

Of the species of Hyalonema, which were somewhat uniformly distributed in all the
three zones, several were found in two zones, viz., Hyalonema (Stylocalyx) depressum
both in the tropics and in the north temperate zone, and Hyalonema (Stylocalyx) elegans
both in the tropics and in the south temperate zone. The species of Pheronema found by
the expedition belong to the tropics, the two species of Poliopogon to the temperate zones
{Poliopogon amadou in the north, Poliopogon gigas in the south), while Semj^erella
schultzei was tropical.

Farrea occa occurred predominantly in the tropics, and Farrea clavigera exclusively
so, while undefined species of Farrea were found in all three zones. AU the five species
of Eurete obtained on the expedition were tropical, while the beautiful Lefroyella decora
occurred in the north temperate zone.

Aphrocallistcs hocagei occurred both in the north temperate and in the tropical zone,
but Aphrocallistes ramosus was found in the latter alone.

Chonelasma lamella and Chonelasma uncinatum occurred in the south temperate
zone, and undefined specimens of the same genus in each of the two other zones.

The tropics yielded Hexactinella lata, Cyrtaidon, Fieldingia, and Sclerothalmnus,
while in the south temperate zone only an undefined specimen of Hexactinella was found.

REPORT ON THE HEXACTINELLIDA. 451

The only representative of the genus Dactylocalyx procured by the Challenger
Expedition was found in the north temperate zone. On the other hand Myliusia
callocyathus and Aulocystis zittelii were found in several localities in the tropics.

I have discussed in some detail the topographical distribution of the Challenger
Hexactinellida, because I believe that the uniformity of the exploring and collecting
methods, as well as the wide extent of the exploration, make the results of great value.

In estimating the number of species obtained by the Challenger in different localities,
it is instructive to compare the statistics with those which are afforded us by the prolonged
and enthusiastic industry of Dr. Doderlein, who made during his lengthened residence
in Japan a collection of Hexactinellids, partly by purchase in Enoshima, and partly by
his own dredging in Misaki Bay near Tokio, which was also explored by the Challenger.
The results of his explorations have been described in detail in this Report along with the
Challenger material. His collection included in all 16 species : —

1. Euplectella oweni, Marshall, dry.

2. Acanthascus cactus, F. E. S., dry.

3. Rhahdocalyptus mollior, F. E. S., dry.

4. Crateromorpha meyeri, Gray, dry.

5. Hyalonema sieboldii, Gray, dry.

6. Farrea occa, Carter, well preserved in spirit.

7. Farrea vosmaeri, F. E. S., dry.

8. Farrea soUasii, F. E. S., dry.

9. Periphragella elisse, Marshall, dry and well preserved in spirit.

10. Aphrocallistes bocagei, Wright, in spirit.

11. Aphrocallistes ramosus, F. E. S., well preserved in spirit.

12. Apihrocallistes vastus, F. E. S., dry.

13. Chonelasma da^derleinii, F. E. S., well preserved in spirit.

14. Clionelasma calyx, F. E. S., dry.

15. Hexactinella tubulosa, F. E. S., dry.

16. Hexactinella cyathus, F. E. S., dry.

On the Challenger Expedition three soundings were made in the Bay of Misaki, at
Stations 232, 236 and 236a, Station 232 yielded, from a depth of 345 fathoms,
several specimens of Hyalonema {Stylocalyx) afertum, and Station 236, from a
depth of 775 fathoms, a fragment of an undefined species of Dictyoniua. In the cursory
exploration of the Challenger only two Hexactinellida were thus obtained at the same
locality where the prolonged search of Dr. Doderlein was rewarded by finding
sixteen species. It is further of interest that among these sixteen species, the species
found by the Challenger — Hyalonema {Stylocalyx) apertum — is not included, though it

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

occurs in a smaller collection of Hexactincllida dredged by Dr. Hilgendorf at the
same locality, and also graciously placed at my disposal. Of course one cannot conclude
from this single instance that at all the localities explored on the Challenger Expedition,
eight times as many species would have rewarded a more prolonged exploration, but the
facts above communicated are calculated at least to convey an impression of the relatively
small proportion of the really existing Hexactinellid species that rewarded the search of an
expedition so splendidly equipped and successfully conducted as that of the Challenger.
And if one further considers that in spite of the large number (276) of dredgings or
trawlings undertaken, immense marine regions, especially the North Arctic and the
greater part of the Indian Ocean, were hardly touched upon, and that in many of the
regions traversed only sample searches were made, one cannot avoid the conviction that,
in spite of the relatively important increase of knowledge obtained by results of the
Challenger Expedition, only a small percentage of the really existing Hexactinellid
species have as yet been discovered.

BATHYMETKICAL DISTRIBUTION.

HISTORY.

In regard to the bathymetrical distribution of Hexactinellida, Marshall^ gave in
1875 the following summary of the few results then available : —

Species.

Depth in Fathoms.

Authority.

Pheronena {Holtenia) carpenteri, .

,550

Wyv. Thomson.

Hi/alonema thomsorii,

83

Wyv. Thomson.

Aphroeallides bocagei, .

700

Wyv. Thomson.

St/mpageUa mix, .

100-120

Pourtales.

Holtenia jwurtalexii, 0. Schmidt,

150-333

Pourtales.

Farrea facunda, 0. Schmidt, .

130-456

Pourtalts.

Dadylocaltjx crispus, 0. Schmidt,

270

Poiirtali^s.

Habrodictyum speciosum,

83

Quoy and Gaimard.

Semper ella schultzei,

200

Ludeking.

This list was extended by 0. Schmidt's ^ report upon the Hexactinellida collected by
A. Agassiz in the Gulf of Mexico, especially near the Lesser Antilles. This list, retaining
0. Schmidt's nomenclature, is given on the next page.

From the above it wiU be seen that the depth varied from 100 to 2410 fathoms, with
the exception of the record of Cystispongia, said to have been found near Yukatan at a
depth of 20 fathoms. On this very divergent number of fathoms much weight must not be
laid, since there is a possibility of error or of misprint, all the more probable since at two
other localities the same species was found at depths of 136 and 292 fathoms. Leaving
this divergent instance out of account, we get from Alexander Agassiz's bathymetrical
statistics an average of about 500 fathoms, while 100 fathoms is the minimum depth at
which any form was found. Without attempting to gather up all the scattered reports
as to depths at which single Hexactinellid forms were found at various localities, I
shall content myself with noting that Dr. Doderlein^ found in the Sagami Bay, Japan,
various Hexactinellids (especially Aiohrocallistes, Fanned, and Hyalonema), at depths of
80 to 200 fathoms, It is also necessary to note some reports which have recently
appeared in the results of the French Expeditious of the " Talisman " and " Travailleur,"
by M. Filhol* and M. Perrier.° According to these Pheronema were dredged from

1 Marshall, Zdisckr. f. wiss. Zool., Bd. sxv. Supplement, p. 142.

- Spong. Meerb. Mexico, 1880. ' Diiderlein Arcliiv f. Nahirgesch., Jahrg. xlix. BJ. i., p. 102, 1883.

* La vie an fond des Mers. ^ Les explorations sous-marines.

454

THE VOYAGE OF H.M.S. CHALLENGER.

depths of GOO to 2200 metres, and one Pheronema fragment from 4789 metres.
Asco)iema setuhalense was dredged from a depth of 410 metres, and A-phrocallistes
hocagei from 860 to 2200 metres. At p. 278 of his work M. Filhol notes : — " Les
Hexactinellids peuvent etrc considerees comme de veritables habitants des grands fonds."
And Edouard Perrier gathers up the results on p. 244 of his work in the following
words : — " Au large des c6tes de France, de Portugal et d'Afrique, c'est a un peu moins
de 1000 metres que Ton commence h rencontrer les Pheronema et les Aphrocallistes.
Ces eponges sont tout h fait communes de 1000 a 1500 metres, et s'associent alors aux
Hyalonema. A partir de 1500 metres elles deviennent rares et sont remplacees jusqu'a
2500 mfetres par les Euplectelles. Aux profondeurs plus grandes les eponges sont
moins communes, et sont surtout representees par les Farrea. Ainsi les Hexactinellides
elles-memes, qui sont les vraies eponges des grands fonds, se maintiennent dans les
parties superieures de la zone abyssale."

Species.

Depth in Fathoms.

Farrea facunda, 0. Schmidt,

300-1000

Farreid, ....

2410

Diaretula cornu, .

805

Diaretula niuretta,

805

Cyathella lutea, .

1591

Rhabdodictyum delicatum,

1591

Syringidimn zittelii,

116-878

Ajihrocallistes hocagei, .

164-400

Cystispongia miperstes, .

[20 and] 136-292

Myliuda zittelii, .

100-756

Dactijlocahjx pumiceus, .

103

Dactglocalyx ml/glohosus.

116-190

Dactocalyx potatoi-um, .

151

Dactylucalyx callocyathits.

116-292

Margaritella cwJoptycMoides,

158

Juanella compressa,

287

Seleroplcgma lanterna, .

292-320

Sdero2)legrna conieum, .

292

Soleroplegma seriatimi, .

200-300

Sderoplegma herculeum,

580

Diplacodium mixtam, .

101-292

Volvulina sigsheei.

100-292

Paehaulidium,

580

Rhahdostauridium retortula,

804

Eupledella jovis, .

416-423

Regadrella phcaiix,

221-288

Hertioigia falcifera,

611

Rhahdopedella tintinnus,

291-994

Hyalonema sp., .

416

Pheronema annx, .

180-248

Leioholidium,

1.507

Asconema \Hyalonemd\ kentii,

338-1507

J

EEPORT ON THE HEXACTINELLIDA. 455

BATHYMETRICAL DISTRIBUTION OF THE CHALLENGER

HEXACTINELLIDA.

In order to afiford a ready survey of the main results of the Challenger Expedition in
regard to the bathymetrical distribution of the Hexactinellida, I have compiled Table
VIII. , in which the horizontal divisions mark out intervals of depth of 100 fathoms each
from 100 to 1000 fathoms, and of 500 fathoms each from 1000 to 3000 fathoms, while
the vertical columns correspond to the four families of Lyssacina, and the two main
divisions of Dictyonina, the Uncinataria and Inermia.

This table shows distinctly that Hexactinellida occur in varying abundance only from
about 100 (more exactly 95) up to 3000 (more exactly 2900) fathoms.

Thus the previously determined character of the Hexactinellida as specifically, and in
fact exclusively, deep-sea forms is indubitably established, and the maximum depth
previously known (2410 fathoms according to Agassiz) is exceeded by several hundred
fathoms, with a maximum of 2900.

In order to express clearly the bathymetrical distribution of the Hexactinellida in
the different oceans, it is convenient to tabulate not only for the whole group, but also
for the several divisions, the statistics of depth and abundance, both in their absolute
and percentage relations. This is done in the following table : —

456

THE VOYAGE OF H.M.S. CHALLENGER.

TABLE VIII.

No. of Dredgings

or Tr.awlings.

w

3

1

a

o

o

^

-tJ

S

g

oj

EnpLECTELLina:.

ASCONEMATID.E.

ROSSELLIDJi.

s

CiO

r^

2

c

^

B2

s

Ti

a
o

'Z

1

^

J3 'T*

OJ

^

1

g

iS

1

CO

ft

lto94

42

f Eupleciella

j asijergilliim, ,

Pull/lophllS

95 to

5

1

20

95

209

phUippitiensis,

95

209

100

1
I

Crateroviiyyplia

7iiey€ri,

95

209

Hossella

autarciica,

Lanuginella
pupa,

(127,
< 140,
(. 150

129-140

196H
145
150

«

192

101 to

26

5

19-2

.

Polylop>hus

200

philipjmieiisis,

129-140

192

Crateroijiorpha

thierfdderi.

129-140

192

murrayi.

129-140

192

Batliydtyrus

'-

stellatus,

140

307

201 to

8

300

f

A ulascm
johnsioni,

310

145a

RliaMocalyphis

r(Bperi, . ,

Acanthascus
dubius, .

400
400

310
310

301 to
400

Craterovwrpha,

13

7

63-9

tuinida, .

360

194a

r

401 to
500

11

6

54-5

Aidochmte
lilium, .

500

214

f

Acanthasciis
grossulana.

550

14Sa

501 to

10

5

50

Rossella

600

antarclica,

600

320

Euplectella

A idochfme

subcrea, .

600

IV.

cylindrica.

600

171

REPOET ON THE HEXACTINELLIDA.

457

TABLE YUl.—contimied.

m

«

tr:

6

s

a

o

o

o

O

ja

.£3

J=!

Htalonematid^.

1
a

Ukcinataria.

.g

J3

iNEEMIA.

i^

a

ct:

o

p
o

^

I

^

a
o

&

ci

£■

rt

P-

rt

&■

"S

Q

m

fi

^

R

ai

Q

02

Semperella

schwltm,

95

209

Pheroneina

Eurete (
schmidtiiy \

102

201

glohosum.

129-140

192

129-140

192

gifjunteum,

129-140

192

se7nperi, .

129-140

192

Semperella

carteri, .

129-140

192

Myliusia

Dictyonine, .

129-140

192

schuUzei,

129-140

192

farreojisis,
Hexactinella

lata,
Cyrtaulon

SolutUSf .

Fieldingia
lagettoides.

129-140
129-140
129-140
129-140

192
192
192
192

callocyathus,
Aidocystis
zittelii, .

129 140
129-140

192
192

Hyalonema

Farrea

Myliusia

( 360
t 390

194a

{Stylocalgx)

clavigera.

360

194a

callocyathuSj

24

apertum,

345

232

Sclerciihamnus

Aulocystis

globus,

3fi0

194

clousii, .

360

194a

zittelii, .

360

194a

toxeres, .

390

24

Aphrocallistes
ramosus, .

Farrea
occaf

Aphrocallistes
bocagei, .

375
390
390

19
24
24

Aphrocallistes
bocagei, .

( 420
< 425
( 450

344

343

23

Dictyonine, .

425

207

Farrea

sp.,

450

23

Chonelasma

sp.,

450

23

Lefroyella

decora, .

435

33

Choutiifsma

htinrUn, .

550

148a

Hexactinella

sp.,

550

148a

Hyalonema

lusitanicum,

600

IV.

(ZOOL. CHALL. KXP. — PART UII. 1887.)

GggSS

458

THE VOYAGE OF H.M.S. CHALLENGER.

Table VIII. — continued.

No.

Df Dredgings

or

Trawliugs.

a

o

S
o

o

S
o

rO

.a

^

-J=

03

^ oi

a

ECPLECTELUD^.

&

ASCONEMATID*:.

2

ROSSELLIDiE.

1

fl

"■§

CB

a

i

3

ma

fig

i

0.

^5

s

33

Q

o

02

1

■ Tcegeria

601 to

pulchra,

610

174c

700

9

4

44-4

- ^VaUerla

Jtemmingii,

630

176a

Euryplegina
aurictdare,

630

170a

701 to

3

800

801 to

a

900

...

901 to
1000

4

2

50

{

' EiqjlecteUa

Sympagella

Aulocalyx

nodosa, ,

1075

66

mix,

1150

94

irregularis,

1075

56

Rhahdodictyum

1001
to

deUcaiuDij

1075

56

31

7

22 "6

Eiiplectella

1500

mbeiea, .
Holascus

1090

V.

fibidalus.

1375

146

Malacosaccus

vastus, .

1375

146

' EuplectcUa

Caulophacus

Baihydoriis

sidierea, .

1600

124

latus,

1600

147

spinosus,

1600

147

Holascus

A ulocalyx

1501

to

2000

Jibulatus,

1600

147

irregularis.

1600

147

35

7

20

Balanites
pipetla, .

1950

156

polejaevii.

1950

157

Polyrhabdtcs

ovi/onnis,

1975

157

'Holascus

Oaulopkacus

Caulocalyx

ridleyi^

2225

211

latus.

2050

246

tetter,

2025

333

Diciyocalyx

elegans,

2300

241

Bathydoriis

gracilis^

2385

261

fivdmatus.

2300

241

Malacosaccus

2001

unguiculatus, .

2450

348

to

38

8

21-1

2500

' Hyalustylics

Trachycauliis

dives.

2550

289

gurlitti, .

2550

292

2501

Jlolascwi

to

S3

6

18-2

film: tus,

2600

163

3000

stellalus,
Eiiplectella

crassislellata, .

2650
2750

325
274

Bathydoriis
Jimbriatm,

2900

248

3001

to

5

4000

REPORT ON THE HEXACTINELLIDA.

459

Table VIII. — contimted.

•i

o3

a

1

g

a

o

o

o

o

5

ja

.a

5

Htalonbmatid^.

.9

n

UUOINATARIA.

1
.3

p

Inebmia.

s

d

o

.a

o

.a

o

ji

o

^

+a

*^

1

m
S

G,

1

1

1

Farrea

( 700
\ 630

207

Hexactinellid,

700

169

occa.

170a

PoUopogon

Chon^lasma

ffiffa^,

630

170a

lamella, .

630

170a

uncinatum.

630

170a

Hyalonema

Farrea

t

(Stylocalyx)

sp.,

1000

135E

thomsmii,

1000

73

Hyalonema

Fairea

sp,,

1050

205

sp..

1075

56

jJociUiwi, .

1375

300

Aphrocallistes

Dactylocalyx

sp.,

1400

184

bocagei, .
C'/unielasma
sp..

1075
1075

56
56

patella, ,

1075

66

PoUopogon

amadou, .

1525

3

Pheronema

Farrea

carpenteri.

1600

124

sp..

1600

147

Byalonema

{Stylocalyx)

irregularis, .

1600

147

JlyaUrnema

emus.

1800

l.-iS

ten lie.

1900

323

HyloTuina

Dictyonine, .

2925

293

sp.,

2025

333

Dictyonine,

2325

286

[Stylocalyx)

depressum

2050

246

gracile, .

2225

211

rofmsttim,

2300

241

dimrgens.

2425

271

{StyhKalyx)

elegans,

2425

271

sp..

2300

241

{Stylocalyx)

depressum.

2425

271

sp..

2425

271

Hyalonema

(Stylocalyx)

elegans,

2550

289

Farreid,

2600

272

460

THE VOYAGE OF H.M.S. CHALLENGER.

This table reveals at once the remarkable fact that a large number of localities
yielding Hexactinellids, and a rich abundance of species occurred in moderate depths of
95 to 200 fathoms, while the depths from 200 to 300 are quite without Hexactinellids,
which become again more abundant both in localities and species at dejDths from 301 to
700 fathoms, disappearing again in depths between 701 and 1000, and finally occurrino-
in moderate abundance between 1000 and 3000 fathoms.

If regard be first had only to the number of Hexactinellid localities and their per-
centage among the total number of dredgings and trawlings undertaken at the relative
depths, the following table represents the bathjanetrical statistics for successively large
ranges of 500 fathoms : —

Depth in

Number of

Number of

Dredgings or

Localities with

Percentage.

Trawlings.

Hexactinellids.

1-500

105

19

18

501-1000

29

10

34-8

1001-1500

31

7

22-6

1501-2000

35

7

20

2001-2500

■ 38

8

21-1

2501-3000

33

6

18-1

3001-3500

4

0

3501-1000

1

0

In the above summary one fact might possibly cause misunderstanding, namely, that
the first region is noted as 1 to 500, although it is well known that in the upper 94
fathoms no Hexactinellids whatever occur. If one excludes the dredgings or trawlings
not deeper than 94 fathoms, and regards the first bathymetrical region as 95 to 500
fathoms, then out of the total 63 dredgings or trawlings, 19 yielded Hexactinellids,
i.e., 30"2 per cent., which is a decidedly larger figure than that given in the table,
18 per cent. It is thus seen that most Hexactinellid localities were between the depths
of 95 and 1000 fathoms, while beyond this the frequency gradually decreases, finaUj-
disappearing below 3000 fathoms.

K the number of species be considered in relation to the number of dredgings or
trawlings at the various depths, it is convenient to follow the plan of the large Table
VIII. (p. 456) and separate the region from 1 to 94 fathoms from that from 95 to 100,
and then to continue up to 1000 fathoms with ranges of 100, and with ranges of 500
up to 3000 fathoms.

REPORT ON THE HEXACTINELLIDA.

461

Percentage of

Depth in
Fathoms.

Number of

Dredgings or

Trawlings.

Number of
Localities with
Hexactinellids.

Number of
Species.

Species in

Drcdgings or

Trawlinga.

1-94

42

95-100

5

i

'4

80

101-200

26

5

20

77

201-300

8

301-400

13

7

14

108

401-500

11

6

6

54-5

501-608

10

5

7

70

601-700

9

4

8

88-9

701-800

3

801-900

3

901-1000

4

2

2

50

1001-1500

31

7

14

45-2

1501-2000

35

7

14

40

2001-2500

38

8

15

39-4

2501-3000

33

6

4

12-4

3001-3500

4

3501-4000

1

...

From this summary it will be seen that at depths from 95 to 200 fathoms the
probability of a dredging including a Hexactinellid is somewhat large {17 to 80 per cent.),
and similarly in depths from 301 to 700 fathoms. Whether the negative results of the
Challenger dredgings or trawlings for the depth 201 to 300 and 701 to 900 are really so
striking as they at first sight appear, I shall leave as a moot point ; the very small
absolute number of dredgings or trawlings in these depths must obviously be taken into
account.

Within the zones from 901 to 2500 fathoms the abundance of Hexactinellida
appears to remain approximately constant at from 40 to 50 per cent. Between
2501 and 3000 a marked decrease in the number of species (to 12*4 per cent.)
takes place, whUe at depths beyond 3000 fathoms no Hexactinellida at all were
dredged.

If a tabular survey, like that given above for the Hexactinellida as a whole, be taken
of the several subdivisions, the following statistic shows the relative distribution in the
first place of Lyssacina and Dictyonina.

From the following table also it will be seen that at the slight depth of 95 fathoms
only Lyssacina were dredged, and that Dictyonina were unrepresented, but that at
depths from 101 to 1000 fathoms Dictyonina occur in equal or slightly greater
abundance.

On the other hand the Dictyonina are far excelled by the Lyssacina in all depths
beyond 1000 fathoms.

462

THE VOYAGE OF H.M.S. CHALLENGER.

Depth in
Fathoms.

Number of

Dredgings or

Trawlings.

Number of
Localities with
Hexactinellids.

Lyssacina.

Dictyonina.

Number of
Speciea

Percentage in

Dredgings or

Trawlings.

Number of
Species.

Percentage in

Dredgings or

Trawlings.

1-94

95-100

101-200

201-300

301-400

401-500

.501-600

601-700

701-800

801-900

901-1000

1001-1500

1501-2000

2001-2500

2501-3000

3001-3500

3501-4000

42

5

26

8

13

11

10

9

3

3

4

31

35

38

33

4

1

"i

5

7
6
5
4

2
7
7
8
6

"i
9

7
1
5
4

1
10
13
13

7

80'
34-6

53'9
9-1
50
44-4

25

32-3

37-1

34-1

21-2

10

"i

5
2
3

1
4
1
2
1

38'-5

53-8
45-4
20
33-3

25

12-9
2-9
5-3
3

If the number of species of both divisions be computed for bathymetrical regions of
500 fathoms, and without reference to the first 94 fathoms, then the statistics read as
follows : —

Depth in

Fathoms.

Number of

Dredgings or

Trawlings.

Lyssacina.

Dictyonina.

Number of
Species.

Percentage in
the Dredgings
or Trawlings.

Number of
Species.

Percentage in
the Dredgings
or Trawlings.

95-500
501-1000
1001-1500
1501-2000
2001-2.500
2501-3000
3001-3500
3501-4000

105
29
31
35
38
33
4
1

21
10
10
13
13
7

20

35

32-2

371

34-1

212

22
6
4

1
2

1

21

20-7

12-9

2-9

6-3

3

From the above this fact is distinctly evident that the Dictyonina show a much
more rapid decrease of species downwards than do the Lyssacina. The percentage
figures of the latter over the whole, from 95 to 3000 fathoms, vary from 20 to 37 per
cent, while the percentages of the Dictyonina over the same range decrease somewhat

REPORT ON THE HEXACTINELLIDA.

463

gradually from 21 to 3 per cent. It is therefore evident that in general the Dictyonina
live at much less depths than do the Lysaacina, which are distributed with much greater
uniformity over the entire bathymetrical range of the HexactineUida.

If the consideration of these relations be extended to the several families (subdivisions
in the case of the Dictyonina), very remarkable differences obtain in regard to the four
families of Lyssacina. In spite of the absolute smallness of the numbers, the diflferences
seem striking enough to merit notice here.

In regard to the Euplectellidse, including Euplectella aspergillum, which lives at
a depth of 95 fathoms, the following table shows the distribution in regions of 500
fathoms : —

Depth iu Fathoms.

Number of Euplectellids.

Percentage in the
Dredgings or TrawUngs.

95-100
101-500
501-1000
1001-1500
1501-2000
2001-2500
2501-3000

1

's

5
3
3

4

20

13 '5

161

8-6

7-9

12-1

Euplectellids are thus widely represented from a depth of 95 fathoms {Euplectella
aspergillum) down to 3000 fathoms, and so far as an estimate with intervals of 500
fathoms is a guide they appear to be distributed with tolerable uniformity at all depths.
That no Euplectellids were found on the Challenger Expedition between 100 and 500
fathoms is probably accidental. It is an interesting fact that the only EuplecteDid
occurring at a depth of less than 100 fathoms, viz., Euplectella aspergillum, belongs to
those forms which, like the Dictyonina, exhibit a connected latticed framework.

Like the Euplectellidse, the Asconematidse are represented down to the lowest zone.
The largest number of species occurs at depths between 1501 and 2000 fathoms, and fi'om
that region the number decreases both upwards and downwards, as is showTi in the
following scale : —

Depth in Fathoms.

Number of Asconematidae.

Percentage in the
Dredgings or Trawlings.

101-500
501-1000
1001-1500
1501-2000
2001-2500
2501-3000

1

i'
3

2

1

1-7

3-2
8-6
5-3
3-3

464

THE VOYAGE OF H.M.S. CHALLENGER.

The Rossellidse exhibit an cntu-ely different distribution. The maximum number of
species occurs in the zone from 101 to 500 fathoms, whence it decreases very markedly
downwards, while two forms were found at a depth of 95 fathoms.

Depth in Fathoms.

Number of Rossellidie.

Percentage in the
Dredgings or Trawlings.

95-100
101-500
501-1000
1001-1500
1501-2000
2001-2500
2500-3000

2

10

4

1
2
2
1

40

17-3

14
3-2
5-7
5-3
3

In contrast to the above, the Hyalonematidse again exhibit a more uniform distribution
over the entire bathymetrical range. At a depth of 95 fathoms, a single member of
this group {Semperella schultzei) was found, while in each of the successive zones of 500
fathoms three to six species occur, until the number sinks to one in the greatest depths,
as the following scale indicates : —

Depth in rathoms.

Number of Hyalonematidae.

Percentage in the
Dredgings or Trawlings.

95-100
101-500
501-1000
1001-1500
1501-2000
2001-2500
2501-3000

1
6
3
3

5
6

1

20

13-4

10-4

9-7
14
15-8

3-3

In regard to the Uncinataria alone, what has been noted in regard to all the
Dictyonina is even more conspicuous, the maximum number of species occurs in the
bathymetrical zone from 101 to 500 fathoms, whence it decreases rapidly downwards to
the smallest figures at the great depths. This is expressed in the following scale : —

Depth in Fathoms.

Number of Uncinataria.

Percentage in the
Dredgings or Trawlings.

101-500
501-1000
1001-1500
1501-2000
2001-2500
2501-3000

15
5
3

1

i

26

17-3
9-7
2-8

REPORT ON THE HEXACTINELLIDA.

465

The Inermia, which only occur in the lesser depths and even there only sparsely,

furnish the following scale : —

Depth in Fathoms.

Number of luermia.

Percentage in tlie
Drcdgings or Trawlings.

101-500
501-1000
1001-1500
1501-2000
2001-2500
2501-3000

2

i

3-4
3'2

Thouo-h the facts expressed above may seem to furnish somewhat scanty material
for collating the bathymetrical range with the characteristics of the several genera
and species, I shall at least attempt to utilise the results of the Expedition in this
direction.

In the genus Evplectella the familiar Euplectella aspergillum was simply found as
before near the Philippine Island, Zebu, at a depth of 95 fathoms, and therefore in
comparatively shallow water ; Euplectella suberea occurred off the Portuguese coast at
depths of 600 to 1090 fathoms, and also to the north-east of Bahia at a depth 1600
fathoms; Euplectella nodosa occurred at a depth of 1075 fathoms near the Bermudas,
and Euplectella crassistellata, on the other hand, at the considerable depth of 2750
fathoms in Mid Pacific.

The four species of the new genus Holascus frequent the greater depths. Several
occurred in the South Indian Ocean between Prince Edward Island and the Crozets, others
south-west of Australia, others near the Philippines, and others again east of Monte
Video, at depths varying from 1375 to 2650 fathoms. With an approximately sunilar
range from 1375 to 2450 fathoms, two species of Malacosaccus were procured in the
Indian Ocean to the west of the Crozets Islands, and in the Atlantic, west of Guinea.
The two beautiful Tsegerinae, Twgeria pulchra and Walteria Jiemmingii, were found in
the South Pacific, near the Kermadec and the Kandavu Islands respectively, at the
moderate depths of 620 and 630 fathoms.

In regard to the Asconematidse, the genus Aulascus, with its single species Aulascus
johnstoni (like the typical genus of the Asconematinse, Asconema itself, which was not
obtained on the Challenger Expedition) occurred at the comparatively slight depth of
310 fathoms, near Prince Edward Island. The branched Sympagella mix occurred
near the Cape Verde Islands at a depth of 1150 fathoms, while the two other Sympagol-
linaj, Pohjrhahdus and Balanites, were found in the southern portion of the Indian Ocean
at depths of 1950 and 1975 fathoms. The stalked Caulophacmse were di-edged from

(ZOOL. CHALL. ESP. — PAET LIII. 1887.)

Ggg59

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

still greater depths ; the two species of the genus Caulophacus itself occurring near the
Crozets Islands and in the northern portion of the Pacific at depths of 1600 to 2300
fathoms, while Trachycaulus gurlittii was dredged in the South Pacific from a depth of
2550 fathoms.

The Rossellidaj were represented at the well-known locality of Euplectella aspergillum,
ofi" the Philipjjine Island, Zebu, at a depth of 9 5 fathoms, by PolylopJms philippinensis
and Crateromorp>ha nieyeri. The former also occurred ofi" the Sunda and Little Ki
Islands at depths of 129 to 140 fathoms, along with two other species of Crateromorpha
and the small Lanuginella pupa; Rossella antarctica was found at similar depths in the
South Indian Ocean near Prince Edward and the Kerguelen Islands ; and one species
of the genus Bathydorus (Bathydorus stellatus), which is otherwise characteristic of the
greater depths, was found in the Magellan Strait at a depth of only 140 fathoms.
Eepresentatives of the genera Acanthascus and Rhahdocalyp>tus occurred along with a
Crateromorpha, Rossella antarctica, previously found in somewhat higher bathymetrical
regions in the Magellan Strait, two species of Aulochone, and the peculiar Euryplegma
auricidare, near the Kermadec Islands, at depths of 300 to 600 fathoms ; Aulocalyx
irregidaris w^as found at the greater depth of 1075 fathoms near the Bermudas, and at
1600 fathoms near the Crozets Islands.

The maximum depth as yet known for any Hexactinellid was that of the genus
Bathydorus, 2900 fathoms. While Bathydorus spinosus occurred to the east of Prince
Edward Island at a depth of 1600 fathoms, Bathydorus fimhriatus was found about
20° to the east of Japan at a depth of 2300, in the middle of the North Pacific down as
far as 2900 fathoms. In the middle of the South Atlantic, at a depth of 2025 fathoms,
the peculiar species Caulocalyx tener occurred.

The Hyalonematid found in shallowest water was Semperella schidtzei, which was
found off the Philij)pine Island, Zebu, in 95 fathoms, at the Moluccas, off Little Ki Island,
in 129 to 140 fsithoms. At the last mentioned locality two species of Pheronema were
also found, while the Atlantic species Pheronema carpenteri occurred to the south of
Pernambuco at a depth of 1600 fathoms. Of the two large species of Poliopogon, the
one occurred at a depth of 630 fathoms near the Kermadec Islands, the other to the
south-west of the Canaries at a depth of 1525 fathoms.

There is a wide interval in the distribution of the different species of Hyalonema.
While several are found at depths of 345 to 390 fathoms, most of them occur below
1000, and thence down as far as 2550 fathoms, which is the depth attained by
Hyalonema (Stylocalyx) elcgans.

The small but peculiar genus Farrea is represented at very various depths. Some
forms occur between 300 and 400, others not above 2600 fathoms. It is different with
the closely allied and externally similar Euretidse. The four different species of the
genus JEurete itself were all procured at the Moluccas, off Little Ki Island, from depths

REPORT ON THE HEXACTINELLIDa. 467

of 129 to 140 fathoms, while the beautiful Lefroyella decora was dredged at the
Bermudas from a depth of 435 fathoms.

AphrocalUstes also occurs at very diverse depths. Some forms were found near the
Philippines at 375 fathoms, and off St. Thomas at 390 fathoms, while others occurred off
Ascension and St. Thomas at depths of 420 to 450, and others finally at the Bermudas
at a depth of 1075 fathoms. Among the Coscinoporidse, the genus Chonelasma was
represented at depths of 450, 630, and 1075 fathoms in the North Atlantic, the South
Indian Ocean, and the South Pacific. Among the Tretodictyidse the genus Hexactinella
occurred near Little Ki Island at a depth of 129 to 140, off the Crozet Islands at a depth
of 550 fathoms. Cijrtaulon and Fieldingia also occurred only near Little Ki Island at
depths of 129 to 140 fathoms, while Sderothamnus was found near the Banda Islands
at slightly greater depths, down to 360 fathoms. Of the remarkable Mseandrospongidse,
Dactylocalyx was found below 1000 fathoms (1075) near the Bermudas. Myliusia
callocyathus, and Aulocystis zittelii, on the other hand, occurred at the comparatively
slight depth of 129 to 140 fathoms off Little Ki Island, and at depths of 360 to 390
fathoms near the Bermudas and St. Thomas.

468

THE VOYAGE OF H.M.S. CHALLENGER.

INFLUENCE OF THE NATURE OF THE GROUND.

It is necessary to consider in the first instance what kind of ground or bottom is
favourable to the occurrence of Hexactinellids. For this purpose the difierent kinds of
sea-bottom may be noted, along with the number of dredgings or trawlings on each
ground. Those searches must then be noted which yielded any Hexactinellida, so that
their proportion to the total number for each kind of bottom may be estimated.

Then the proportionate abundance of species on each kind of bottom must l^e collated
with the number of dredgings or trawlings. In Table IX. (p. 470) the data for this
purpose are given and the general results may be summarised thus : —

Nature of the Bottom.

Number of

Dredgings or

Trawlings

Number of

Dredgings or

Trawlings witli

Hexactinellids.

Percentage
with Hexac-
tinellids.

Number of
Species.

Percentage of

Species.

1. Sand, ....

22

5

22-7

5

22-7

2. Gravel and stones,

3

2

66-6

2

66-6

3. Hard ground, .

17

4

23-5

6

35-3

4. Coral mud.

16

2

12-5

7

43-8

.5. Volcanic mud, .

24

4

16-6

14

58-3

6. Green mud.

17

1

5-9

1

5 '9

7. Red mud, ....

11

1

91

2

18-2

8. Mud (including blue mud),

53

12

22-6

32

60-4

9. Red clay, ....

39

6

15-4

11

28-2

10. Globigerina ooze,

53

10

19

13

24-5

11. Pteropod ooze, .

10

3

30

7

70

12. Eadiolarian ooze.

4

2

50

2

50

1.3. Diatom ooze.

4

3

75

9

225

From this summary it is seen that the smallest number of localities yielding Hexac-
tinellida occurred on those bottoms which are described as green mud and red mud. A
somewhat small number were also dredged from Coral mud, volcanic mud, and red clay,
and then from Globigerina ooze and mud (including blue mud), on all of which the
percentage of Hexactinellid localities is over twenty-two. The same may be said of the
sandy and rocky ground. The localities on Pteropod ooze were somewhat more
abundant, viz., 30 per cent., and on Eadiolarian ooze yet more so, viz., 50 per cent.
The highest numbers refer to gravel and stones, 66 '6 per cent., and to Diatom ooze,
75 per cent.

The statistics are somewhat different when the number of species found on the different
bottoms is taken into consideration. Here again green mud and red mud yielded the
smallest percentage of forms, viz., 5 '9 per cent, and 18*2 per cent, respectively. Then

REPORT ON THE HEXACTINELLIDA. 469

follow sand with 22*7, Globigcrinca ooze with 24"5, and clay with 28'2 per cent., while
hard ground yielded 35"3 per cent., Coral mud 43'8 per cent., and mud (including blue
mud) as much as 60*4 per cent. Radiolarian ooze yielded 50 per cent., gravel and
stones 6 6 '6 per cent., Pteropod ooze 70 per cent., and Diatom ooze as many as
225 per cent.

As one would naturally expect, the forms equipped with root tuft were principally
found on soft muddy ground, on the various kinds of mud and ooze, and on red clay,
while those adherent to solid bodies occurred partly on hard ground, gravel and stones,
partly also on mud, both volcanic and Coral, on Pteropod ooze and on sand, being in
such cases usually found attached to projecting stones or other solid bodies.

Most of the Euplectellidse, which possess a developed basal tuft, occurred on mud
(including blue mud), several also on Globigeriaa and Diatom ooze, and one species on
each of the following — Coral mud, red clay, and Eadiolarian ooze. Of the firmly attached
Tsegerinse and less definitely classifiable forms, several occurred on Coral mud, one form
on volcanic mud and another on red clay.

Of Asconematidse three species occurred in Diatom ooze, two in Globigeriaa ooze,
and one species on each of the following — sand, volcanic mud, and red clay.

The Rossellidse were represented on very diverse bottoms, the majority (nine species)
on sand, several (three on each) on red clay and volcanic mud ; two on sand, two on
hard ground, and two on Diatom ooze ; one on gravel and stones, and another on Coral
mud.

The Hyalonematidae (Amphidiscophora) were altogether absent from sand, gravel
and stones, Coral mud, and Radiolarian ooze ; while hard ground, green mud, red mud,
and Diatom ooze, each yielded one species ; volcanic mud and Pteropod ooze each two ;
red clay, three ; mud (including blue mud), six ; and Globigeriua ooze, the majority,
namely, eight.

In regard to Dictyonina, the Uncinataria were altogether absent from green and red
mud; one species occurred on each of the following — gravel and stones, red clay,
Globigerina ooze, Radiolarian ooze, and Diatom ooze ; two were found on sand, three
on hard ground, and the same number on Coral mud, four on Pteropod ooze, five on
volcanic mud, and ten on mud (including blue mud).

Of the few Inermia which were obtained by the Challenger, Coral mud and Pteropod
ooze each yielded one, whde volcanic mud and mud (including blue mud) each yielded
two species.

470

THE VOYAGE OF H.M.S. CHALLENGER.

TABLE IX. — Eelation of the Hexactinellida of the

Sand.

Gravel and Stones.

Hard Ground.

Coral Mud.

Volcanic Mud.

Green Mud.

Station.

Station.

Station.

EUPLECTEL-

Statioa
EupUcteUa

nodosa, . 56
Tcegeria

pidchra, . 174c
Wuibdodictyuvi

delicatwm, 56

ASCONEMA-
TID^.

Aidasciis

johnstoni, 145a

ROSSELLID^. ■

(14!

antarctica, \ 320
Aidocalyx

irregularis, 145a

Russella

antarctica, 150

A ulochone

cylindrica, 171
Acanthascus

grossularia, 148a

Aulocalyx

irregularis,

56

Station.
Walteria

flemmingii, 170a

Station.

Sympagella
mix.

94

Rossella

antarctica,
Orateromorpha

tumidtt, .
Euryple^ma

auricidare.

149H
194a

170a

Amphidis-

COPHORA.

s. Htalone-

MATID.E.

Poliopogon
amadou,

Poliopogon

gigas, . 170a
Stylocalyx

globus, . 194a

Stylocalyx
apertus,

232

Uncinataria.

Dictyonine, . 343
Aphrocalliste

hocagei, . 344

Eurete

schmidiii, 201

Chonelasma
lamella,

Farrea
sp.,

Rexactindla
sp.,

148a
135E

148

Farrea

sp.,
Aphrocallutes

bocagei, .
Chonelasma

sp.,

56
56
56

clavigera,

194a

occa.

17()A

Chonelasma

lamella, .

170a

uncinatum,

170a

Sclerothamnus

clausii, .

194a

Inkbmia.

Dactylocalyx
patella.

56

Myhiisia

callocyathus, 194a
Auloet/sfis

zUlelii, . 194a

5 Localities.
5 Species.

2 Localities.
2 Species.

4 Localities
6 Species.

2 Localities.
7 Species.

4 Localities.
14 Species.

1 Locality.
1 Species.

REPORT ON THE HEXACTINELLIDA.

471

Challenger Collection to the nature of the Bottom.

Red Mud.

Mud and Blue Mud.

Red Clay.

Globigerine Ooze.

Pteropod Ooze.

Radiolarian Ooze.

Dia,tom Ooze.

Station.

Station.

Station.

Station.

Station.

Station.

Station.

Eupledella

Hexactinellid ?, 169

Bolascus

Eupledella

Eupledella

Bolascus

tuberea, . 124

Etqilectdla

fibulatus, 160

suberea, . V.

crassistellata, 274

fibulatus, 147

subrrea, IV,

Eyalostylus

Solascus

polejaevii, 157

aspergillum, 209

dives, . 289

fibidatus, 146

Holascus

Didyocalyx

.3/aZ«cosrtccw5

stellatus, 325

gracilWf 281

!{Ji</MiC!t-

ndleyi, . 211

Za<KS, . 348
vastus, . 143

Caidophacics

elegans, . 241

Caulophacus

sp., . 246
<ener, . 333

Balanites

pipetta, , 157
Polyrhabdus,

ovifonnis, 156
Caulophacus

latus, . 147

Lanngindla

Bathydorus ( 241
fimbriatus, \ 248

Bathydorus

pupa, . 192

spinosui, 147

Polylojihus

baculifer, 286

yi ulocalyx

jih ilippin-

Trachycoiilus

irregularis, 147

ensis, . 209

gurlitli, . 289

Acanthascus

duhius, . 310

Batln/'lonfs

SirllulMS, 307

lihabdocali/ptus

rosperi, . 310

Crateromorpha

thierfelderi, 192

viurrayi, 192

meyeri, . 209

Aulochone

lilium, . 214

PheT07iema

Hyalonema

Hyalonema

Hyalonema,

Stylocalyx

Stylocalyx

carpaUeri, 124

gracile, . 211

sp., . 241

sp., . 184

thomsoni, 73

claviger, 147

tenue, . 323

robwstum, 241

sp., . 271

Hyalonema

sp., . 205

Stylocalyx

sp., . 333

toxeres, . 24

Pheronema 211

elegans, . 289

comts, . 158

gigatiteum, 192

divergens. 111

globosum, 192

poculum 300

Semperella (192
schuUsei, \ 209

Stylocalyx (246
depressus, \ 271

elegans, . 271

Dictyonine, . 192
Farrea

( ^Sl
Dictyonine, . -j Jgg

Dictyonine, . 293

sp., . 23

Farreid, . 272

/Ta?Tca

sp., . 147

occa, . 207

occa, . 24

£ure<e

Aprocallistes ( 23
bocagei, \ 24

scmperi, 192

marshalli, 192

Chonelasma

carteri, . 192

sp., . 23

farreopsis, 192

Aphmcallisles

ramosus, 210

Bexactinella

lata, . 192

Crytaulim

solutus, . 192

Fieldingia

lagettoides, 192

Myliu^ia

Myliusia

callocyathus, 192

callocyathus, 24

jl)rfo(-vs(is

ii-«rf8(', . 192

1 Locality.

12 Localities.

6 Localities.

10 Localities.

3 Localities

2 Localities.

3 Localities.

2 Species.

32 Species.

11 Species.

13 Species.

7 Species.

2 Species.

9 Species.

472

THE VOYAGE OF H.M.S. CHALLENGER.

TABLE X. — Survey of the Hexactinellida in the Challenger Collection, and of other

SUFFICIENTLY DESCRIBED FORMS, WITH THEIR LOCALITIES, AND THE NUMBER OF SPECIMENS OBTAINED.

Species.

CHALLENGER COLLECTION.

Locality.

Nature

of the

Bottom.

I. LYSSACINA.

I. Hexasterophora.

I. EuplectellidtE.
1. EupleotelliiiEe.

1. Euplectdla, Owen.

1. EupUctella aspergillum, Owen,

2. EupUctella svherea, Wyv. Thomson,

3. EuplecteUa cucumer, Owen, .

4. EuplecteUa jovis, 0. Schmidt,

5. EuplecteUa oioeni, Herklots and Marshall, .

6. EupUctella crassistellata, a. sp.,

7. EuplecteUa [f) nodosa, n. sp., .

2. Regadrella, O. Schmidt.
Regadrella phcenix, 0. Schmidt,

2. Holascin^.

1. Hohiscus, n. gen.

1. Holascus stdlatu.1, n. sp.,

2. Holasctts Jiiulatus, n. sp.,

3. Holascus 2'oUjaeva, n, sp.,

4. Holascus ridkyi, n. sp.,

2. Malacosaccus,

1. Malacosaccus vastus, n. sp.,

2. Malacosaccus unguiculatus, n. sp.,

3. Tsegerinse.

1. Tayeria, n. gen.

Station.

Zebu, Philippines,

South-east of Cape St.

Vincent,
West of Gibraltar,
North-east of the Rio San

Francisco,

North-west of the CaroUne
Islands,

Off Bermudas, .

bl. m.
gl. oz.

rad. oz.

East of Monte Video,

Marion Island to Crozets,
West of the Crozet Islands,
West of Tasmania,

Southern Ocean,

West of Mindanao,

Marion Island to the Crozets
South of Sierra Leone,

Tasgeria pulchra, n. sp. ,

Fiji Islands,

bl. m.

gl. 02.

di. oz.
r. cl.

di. oz.

bl. m.

gl. oz.
gl. oz.

crl. m.

209

IV.
V.

124

274

325

146
147
160

157

211

146
348

174c

OTHER COLLECTIONS.

Locality.

Philippines,

Bohol and Zebu, Philippines,

Philippines,

Zebu, . . . .

Cape St. Vincent, off Portugal

North of Scotland,

North Atlantic,

Seychelles Islands,

St. Lucia (423) and Grenada
(416), Gulf of Mexico,

Authority.

Depth in
Fathoms.

Japan,

Barbados (221, 288), Santa
Cruz (248, stones), Gulf of
Mexico,

Owen,
Bowerbank,
Semper,
Claus,
Higgins, .

Wyv. Thorn

son,
Filhol,

95-100

600
1090

1600

Owen,

0. Schmidt,

433, 416

MarshaU, .

2750

0. Schmidt,

221, 288,
248

2650

1375
1600
2600

1950

2225

1375

2450

61

Number
OF Speci-
mens.

many

1

several
several

1
1

several

1

several

1

1

1

REPORT ON THE HEXACTINELLIDA.

473

Table X. — continued.

Species.

CHALLENGER COLLECTION.

Locality.

2. Waltcria, n. gen.
Walteria fiemmingii, n. sp.,

IlaJbrodictyum., Wyv. Thomson.
Eabrodiclywn speciostim, Quoy et Gaimard, .

Eiidiclyum, Marshall.
Eudiciyvm- elegans, Marshall,

Dietyocalyx, n. gen.
Dictyocalyx gracilis, n. sp. ,

RhaModictyum, 0. Schmidt.
Rhahdodictywm delicatum, 0. Schmidt,

Rhahdopectella, O. Schmidt.
Ithabdopectella tintinmcs, 0. Schmidt, .

Sertmgia, 0. Schmidt.
Hertwigia/alcifera, 0. Schmidt,

Hyalostyhts, n. gen,
Ilyalostylws dives, n. sp., .

IL Asconematidas.
1. Asconematinae.

1. Asconema, Sav. Kent.

Asconeina setubalense., Sav. Kent,

2. A^Uascvs, n. gen.
lulascnsjohnstoni,u.%^..

2. Sympagellinse.

1. Sympagella, 0. Schmidt.

Sympagella mix, 0. Schmidt,

2. PolyrlwMxis, n. gen.
Pnlyrhabdus orn/ormis, n. sp., .

3. Balanites, n. gen.
Balanites pipetta, n. sp., .

Off the Kermadec Islands, .

South of Tahiti,

South-west of the Bermudas,

Nature
of the
Bottom

vole. m.

r. cl.

Station.

South Pacific,

South-east of Prince Ed-
ward Island,

Porto Prayo and St lago,
Cape Verde Islands,

Southern Ocean,

Southern Ocean,

di. oz.

di. oz.

OTHER COLLECTIONS.

Locality.

Authority.

:70a

281

56

Moluccas,
Unknown,

Bequia, Gulf of Mexico,

Bequia, Gulf of Mexico,
Grenada,

Dominica,

0. Schmidt,

0. Schmidt,

0. Schmidt,

289

145a

156

157

Off coast of Portugal, Fffiroe
Channel, " Triton " Ex-
pedition, Station 4,

Florida,

Depth in
Fathoms.

630

Number
OF Speci-
mens.

2385
1591

1075
1591

291,994

O. Schmidt, 611

2550

Murray,

0. Schmidt,

327^30

310

100-128
98,123

1975

1950

several

(ZOOL. CHALL. EXP. — PAET LIII. 1887.)

Ggg 60

474

THE VOYAGE OF H.M.S. CHALLENGER.

Table X. — continued.

CHALLKNGEE COLLECTION.

OTHER COLLECTIONS.

Number
OF Speci-
mens.

Species.

Locality.

Nature

of the

Bottom.

Station.

Locality.

Authority.

Depth in
Fathoms.

3. Caulophaciuie.

1. Caulojjhacus, n. gen.

1. Cavlophacxislatus,-!!. s\^.,

West of the Crozet Islands, .

di. oz.

147

1600

1

2. Caulophacus elegans, n. .sp., .

East of Yokohama,

r. cl.

241

2300

3

2. Trachyccttdus, n. gen.

Trachycaulus gurlittii, 11. sp., .

South Atlantic,

r. cl.

289

...

2550

1

III. Ross ell ids.

1. Lamiffinellu, 0. Schmidt.

Lamiginella inipa, 0. Schmidt, .

2. Polylojihus, D. gen.

Little Ki Island,

1)1. m.

192

St. lago, Cape Verde Islands,

0. Schmidt,

129-140

several

Polylophus philippinensis, Gray,

Zebu, Philippmes,

209

Zebu, Philippines,

Gray," .

95-100

many

3. Rossella, Carter.

1. Rossella antarclica. Carter, .

Off Marion Island,

Off Cumberland Bay, Ker-

guelen Island,
Heard Island, .
South-east of Monte Video, .

vole. sd.

vole. m.
co.gr.
gr. & sd.

145

149H

150

320

Antarctic Ocean, lat. 74°-77°
S., long. 175° W., .

Carter,

310

120
150
600

206-300

several

several

many

2

2. Rossella velata, Wyv. Thomson,

Strait of Gibraltar, "Porcu-
pine " Expedition, .

Wyv. Thom-
son,

651

4. Acanthascus, n. sp.

h Acanlhasctts grossidaria, n. !i[).,

Possession Island, Crozets, .

hd. gr.

148

210

1

2. Acanthascus dubius, n. sp., .

South of Puerto Bueno,
Magellan Strait,

bl. ni.

310

400

1

3. Acanthascus cactus, n. sp. ,

Sagami Bay, Japan, .

Ddderlein,

1

5. Bathydorus, n. geu.

1. Bathydorus fimbriatus, n. sp.,

North Pacific, .
North Pacific, .

r.cl.
r. cL

241
248

2300
2900

1

1

2. Bathydorus stellatus, n. sp. , .

Port Grappler, Macjellan
Strait, . . ' .

bl. m.

307

140

1

3. Bathydorus sjnnosus, n. sp., .

West of the Crozets, .

di. oz.

147

1600

4. Bathydorus bacuH/er, n. sp., .

South Pacific, .

r.cl.

286

AS35

1

6. Rhahdocalyptus, n. gen.

1. Rhabdocalyptws mollis, n. sp.,

Sagami Bay, Japan, .

Dbderlein,

2

2. Rhabdocalyptus roeperi, n. sp..

South of Puerto Bello, Magel-
lan Strait,

bl. m.

310

400

1

7. Crateromorpha, Gray.

1. Cratero7norpha meyeri, Gray,

Zebu, Philippines,

bl. m.

near 209

Zebu, Philippines,

Gray and
Carter, .

95

several

REPORT ON THE HEXACTINELLIDA.

475

Table X. — continued.

Species,

2. Crateromorpha thierfelderi, n. sp., .

3. Crateromorpha murrayi, n. sp.,

4. Crateromorpha tumida,

8. A ulochmie, n. gen.

1. Aulochone cylindrica, n. sp. , .

2. Aulochone lilium, n. sp.,

9. CaiUocalt/x, n. gen.
Catilocalyx tener, n. aTp., .

10. Aidocahjx, n. gen.
Aulocalyx irregularis, n. sp.,

11. Enrypleijma, n. gen.
Euryplegma aurictUare, n. sp. , .

TI. AMPHIDISrOPHORA.
Uyalonematidse.
1. Hyalonematinse.
1. llyalonema.

i. Hyalonema, s. xtr

1 . Hyalonema Heboldi, Gray,

2. llyalonema gracUe, n. sp. ,

•3. Eyalonema divergens, n. sp.,
4. Hyaloneina kentU, 0. Schmidt,

5. Hyalonema, toxeres, Wyv. Thomson,

6. Hyalonema poculum, ii. .sp., .

7. Hyalonema cmtus, n. sp. ,

2. Stylocalyx.

1. Stylocalyx thomsonii, Marshall,

2. Stylocalyx apertus, n. sp. ,

3. Stylocalyx depressus, n. %Y., .

CHALLENGER COLLECTION.

Locality.

Off Little Ki Island,
Off Little Ki Island,
Off Banda,

North of Raouland Kermadec
Islands,

South-east of Mindanao,
Philippines, .

South Atlantic,

South-west of the Bermudas,
South-east of Prince Ed-
ward I.sland,
West of Crozets,

North-east of Kermadec
Islands,

Nature
of the
Bottom.

bl. m.
bl. m.
bl. m.

hd. gd.
bl. m.

gL 07.

vole, sd,

vole, sd
di. oz.

West of Mindanao, Philip-
pines,

Mid Pacific,

North of St. Thomas, West
Indies,

Juan Fernandez,

South Pacific, .

West of the Azores Islands, .

Sagami Bay, Japan,

North Pacific, .
Mid Pacific,

Station.

bl. m.
gl. oz.

pt. oz.

gl. 02,
gl. OZ.

pt oz.

gr. m.
gL 02.

192
192
194

171

214

333

66

145a
147

170a

OTHER COLLECTIONS.

Locality.

211
271

24
300
158

73

232

246
271

Japan, .

Sagami Bay, Japan,

Authority.

Depth is

Fathoms.

Grenada (338, 576), Mar-
tinique (565), Guadeloupe
(583), Bequia (1507), Gulf
of Mexico,

North of Shetland Islands, .

129-140

129-140

360

600
500

2025

1075

310
1600

630

NUMBEIt

OF Speci-

.MENS.

Gray,
Diiderlein

and
Hilgendorf,

O. Schmidt,

Marshall,

circa 200

2225

2425

33&-1507

390
1375
1800

1000
550

345

2050
2425

Sf^veral
1

several

several

several
2

476

THE VOYAGE OF H.M.S. CHALLENGER.
Table X. — continued.

CHALLENGER COLLF.CTION

OTHER COLLECTIONS.

Species.

Defth in
Fathoms.

Number
OF Speci-
mens.

Locality.

Nature
of the
Bottom.

Station.

Locality.

Authority.

4. Stylocalyx claviger, n. sp..

West of the Crozets, .

di. oz.

147

1600

1

5. Stylocalyx glohm, n. sp..

OffBanda,

vole. m.

194a

360

1

6. Stylocalyx elegans, n. sp.,

South of Hawaii,

gl. oz.

271

2425

1

7. Stylocalyx tener, n. sp..

South Pacific, .

r. cl.

289

2550

1

Hyalonema lusitanicwn (?), Barboza du Bocage,

South-east of Cape St. Vin-
cent ; west of Gibraltar, .

m.

IV.

Portugal,

Barboza, .

600

1

Hyalonema cebuense, Higgin,

...

...

Zebu, Philippines,

Higgin, .

1

Byal<mematenue,n. s,}^., ....

East of Buenos AjTes,

bl. m.

323

1900

1

Byalonema robust um, u. sp..

North Pacific, .

r. cl.

241

2300

1

Hyalonema sp., .....

South-east of Cape York,
Torres Strait,

gl. oz.

384

1400

1

Hyalonema sp., .....

West of Luzon, Philippines,

bl. m.

205

1075

1

Uyaloiieina sp. . .

Mid Pacific, .

gl. oz.

271

2425

1

Hyalonanasp., .....

South Atlantic,

gl. oz.

333

2025

1

2. PheroneitM, Leidy.

1. Pferonemaonna!, Leidy.,

Santa Cruz, West Indies,
Santa Cruz,

Leidy,

0. Schmidt,

lsd-248

1
several

2. Pheronetna carpenteri, Wyv. Thomson,

Off Macio, Brazil,

124

North-west of Scotland,

West of the English Channel,

" Lightning " Expedition, No.

" Porcupine," Station 42,
"Triton Expedition, Station
13

Wyv! Thom-
son,

Wyv. Thom-
son,

1600

530
862

570

several

4
several

1

3. Pheronema grayi, Sav. Kent,

Off Setubal, Portugal,

Sav. Kent,

several

4. Plieronema litmisphcericum. Gray, .

... , ...

Singapore,

Gray,

1

5. Pheronema gloiosum, n. sp., .

Off Little Ki Island, .

bl. m.

192

140

several

6. Pheronema giganteum, n. sp.,

Off Little Ki Island, .

bl. m.

192

140

1

3. Poliopogon, Wyv. Thomson.

1. Pdiopogon amadou, Wyv. Thomson,

South-west of Canary Islands,

hd. gd.

3

1525

2

2. Poliopogon gigas, n. sp.,

Between Kermadec and
Raoul Islands,

vole. m.

170a

630

1

2. Semperellinse.

Semperella, Gray.

Semperella schuUzei, Semper,

Off Zebu, Philippines,
Off Little Ki l.'sland, .

209
325

Philippines,

Semper and
Gray, .

100
140

1

1

REPORT ON THE HEXACTINELLIDA.
Table X. — continued.

47'

CHALLENGER COLLECTION.

OTHER COLLECTIONS.

Number

OF Speci-

MEKS.

Species.

Locality.

Nature

of the

Bottom.

Station.

Locality.

Authority.

Depth in

Fathoms.

II. DICTYONINA, Zittel.

-I

1

I. Uncinataria

I. CLAVULAKIA.

F a r r e i d oe.

Farrea, Bowerbank.

1. Farrea occa, Carter. ....

North of St. Thomas, West

Indies,
North - east of Kermadeo

Islands,
Philippines,

pt. oz.

vole. ra.
bl. m.

24

170
207

Off Misaki, Japan,
Sagami Bay, Japan, .

Dbderlein

& Carter,
Doderlein,

390

520
700

1

]
1

several

2. Farrea sollasii, n. sp., ....

Sagami Bay, Japan, .

Doderlein,

1

0. Farrea vosTnaeri, n. sp.,

Sagami Bay, Japan, .

Doderlein,

1

4. Farrea clavigera, n. sp.,

OffBanda,

vole. m.

194a

200-360

1

Farrea sp. (.? woodwardi, Sav. Kent), .

Portugal,

Fan-ea sp., .

Off Sombrero, West Indies,

pt. oz.

23

450

1

Farrea sp., ......

South-west of Bermuda,

crl. m.

56

1075

1

Farrea sp. , .

Off Tristan da Cunha,

hd. gd.

135e

1000

I

Farrea sp., ......

West of the Crozets, .

di. oz.

147

1600

1

2. SCOPULABIA.

1. Euretida?. F. E. S.
1. Eurete, Carter.

1. Etiretesem})eri,u. sp., .

Off Little Ki Island, .

bl. m.

192

140

1

2. Eurete sehmidtii, n. sp. ,

Off Samboangan, Philippines,

St. gr.

201

102

2

3. Eurete farreopsis, Carter,

Off Little Ki Island, .

bl. m.

192

140

2

4. Eurete carteri, n. sp., .

OffLittleKi Island, .

bLm.

192

140

2

5. Em-ete tnarshalli, n. sp., • » •

Off Little Ki Island, .

bL ra.

192

140

1

6. Eurete bmoerbankii, n. sp., .

Sagami Bay, Japan, .

Doderlein,

1

Eurete sp. (simplicissima. Semper),

Zebu, Philippines,

Semper, .

2. Periphragella, Marshall.

PeriphrageUa elism, Marshall, .

Moluccas,

Sagami Bay, Japan, .

Marshall, .
Doderlein,

Gottsche,
Hilgendorf,

5

3. Lefroyella, Wyv. Thomson.

Lefroyetla decora, Wyr. Thomson,

Off Bermudas, .

2

478

THE VOYAGE OF H.M.S. CHALLENGER.

Table X. — continued.

CHALLENGER COLLECTION.

OTHER COLLECTIONS.

NCMBKR

OF Speci-
mens.

Species.

Locality.

Nature
of the
Bottom.

Station.

Locality.

Authority.

Depth in
Fathoms.

II. M e 1 i 1 1 i 0 n i d oe, Zittel. ,
Aphrocallisies, Gray.

1. Aphrocalliates bocagei, Perceval Wright,

South - west of Ascension

Island,
Ofif St. Thoma.s, West Indies,
Off St Thomas, West Indies,
South-west of Bermudas,
West of Ascension Island, .

vole. sd.

pt. oz.

pt. oz.

crl. m.
vole. sd.

343
23

24
56
344

South-east of coast of Spain,

' ' Porcupine " Exp. ,
Cape Verde Islands, .
South-east of Florida,
Off Spain and Portugal,
Off France and Portugal,

Wright,
0. Schmidt,
Sav. Kent.
Milne - Ed-
wards, .

425
450
390
1075
420

...

1

1

1
several
several

2. Aphrocallisies ramosus, n. sp.,

Philippines,

bl. ni.

210

Sagami Bay, .

Doderlein,

375

3. Aphrocallisies vastus, n. sp. , .

Sagami Bay, .

Doderlein,
Gottsche,

4. Aphrocallisies beatnx, Gray, .

Off Malacca, . .

Gray,

1

III. CoscinoporidsB, Zittel.
Chonelasma n. gen.

\. Chonelasma lamella, Tt. si^., .

South of Possession Island,

Crozets,
North - east of Kermadec

Islands,

hd. gd.
vole. ni.

148a
170a

550
630

]
1

2. Chonelasma hamatum, n. sp., .

North - east of Kermadec
Islands,

vole. m.

170a

630

1

3. Chonelasma doederleinii, n. sp.,

Sagami Bay, Japan, .

Doderlien,

1

i. Chonelasma calyx, n. sp..

Sagami Bay, Japan, .

Doderlien,

1

Chonelasma sp., .

Off St. Thomas, West Indies,
South-west of Bermudas,
South of Possession Island, .
Off coast of Portugal,

pt. oz.
crl. m.

23
56
148a

450
1075
550

1
1

1

IV, Tretodictyidffi.

1. Hexactinella, Carter.

1. Hexactinella tubulosa, n. sp..

Sagami Bay, .

Doederlein,

several

2. Hexactinella lata, n. sp.,

Off Little Ki Island, .

bhm.

192

140

several

3. Hexactinella ventilabrum, Carter,

Japan, ....
Sagami Bay, .

Carter,
Doderlein,

l'

Hexactinella sp. , .

South of Possession Island,
Crozets,

hd. gd.

148a

550

1

2. Cyrtaulon, n. gen.

1. Cyrtaulon sigsbeei, 0. Schmidt,

Barbados,
Morrolight,
St. Vincent, .

0. Schmidt,

158. 100
292
124

■3

2. Cyrtaulon solutus, n. sp.,

Off Little Ki Island, .

bl. m.

192

140

1

REPORT ON THE HEXACTINELLIDA.
Table X. — continued.

479

Species.

4. Fieldingia, Sav. Kent.
Fieldingia lageltoides , Sav. Kent,

5. Sclerothamnus, Marsliall.
Sclerothamnus clmtsii, Marshall,

V. M a' a n d r 0 s p 0 11 g i d le, Zittel.

1. Dactylocalijx, Stutchbury.

1. Dactyhcalyx pumiceus, Stutchbury, .

2. Dactylocalyx subglobosus, Gray,

3. Dactylocalyx patella, n. sp., .

2. Scleroplegma, 0. Schmidt.
Scleroplegma conicum, 0. Schmidt,

3. Myliusia, Gray,
Myliusia callocyathus, Gray,

4. Aulocystis, n. geu.

1. AulocyslU grayi,BoweTba,nk,

2. Aulocystis zittelii, Marshall, .

CHALLENGER COLLECTION.

Locality.

Off Little Ki Island,

Off Banda,

Nature

of the Station.
Bottom.

OTHER COLLECTIONS.

Locality.

Authority.

Depth in
Fathoms.

South-west of Bermuda,
Oil coast of Portugal,

Off St. Thomas, West Indies,
Off Little Kj Island, ,
Off Banda,

Off Little Ki Island,
Off Banda,

Number
op Speci-
mens.

i

1
bl. m.

192

vole. m.

194a

56

pt. oz.

bl. m.

vole. ni.

24

192
194a

bl. m.
vole. m.

192
194a

OffCeziinbra, Portugal,

Off Negros and Zebu, Philip-
pines,

Off Barbados, St. Vincent,

Malacca ?,
St Lucia,
North-west of Havanna,

llorrolight, West Indies,

West Indies,

Off St. Vincent, West Indies,

Zebu, Philippines,

Sav. Kent,

140

500

360

Murie,

80,100

Stutchbury,
0. Schmidt,

103

Gray,

0. Schmidt,

116
190

1075

0. Schmidt

292

Gray," .

390
140
360

Bowerbank,

::: 1

Meyer and
Marshall,

140
360

several

480

THE VOYAGE OF H.l\r.S. CHALLENGEE.

List of Species of recent Hexactinellida arranged Alphabetically.

Acanthascus cactus, n. gen. et sp., this Keport, p. 148,

„ iJuhiiis, n. gen. ct sp., this Report, p. 147,

,, f/rossularia, n. gen. et sp., this Eeport, p. 145, .

Adrasta infundibulum, Wyv. Thomson, Phil. Trans., clix., .
Alcyonellum aspergilhun (Owen, sp.) Ewbk., Phil. Trans.,

,, corbicula, Bwhk., Phil. Trans., ....

„ gelatinosum, Blainville, Diet. d. Sci. Nat, torn. Ix. p. 492,

,, rohustum, Bwbk., Pliil. Trans., ....

,, specioaum, Quoy et Gaimard, Voyage de rAstrolabe,

AmpMdiscushipileatushyalonematis, Ehrenberg, Monatsber. d. k. preuss. Akad. d. Wiss. Berlin, p. 851
AplirocalMstes beatrix, Gray, Proc. Zool. Soc. Loud., vol. xxvi. p. 114,

„ hocagei, Perc. Wright, Quart. Journ. Micr. Sci., vol. x. p. 1,

„ ramosus, n. sp., this Report, p. 319, .

„ vastus, n. sp., this Eepgrt, p. 317,

Asconema kentii, 0. Schmidt, Spong. Meerb. v. Mexico, p. 65,

„ setHbalenee, Sav. Kent. Monthly Mior. Journ., p. 205,
Aulascus johnstoni, n. gen. et sp., this Report, p. 118,
Aulocalyx in-egularis, n. gen. et sp., this Report, p. 174,
Aulochone ojlindrica, n. gen. et sp., this Report, p. 168,

„ lilium, n. gen. et sp., this Report, p. 171, .
Aulocystis grayi (Bwbk., sp.), Proc. Zool. Soc. Lond., p. 335,

„ ziftelii (jMarshall, sp.), INIarsh. u. Meyer, Mittheil. Zool. Mus. Dresden
Aulodictyum costiferum, Marsh, u. Meyer, Jlittheil. Zool. Mus. Dresden,

„ facundum, 0. Schmidt, Spong. Meerb. v. Mexico, p. 43,

„ intermedium, Marsh, u. Meyer, Mittheil. Zool. Mus. Dresden,

„ woodivardi, Sav. Kent, Monthly Micr. Jouru., .

Bcdanites jnpetta, n. gen. et sp., this Report, p. 122,
Batliydorus baculifer, n. gen. et sp., this Report, p. 154,
,, fimhriatus, n. gen. et sp., this Report, p. 151,

,, spinosus, n. gen. et sp., tliis Report, p. 153,

„ stellatus, n. gen. et sp., this Report, p. 152,

Culiptera, Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. ix. p. 450,
Callisphsra grayi, Sav. Kent (sp.), Ann. and IMag. Nat. Hist., ser. 4, vol. vi. p. 311
Carteria japonica. Gray, Proc. Zool. Soc. Lond., 1867, p. 540,
Caulocalyx tener, n. gen. et sp., this Report, p. 172, .
Cauloplwtcus elegans, n. gen. et sp., this Report, p. 126,
- „ latus, n. gen. et sp., this Report, p. 124,

Chonelasma calyx, n. gen. et sp., this Report, p. 326,

„ dcederleinii, n. gen. et sp., this Eeport, p. 324, .

„ hamatum, n. gen. et sp., this Report, p. 323,

„ lamella, n. gen. et sp., this Report, p. 321,

Corbitella speciosa. Gray, Proc. Zool. Soc. Lond., p. 530,

Crateromorpha meyeri. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. x. p. 110

„ murrayi, n. sp., this Report, p. 1G4, .

„ thierfelderi, n. sp., this Report, p. 164,

„ tumida,-a. sp., this Report, p. 166, .

Cyathella lutea, 0. Schmidt, Spong. Meerb. v. iMexico, p. 46,

1887
1887
1887

1869
1862
1862
1830
1862
1833
1867
1858
1870
1887
1887
1880
1870
1887
1887
1887
1887
1869
1877
1877
1880
1877
1870
1887
1887
1887
1887
1887
1872
1870
1867
1887
1887
1887
1887
1887
1887
1887
1867
1872
1887
1887
1887
1880

REPORT ON THE HEXACTINELLIDA.

481

Cyrtaulon sigsbeei, 0. Schmidt, n. gen., Spong. Meerb. v. Mexico, p. 58,

,, solutus, n. gen. et sp., this Ecport, p. 332, .
Oystispongia supersfes, 0. Schmidt, Spong. INIeerb. v. Jlcxico,
Dadylocalyx azorica, Wyv. Thomson, PhQ. Trans., vol. clix. p. 701,
,, hoicerhanldi, Johnson, Proc. Zool. Soc. Lond., p. 257, .

„ callocyathus, Wyv. Thomson, Phil. Trans., vol. clix. p. 701,

,, crispus, 0. Schmidt., Spong. Atlant. Gebiet.,

„ lieteroformis, Bowerbank, Proc. ZooL Soc. Lend., 1869, p. 69,

„ ingalli. Gray, Ann. and JIag. Nat. Hist., ser. 4, vol. i.,

„ macandreicii, Bowerbank, Proc. Zool. Soc. Lond., 1869, p. 69,

,, masoni, Bowerbank, Proc. ZooL Soc. Lond., 1869, p. 69,

„ patella, n. sp., this Report, p. 350,

„ polydiscus, Bowerbank, Proc. Zool. Soc. Lond., 1869, p. 96,

,, potatorum, O. Schmidt, Spong. Meerb. v. Mexico,

„ prattii, Bowerbank, Phil. Trans., vol. clii.,

„ pujniceus, Stutchbury, Proc. Zool. Soc. Lond., 1841, p. 86,

„ stutchhuryi, Sollas, Journ. Roy. Micr. Soc, vol. ii. p. 22,

„ subglobosus. Gray, Proc. Zool. Soc. Lond., 1867, p. 492,

„ tm-va, "Wyv. Thomson, Phil. Trans., vol. clix. p. 701, .

Deanea favoides, Bowerbank, Proc. Zool. Soc. Lond., 1876, p. 535, .
„ virgultosa, Bowerbank, Proc. Zool. Soc. Lond., 1875, p. 272,
Dendrosjpongia steerii, Murie, Trans. Linn. Soc. Lond. (Zool.), ser. 2, vol. i. pt. 5, p,
Diarefula eoi'tm, 0. Schmidt, Spong. Meerb. v. Mexico,

,, murefta, 0. Schmidt, Spong. Meerb. v. Mexico,
Dictyocalyx gracilis, n. gen. et. sp., this Report, p. 105,
Diplacodium mixtum, O. Schmidt, Spong. Meerb. v. Mexico,
Evdictyurti elegans, Marshall, Zeitschr. f. wiss. ZooL, 1875, Bd. xxv. p. 211,
Eiiplectella aspergillum, Owen, Proc. Zool. Soc. Lond., vol. ix. p. 3, .
„ corbicula, Claus, Grundziige d. Zool., 4 Aufi. p. 220,

,, crassistellata, n. sp., this Report, p. 81, .

,, cucumer, Owen, Trans. Linn. Soc. Lond., xxii. 2, p. 17, .

„ jovis, O. Schmidt, Spong. !Meerb. v. Mexico,

„ mammillaris, Bowerbank, Rep. Thirtieth Meet. Brit. Assoc, p. '2

,, (?) nodosa, n. sp., this Report, p. 82, .

„ owenii, Herklots and Marshall, Arch. NeerL, iii. p. 435, .

,, robusta, Bowerbank, Rep. Thirtieth Meet. Brit. Assoc, p. 217,

,, speciosa. Gray, Ann. and Mag. Nat. Hist., ser. 3, vol. xviii. p. 48

„ suberea, Wyv. Thomson, Nature, xiv. p. 93, fig. 1,

Eurete bowerbankii, n. sp. , this Report, p. 297, ...

carteri, n. sp. this Report, p. 296, ....
farreopsis, Carter, Ann. and Mag. Nat. Hist., ser 4, voL six. p. 121,
marshaUi, n. sp., this Report, p. 297,

schmidtii, n. sp., this Report, p. 293, ....
schultzei, Claus, Grundziige d. Zool., 4 Aufl. p. 220, .
semperi, n. sp., this Report, p. 292, ....
simplicissima, Semper, Verhandl. Wlirzb. GeseUsch., Bd. i. p. 29,
Euryplegma auriculare, n. gen. et sp., this Report, p. 176, .
Farrea aeuhata, Bowerbank, Proc. ZooL Soc. Lond., 1875, p. 272, .
„ clarigera, n. sp., this Report, p. 287,
„ deanea, O. Schmidt, Spong. Meerb. v. Mexico, p. 45,

(zool. CHALL. EXP. — PART LIII. — 1887.)

17,

I,

219,

1880

1887

1880

1869

1863

1869

1870

1869

1868

1869

1869

1887

1869

1880

1862

1841

1879

1867

1869

1876

1875

1877

1880

1880

1887

1880

1875

1841

1880

1887

1857

1880

1861

1887

1868

1861

1866

1876

1887

1887

1877

1887

1887

1880

1887

1868

1887

1875

1887

1880

61

482

THE VOYAGE OF H.M.S. CHALLENGER.

Farrea detisa, Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. xii. p. 349,
facunda, O. Schmidt, Spong. Atlant. Gebiet.,
farreopsis, O. Sclimidt, Spong. Meerb. v. Mexico, p. 44,
fistulata, Bowerbank, Proc. Zool. Soo. Lend., 1875, p. 276, .
gassioti, Bowerbank, Proc. Zool. Soc. Lond., 1875, p. 272,
inermis, Bowerbank, Proc. Zool. Soc. Lond., 1876,
infumlihidaris, Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. xii.,
infundihuUfonnis, Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. xii
irregularis, Bowerbank, Proc. Zool. Soc. Lond., 1876,
Ixvis, Bowerbank, Proc. Zool. Soc. Lond., 1875, p. 278,
occa, Bowerbank, Phil. Trans., vol. clii.,
parasitica, Bowerbank, Proc. Zool. Soo. Lond., 1875, p. 279,
perarmata, Bowerbank, Proc. Zool. Soc. Lond., 1876,
pocillum, Bowerbank, Proc. Zool. Soc. Loud., 1875, p. 273,
rohusta, Bowerbank, Proc. Zool. Soc. Lend., 1875, p. 562,
soUasii, n. sp., this Eeport, p. 286, .

spinifera, Bowerbank, Proc. Zool. Soc. Lond., 1875, p. 509, .
spinosissima, Bowerbank, Proc. Zool. Soc. Lond., 1875, p. 508,
spinulenta, Bowerbank, Proc. Zool. Soc. Lond., 1875, p. 279,
iulndata, S. Kent, Monthly Micr. Journ., vol. iv.,
valida, Bowerbank, Proc. Zool. Soc. Lond., 1875, p. 507,
vosmaeri, n. sp., this Report, p. 286, .
woodwardi, Sav. Kent, this Report, p. 477, .
Fieldingia lagettoides, Sav. Kent, Ann. and Mag. Nat. Hist., ser. 4, vol. vi. p. 219,
Habrodidywn corhicula, Wyv. Thomson, Ann. and Mag. Nat. Hist., ser. 4, vol. i. p. 114,

,, speciosum (Quoy and Gaimard, sp.),"Wyv. Thomson, Ann. and Mag. Nat. Hist,

vol. i p. 114,
Herlwigia falcifera, 0. Schmidt, Spong. Meerb. v. Mexico, p. 62,
Heterotella corhicula, Gray, Proc. Zool. Soc. Lend., p. 531, .
Hexactinella lata, n. sp., this Report, p. 327,

„ tubulosa, n. sp., this Report, p. 328,

„ ventilahrum, Carter, Ann. and Mag. Nat. Hist., ser. 5, vol. xv. p. 397

Holascus jibulatus, n. gen. et sp., this Report, p. 87, .
„ polejaevii, n. gen. et sp., this Report, p. 89,
„ ridleyi, n. gen. et sp., this Report, p. 90,
„ stellatus, n. gen. et sp., this Eeport, p. 86, .
Holtenia carpenteri, Wyv. Thomson, Phil. Trans., vol. clix. p. 701, .

,, grayi, Marshall, Zeitschr. f. wiss. Zool., Bd. xxv., Suppl., p. 142,
,, pimrtalesii, 0. Schmidt, Spong. Atlant. Gebiet., p. 14,
,, saccus, 0. Schmidt, Spong. Atlant. Gebiet., p. 15, .
Hyalocaulus simplex, Marshall and Meyer, Mittheil. Zool. Mus. Dresden, p. 264
Hyalocheda possieti, Brandt, Bull. Acad. St. Petersb., vol. xvi. p. 65,
Hyalonema ajjine, Brandt, Symb. ad polyp Hyaloch. spect., .

,, Marshall, Zeitschr. f. wiss. Zool., Bd. xxv., Suppl., p. 142,
anomalum, Bowerbank, Proc. Zool. Soc. Lond., 1877, p. 456,
apertum, n. sp., this Report, p. 214,

arcticum, Armauer Hansen, Norske Nordhavs Exp., xiii., Spong., p. 19,
cebuense, Higgin, Ann. and Mag. Nat. Hist., ser. 4, vol. xv. p. 377,
clavigerum, n. sp., this Report, p. 220, .
conus, n. sp., this Eeport, p. 209,

, ser. 4,

1873
1870
1880
1875
1875
1876
1873
1873
1876
1875
1862
1875
1876
1875
1875
1887
1875
1875
1875
1870
1875
1887
1887
1870
1868

1868
'1880
1867
1887
1887
1885
1887
1887
1887
1887
1869
1875
1870
1870
1877
1857
1859
1875
1877
1887
1885
1875
1887
1887

KEPORT ON THE HEXACTINELLIDA.

483

Hijalonetna depressum, n. sp., this Report, p. 217,

„ divergens, n. sp., this Report, p. 199,

,, elegans, n. sp., this Report, p. 223,

„ globus, n. sp., this Report, p. 221,

,, graeile, n. sp., this Report, p. 196,

„ kentii, 0. Schmidt, Spong. Meerb. v. Mexico, p. 65, and this Report, p.

„ loveni, Wyv. Tliomson, Phil, Trans., vol. clix. p. 701,

,, ludtaaiaum, Barb, du Boc, Proc. Zool. Soc. Lond., 1864, p. 265,

„ mirahile, Gray, Proc. Zool. Soc. Lond., 1857, p. 278,

,, parvuni, Sars, On some Remarkable Forms from Great Depths, .

„ poculum, n. sp., this Report, p. 208, ....

„ robusttim, n. sp., this Report, p. 229, ....

,, scliidtzei, Semper, Verhandl. Gesellsch. Wiirzb., Bd. i. p. 29,

,, sieholdii, Gray, Synops. Cont. Brit. Mus., p. 77, 1834; Proc. Zool. Soc,
p. 63,

„ tenerum, n. sp., this Report, p. 224,

,, tenue, n. sp., this Report, p. 228,

„ thomsoni, Marshall, Zeitschr. f. wiss. Zool., Bd. xxv. Suppl.,

,, toxeres, "Wyv. Thom.son, Nature, vol. viii. p. 247, fig. 34, 1873; Atlantic,

Hyalostylus dives, n. gen. et sp., this Report, p. 110,
Hyahithauma hidekingii, Herklots and j\Iarshall, Arch. Neerl., iii. p. 435,
Hyalothrijc lusitanica. Gray, Proc. Zool. Soc. Lond., 1867, p. 117, .
Iphiteon heatrix, Bowerbank, Proc. Zool. Soc. Lond., p. 66, .

„ callocyathes, Bowerbank, Proc. Zool. Soc. Lond., p. 66,

„ ingcdli, Bowerbank, Proc. Zool. Soc. Lond., p. 66, .

,, paniceum, Bowerbank, Phil. Trans., vol. clii. p. 147,

„ pumicea, Bowerbank, Phil. Trans., vol. cxlviii. p. 279,
- „ subglobosa, Bowerbank, Proc. Zool. Soc. Lond., 1869, p. 66,
Joanella compressa, 0. Schmidt, Spong. Meerb. v. Mexico, .
Labaria heniisphxrica, Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. xi. p. 235,
Lanuginella pupa, O. Schmidt, Spong. Atlant. Gebiet., p. 13,
Lefroyella decora, Wyv. Thomson, Atlantic, voL i. p. 403.
Leiobolidium, 0. Schmidt, Spong. Meerb. v. Mexico, p. 65, .
Malacosaccus unguiculatii.% n. gen. et sp., this Report, p. 93, .

,, vastus, n. gen. et sp., this Report, p. 91,

Margaritella coslojitychioides, 0. Schmidt, Spong. Meerb. v. Mexico, p. 54, .
McyereUa davifonnis. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. x. p. 76, .
Meyerina damformis. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. x. p. 134,
Myliuda caJlocyathus, Gray, Proc. Zool. Soc. Lond., 1859, p. 437,

,, grayi, Bowerbank, Proc. Zool. Soc. Lond., 1869, p. 66,

„ hassleri, O. Schmidt, Spong. Meerb. v. Mexico, p. 52,

„ zittelii, Marshall and Meyer, Mittheil. d. Mus. Dresd.,
Padiaulidium, 0. Schmidt, Spong. Meerb. v. Mexico, p. 59,
Parafieldingia sonalis, Vaillant, Comptes rendu?, Acad. Paris, xciii. p. 931, .
Peripliragella elisx, Marshall, Zeitschr. f. wiss. Zool., Bd. xxv., Suppl.,
Pheronema annx, Leidy, Proc. Acad. Nat. Sci. Philad., 1869, p. 9, .

„ carpenter i, Sav. Kent, Ann. and Mag. Nat. Hist., ser. 4, vol. vi. p. 182

,, giganteum, n. sp., this Report, p. 250, ....

,, globosum, n. sp., this Report, p. 248, ....
,, gi'ay'h Sav. Kent, Ann. and Mag. Nat. Hist., ser. 4, vol. vi. p. 182,

207,

Lond..

vol. iii.

vol. i.

p. 273,

1887

1S87
1887
188J.
1887
1880
1869
1864
1857
1872
1887
1887
1868

1835
1887
1887
1875
1877
1887
1868
1867
1869
1869
1869
1862
1858
1869
1880
1873
1870
1877
1880
1887
1887
1880
1872
1872
1859
1869
1880
1877
1880
1881
1875
1869
1870
1887
1887
1870

484

THE VOYAGE OF H.M.S. CHALLENGER.

p. 234,
342,

Pheronema hemisj)hsericum, Graj-, Ann. and Mag. Nat. Hist., ser. 5, vol. xi. p. 23.5,

,, parfaiti, Filhol, La vie au fond des mers, p. 283,

„ velatum, Carpenter, Not. Proc. Roy. Inst. Great Brit., vi. p. 236,

PoKopogon amadou, Wyv. Thomson, Nature, vol. viii. p. 29, fig. 1, 1873 ; Atlantic,

„ oif/as, n. sp., this Report, p. 2.57,

Polylophus philippineTisis, n. gen. et sp., this Report, p. 133,
Polyrhabdus oviformis, n. gen. et sp., this Report, p. 121,
Psetalia globulosa, Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. xi
Purisiplionia clarlcei, Bowerhank, Proc. Zool. Soc. Lend., 1869, p
Regadrella phcenix, 0. Schmidt, Spong. Meerb. v. Mexico, p. 61,
Rliahdocalyjitiis mollis, n. gen. et sp., this Report, p. 155,

„ roeperi, n. gen. et sp., this Report, p. 158, .

Rhabdodictyum delicatum, O. Schmidt, Spong. Meerb. v. Mexico, p. 46,
Rhabdopedella tintinnus, 0. Schmidt, Spong. Meerb. v. Mexico, p. 62,
Rhabdostauridium retortula, O. Schmidt, Spong. Meerb. v. Mexico, p. 59,
Rossella antardica, Carter, Ann. and Mag. Nat. Hist., ser. 4, vol. ix. p. 414,
,, philippinensis. Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. x. p. 134,
„ velata, Wyv. Thomson, Depths of the Sea, p. 419, .
Sderoplegma contemn, 0. Schmidt, Spong. Meerb. v. Mexico, p. 57,
„ heixuleum, O. Schmidt, Spong. Meerb. v. Mexico, p. 57,

„ lanterna, 0. Schmidt, Spong. Meerb. v. Mexico, p. 56, .

„ seriatum, 0. Schmidt, Spong. Meerb. v. Mexico, p. 57, .

Sderothamnm dausii, Marshall, Zeitschr. f. wiss. Zool., Bd. xxv., Suppl., p. 171,
Semperella sdiidtzei, Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. ii. p. 376,
Spongia odancyrx, Brandt, Symbolae ad pol. Hyal., .
„ spinicrux, Brandt, Symbolse ad pol. Hyal., .
Styloealyx apertus, n. subgen. et sp., this Report, p. 214,

„ davigei; n. subgen. et sp., this Report, p. 220,

,, depressus, n. subgen. et sp., this Report, p. 217, .

„ degans, n. subgen. et sp., this Report, p. 223,

,, globus, n. subgen. et sp., this Report, p. 221,

,, thomsoni, Marshall, sp., n. subgen., Marshall, Zeitschr. f. wiss. Zool., Bd,

„ tener, n. subgen. et sp., this Report, p. 224,
Sympagella nux, 0. Schmidt, Spong. Atlant. Gebiet., p. 15, .
Syringidium zittelii, 0. Schmidt, Spong. Meerb. v. Mexico, p. 46
Txgeria puldira, n. gen. et sp., this Report, p. 94,
Tradiycmdus gurlitlii, n. gen. et sp., this Report, p. 128,
Tridiaptella elegans, Filhol, La vie au fond des mers, pi. viii.,
Vazella pourtalesii (0. Schmidt, sp.), Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. vi

,, saccus (0. Schmidt, sp.). Gray, Ann. and Mag. Nat. Hist., ser. 4, vol. vi. p,
Volvidina sigsbei, O. Schmidt, Spong. Meerb. v. Mexico, p. 58
Walteria flemmingii, n. gen. et sp., this Report, p. 96,

vol. i. p

p. 309^
309,

175,

Suppl.,

1873
1885
1871
1877
1887
1887
1887
1873
1869
1880
1887
1887
1880
1880
1880
1872
1872
1873
1880
1880
1880
1880
1875
1868
1859
1859
1887
1887
1887
1887
1887
1875
1887
1870
1880
1886
1887
1885
1870
1870
1880
1887

THE PHYLOGENY OF THE HEXACTINELLIDA.

After a detailed investigation of a group of animals, it is incumbent on every
naturalist who accepts the evolution theory to attempt the appreciation of his results
in theii- relation to the phylogeny of the group.

In attempting to draw conclusions from the results of my investigation of living
forms, I am indeed conscious that such deductions as to pedigree cannot claim to have
any absolute certainty, and that the less since, in spite of the splendid work of Zittel,
the anatomical and embryological facts can be but slightly supplemented by the results of
palseontological research. Certain skeletal portions of fossil Hexactinellida, and especially
the connected framework, have indeed sometimes been found wonderfully preserved, and,
after separation from adjacent material by careful maceration in acids, have even
admitted of very intimate study. But the siliceous spicules occurring loosely in the
parenchyma, and specially important for diagnosis, are as a rule not preserved at all.
Of the soft parts there is furthermore no trace ; and, finally, the fossil remains that have
been found represent only some small divisions of the entire class. I must therefore
restrict myself simply to collating the conclusions drawn from the living foi'ms wdth
the few results of pal?eontological research, to show at least to what extent they
harmonise.

In the discussion of the phylogenetic relations of living Hexactinellids to one
another and to known fossil forms, I shall have repeatedly to refer to the papers which
I have in the past year laid before the Berlin Academy.^ There can be no doubt,
to any evolutionist, that the close resemblance, both in general structure and
details of organisation, which is expressed in the grouping of difierent forms in
the systematic unities of species or even genera must be based upon, and explained
in terms of real relationship. But the greater the extent of the systematic cate-
gories, the greater usually are the gaps which are seen to exist in the living fauna,
and the more difficult does it become to determine the actual relationships. The con-
ventional method of representing the various forms in a continuous series is felt to be
much more insufficient when dealing with the larger divisions than when expressing the
relations of the usually simpler varieties of a species or of the members of a genus.

' Abhandl. d. Kbnvjl. Preuss. Akad., 1886.

(ZOOL. CIIALL. EXP. — PART LIII. — 1887.) Ggg 62

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

' This becomes at once very evident when we consider the two great subdivisions of
Lyssaciua and Dictyonina, which, according to most mo'dern investigators of sponges,
and in my opinion also, may be recognised in the Hexactinellida. The question at
once arises whether these two divisions, regarded in classification as of approximately
equal importance, are to be expressed by the forking of a common stem, or are not
rather to be considered as the continuous but successive divisions of one ramified tree.
In the first case we must suppose that the two divisions with divergent forms originated
almost or quite contemporaneously from a common ancestral form, and that each
developing by itself in a special direction gradually exhibited the distinctive characters
of the modern forms. In the second case we have to suppose that the ancestors of the
hio-her division must at first have had the characters of the lower, and have been
systematically included within it, — we must suppose, that is to say, that the one
division sprang from the other.

But before I proceed to the discussion of this and similar questions, I shall briefly
review the relative opinions of previous investigators.

In his researches on Hexactinellida,^ published in 1875, Marshall says: — " The state
of skeletal coalescence " (in which the axial canals of the framework beams are said to
form one connected anastomosing system) " I regard as phylogenetically oldest, as
that from which have developed the Hexactinellida with free siliceous elements, and
especially those with predominantly hexradiate spicules, which may be regarded as
simply inherited. By adaptation we account for the large series of frequently very
beautiful forms, for which Bowerbank has invented such an elaborate nomenclature.
The third state, that of fusion, appears to arise in diS'erent ways : ( 1 ) by the simple
union of the ensheathing substance of two adjacent spicules ; (2) by lamella-like
structures which extend like bridges between two adjacent but not directly apposed
spicules ; (3) by the development of lamellar layers of flinty material round two
parallel and adjacent spicules."

In the memoir entitled Ideen fiber die Verwandtschaftsverhaltnisse der Hexact-
inelliden,- Marshall has more definitely expressed his conclusions as to the phylogeny
of the Hexactinellida. Starting from a Chalynthus-like ancestral form without skeleton,
he regards it as probable that in the w^all of this simple sack somewhat firmer longi-
tudinal, circular, and radial strands of " hardened protoplasmic material " were developed,
intersecting at right angles, and forming a connected fibrous framework with square or
cubical meshes. "The next form," Marshall continues {loc. cit, p. 119), "is thus a
sponge with simple connected siliceous lattice work, in which the central canals are
also connected, and which has not yet acquired any functionally important free spicules.
From such a simple Protohexactinellid there have developed, on the one hand, forms
like Sclerothamniis, with single free spicules, and, on the other hand, forms in which

1 Zeitschr. f. iviss. Zool., Bd. xxv., Supplement. - O/j. cit., Bd. xxvii. p. Ill

REPORT ON THE HEXACTINELLIDA. 487

the connectiou of the axial (■aiials no longer occurs, though secondary development still
results in the formation of a connected skeleton, and in which there are no special
needles developed besides the regular hexacts. On this phylogenetic theory, Eurete "
(which according to Marshall is entirely without isolated siliceous elements) " is of im-
portance, while an ontogenetic recapitulation of the development is illustrated ])y the
extremely simple embryo of Hyalonema, wdiich I (Marshall) have described."

From the still extant genus Sclerothamnus, representing the ancestral form,
Marshall derives the group of Synauloidse, in which " the lumina of the spicules, as well
as the spicules themselves, are continuously connected with one another, so that the
whole lattice-work of the sponge is penetrated by a uniformly connected system of tubes."
The other modern Hexactinellida are united by Marshall into the group Asynauloidte,
" in which the lumina of the stalks of the various spicules are never connected, but each
spicule, so far as its central filaments are concerned, forms an independently developed
individual. Where a lattice-work is developed, that is exclusively the result of a
syncytium formed by the secretion of layers of siliceous material."

We thus see that Marshall regards the presence of a special form of dictyonal
framework as the oldest type of Hexactinellid skeleton, from which the forms described
as Lyssacina (with isolated hexradiate spicules) have afterwards developed. In many
of the latter a secondary union of the hexradiate spicules again results in the formation
of a connected siliceous framework (the dictyonal framework of our Dictyonina), and at
first of this alone, as in Eurete, but afterwards also with the association of adjacent
isolated spicules. In other derived forms the secondary fusion of the spicules has not
occurred, but the form of the isolated spicules has become more or less complicated.

In 1877, in his Studies on ossil Sponges,^ Zittel apjjlied the results of his investi-
gation of abundant palseontological material to the elucidation of the system and
phylogeny of the Hexactinellida. His conclusions agree in several points with those of
Marshall, but difier not inconsiderably in others. "Were it true," Zittel says [loc. cit.,
p. 19), "that the compact Hexactinellid skeleton was developed from a special frame-
work of delicate protoplasmic strands, then the older fossil forms must necessarily, on
Marshall's theory, belong to the Synauloidfe. This is not, however, by any means the
case. My researches on fossil forms have shown that the connected lattice-works
consist without exception of fused hexradiate spicules, in which the axial canals are indeed
frequently apposed and seem to form closed and connected tubes, but are in reality
always separate, while they usually lie, as in the living genera Farrea, Eurete, and
Aphrocallistes so that the axial filaments of the different hexacts are. seen to be
distinctly separate."

After Zittel had shown that, even in Sclerothamnus, the axial canals of the lattice
framework do not form an open anastomosis, but are referable throughout to single

' Ahhandl. d. 11. CI. k. baier. Alxid. d. U'iss., BcL xiii. Abth. 1.

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

fused hexacts with blind processes lying side by side, he was led altogether to deny the
accuracy of Marshall's group of Synauloidse. He also threw doubt on the correctness
of Marshall's representation of Eurete, Semper, as a Monacid. On this account Zittel
divides all the Hexactinellida into two main divisions, — Lyssacina, in which " the skeletal
spicules generally remain isolated, and only united by protoplasm" and Dictyonina, in
which " the skeletal spicules are fused in a regular way, and form a connected lattice-
work with cubical or polyhedral meshes. " The soldering of the spicules into a compact
connected framework, as occurs in some Lyssacina such as Euplectella aspergillum,
cannot be identified in mode or nature of union with the regular fusion of a distinct
dictyonal skeleton. The irregularity of arrangement, and the inhibition of further
development of the spicules as the consequence of this external union, demonstrate the
secondary importance of the former case.

For the further classification of the Lyssacina, Zittel emphasises, like Marshall, the
degree of differentiation exhibited by the loose spicules, and forms three families : — (1)
Monacidse (Zittel), with only a single form with loose skeletal elements, including also
the few known fossil Lyssacina (like Astrseospoyigia, Eom, and Stcmractinella, Zittel);
(2) Pleionacidas, Marshall, in which the main mass of the skeleton consists of regular
hexacts, in association with broom-forks and rosettes {Asconema, S. Kent, and
Lanuginella, 0. Schmidt); and (3) PoUacidaj, Marshall, in which the form of the skeletal
and loose spicules is very manifold, especially in the dermal skeleton and in the lining
of the enteric cavities, while the base usually bears a root-tuft of long siliceous spicules
(numerous living and some fossil forms). The Dictyonina, on the other hand, are
divided by Zittel into a large number of families with complicated characteristics.

In an essay by Marshall and Meyer ^ on some new or slightly known Philippine
Hexactinellids, Marshall withdraws his opinion as to the continuity of the axial canal
system in the framework of Sclerothamnus, and thus abandons the group of synauloid
Hexactinellids.

In 1880,^ 0. Schmidt discusses these systematic questions in detail. He says (p. 41)
that " the conception which one is apt to associate with the division of the Hexactinellids
into Dictyonina and Lyssacina, namely, that each group for itself has originated from one
or several common ancestral forms, and that all the Dictyonina, and especially the recent
forms, are more closely related to one another than to the Lyssacina, is certainly not
in accordance with the facts. The relation between these two divisions appears indeed
to be much closer ; the phylogenetic branches have probably repeatedly crossed, and
are interlaced by connecting twigs. In one of the new genera, Hertwigia, this insepar-
able relationship is expressed in the most convincing way, since this sponge at the
branched base is distinctly dictyonal, but forms further up, where it consists of irregular
tubes and plates, a transitional type, and finally, still further up and towards the exterior,

1 Mitth. a. d. k. zoolog. Mus. Dnsden, 1877. - Spongieu des MeerLusens von Mexico.

REPORT ON THE HEXACTINELLIDA. 489

exhibits a beautiful Lyssacine structure. The same transition is exhibited by Rhahdo-
dictyum and Rhabdostauridium."

" At any rate the Lyssacina were once the sole representatives of the Hexactinellida.
As a Lyssacine every Dictyonal form must still begin its development, though this stage
may only perhaps last for a very short time. And thus there was at any stage the
possibility of the Dictyoninal form becoming again reduced to a Lyssacine. The stifFest
and most brittle Dictyoninal framework differs after all only in degree from that of the
loosest of the Lyssacina."

My own researches have convinced me that there is indeed a certain antithesis
between Lyssacina and Dictyonina, which may be justly recognised in the distinction
of these two systematic divisions of the order Hexactinellida, but that this difference is
not fundamental, nor involving the separate origin of the two divisions, but that it is
rather one of degree and in no way suggesting a hard and fast separation.

In distinguishing the two groups, Zittel laid stress, however, not merely on the
simple fact whether the principal spicules were or were not united into a firm connected
framework, but rather on the manner in which this union was effected. It must be
allowed that what is regarded as the characteristic dictyonal mode of framework
formation, viz., by close apposition of the two corresponding arms of adjacent hexradiate
spicules, and the formation of a common sheath, is indeed very frequent, and in some
Dictyonina, like Aulocystis, or in the youngest portions of Farrea stocks is even normal
or perhaps constant, but it cannot be overlooked that it is extremely common to find that
the mode of union of the dictyonalia is in part, or here and there throughout, essentially
different. Zittel indeed eaUed attention to the fact that beside the hexradiate spicules
united as above indicated there were others " which left the series and had their rays
soldered arbitrarily to the rest of the framework. When one or two rays of such
irregularly disposed spicules become united by chance to the thickened centre of a
hexact, the result is obviously the apparent origin of more than six rays from one centre
of intersection. Other irregularities may also arise by the curvature or direction of
individual rays, so that two rays in one axis come to lie no longer in a straight line."
If these deviations, which may be readily detected in most Dictyonina, are really only
exceptions to the tj^pical mode of union which Zittel supposed to obtain, then the sharp
and certain separation of the Dictyonina from the Lyssacina with connected framework
cannot be said to be seriously affected. But there are, besides, not a few Hexactinellida in
which the supposed typical mode of union is either not to be detected at all, or only here
and there after careful search, and even then in a fashion so far from characteristic that
among the countless deviating and arbitrary intersections of the majority of the rays the
exceptional occurrence of the typical mode may be indeed regarded as a matter of
chance. Thus, for example, in the different species of the genus Aphrocallistes known to

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

me, the t5''pical union of the dictyonalia is wholly absent, although no one has any doubt
that AphrocaUistes, in the general structure of its skeleton, and especially in the
structure and disposition of its isolated spicules, belongs to the Dictyonina, and is to
be referred to the neighbourhood of Euretidse and Coscinoporidse. As an instance of the
second case, where a skeleton which is macroscopically like many dictyonal frameworks
contains only isolated examples of two hexradiate spicules uniting in the typical fashion,
I may refer to Euryplegma auriculare.

In regard to this form I have indeed hesitated for a long time whether I should refer
it to the Lyssacina or to the Dictyonina. At first, in my memoir Ueber den Bau und
das System der HexactineUiden,^ I regarded it as a Dictyonine, but have finally preferred
to place it among the Lyssacina, and beside the Rossellida^, which, both in regard to the
structure of the loose needles and the absence of uncinates, it resembles more closely
than it could any Dictyonine form. In its other characters it is, however, Dictyonine-
like, though the absence of uncinates and scopulse give it a peculiar appearance and must
always make its position exceptional.

It was a fact of much interest to me that my respected colleague Professor
Zittel unhesitatingly referred a macerated skeleton of Euryplegma, which I asked him to
examine in the Berlin Zoological Institute, to the Dictyoninal type, while allowing that
various microscopical sections of the same skeleton comj^letely resembled the Lyssacina
in the spicular union of the framework.

There are, indeed, other distinctive characters between Lyssacina and Dictyonina
than the mode of union of the spicules, but these also establish, not a fundamental
separation, but only a difierence of degree. Previous investigators have noted the readily
verified fact that in all emphatically Lyssacine types, which form a connected skeletal
framework, there occurs, besides the simple soldering of the spicule branches, another
very frequent mode of union by means of short connecting bridges, the so-called
sjmapticula, which bind together more or less approximated rays of adjacent spicules,
which are not, however, in actual contact. By the development of numerous synapticula
at approximately uniform intervals, there arises a scalariform structure, which is regarded
by some as characteristic of the Lyssacina with connected framework, and as contrasting
them with the Dictyonina. But while it is indisputable that such scalariform structures
occur with great frequency, indeed quite regularly in the framework of Lj^ssacina, it is
incorrect to assume that they are absent from all Dictyonina. I have observed them,
though not frequently, yet quite typically developed in indisputable Dictyonina, such as
Fieldingia lagettoides, S. Kent, which in its uncinates and scopulse is certainly one of the
Dictyonine series.

Another fact, to which I first called attention, is the early union of the dictyonalia
simultaneously with the development of the associated portion of the body, which is, so

1 Ahhandl. Kiiiiigl. Preuss. Akad., 1886.

EEPOPvT ON THE HEXACTINELLIDA. 491

far as we know, a contrast to the subsequent, and late union of the various spicules in
certain Lyssacina. In the latter it is deferred till after the differentiation of the body
has made considerable progress, and then only gradually develops from a given centre in
the already perfected body.

But although I regard this difference as most conveniently diagnostic of the two
groups, I must allow that it also represents only a qualitative difference. The dictyoninal
spicules of the Dictyonina were once, as Marshall first clearly demonstrated for
Aulocystis zittelii, free spicules, and the union gradually progresses from the younger
to the older portions of the body. We also know of many Lyssacina with connected
skeletal framework, like Ehahdodictyum delineatum, 0. Schmidt, Aulocalyx irregidaris,
F. E. S., and especially the above-mentioned Euryplegma auricidare, F. E. S., in regard
to which we are by no means certain, whether, as in species of Eu2olecteUa and some
other Lyssacina, the process of fusion or the formation of synapticula take place some time
after the differentiation of the sponge-body, or shortly after its incipient formation. In
regard to those Lyssacina in which an abundant fusion of spicules forms a compact base
attached to solid bodies, it is very probable that the fusion of the spicules begins somewhat
early at the lower end, and progresses gradually upwards to the upper margin. The upper
end of many Lyssacina with latticed framework has been long since carefully described, as
0. Schmidt has done for his Hertioigia falcifera and other Hexactinellids, and known to
consist solely of loose isolated spicules, while the lower end exhibits a well-developed fusion.

We thus see that the relation at present demonstrable between extant Lyssacina and
Dictyonina does not suggest a long-established separation of the two groups, but rather
a gradual modification of certain Lyssacine groups into Dictyonina, while others have
remained at the undoubtedly older Lyssacine stage.

We do not, indeed, know the development of the Dictyonina, but it is to be expected
that they possess at first only isolated spicules, and recapitulate probably for a short time
the Lyssacine stage, before a regular fusion of hexacts occurs, and the typical dictyonal
characteristics set in.

It appears to me further a very noteworthy fact that the results of my bathymetrical
statistics show that the Lyssacina occur predominantly in the greater depths and far
from the coast in Mid Ocean, while the Dictyonina are chiefly found in moderate depths
and near the coast. The Hexactinellids of the Challenger Expedition, which were
obtained from the greatest depths belong to the simplest and most typical Lyssacina,
as for example certain species of Holascus and Bathydorus, especially Bathydorus
Jimhriatus, which was dredged in Mid Pacific from a depth of 2900 fathoms.

Furthermore, in some species of certain Lyssacine families, such as Euplectellidfe and
Eossellidas, we find in the very irregular and long deferred fusion of the larger spicules
into a compact connected framework, what may be regarded as an incipient development
of a dictyonal skeleton.

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

The dictyonal character is, therefore, regarded as acquired, by some groups in the far
past, by others at a later stage, while many do not in any way exhibit it. Thus we may
explain that in former epochs Dictyouina and Lyssacina are found to have occurred
together as they now do.

If we now take a survey of the various main and side branches of the hypothetical
genealosical tree of the Hexactinellids, so far as that can be sketched out from the results
of living forms, we are at once brought to face a deep division, aflfecting both soft jiarts
and skeleton, between the Amphidiscophora or Hyalonematidge on the one hand, and all
the rest of the Hexasterida on the other.

While in the latter the membrana reticularis, wdiich is doubtless so important in
relation to the nutritive process, appears to form throughout appproximately equal thimble-
shaped chambers, longitudinally apposed to one another, in the Hyalonematidse it is more
or less irregular in its contour, and forms chambers not so sharply separated and without
any typical structure and of approximately equal size. It seem to me that this peculiar
condition of the membrana reticularis in the Hyalonematidse perhaps suggests a relatively
lower grade of differentiation, and is at any rate a not unimportant deviation in the
general structure, which otherwise closely resembles the other Hexactinellida. But the
Hyalonematidse are yet more distinctly separated from the others in the constant and
peculiar possession of the siliceous elements known as amphidiscs (or birotulae), as also
in the complete absence of hexasters which occur in all the other Hexactinellids. While
these facts point to a marked independence of the Hyalonematids, and to a distinct
separation from all other Hexactinellids, there are also certain other characters which
occur with great constancy and uniformity within the whole group, but less so in other
divisions. Thus we note the constant mode of attachment in the muddy bottom by
means of a basal tuft, and the way in which the entire outer surface is covered with
pinuli.

One cannot therefore but suppose an early separation and an independent develop-
ment of the Hyalonematidae or Amphidiscophora, as is represented in the genealogical
tree by the deep cleft separating this important and at present richly developed branch
from the other Hexactinellids.

Among the other Hexactinellids, which are without amphidiscs but contain hex-
asters, and may therefore be conveniently designated as Hexasterida, one group of
families distinguished by the presence of uncinates may be somewhat sharply separated
off from the others. These Uncinataria are all Dictyonina, and have apjaarently at an
early stage separated into two divergent branches, namely, on the one hand, the small,
but sharply defined family of Farreidse, distinguished by the single-layered structure of
their square-meshed lattice-work in its youngest growths, and also by the exclusive
possession of the remarkable clavulse in their limiting membranes ; and, on the other
hand, the Scopularia, which exhibit in their scopulse spicules so peculiar and charac-

REPORT ON THE HEXACTINELLIDA. 493

teristic that one can hardly find a more striking proof of close relationship. Of less
importance, though in detail often not without striking peculiarities, are the characters
which distinguish the four living families of Scopularia. Thus, for example, the family
of Melittionidae, though including only the single genus AphrocaUistes, with a few
species, is so sharply defined and removed fi'om the other families by the hexagonal,
prismatic, honeycomb-like radial perforations of its fiat dictyonal framework, that one
must assume a long, independent, ancestral series, that is to say, a somewhat early
separation of the twig from the common branch of the Scopularia. The characters of
the primitive portion of this branch appear to me to be represented in their least
modified form by the family of the Euretidae, where the structure is comparatively
simple and slightly ditferentiated, especially in regard to the aflferent and efferent canals
penetrating the body-wall. For while in the Euretidse the afferent and efferent canals,
which traverse the thin wall of the tubes forming the entire sponge, are quite short and
usually sack-shaped, of simple uncharacteristic form, the efferent canals of the Melittionidse
are straight, hexagonal prisms, those of the Coscinoporidee straight, narrow, and usually
long, alternating funnels penetrating the body-wall at right angles, and those of the
Tretodictyidae, finally, are of irregular course. And as to the other Hexasterida
which contain no uncinates, the family of the Masandrospongidaj, which flourished in
the Cretaceous period, is distinguished by the structure of the body, which consists of a
system of anastomosing, meandering, thin-walled tubes, with an interjacent system of
anastomosing canals.

But while these Mseandrospongidse have long since become true Dictyonina, the
other families of Hexasterophora, without uncinates, have either entirely preserved the
Lyssacine character, or they nearly approach the Dictyonine type by the more or less
marked soldering of the larger spicules into a connected framework, which develops
with increasing age. Of the three families — Euplectellidse, Asconematidse, and Rossel-
lidse, the first has a markedly separate position, due to the hexradiate character of the
spicules supporting the skin (hyjiodermalia), while the others exhibit only pentact hypo-
dermalia, with an internal radial ray. While in the Asconematidse autodermal piuuli
are also developed within the skin, giving to this Hexasterid family a very peculiar
character somewhat resembling that of the Hyalonematidse, in the Rossellidse such
autodermalia, with a freely projecting fir-tree-like ray, do not occur. It is difficult, if
not indeed impossible, to determine, with any certainty, the relationship of these three
allied families. It is obvious that the Euplectellidse, both in their relatively simple,
saccular or tubular form, and also in their hexradiate dermal sjjicules, have retained
primitive characters, so that one need not wonder that very simple related forms are
found at a comparatively early period. On the other hand, both in the parenchyma
of several species, such as Dictyoccdyx gracilis, and also in the extremities of
the dermal projections of many genera, e.g., Euplectella, Twgeria, Walteria, &c.,

(ZOOL. CHALL. EXP. — PART LIII. — 1887.) Ggg 63

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

very complicated, highly differentiated spicules (discohexasters, floricomes, &c.) occur,
and so many species incline towards fusion of the large spicvdes and formation of a
connected framework, that one cannot regard these forms at least as primitive.

A very simple structure is exhibited by certain tubular or saccular Eossellidse, as
for instance, Bathydorus Jimhriatus, from the great depths. On the other hand, there
are species, like Aulocalyx irregularis, which, both in the formation of a connected
supporting framework, and in the complicated structure of the isolated spicules, occupy
a decidedly higher grade. Such are the members of the entire division of Cratero-
morphinse, in which a more or less firm long stalk is formed, and the chamber-layer
exhibits a complicated folding. The Asconematidas exhibit a remarkable affinity ^\-ith
the otherwise indisputably far-removed Hyalonematidse, in the possession of autodermal
pinuli. Here there can hardly be any direct inheritance, either of one family from the
other, or of both from a common ancestor. I am rather of opinion that the tendency to
form lateral teeth is very generally distributed among Hexactinellid spicules, and is
expressed in development whenever such would be physiologically advantageous or
specially useful for the preser\-ation and strengthening of the organism, as for instance
in the formation of weapons of capture or defence on the autodermaha projecting from
the skin. I believe that the tendency may be expressed in far-removed Hexactinellids,
without there being any transmission by inheritance. And this opinion is supported by
the fact that, among the typical Scopularia, there are isolated cases in the various genera
and species, e.g., in ApJirocallistes and again in Chonelasnm dcederleinii (but not in
Clionelasma lamella), in which the distal radial ray of the hypodermalia, projecting more
or less beyond the skin, exhibits exactly similar teeth, and thus most distinctly evidences
the formation of a pinule.

In the following sketch of a genealogical tree, including those Hexactinellids which I
have investigated, the opinions above expressed are graphically represented.

When an attempt is made to collate the results of the investigation of li%Tng
Hexactinellids -n-ith what is known of fossil forms, an array of difficulties beset the task.
Above all it is unfortunate that of the great majority of fossil forms only the dictyonal
framework is known, so that there is no possibihty of induction as to the configuration of
the soft parts, nor as to the form, number and disposition of the isolated skeletal elements
which are so pre-eminently characteristic, and so extremely important, in spite of Zittel's
opinion, in determination of relationship. Again, the blanks in the geological record are
here even more serious than in most of the other groups. From entire geological
formations either no Hexactinellids are known, or only slight traces. Zittel indicates an
obvious explanation in the following sentences:^ — "Our knowledge of the fossil
Hexactinellids is limited to isolated remains of a developmental series, widely separated
(both in time and space) the links of which perhaps lie buried in the deposits now sunk

> Pateontologie, i. pp. 199, 200.

REPORT ON THE HEXACTINELLIDA.

495

Fig. S. — Genealogical Tree of the Hexactiuellida.

496 THE VOYAGE OF H.M.S. CHALLENGEr..

beneath the sea-level, or in still unexplored regions. Under these circumstances it is not
yet possible to construct a genealogical tree for the various genera, so we must dismiss to
a pre-Silurian time, where the light of experience no longer guides us, all the hypotheses
which would derive the Hexactinellida from the Tetractinellida or Monactincllida."

An explanation of the sudden appearance of the fossil Lithistida3 and Hexactinellida
may be reasonably sought in the mode of life exliibited by their modern representatives.
Both groups are especially inhabitants of the deep-sea, and only in the former deep-sea
deposits can one hope to find their fossil remains in any abundance.

If therefore we have to refer the origin of the Hexactinellida to a pre-Silurian period
of which no certain representatives are preserved, we cannot obviously expect from
palaeontology any conclusion as to the manner of their origin or the nature of the primitive
forms. It is, however, of importance, that even in the Silurian period both Lyssacine and
Dictyonine forms appear to be present. In the later mesozoic formations which are so
rich in Dictyonina (especially the Jurassic and Cretaceous), hardly a hint of Lyssacina is
to be found, while now they form among living forms a large majority — a fact which
appears to find its explanation in the improbability of the preservation of foi-ms with a
loose spicular framework which would fall to pieces with the destruction of the soft parts.
The following considerations may, however, appear not unworthy of close examination.

From the fact that the modern Hexactinellids all live at a considerable depth, always
below 95 fathoms, Zittel has already drawn the conclusion that they were probably from
the first deep-sea forms.

From the bathymetrical statistics of the Challenger Hexactinellida, it is seen that the
Hexactinellids now living in the greater oceanic depths are almost exclusively Lyssacina,
and that the modern Dictyonina are, with the exception of one very simple genus, all
confined to the relatively lesser depths, between 100 and 1500 fathoms. The conclusion,
. therefore, seems warranted, that in ancient times also the Lyssacina predominantly occurred
in the greater depths, while the more differentiated Dictyonina inhabited, as they now do,
relatively shallower waters, at no very great distance from the coasts. Now if one
may assume that the deej^est regions of the great oceans have remained permanently
covered with water since the Palteozoic period, while only the shallower regions near the
continents were here and there raised above water, and thus became now accessible to the
hammer of the palseontologist, we can understand why we find in certain Jurassic and
Cretaceous deposits so many and highly differentiated Dictyonina, but only slight hints
of Lyssacina, even in circumstances which would not exclude their preservation, or at least
that of their characteristic spicules.

Hitherto I have restricted myself to the Hexactinellida themselves ; I shall now consider
their relation to the other groups, beginning with the calcareous sponges.

It may be assumed as a generally acknowledged fact that tlie calcareous sponges are

REPORT ON THE HEXACTINELLIDA. 497

Jistiuguislied as a separate group, not ouly in the substance out of which the skeleton
consists, but also in the form of the component spicules.

One is therefore inclined to suppose their very early divergence from the great sponge
stem. On the other hand, the marked uniformity of their skeletal elements lead one to
suppose that they have had a common starting point, i.e., a monophyletic origin. This
supposition is confirmed by the certain fact that the Sycones in their ontogenetic
development pass through a distinct Ascon stage, and that between the Sycones and
Leucones recent investigations have discovered many connecting links, as has been shown
especially in the Challenger Eeport on Calcareous Sponges by Polejaeff, and von Lenden-
felcVs researches on Austi'alian forms.

A closer relationship connects the siliceous, horny and soft sponges. In regard to
the last I have previously shown, that both on account of the incongruity between the
different members of the group, and the obvious relationship between certain forms and
indisputable horny and flinty sponges, the group cannot be regarded as independent,
closed and natural, but must be split up and its members referred to different positions
on the genealogical tree near their various congeners, and regarded as twigs degenerate
in respect to their skeleton.

In regard to the horny sponges there seems to me no other supposition possible, but
that of regarding them as originating from flinty or flinty-horny sponges by the gradual
reduction and final disappearance of the siliceous spicules.

The more abundant and differentiated the horny substance the more degenerate the
flinty skeleton, until finally, as in many Chalinidee, which approach the true Keratosa,
we find only very simple smooth spindles, which I am compelled to regard as the
extreme of the phylogenetic series of siliceous s^aicule modification.

In my memoir on the family of Plakinidse, I have shown in detail why it is that in
"the long and continuous series of transitions between the typical regular tetracts and
the simple straight spindles, exhibited both in individual species and often in
one individual, as weU as in the skeletons of nearly related species, it is impossible
to regard the straight spindle as the primitive form from which the triacts and tetracts
have been formed by the growth of new rays, but necessary to regard the tetracts as
primitive and ancestral, from which the triacts and diacts have arisen by atrophy and
degeneration of the various rays. Oscar Schmidt^ was led to the same conclusion by a
detailed investigation of other Tetractinellida, and especially of the Ancorinidse, where
it may be very readily seen how gradual reduction of the typical tetracts, only modified
into anchors by the elongation of one ray, leads finally to simple rod-like spicules, and
further how within those genera, which, like Caminus, have acquired rod-like spicules,
degenerate anchors here and there persist, showing the mode in which the rods have,
originated from tetracts.

■ Entstehung iieuer Arten durch VerfoU uiid Scliwund alterer Merkniale, Zeitschr.f. wiss. ZooL, Ed. .\lii. p. 039.

498 THE VOYAGE O'F H.M.S. CHALLENGER.

I may therefore regard it as indubitable that in many cases the regular and typical
tetraets have, by gradual reduction of individual rays, given rise to diacts and even
monacts. I do not mean to say that all diacts and monacts have originated in tetraets.
On the contrarj^, previous investigators of the Hexactinellida have shown, what I think
I have also clearly demonstrated, that in this group at least the very abundant and
richly developed diacts and monacts have arisen, not from the regular tetraets of
the Tetraxonia (the so-called chevaux de frise), but from the regular hexacts of the
Triaxonia. But while there are numberless extant transitions from the Tetraxonia with
typical tetraets to the strict Monaxonia with only straight diacts or monacts, there are
among living, and, so far as I know, among fossil sponges no transitions from the
Triaxonia to strict Monaxonia, so that we have no reason for the supposition that the
latter have been evolved from the former. The case is somewhat different in regard to
some sponges without a skeleton, lately discovered in Australia by von Lendenfeld (such
as Bajalus') in which the structure agrees so closely with the soft parts of several
Hexactinellids that one is inclined to suppose their derivation from the latter by a total
loss of siliceous spicules.

Under these circumstances, the supposition is legitimate, that all the Monaxonia, and
the Keratosa which have probably developed from them, have originated from the stem
of the Tetraxonia. And since the spicules of the Lithistidae, as 0. Schmidt, Zittel, and
others have conclusively shown, are derivable from the regular tetract type, we may thus
regard the Tetraxonia with simple regular tetraets as the starting point for all the flinty
and horny sponges except the Hexactinellida.

For the possibility that the Hexactinellida also stand in genetic relationship with the
Tetraxonia, I find no basis of facts. In 1870 0. Schmidt^ expressed the same opinion in
the following words : — " Between the type of spicule in which the rays are determined
by the three-sided pyramid and the triaxonial type, there are, so far as the forms go, no
relations. The sponges in which these two types occur appear to be distinct and separate
independent twigs, in regard to which one must distinguish clearly between the general
homologies and the adaptive analogies."

We thus reach this conclusion, that the sponges may be grouped along three main
stems, which may indeed be regarded as springing from a common root — a very simple
primitive sponge without a skeleton — but which remain separate from this root onwards,
without exhibiting any connecting links.

This may be expressed in the form of a genealogical tree (see fig. 9).

Now if we may regard it as probable that each of these three main stems, which

represent the divisions of the great crowd of sponges, namely (1) Calcarea, (2) Tetraxonia

« with Monaxonia and Keratosa, and (3) Triaxonia, possessed to start with, either exclusively

' Proc. Linn. Soc. N.S. IF., vol. s. p. 5.

^ Grundziige einer Spotigienfauna des atlantischen Gebietes, 1870, p. 5.

REPORT ON THE HEXACTINELLIDA.

499

Cera.tosa.

MoJUKeotiia.

Tetmetine^t/a,

LitA/^fiWft.

or predominantly, that form of skeletal element from which the rest have developed, we
must, after determining this fundamental form, strive to understand why in each
individual stem the specific form of spicule must be what it is.

In regard to the calcareous sponges, Haeckel expresses, in his epoch-making monograph,
the conviction, based both ou anatomical and em-
bryological facts, that the fundamental and primi-
tive form of all calcareous sponge spicules is the
regular triradiate form, and in association with that,
perhaps, though less widely distributed, the simple
rod-like form. But by " regular triradiate "
Haeckel means a three-rayed spicule, in which the
rays, perfectly similar in size and form, meet at
equal angles of 120°, and all lie in one plane.
Haeckel ^ shows that even yet the triradiate spicules
are throughout the firm supports of the body, while
the tetracts occur as the protective weapons of the
gastral surface, or the rods of the dermal ; and he
draws the conclusion " that the triradiate forms
originally and primarily played the principal role,
while the four-rayed spicules first arose only as
internal adaptive modifications on the gastral sur-
face, and the rod-like forms as external and therefore
secondary diS"erentiations."

As of special importance in regard to the forma-
tion of this specific form of spicule (triradiate) in
the soft parts of the primitive calcareous sponges,
Haeckel notes (p. 377) a peculiar process of
" bioerystallisation, i.e., a compromise of the
crystallising tendency of carbonate of lime and
the organic processes of the protoplasm." The

calcareous spicules of the Calcispongias are, according to Haeckel, to be regarded as
" biocrystals or form -individuals occupying a median position between an inorganic crystal
and an organic secretion, and in their origin expressing a compromise between the
crystallisation of carbonate of lime and the formative activity of the fused cells of the
syncytium." " The original and fundamental form of all three-rayed and four-rayed
spicules is the absolutely regular triact, which may be considered as a hemiaxonial
form of the hexagonal crystal system, in which the carbonate of lime crystallises as
calc-spar."

' Lot. cif., Bd. i. p. 352.

CaCcaTva

ITexaaUnellida .

Fig. 9. — Genealogical Tree of the Porifera.

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

0. Schmidt has also occupied himself with the notion of explaining the form of the
typical sponge spicules in terms of the crystallising tendencies of the different substances.
He says : " The question is a difficult one, how far the nature of the lime and the flint
determines the rayed and anchor-shaped forms related to the three-sided prism. For
the flint the explanation holds good, but not for the lime. But while in the group
of triaxial siliceous spicules (hexacts and their derivatives) we have only to choose
between the triaxial and the binaxial and monaxial system, and the hexagonal does
not come into account, we must remember the fact, inconvenient to mineralogists, that
quartz crystals arising in amorphous matrix not unfrequently exhibit axial deviations from
the hexagonal system, and that we may the more readily expect in our sponge spicules
with organic basis and admixture other forms than those of the crystallographic systems." *

For my own part I must pronounce against such an attempt to bring the form of the
sponge spicules, whether they consist of carbonate of lime or hydrated silica, into relation
with the crystallising tendencies of these substances in the way of origin or explanation.
In the first place, in regard to the siliceous spicules, the fact has to be noted that the
silica in them never occurs in a crystalline state, but is always present as completely
amorphous hydrated silica or opal, as is shown, for instance, by the fact that they are not
doubly but only simply refractive. This is therefore against the supposition that the
various skeletal elements can be reduced to or derived from the crystallisations of the
substances of which they consist. Furthermore, we cannot ignore the extraordinary
abundance and frequent importance of the deviations of the rays from the typical angle
at which they ought to stand to one another, nor does the marked curvatures of the rays
accord with the supposition of defined crystalline axes.

I must rather maintain that the form of all the sponge spicules is determined by the
organic matrix in and from which they originate, and that the formative forces are in no
essential way different from those Avhich are everywhere exhibited in the shaping of the
living organism and its parts.

And though we still know very little about the forces determining form, yet in these
skeletal formations it is possible here and there to detect factors which, though they do
not indeed explain everjrthing, yet make much at least more intelligible.

If one can distinctly demonstrate a natural and necessary connection between
the form and disposition of a skeletal element and the function which it discharges, one
has, from the standpoint of utility and natural selection, rationalised the appearance of that
form and disposition.

In regard to the question why the typical and primitive spicule in the calcareous
sponges should be the plane, regular, triradiate form, in the Tetraxonia with their

• 1 Gruudziige einer Spongienfauna des atlantiscben Gebietes, p.- 4.

REPORT ON THE HEXACTINELLIDA. 501

derived allies the Mouaxonia and horny sponges the regular four-rayed form (spanischc
Reiter), and in the Triaxonia (Hexactiuellida) the regular six-rayed spicule, it seems to
me of essential importance to note the diflerencc of tlic architecture of the soft paits in
these three principal groups.

The Ascones, which may be regarded as the starting group in the calcareous sponge
series, exhibit, as is well known, in the simplest instance, a fixed thin-walled tube open at
the free end, and with its side wall penetrated by uniformly distributed circular pores.

In the much-varied group of Tetraxonia and their descendants we may consider the
typical form as that of a thick-walled cup, in the compact wall of which roundish or
quite spherical ciliated chambers lie crowded together like the acini in an acinose gland.
Such at least are most of the Tetractinellida and Lithistida, numerous Monactinellida and
horny sponges, though deviations also occur, as in the flat crusts of many Plakinidae,
which can hardly be regarded, however, as typical or primitive forms.

The typical structure of the Hexactinellida is very diff"erent.

The exceedingly loose wall of the typically .sack-shaped body exhibits, between two
abundantly penetrated, thin, parallel bounding lamellae, a single layer of large, sack-shaped,
ciliated chambers, connected both with the outer dermal and inner gastral membrane
by a trabecular system of thin thread-like strands. In simple and young forms these
thin connective-tissue trabeculae run predominantly in a radial direction between the
cliamber-layer and the two limiting lamellse, or directly between the latter, and are
usually so closely united with one another by tangential uniting strands that one
usually observes six threads at right angles to one another at each node of intersection.
This simplest structure is not of course persistent throughout. And in many cases,
doubtless, the insufficient preservation of the soft parts does not permit of the recogni-
tion of the original disposition of the trabecular.

Let us now consider how the spicules, regarded as typical and primitively charac-
teristic, are disposed in the A^ry variously con.stituted tissue of these three divisions of
sponges.

The regular three-rayed spicules of the Ascones are well known to occur tangentially
embedded in the wall of the tube, and in such a way that one ray lies parallel to the
axis dii'ected backwards towards the base, while the two others run ol)liquely forwards
and to the side, each usually embracing a parietal pore posteriorly.

In the Tetraxonia the typical regular tetracts lie in their simplest and most normal
form between the spherical ciliated chambers, while, as a rule, the regions without
chambers, viz., the margin, basis, and neighbourhood of the large canals, contain more
or less markedly differentiated spicules.

In the Triaxonia (Hexactinellida), finally, the typical regular hexacts arc found
almost exclusively in the strands of the trabecular framework, while in the chamber- wall

(ZOOL. CHALL. EXP. — PART LIII. — 1887.) Ggg 64

50-2

THE VOYAGE OF H.M.S. CHALLENGER.

110 spicules occur, and only iiuich modified elements in the two limiting lamella;, the
basis, the oscular margin, &e.

If we assume that the spicules in the body- wall serve essentially only for the support
or stiifening of the soft parts, it is to be expected that that form and disposition of the
skeletal elements will in eacli case be developed which in the given circumstances is
best fitted to give the necessary firmness to the body-wall.

I am decidedly of opinion that it can be shown with convincing probability that
such a necessary relation does exist between the structure of the soft parts in each of
the three principal groups of sponges and the characteristically typical forms of spicule

rQf

Fic. 10. — Triacts with each of the three rays lying at
a uniform angle between two adjacent pores.

Fig. 11.— Triacts di-sposed so that half of the interspaces
are occupied by their central portions and tlie other
lialf by their convergent rays.

which we regard on anatomical and developmental ground as primitive and fundamental
for each group.

If a plate is to be perforated by the maximum number of uniformly large spherical
pores in such a way that the lumina of the pores have a certain scope for expansion or
contraction, these pores can only exhibit one definite mode of arrangement, namely, that
of the cells of a honeycomb, and will leave a network with somewhat broad beams
between them.

If the plate consist of a mass which requires to be supported by the deposition of
hard parts, and if these are, on the one hand, to preserve the maximum of firmness, and on
the other to allow of a certain degree of expansion both to the entire tube and also to the

REPORT ON THE HEXACTINELLIDA. i,0'i

interjacent pores, then the triradiate spicule is the fittest form. These can be distributed
in such a way that each interspace between three adjacent pores is occupied by a regular
triact with each of the three rays lying at uniform angle between two adjacent ])ores
(fig. 10), or they may be disposed so that only half of the interspaces between the pores
are occupied by the central portions of the triacts, while the other half contain the points
of convergence of the extremities of three rays of three adjacent triacts (fig. 11). This
latter mode is exhibited by numerous very simple calcareous sponges of the Ascon type.
For the case of an open tube, fixed at one end, and with the other (oscular) extremity
free, careful consideration will show that the latter mode of disposition is the most
advantageous. In this way the boundary of each pore, especially on the lower mai-gin, is
strengthened by the forking of the triact which embraces it posteriorly, and the ^\'hole
sponge-tube is better strengthened by the relatively longer spicules than it would have
been on the former plan. We may therefore regard the development of the regular
triact as that conditioned and demanded by the structure of the soft parts of the
primitive calcareous sponges.

In regard to the Tetraxouia, with their regular tetracts, 1 submit the folluwing con-
sideration. When a number of spheres of equal size are uniformly pressed together on
all sides, they become disposed to one another in such a way that between each four
adjacent and directly contiguous spheres a regularly formed cavity is left, which is
continued in four three-sided clefts disposed at a uniform angle, and is thus connected
with the adjacent interspaces of similar form. One can best compare the form of these
spaces to regular tetrahedra with inpushcd walls and drawn-out
angles, which pass into the similarly elongated angles of adjacent
tetrahedral spaces, and thus secure the connection of aU the
cavities. Now, if one supposes this entire system of cavities
to be filled with a semi-soM mass, and the spheres to be empty
spaces, there is an obvious necessity for a supporting framework.
And if the skeletal system necessary for the support of this
framework consists of uniformly movable skeletal elements with
cylindrical branches, then each of these bodies must necessarily '°' tacrwuirfnur'spher^ ™"
liave its centre in the middle of each tetrahedral mass between

each four adjacent hollow spheres, and from this centre four strands must run out along
the four elongated angles of the tetrahedron.

The best supporting element for such a mass is aflbrded by such regular tetracts as
we find in the similarly constituted parenchyma between the ciliated chambers of
Tetraxonia, and known to be typical for this group of sponges.

Although the almost wholly unknown development of the HexactineUida gives us
as yet no basis for framing a conception of the architecture of the primitive Hexac-
tineUida, it is possible, from the close resemblance in the essentials of structure exhibited

504

THE VOYAGE OF H.M.S. CHALLENGER.

by all the known representatives of the group, to conclude that the jjrimitive ancestors
must have had a similar structure, though in very simple form.

As I have already noted, the generally sack-like, extremely loose body (consists
essentially of two approximately parallel bounding lamellte, the dermal and the gastrnl
membranes, l)etween which there extends the variously sinuous memlu-ana reticularis,
usually forming a folded chamber-layer, and supported by a framework of fine Ijeams.
The principal strands of this trabecular framework enclosing the parenchymal skeleton

extend at right angles from each of the two limiting
lamellae, and usually meet one another in such a.
way that they form beams penetrating the body-
wall transversely, while they are also laterally con-
nected by numerous trabeculse, which, running in
another direction, form a somewhat irregular frame-
work, though longitudinal and transverse strands
predominate.

It thus appears to me evident that, in these
circumstances, no more advantageous form of spicule
for the support of such a simple, loose Hexactinellid
l)ody, could be devised than the regular hexacts,
disposed in such a way that one radial ray unites
the two bounding lamellae, while the second is

tangential, and the third longitudinal (fig. 13), just
as they do indeed occur in the simplest Lyssacina,
Ilolascus, Bathydorus, &c.

By the firm union of all such hexacts a lattice-
work is formed such as we find to be developed with
almost ideal regularity in the younger portions of
Farrea. As the wall becomes thicker new layers
of similar regular hexacts are laid down, as we find
best developed among the Lyssacina in Holascus Jihxdatus, and among the Dictyonina
in the older portions of i^arrea-stocks, but more or less distinctly in all Dictyonina and
most Lyssacina.

Thus it appears that a consideration of the mechanical conditions of the soft parts
to be supported affords in this case also an insight as to the utility of the specific
structure of the skeletal elements as here represented by the regular hexacts.

Flo. 13. — Arrangement of Hexacts between
the chambers.

INDEX.

The more important references are indicated by heavier figures.

Acanthascus, n. gen., 145, 375.

„ cactus, n. sp., 148, 375 (PI. Ivii.).

„ duhius, n. sp., 147, 375 (PI. Ivii.).

,, (jrossularia, n. sp., 145, 375 (PI. Ivi.).

Acantlwspomjia, Salt, 14, 184.
Acanthospongioe, 56.
Actiniae, Eurete marshallt, 297.

„ Eurctc semperi, 292.
Adrasta, 7.

Aega spongiopliila, 56.
Afferent passages, 20.
Agassiz, 61.

,, exploration by, 427.
Aleyoncellum speciosum, 5, 54.
Amphidiscs, 36, 48; firtree-Iike, 212; largest, 202;

peculiar, 192.
Amphidiscophora, 178, 380 ; phylogeny, 492.
Ancora, 48.

Anderson, T., 18; on Japanese sponges, 276.
Annular thickening, 171, 173, 192.
Anomocladina, 16.
Antarctic Ocean, Balanifes, 122; Holascus, 368;

Polyrhahdus, 121.
Antilles Islands, Aplirocallistes, 395; Cyrtaulon, 400;
Euplectella, 366; Regadrella, 367; Rliahdo-
peetella, 371.
Aphrocallisfes, Gray, 305, 306, 310, 395.

heatru; Gray, 305, 311, 396 (PI.
Ixxxiv.).
„ hocagei, Wright, 306, 313, 395 (Pis.

Ixxxiii., Ixxxiv.).
„ ramosus, n. sp., 319, 396 (PL l.xxxvi.).

Aplirocallistes vastus, n. sp., 311, 315, 317, 396 (PI.
Ixxxv.).
„ tubulosus, 311.

Apical cone, Hyalonema tkomsoni, 213.
Ascension, Aplirocallistes, 313.
Asconema, Sav. Kent, 113.
„ kentii, 207.

„ setuhalense, Sav. Kent, 116, 371 (PI. xxi.).
Asconematidse, 371.

„ defined, 113.

„ bathymetrical distribution, 456, 458.

,, geographical, 436.

„ phylogeny, 494.

Asconematins, 113, 371.
Astylospongidfe, Zittel, 12, 18.
Asynauloidae, 10, 487.
Aulascus, n. gen., 118, 372.

„ johnstoni, n. sp., 118, 372 (PI. xxii.).
Aulocalyx, u. gen., 174, 380.

„ irregularis, n. sp., 174, 380, 491 (PL Lx.).
Aulochone, n. gen., 168, 378.

„ cylindrica, n. sp., 168, 171, 379 (Pie.
Ixvi., Ixviii.).
Uliuw, 170, 171, 379 (PI. Ixviii.).
Aulocystis, n. gen., 405.

„ grayi, Bwk., 357, 405 (PL civ.).

zitteUi (Marshall), 361, 406 (PL civ.).
Aulodictyon, Saville Kent, 269.

,, woodward a, 286.

Australia, Holascus, 87, 89, 368; Hycdonema, 209,

383.
Autodermalia, 43, 49.

50(5

THE VOYAGE OF H.M.S. CHALLENGED.

Autogastralia, 49.
Axial canal, 27, 28.
Azores, Sti/locali/j; 187, 38.'{.
r.abia, EiipJedella, 76, :?66.
Hajalus, 502.
Balanella, n. gen , .'J14.
Balanites, n. gen., 122, 372.

„ pipetta, n. sp., 122, 372 (PI. .\xiii.).
Ban da Islands, Aulocydiis, 301 ; Crateromorpha, IGG,

378; Farrca, 287, 391 ; Mi/Uima, 404; Srlero-

tliamims, 339; Siylomlyx, 221, 384.
Barbados, Dadijlocahju; 403 ; Regadrella, 85.
Barboza du Bocage, Htiahmuma Jusitaninim, 183, 185,

225.
BarbukB, 37, 276.
Basalia, 40, 48.
Basal collar-pad, Ili/alonema deholdii, 195, 197,

206.
Jlat/iydoni,-!, n. gen., 150, 376.

haculifer, n. sp., 154, 376 (PI. li.\.).
„ fimhriatua, n. sp., 20, 151, 376, 491

(PI. Iviii.).
„ spinosus, n. sp., 153, 376 (PI. lix.).

dellaUi.", n. sp., 152, 376 (PI. lix.).
Bathymetrical distribution, 453.
Borlingues, Euplectella, 366.
Bermudas, Aphrocalli^ifex, 313; Auloralyx, 174 ; Chone-

Icu^ma, 397 ; Dartylocaly.r, 350, 403 ; Euplec-

fella, 82, 367; Farrea, 288; Lefroyella, 301,

302, 394; Rhahh„lkfyum, 107.
Bequia, Hyalonema, 207 ; Rliabdoilkfyuiii, 370.
Birotulifera, Carter, 9, 260.
Blainville, 99.
Bottom, nature of, 468.
Bowerbank, 6, 8; AphrocalUstcs, 306; Dactyl ocalyj:,

343; Euplectella, 54, 55, 59; Farrea, 267, 268,

271, 272; Hyalonema, 183, 184, 188; fyhiteon,

345; Rossella, 136.
Branching, 22.
Brandt, 182.

Brazil, Eupdedella t<uberea, 73; PherovniK, 241.
Budding, 25; oi Polylophus, 133.
Buenos Ayres, Holascus, 86, 368; Rossella, 375.
Calcarea, pbylogeny, 497.
CallicispongiEe, 9.

CalUsphsera { = Pheronema grayi), 237.
CaUodictyonidse, Zittel, 13.
Carninm, 497.
Canalaria, 45, 49.
Canary Islands, Polioporjon, 254, 389.

Cape of Good Hope, Holascus, 88.

Cape St. Vincent, Eitpleclella, 73; Hyalonema, 183.

Cape Verde Islands, A].'hro/-allistcs, 395; Lunnip.nclta,

130; Sympagella, 120, 372.
Cape York, Hyalonema, 232.
Caroline Islands, Euplectella, 81.

Carter, 18; Alcyoncelhirn, 103; Aphroeall isles, 308;
Asconema, 115; Birotulifera, 260; Classification,
9; €rateromorpha, 115, 160; Dactylocalyx,
346 ; Distribution, 421 ; Euplectella, 58 ;
.g'M/'efc, 290; /''«/•*(«, 269, 270, 276; //airo-
dictyum, 103; Hexactinella, 328; Hyalonema
cehuense, 228; Labaria, 238; Meyerina =
Semperella, 260; Pheronema, 237; Polylo-
phus, 133; Rossella, 136, 137; Sympagella,
119.
Uurteria japonka, Gray, 184.
,, lusitanka, Gray, 185.
Cauluealyx, n. gen., 172, 379.

^CTier, n. sp., 172, 379 (PI. Ixix.).
Caulopliacina?, 124, 373.
Caulophacus, n. gen., 124, 373.

„ elegans, n. sp., 126, 373 (Pis. xxv.,

xxvi.).
„ Zni«s, n. sp., 124, 373 (PI. xxiv.).

Celebes, Aulochone, 379.
Central boss, Hyalonema, 189.
Central column, Hyalonema sieboldii, 191.
Central cone, Hyalonema apertum, 204, 215.
Chalinida;, 497.

Cballengcr Hcxactinellida, List of, 472.
Chamber cells, 23.

„ layer, 20, 47.
„ pores, 19, 47.
Chonelasma, n. gen., 320, 397

,, calyx, n. sp., 326, 398 (PI. Ixxxix.).

daiderleinii, 324, 398, 498 (PI. xc).
,, hamatum, n. sp., 323, 397 (PI. xci.).

,, lamella, n. sp., 321, 397 (Pis. Ixxxvii.,

Ixxxviii.).
sp., 326 (PI. xc).
Christina Islands, Stylocrdyx, 223, 385.
Classifications, 8, 9, 10.

Claus, 7, 9 ; Dactylocalyx, 344 ; Euj^leetella, 56.
Clavute, 37, 48.
Clavularia, 266.
Caloptychidoe, Zittel, 14.
Collar, 23.

Columella, Hyalonema yracile, 197.
Comitalia, 46, 49.

REPORT ON THE HEXAUTINELLIDA.

507

Commensals — Artinix, 292, 297 ; Aeya sporujiophila,
56 ; Anthozoa on Eurefe, 298 ; on Hijalnnema sic-
hohlii, 191 ; Ci'ustacnans, 56, 61; HyJrozoon, 96;
PaKomonid, 56 ; Palijthoa fafua on Hijalonenia,
183, 189; SfepJtOfiri/pJms w/'rahi/i'!, 98.
Connecting membrane, 47.
Connefttive -substance, 24.
Corhitella, 101.

„ spociosa, 101.
(JoralliospongiK, 8, 57.

CoscinoiJorida;, 12, 320, 396 ; distribution, 438.
Grateromorpha, Gray, 159, 161, 377.
Crateroimirplta meyeri. Gray, 161, 377 (PI. Ixi.).

„ murraiji, n. sp., 164, 378 (PI. Ixiii.).

„ iJderfelderi, n. sp., 164, 378 (Pi. Ixii.).

„ fiimuht, n. sp., 166, 378 (Pis. Ixvii.,

Ixviii.).
Crozets Islands, Aranfhasais, 375 ; Aulocalyx, 176 ;
Baf/njdoru.% 153 ; Gaulophacus, 373 ; Cltone-
lasma lamella, 321; Farrea sp., 288; Holasciis,
368 ; Malacosarcus, 368 ; Stylocalyx, 384.
Cufl' of spicules, 65.
Curvature of rays, 170.
Cijathella, O. Schmidt, 14, 15.
Cyatlwphycus, 63.
Cyrtaulon, n. gen., 332, 400.

„ su/sbeet, O. Schmidt, 333, 400 (PI. xcii.).
,, solutus, n. sp., 333, 400 (PI. xcii.).
Cystispongia siiperstes, 0. Schmidt, 15.
Dadylocalyx (Stutchbury), 341, 342, 348, 402.
„ hoirerhanlni, Johnson, 343.

,, i-fispiis, O. Schmidt, 346.

,, heteroformis, Bvvk., 344.

„ inasonis, Bwk., 344.

{V) patella, n. sp., 350, 403 (PL c).
,, polyliscus, Bwk., 344.

,, iMtaforum (O. Schmidt), 348.

,, prattii, Bwk., 343.

,, piimicea, Gray, 344.

,, pumiceuti, Stutchbury, 348, 403.

■•.ubffloboxiis. Gray, 344, 349, 403 (PL
xcix.).
,, stufchburyi (Sollas), 347.

Death, processes of, 26.
I )endroq)ongia steerei (Scleruihamnus), 337.
Dermalia, 43, 48.
Dermal membrane, 19, 47.
Dermal pores, 47.
Development, 26.
Diacts, 36.

Diaretula, 0. Schmidt, 15.
Dirfyocaly.r, u. gen., 105, 370.

,, grwiliif, n. s])., 105, 370 (PL xii.).

Dictyonalia, 46, 49.
Dictyonina, Zittcl, 12,39, 265, 390; bathymelrical

distribution, 462 ; localities of, 435.
Diciyitphyion, 63.
Diplacodiwm, O. Schmidt, 15.
Dvplohelia profunda, 203.
Directalia, 46.
Di.sc of ro.settes, 173.
Discohexact, 48.
Discohexaster, 32, 48.
Discodermia, Gray, 346.
Doderlein's collection, 2, 427, 451.
Duncan, P. M., Aphrocallitfes, 309.
Eiierent passages, 20.
Ehrenberg, Hyalonema, 182, 184.
English Channel, Aphroeallistes, 395.
Enoshima, Aplmical/istes, 314; Chonelasma, 326;
Eurcte, 294 ; Ilpxadinella, 329 : PeriprageUa,
299.
Eospongia, 18.
Epithelium, 23.
Esperiada?, Gray, 184.
Etheridge, Capt., Eupledella cucumei; 266.
Eudictynm, Marshall, 104, 370.
Eiidietyum elegans, Marshall, 103, 104, 370.
Eupledella, Owen, 64, 365 (Pis. i.-vi., xiii., xiv.) ;

bibliography, 53 ; history, 53.
Eupledella aspergillum, R. Owen, 67, 64, 67, 366.
,, crassistellata, n. sp., 81, 367 (PL xiii.).

„ cueumer, Owen, 76, 366.

,, jovis, Oscar Schmidt, 72, 366 (PL vi.).

(?) nodosa, n. sp., 82, 367 (PL xiv.).
,, owenii, Herklots and Marshall, 56, 60, 78,

366 (PL vi.).
„ siiberea, Wyv. Thomson, 60, 73, 360

(Pis. v., vi.).
Eupleetellida', 51 ; batliymetrical distribution, 456,

458 ; geogTaphical, 436.
Euplectellinae, 52, 365.
Eh ret e (Semper), Carter, 289, 291, 392.

„ bowerbankii, n. sp., 297, 393 (PL Ixxix.).

„ eaderi, n. sp., 296, 393 (PL Lxxviii.).

,, farreopsis, Carter, 275, 290, 295, 392 (PI.

Ixxix.).
,, marshaUi, n. sp., 297, 393 (I'L bcxix.).
„ schmidtii, n. sp., 293, 392 (PL lxxviii.)
., .semper!, n. sp., 292, 392 (PL IxxviL).

508

THE VOYAGE OF H.M.S. CHALLENGER.

Eurelo simplicinahna, Semper, 298.

Euretitlie, 12, 289, 391 ; distiihutioii, 438 ; phylogeny,

193.
Euryple<jma, n. gen., 176, 380.

„ auriculare, n. sp., 176, 380, 490 (PI.

cii.).
Eversion of cup of Etwijplegma, 169.
Farre, Dr. A., Farrea, 266.
Farrid, Bowerbank, 266, 276, 390.
,, aculeafa, Bowerbank, 271, 274.
„ davigem, n. sp. 287, 391 (PI. Ixxv.).
„ densa, Carter, 270.
„ facimda, 0. Schmidt, 268, 275.
,, fistiilata, Bowerbank, 271, 272.
„ gassioti, Bowerbank, 271.
,, haeckelii, 276.
„ inermis, Bowerbank, 271, 274.
„ infundibidiforinis, Carter, 270.
,, iiTegidaris, Bowerbank, 271, 374.
„ Ixvis, Bowerbank, 271, 272.

occa (Bwk.), Carter, 277, 390 (Pis. Ixxi.,
Ixxii., Ixxiii., Ixxvi.).
„ parasitica, Bwk., 271, 273.
,, perannata, Bowerbank, 271, 274.
„ pocillum, Bowerbank, 271.
„ solladi, n. sp., 281, 286, 391 (PI. Ixxiv.).
„ sp. (?), 288 (PL Lxxvi.).
,, "pini/era, 271, 273.
„ spinosissima, Bowerbank, 271, 273.
,, spinulenta, Bowerbank, 271, 273.
,, robusta, Bowerbank, 271, 274.
,, iuhulata, 272.
,, valida, Bowerbank, 271, 273.
„ msmaeri, n. sp., 281, 286, 391 (PI. Jxxiv.).
Farreidse, 266 ; distribution, 438.
Fieldingia, Sav. Kent, 335, 401.

„ lagettoides, Sav. Kent, 335 (PI. xcvii.).

Fiji Islands, Txgeria, 94, 369.

Filhol, H., 17; Asconema, 116; bathymetrical dis-
tribution, 454; divergent results, 433; Euplec-
tella suherea, 76; Pheronema, 239; flagellum,
23.
Flesb spicules, 10.
Floricome, 32, 48, 170.

Florida, AphrocalUgtes, 395; Sympiagdla, 372.
Form and structure, 19.
Fossil Euplectellids, 63.
I'rance, AphrocalHstes, 395.
Gastralia, 44, 49.
Gastral cone, Hyalonema, 189.

Gastral membrane, 19, 47.
pores, 47.

septa, Hjialonema sieboldii, 191.
Gemmul.'e, 60.
Genealogical tree, 495, 499.
General structure, 19.
Genital products, 24.
Geograpliical di.stributiou, 420.
Gibraltar, Eupilcddla, 76, 366 ; Ros.:<dla, 143, 375.
Glossary, 47.
Glassrope, 202.
Granules, 24.
Graphiohexasters, 31, 48.

Gray, 6, 56 ; AphrocalHstes bvatrix, 305 ; Asmnenia,
114; Crateromorphida>, 160; Dactylocalycidw,
343, 346 ; Eup)ledella, 55, 57 ; Hyalonema,
181, 184, 185; Labaria hemisphserica, 237;
Meyerella dariformis = Semperella, 260;
Myliusia, 352; Pheronemadse, 237; Pohjloplnis,
132, 133; Rossella, 136; Sympagella, 119.
CIrenada, Eupledella, 77 ; Hyalonema kentii, 207.
Guadeloupe, Hyalonema toxeres, 207.
Habi-odidyum, Wyv. Thomson, 99, 370.

„ sjieciosiim, Quoy and Gaimard, 99, 370.

Haeckel, spicules, 499.
Hansen, A., Hyalonema arctieum, 188.
Havana, Margaritella, 404.
Herklots and Marshall, Eupledella, 56.
Hertwigia, 0. Schmidt, 109, 371.

„ faldfera, O. Schmidt, 62, 109. 371, 491.
„ faldformis, 16.
Heterotetla, 101.
HexactinelUda — origin of term, 5. 51, 365; List of,

480.
Hexadinella, Carter, 327 ; hi.story, 327, 328, 398.
lata, n. sp., 229, 399 (Pis. xciv., xcv.).
fubulosa, n. sp., 328, 398 (PI. xciii.).
ventilahrum. Carter, 331 (PI. xcvi.).

„ Carter ( = Tre.todidyum

cyathus), 328, 399.
Hexacts, 29 ; irregular, 32 ; regular. 48 : nodular,

119. ^

Hexasterida, phylogeny, 492.
HexasieropJiora, 51, 365.

Higgin, Eujyleddla, 58 ; Hyalonema cebuense, 186,
228 ; Labaria hemisphxrica, 238.
,, Fheronema hemisplixrirmn, 247.
Hinde, G. J., 18.
Histology, 23.
History of work, 1 ; of Hexactinellida, 5.

EEPORT ON THE HEXACTINELLIDA.

yOO

Holascinre, 85, 3G7.
Holascus, n. gen., 85, 367.

„ fibulatas, n. sp., 87, 368 (Pis. xv., xvi.).
„ poleja'evii, n. sp., 89, 368 (PI. xvii.).
„ ridleyi, n. sp., 90, 368 (PI. xvii.).
„ dellatus, n. sp., 86, 36S (PI. xiv., xv.).
Holtenia, Wyv. Thomson, 115, 234.

„ catyenteri, Wyv. Thomson, 235.
,, jwurtalesii, (O. Schmidt), 236, 238.
H yaloclixte possieti^ Brandt, 182.
Ilijalonema, Gray, 179, 189, 381.

„ sens, strict., subgen., 190.

affine (Mar.shall), 187, 216, 217.
„ anonialum (Bowerbank), 188.

,, arcticum (Hansen), 189.

„ eebuense, Higgin, 186, 228, 385.

eonus, 209, 383 (PI. xxxiii.).
„ divergens, n. sp., 199, 382 (PI. xxviii.).

gracile, 196, 382 (PL xxvii.).
l-entii, 0. Schmidt, 207, 382 (Pi. xxx.).
„ loveiii, Wyv. Thomson, 186.

,, lusitanicum, Barboza du Bocage, 183, 214,

225, 385 (Pl.s. xxxviii., xxxix.).
„ mirabile, Bowerbank, 184.

„ mirabilis, 181.

parallelum, M'Coy, 183, 188.
„ pocidum, n. sp., 208, 383 (PI. xxxiii.).

„ rohmtum, n. sp., 229, 386 (PI. xxxii.).

„ schultzei ( = Semperella), 259.

sieholdii. Gray, 181, 190, 381 (PI. xxvii.);
spicules, 192.
„ smithii, Young and Young, 188.

„ tenue, n. sp., 228, 386 (PL xxx.).

„ thomsoni, 187.

„ thomsoni, var. exiguum, 214.

toxeres, Wyv. Thomson, 188, 201, 202,
203, 382 (PL xxix.).
„ (Stylocalyx) apertum, n. sp., 214.

„ [Stylocalyx) clavirjerum , n. sp., 220 (PI.

xli.).
„ {Stylocalyx) depressum, n. sp., 217, 219.

„ {Stylocalyx) elegam, n. sp., 223 (PI. xsxi.).

„ {Stylocalyx) glohiis, n. sp., 221 (PL xL).

„ {Stylocalyx) tenerum, n. sp., 224 (PL

sxxi.).
„ {Stylocalyx) thomsoni, Marshall, 211.

„ species diversae indefinite, 231 (Pis. xxxii.,

xxxix.).
Hyalonematidae, Gray, 178, 381; bathymetrical dis-
tribution, 457, 459; geographical, 438.

(ZOOL. CHALL. EXP. — PART LIII. 1887.)

Hyalonematina?, F. E. S., 178, 381.
Ilyalostylus, n. gen., 110, 371.

,, dives, n. sp., 110, 371 (PL Ixx.).

Hyalothauma ludelcingi { = Semperella), 260.
Hyalothrix lusifanica. Gray, 185.
Hymedesmia johnsuni (Bowerbank), 272.
Hypodermalia, 44, 49.
Hypogastralia, 45, 49.

Indian Ocean, 440; Aidascus, 118; Malacosaccm, 91.
Inermia, 341, 402; bathymetrical distribution, 457, 459.
Jphiteon heatrix (Bowerbank), 306.
Gray, 345.

caUocyatlms, Gray, 345.
ingalli, Bwk., 345.
panicea, Valenciennes, 306, 344.
suhglohosa. Gray, 345.
Intermedia, 49.
Japan, Acanthascus, 376 ; Aphrocallistes, 396 ; Bathy-

dorus, 151; Cavlophacus, 126, 373; Chone-

lasma, 398; JSuplectella, 78, 366; Eurete, 392;

Hexactinella, 399, 400; Hyalonema, 191, 381 ;

Periphragella, 299, 394 ; Rhabdoccdyptus, 377.
Japanese composition of sponge, 184.
Jeffreys, Gwyn, Aphrocallistes hocagei, 308 ; Rossella,

143.
JoaneUa, O. Schmidt, 16.
Juan Fernandez, Hyalonema, 208, 383 ; Hyalonema

poculum, 208.
Kaliapsis (Gray), 346.
Kent, Saville, Asconema, 113; Aphrocallistes, 307;

classification, 8, 9 ; Farrea, 268, 269 ; Field-

irigia, 335 ; Holtenia, 236 ; Pheronema, 236.
Keratosa, 497, 498.
Kerguelen Island, Caulophaeus, 124; Clionelasvia, 391;

Holascus, 88; Rossella, 139, 375.
Kermadec Islands, Aulochone, 168, 379; CJwnelasnta,

397; Euryplegma, 380; Farrea, 277; Polio-

pogon, 257 ; Walteria, 96, 369.
Key to Genera and Species, 407.
Klistermann, Hyaloiiema sieholdii, 187.
Labaria { = Pheronema), 246.

„ hemisphxrica (Gray), 237, 238.
Lanuginella, O. Schmidt, 129, 374.
Lartuginella pupa, O. Schmidt, 130, 374 (PL liii.).
Largest forms, 40.
Lefroyella, Wyv. Thomson, 301, 394.

„ decora, Wyv. Thomson, 302, 394 (PL

Ixxxii.).
Leidy, Pheronema ann«, 234.
Leiobolidium, 0. Schmidt, 15, 238.

Ggg65

510

THE VOYAGE OF H.M.S. CHALLENGER.

Lendenfeld, Calcarea, 497.

List of Challenger specimens, 472.

Little Ki Island, Auloojstis, 361, 406; Crateromorpha.
164, 378; Cijrtaulon, 333, iO\; Eurete, 292,
295, 296, 297, 392, 393 ; Fieldmgia, 335, 401
Hexadinella, 399 ; Lanuginella, 130, 374
Myliuda, 405; Pheronema, 248, 250, 388
Polyloplms, 133, 374; Semjm-ella, 265.

Lithistidse, 8, 496, 498.

Localities, list of, 425, 426.

Lopliocalijx, n. gen., 514.

Lophohelia (Farrea), 147, 269.

Lophospongiae (Schultze, M.), 40, 184.

Lov(5n, Hyalonema horeale, 185.

Lovenia (Bocage), 185.

Luzon, Hyalonema, 233.

Lyssacina, 14, 39, 51, 365 ; bathymetrical distribution,
462 ; localities of, 434.

Macandrewia azorica, Gray, 343.

Macaulay Islands, Polioporjon, 257, 389.

Macio, Phei-onema caiyenteri, 242.

Magellan ^ixaXt, Acanthascus, 147; Bathydorus, 152;
Rliabdocalyptus, 158.

Maeandrospongidfe, Zittel, 13, 34, 402; distribution,
438 ; pliylogeny, 493.

Malacca, Aphrocallistes heafrix, 311, 396.

Malacosacais, n. gen., 91, 368.

„ unguiadatus, n. sp., 93, 369 (PL xix.).

„ vastus, n. sp., 91, 368 (PI. xviii.).

Maldon Island, Hyalonema, 199.

Maly, analysis of spicules, 28.

Manila, Farrea, 278, 391.

Margaritella, 0. Schmidt, 351, 404.

Margaritella cwloptychioides, O. Schmidt, 351, 404
(PL ci.).

Marginal fringe, Hyalonema, 189, 190.

Marginalia, 43, 48.

Marginal teeth, 32.

Marion Island, Aidoealyx, 174, 380.

Marshall, 3; Aphrocallistes, 309; Asconema, 115; bathy-
metrical distribution, 453; classification, 10;
Crateromorpha, 1 60 ; Dactylocalyx, 347 ;
diagnosis of Euplectellidae, 103 ; distribution,
422; Eupledella, 60, 77; Eurete, 220 ; Habro-
dictyum, 103; Hyalonema, 187, 216; Laharia =
Pheronema, 246; Periphragella, 299; pliylogeny,
486; Possella, 138; Sclerothamnus, 337;
Semperella, 260; 8tylocalyx, 211.

Marshall and Herklota, Hyalothaum.a ludekingi =
Semperella, 260.

Slarshall and Meyer, Aulocystis, 363; Polylophus, 132,

133.
Martens, von, 3; Hyalonema, 183.
Martinique, Hyalonema toxeres, 207.
Meangis Islands, Avloehone, 171, 379.
MelittionidiB, Zittel, 13, 304, 309, 395; distribution,

438 ; phylogeny, 493.
IMellish Islands, Hyalonema, 217.
Membrana reticularis, 19, 492.
Membrane of parietal gaps, 47.
Messier channel, Bathydorus, 152, 376.
Meyer (see IVIarshall), 3.
Meyerella claviformi.s = Semperella, 260.
Meyeiina, Carter ( = Semperella), 260.

„ claviformis, Gray, 237.
Meyerinid®, Gray, 260.

Mid Pacific, Eupledella, 81; Hyalonema, 231, 382.
Milne-Edwards, Alcyoneellum, 100 ; Aphrocallistes,

310 ; Eupledella suberea, 76 ; Habrodidyum,

99.
Mindanao Island, Hyalonema gracile, 190.
Misaki Bay (Doderlein), 451.
Molucca Islands, Eurete carteri, 296; Habrodidyum,

370; Periphragella, 299, 394; Monacidse, 10,

14.
MonacteUidse, 496.
Monacts, 37.

Monte Video, Holascus, 86; Hyalonema, 228.
Moore, T. J., Eupledella, 57.
Morrolight, Scleroplegma, 351.
Morocco, Asconema, 372.
Moselej^ 3.

Murie, J., Sderothamnus, 337.
Murray, J., 3; Pheronema carpenteri, 241.
Myliusia, Gray, 352, 404; bibliography, 352; liistory,
353.
„ callocyathus. Gray, 346, 404.
Myxilla, 98.
Myxospongise, 498,
Nature of bottom, 428.
Nature of ground, influence of, 468.
Nidos de Mer { = Pheronema grayi), 236.
Norman, Asconema, 115; Holtenia carpenteri, 238;

Hyalonema lusitanicum, 188.
North Atlantic, 440.
North Pacific, 440; Bathtjdoms, 151, 376; Hyalonema,

229, 386 ; Stylocalyx, 384.
North Temperate Zone (distribution), 445.
North Tropical Zone (distribution), 446.
" Oscular plate," Sollas, 10.

REPOET ON THE HEXACTINELLIDA.

511

Ova, 24.

Owen, Eupledella, 53 ; Eiipledella cummer, 266 ;

Farrea, 267.
Oxyhexacts, 30, 31, 48.
Oxyliexasters, 48.
Pttchastrella ahyssi, 115.
Pachaulidiwn (0. Schmidt), 15.
Palxomanon, 18.

Pahjthoa, 183, 189, 190, 203, 211, 214.
Parenchymalia, 45, 49.
Parietal gaps, 47.
Pathological modifications, 170.
Patulina (Carter), 9.
PearsaU, T. J., 181.
Penguin Island, Bathi/dorus, 153, 376; CJionelasma,

327; Stylocalyx, 200, 384.
Pentacts, 33.
Periphragella, Marshall, 299, 394.

elisx, Marshall, 174, 299, 394 (Pis.
Ixxx., Ixxxi.).
Perrier, E. (bathymetrical distribution), 454.
Phalcellia ventilahrum, Bwk., 328.
Pheronema, Leidy, 234, 239, 386.

an9ix, Leidy, 239, 386 (PL xlii.).
„ carpenteri (Wyv. Thomson), 241, 387

(PI. xliii.).
„ giganteum, n. sp., 250, 388 (Pis. xlv.,

xlvi.).
„ glohosum, n. sp., 248, 387 (PI. xliv.).

grayi, Sav. Kent, 236, 246, 387.
„ hemispheericum (Gray), 246, 249, 387.

„ parfaiti, 17.

Pheroneniadse (Gray), 237.

PhiUppine Islands, Aplirocallistes, 319, 396; Aulocystis,
406; Crateromorpha, 161; Eupledella, 64;
Eurete, 293, 392; Farrea,, 277; Holascus, 90,
368; Hyalonema, 196, 382; Polylophus, 133;
Sderofliamnus, 402; Semperella, 261.
Phylogeny, 485.
Pinulus, 33, 48.
Placodidyum cucumaria, O. Schmidt = (Holothurian,

Thyone), 8.
Plakinidse, 497.

Plates— Sir WyviUe Thomson'-s, 3.
Pleionacidfe, 10, 14. ,

Pleorhabdus, n. gen., 514.
Pleuralia, 42, 48.
Plumicomes, 31, 48, 173.
Pol^jaefi', Calcarea, 497.
Poliopogon, Wyv. Thomson, 254, 388.

Poliopogon amadou, Wyv. Thomson, 254, 388 (Pis.
xlix., 1.); analysis of spicules, 28;
young forms, 256.
„ gigan, n. sp. 257, 389 (Pis. xlvii., xlviii.).

PoIIacidoe, 10, 14.
Polylophus, n. gen., 132, 374.

„ philippinensis (Gray), 133, 374 (Pis. liii.,
liv.) ; budding, 25.
Polyrltabdus, n. gen., 121, 372.

,, oviformis, n. sp., 121, 372 (PI. xxiii.).

Portugal, Asconema, 372 ; Dadylocalyx, 350, 403 ;

Fieldingia, 401 ; Pheronema, 387 ; Sympagella,

372.
Possession Island, Acanthascus, 145, 375.
Pouch-like depressions, 171.
Pourtal6s, Count, 268 ; exploration, 427.
Principalia, 46, 47, 49.
Prince Edward Island, Aulascus, 372 ; Aulocalyx, 174 ;

Rossella, 139, 375.
Prostalia, 40, 48.
Protaddlleum, 18.
Prout, W., 181.
Psdalia globulosa, 132.
Pseudogasters, Semperella, 261.
Pseudogastral cavity, Cyrtaulon, 332.
Pseudo-sieve plates, Semperella, 261.
Puerto Bueno, Acanthascus, 147, 375; Rhahdocalyptus,

158, 377.
Quoy and Gaimard, 55 ; Alcyoncellum, 99 ; Hahro-

didyum, 99.
Kadial fold's, 169, 171.
Raoul or Sunday Islands, Euryplegma, 176, 380;

Poliopogon, 257, 389.
Regadrella, O. Schmidt, 16, 84, 367.

phamix, O. Schmidt, 62, 84, 367 (PI. xiii.).
Reproduction in Farrea, 285.
Reserve nutriment, 24.
Rhahdocalyptus, n. gen., 155, 377.

„ mollis, n. sp., 155, 377 (PI. Ixiv.).

,, roeperi, n. sp., 158, 377 (PI. Lxv.).

Rhabdodidyum, O. Schmidt, 107, 370.

delicatum, 0. Schmidt, 106, 107,
370 (PI. XX.).

,, delineatum, 491.

Rhabdopedella, O. Schmidt, 108, 370.

„ tintinnus, O. Schmidt, 16, 62, 108,

370 (PL xii.).
Rhabdostauridium, 0. Schmidt, 15, 489.
Ridges, 65, 177.
Ridley, S. 0., 181.

512

THE VOYAGE OF H.5I.S. CHALLENGER.

Rigid balls, 17.

Rio de la Plata, Hyalonema, 228, 38G.

Roller-stars, 120.

Root-tuft, 40.

Rosettes, 30, 48, 128, 175.

Rosettifera (Carter), 9, 138.

Rosselki, Carter, 115, 136, 138, 374.

antardica. Carter, 138, 139, 374 (PI. Iv.).
„ phiUppinengis, 132.

celata, Wyv. Thomson, 143, 375.
RosselliJfe, 129, 374 ; batliymetrical distribution, 456,
458 ; geograpliical distribution, 436 ; phylo-
geny, 494.
Rudiments of rays, 170.

Sagami Bay, 3 ; AphrocolUstes, 317, 319; Clionelasma,
324 ; Eurete, 296, 297, 393 ; Fairea, 277, 278,
391 ; Sti/Iocalyx, 214, 383.
Santa Cruz, Pheronema, 85 ; Reijadrella, 239, 387.
Sarcohexactinellida, 9.
Sarco-vitreobexactinellida, 9.
Sars, Hyalonema longissimum, 186; Hyalonema

parvum, 186.
Schmidt, 0., 8; Aphrocallistes, 307, 309; Asconema,
115; Cyrtaulon, 332; Dactyloealyx, 346, 347;
Eiipledella, 62; Fansa, 268; Hertwigia, 109;
Holtenia jMurtalesii, 236 ; Hyalonema kentli,
207; Hyalonema sieboldii, 188; Lannginella,
129 ; Lefroyella, 302 ; Margaritella, 351 ;
Mexican Sponges, 15; Phei-onema, 245 ; Pher-
onema annse, 238; phylogeny, 488, 497; Re-
gadrella, 84; Rhahdodkfyum, 107; Rhahdo-
pedella, 108; Sympagella, 119.
Schultze, M,, 6 ; Eupledella, 55 ; Hyalonemata, 182 ;

Lophospongise, 184.
Scleroplegma, 0. Schmidt, 350, 404.

conicum, 0. Schmidt, 351, 404 (PI. ci.).
,, laterna, 16.

Sderothamnm, Marshall, 337, 401.
Sderoihamnus daudi, Marshall, 339 (PI. xcviii.).
ScopulK, 34, 48.
Seopularia (Carter), 289, 391.
Scopulifen-a (Carter), 9.

Scotland, Aaconema, 372 ; Pheronema, 241, 387.
Scyphia, 8.
Secondary rays, 30.
Semper, 3 ; Eupledella, 55 ; Eureie, 289 ; Hyalonema

sdiuUzei { = Semperella), 185, 259.
Semperella, Gray, 259.

„ sdiultzei, Semper, 261 (Pis. li., lii.).

Semperellinse, 259, 389.

Setubal, Hyalonema, 385.
SeycheUe Islands, Eupledella, 366.
Shetland Islands, Stylocalyx, 383.
Sierra Leone, Malacosaceus, 93, 369.
Sieve-plate, 66, 77, 190, 290.
Silicea (Bowerbank), 183.
Skeletal spicules, 27.
Skeleton, 27 ; evolution of, 502.
South Atlantic, 440.
South Indian Ocean, Balanites, 372.
South Pacific, 440; Balhydoriis, 155, 376; Didyo-
calyx, 105, 370; Hyalostylus, 371; Stylocalyx,
224, 385; Traclujcaulus, 128, 373.
South Temperate Zone, distribution, 445.
South Tropical Zone, distribution, 447.
SoUas, classification, 10 ; Dadylocalyx, 347.
Spain, Aphrocallides, 395 ; Sympiagella, 372.
Sperms, 24.
Sphierohexact, 48.
Sphffirohexasters, 31, 48.
Spicular sheath, 24.

Spicules, analysis, 28 ; anchor, 42, 132 ; appendicular,
132 ; arrangement, 40, 41 ; axes of, 29
birotulate, 193 ; broom-like, 34 ; clasps of, 78
comital, 68 ; compass, 80 ; cruciform, 1 86
curvature of, 30; extrinsic, 75, 110; flori-
como-hexradiate, 59 ; defensive, 71 ; formation,
499; lamination, 205; orientation, 502; porrecto
multiradiate, 307 ; position, 40, 41 ; raphide-
like, 81 ; rationalised, 501 ; spindle-shaped,
192; steUate, 109; structure, 27; union, 38.
Spongicolffi, Gray, 182.
Spongia odancyra, Brandt, 182, 184.
„ spinicrux, Brandt, 182, 184.
Sponge, dredging, 61.
Stalk, 169, 171, 172.
Stauradinella, Zittel, 14.
Staurodermidse, Zittel, 13.
Stauronemata (SoUas), 10, 347.
Steere's sponge {Sderothamnus), 337.
StephoscypJius mirahilis (commensal), 98.
St. Lucia, Eupiledclla, 77.

St. Thomas, Aphrocallistes, 305 ; Clionelasma, 327 ;
Fan-ea, 277, 288 ; Hyalonema, 201 ; Myliusia,
405.
St. Vincent, Aulocystis, 406.
Stutchbury, Dadylocalyx, 342.
Stylocalyx, n. subgen. (see Hyalonema), 211, 383.

„ apertum, n. sp., 214, 383 (Pk. xxxvii.,
xxxrai.).

REPORT ON THE HEXACTINELLIDA.

513

Sfylocalyx davigerum, ii. sp., 220, 384 (PI. xli.).

,, depressum, n. sp., 217, 383 (PLs. xxxv.,
xxsvi.).
elegans, n. sp., 223, 384 (PI. xxx.).
globus, n. sp., 221, 384 (PI. xl.).
tenemm, n. sp., 224, 385 (PI. xxxi.).
„ thomsoni, Marshall, 211, 383 (PI. xxxiv.).
Superimposed individuals, 62.
Superior funnel, 171.
Subdermal trabecular space, 19.

„ trabeculfe, 47.

Subgastral trabecular space, 1 9.

„ trabeculse, 47.
Survey of genera and species, 365.
Siiss, Hyalonema parallel um, 183.
Sympagella, O. Schmidt, 119, 372.

nuj:, O. Schmidt, 120, 372 (PI. xxii.).
SympageUinse, 119, 372.
Synapticula, 28, 38, 172.
Synauloidffi, 10, 39, 487.
Syrim/idium (0. Schmidt), 15, 302.

ziftelii (O. Schmidt), 347.
Txgeria, n. gen., 94, 369.

„ piidchra, n. sp., 94, 369 (Pis. vii., viii., xi.).
Tregerina, 94, 369.
Tahiti, Didyocalyx, 105.
"Talisman," 17.
Terminal rays, 47, 172, &c.
Terminal sieve-plate, 21, 47.
Tetilla polyura, 132.
Tetractinellida, 496.
Tetracts, 35.

Thomson, Wyville, 7; Aphrocallistes, 306; Asconema,
115; Corbitella, 101; Eupledella, 60; Farrea,
268; Habfodidyum, 101; Heterotella, 101;
Holtenia, 186, 235; Hyalonema, 186, 201;
Lefroyellci, 301 ; Pheronema, 241 ; Polio-
pogon, 254; Rossella, 137, 144.
Timor, Myliusia, 405; Sde.rofhamnus, 339, 402.
Tissue, 23 ; flake-like, 69.
Tradiycaulus, n. gen., 128, 373.

gurlitti, n. sp., 128, 373 (PI. xxvi.).
" Travailleur," 17.

Tretodictyidae, 327, 398; distribution, 438.
Tretodidyum = Hej-adinella, 328.

„ ryathus ( = Hexadinella ventilabrum), 328.

,, latum { = Hexadinella lata), 327.

Treiodidywn tubulositm ( = Hexadinella tubulosa),
328.

Triads, 35.

Triaxonia ( = Hexactinellida), 51.

TridiapteUa elegans, 17.

Tristan da Cunha, Caulocalyx, 172, 319 ; Farrea, 288;
Hyalonema, 232.

Tropical Zone, distribution, 445.

Tuhuliiia, Carter, 9.

Umbels, 36.

Umbelluluria grmnlandica, 352.

Uncinata, 37, 48.

Uncinataria, 266 ; bathymetrical distribution, 457,
458.

Vphantsenia daicsoni, 63.

Valenciennes, Hyalonema, 181.

Vaxella ( = Holtenia saccus), 237.

Ventriculites, 8.

Ventriculitidoe, 13.

Verticils, double, 207.

Vitrea, 7.

Vitreo-hexactinellida, 9.

Vitreous fibre, 27.

Volvidina, 0. Schmidt ( = Cyrtaulon), 332.
„ sigsbeei, 16.

Vosmaer, 3.

Walcott, fossil EuplecteUid, 63.

Walferia, n. gen., 96, 369.

„ Hemmingii, n. sp., 96, 369 (Pis. ix., x.,
xi.).

"Weltner, Ajjhrocallistes, 310; fossil forms, 16.

West Hebrides, Hyalonema, 213, 382.

West Indies, Aulocystis, 406; DadyJocalyx, 403;
Myliusia, 405 ; Sderoplegma, 401.

Witefield, R. P., fossil EuplecteUid, 63.

Willemoes Suhm, Eupledella, 61.

Wright, Perceval, Aphrocallistes bocagei, 306; Hyalo-
nema mirabiles, 186.

Young forms, Poliopogon, 256.

Young, M. and I., fossil Hyalonema, 188.

Zebu, Crateromorplia, 161, 378 ; Eupledella, 366
Euplectella aspergillum, 72; Etivete, 298
Hyalonema, 385 ; Plieronema, 246, 387
Polylophus, 133; Sclerotkammis, 337; Sem-
per ella, 261.

Zittel, Aphrocallistes, 309, 310; classification, 11:
Dadylocalyx, 347; phylogeny, 491.

POSTSCRIPT.

Balanites, Polyrhabdus and PolylopMs are preoccupied as generic names. 1 propose, therefore, to
substitute Balanclla for Balanites, Pleorhahdus for Polyrhabdus and Lophocalyx for Polylophtis.

tu

_J

•I?

^*;/;y ;.,

:-«.:.

'. >';""■»,-

'Ai

;;i^.

M

S'A

V!

■i-'i: