Li V

HARVARD UNIVERSITY

LIBRARY

OF THE
MUSEUM OF COMPARATIVE ZOOLOGY

GIFT OF

Linnean Soc. London.
1831-1889

THE

TRANSACTIONS

**-£

OF

THE LINNEAN SOCIETY

OF

LONDON.

SECOND SERIES— VOLUME II.

ZOOLOGY.

LONDON:

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

SOLD AT THE SOCIETY'S APARTMENTS, BURLINGTON-HOUSE.
AND BY LONGMANS, GREEN, AND CO., PATERNOSTER -ROW.

1879-1888.

S-E.S-L

CONTENTS.

PART I.— December, 1879.

I On the Genus Actinometra, Mull., with a Morphological Account of a New Species
(A. polymorph*) from the Philippine Islands. -Tart I. By P. Herbert Car-
penter MA, Assistant Master at Eton College. {Communicated by W. B.
Carpenter, C.B., M.D., BB.B., F.B.S., F.B.S.) (Plates I.-VIIL). . page 1

PART IL— March, 1881.

II. On some New Species of Nudibranchiate Mollusca from the Eastern Seas. By

Cuthbert Collingwood, M.A., M.B., F.B.S., 8fc. (Plates IX. & X. and Wood-
cut.)

III. On the Anatomy of Ants. By Sir John Lubbock, Bart., M.P., F.B.S. , F.L.S.,

B.C.B., LL.U., rice-Chancellor of the University of London, President of the
Entomological Society. (Plates XL & XII.) 141

IV . On the Extinct Band- Tortoises of Mauritius and Rodriguez. By Alfred C. Haddon,

B.A., Scholar of Christ's College, and Curator in the Museum of Zoology and
Comparative Anatomy of the University of Cambridge. {Communicated by Prof.
Newton, M.A., F.B.S.) (Plate XIII.) 155

PART III.— February, 1S82.

V. On the Morphology of the Skull in the Amphibia Urodelia. By Prof W. Kitchen

Parker, F.B.S., F.L.S. (Plates XIV.-XXI.) 165

VI. On the Tusks of the Fossil Walrus, found in the Bed Crag of Suffolk. By E. Ray

Lankester, M.A., F.B.S., F.B.S., Professor of Zoology and Comparative Anatomy
University College, London. (Plate XXII. and Woodcut.) 213

in

iv

J

PART IV.— March, 1882.

VII. The Parasites of Elephants. By T. Spencer Cobbold, M.D., F.R.S , F.L.S., fyc.
(Plates XXIII. & XXIV.) page 223

PART V.— July, 1882.

VIII. On the Digastric Muscle, its Modifications and Functions. By G. E. Dobson,
M.A., M.B., F.L.S. (Plate XXV.) 259

PART VI.— April, 1883.

[X. On the Clasping-Organs Ancillary to Generation in Certain Groups of the Lepi-
doptera. By Philip Henry Gosse, F.R.S. {Communicated by R. M'Lachlan,
F.R.S., F.L.S.) (Plates XXVI.-XXXIII.) 265

PART VII.— September, 1883.

X. On certain Points in the Anatomy of the Polynoina, and on the Polynoe (Lepidonotus,
Leach) clava of Montagu. By Alfred Gibbs Bourne, B.Sc. Loud., Univ.
Scholar in Zoology, and Assist, in the Zoological Laboratory, University College,
Loudon. (Communicated by Prof. E. Ray Lankester, M.A., F.L.S.) (Plates
XXXIV-XXXVI.) 347

PART VIII.— August, 1883.

XI. On Simondsia paradoxa and on its Probable Affinity icith Sphserularia bombi. By
T. Spencer Cobbold, M.lJ., F.R.S., F.L.S., Correspondent of the Academy of
Sciences of Philadelphia. (Plate XXXVII.) 357

PART IX.— October, 1883.

XII. On the Testis of Limulus. By W. B. S. Benham, Esq. (Communicated by Prof.
E. Ray Lankester, M.A., F.L.S.) (Plate XXXVIII.) 3G3

PART X.— April, 1884.

XIII. The Metamorphosis of Filaria sanguinis bominis in the Mosquito. By Patrick
Manson, M.I)., Hong Kong. (Communicated by Br. Cobbold, F.R.S., F.L.S.)
(Plate XXXIX.) 307

[ v ]

PART XI.— December, 1884.

XIV. On the Compound Vision and the Morphology of the Eye in Insects. By B.
Thompson Lowne, F.B.C.S., F.L.S., Lecturer on Physiology, Middlesex Hospital,
formerly Arris and Gale Lecturer, Royal College of Surgeons. (Plates XL.-

XLIII.) P a § e 389

PART XII.— November, 1885.

XV. Contributions to the Knowledge of the Genus Anaphe, Walker. By Lord Wal-

singham, M.A., F.L.S. (Plates XLIV. & XLV.) 421

PART XIII.— August, 1884.

XVI. On a new Species of Coelacanthus (C. Tingleyensis)/ro?ra the Yorkshire Cannel Coal.
5y James W.Davis, M.S., .F.G.S. (Plates XLV I. -XLIX.) 427

PART XIV.— January, 1885.

XVII. On three new Species of Metacrinus. By P. Herbert Carpenter, D.Sc,
Assistant-Master at Eton College. With a Note on a new Myzostoma, by Prof
L. von Graff, Ph.D. (Communicated by Dr. W. B. Carpenter, F.B.S., F.L.S.)
(Plates L.-LII.) 435

PART XV.— August, 1885.

XVIII. On the Breeding of Salmon from Parents which hare never descended to the Sea.
By Francis Day, F.L.S. (Plates LIU. & LIV.) 447

PART XVI.— October, 1885.

XIX. Golfingia Maclntosliii, a new Sipuncu lid from the Coast of Scotland. By~E. Ray
Lankester, M.A., LL.D., F.B.S., Jodrell Professor of Zoology in University
College, London, Fellow of Exeter College, Oxford. (Plates LV. & LVI.) . 469

PART XVII. — February, 1886.

XX. On the Variations in the Form of the Cirri in certain Comatula?. By P. Herbert

Carpenter, D.Sc, F.R.S., Assistant Master at Eton College. (Communicated by
W. Percy Sladen, Sec. Linn. Soc). (Plate LVII.) -475

PAP^T XVIII.— November, 1888.
Title-page, Contents, and Index.

second series. — zoology, vol. II.

TRANSACTIONS

OF

THE LINNEAN SOCIETY.

I. On the Genus Actinornetra, Mull., with a Morphological Account of a new Species
(A.) polymorpha from the Philippine Islands. — Parti. By P. Herbert Carpenter,
M.A., Assistant Master at Eton College. {Communicated byW. B. Carpenter,
C.B., M.B., LL.D., F.B.S., F.Z.S.)

(Plates I.-VIII.)

Read June 21st, 1877.

I HE principal part of the investigations, of which the results are set forth in the fol-
lowing pages, were carried on during the year 1876, in the Zool.-Zootomical Institute of
the University of Wiirzburg, under the superintendence of its director, Prof. C. Semper.
I would here express my most sincere thanks to Prof. Semper, not only for the
generous manner in which he placed at my disposal all his specimens of Act. polymorpha,
which he had himself collected in the Philippine Islands, hut also for the ready and
valuable advice which he constantly afforded me during the progress of my work, and
for the free use which he permitted me to make of his extensive library. I am also
greatly indebted to Dr. Sandberger, Professor of Geology in the University of vYiirzburg,
to the authorities of the University Library, and to Professor Dr. Halm, of the Royal
Library at Munich, for the means of reference to many books which would otberwise
have remained inaccessible to me ; and I desire to record my best thanks to all these
gentlemen for the ready kindness with which they met my frequent and numerous
wants.

I. Historical.
(§ 1) In the remarkable work of Linckius 1 upon the " Sea Stars," the modern family
Comatulidce (J. Miiller) is described under the name of Crinita), or Comatse Stelloe, as a
group distinct not only from the Asterida?, but also from the Ophiuridaa and from
Astrophyton {Euryale, Lamarck), with which they have been united by many later
systematists.

1 Johannis Henkici Lisckii Lipsiensis ' De Stellis Harinis liter singularis,' Lipsise, 1733, p. 53.
SECOND SERIES. — ZOOLOGY, VOL. II. 1

2 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

Liuck included three genera in this family, or, as he called it, " Classis." The first of
these he named Ae/caKvij/ioq, to indicate " stellam marinam decern caudis crinitis radi-
antem ; " and he referred to it three species : — (1) The " Crocea zaffarcma Neapolitanorum"
or 8e/caSa<TuaKT(>'oei£)1c of Fabius Columna x , whose description he quotes ; (2) the Decempeda
Cornubiensium of Llliuyd 3 , Avhich Linck figured and named " Stella Se/ca/ci')),uoc rosacea ; "
and (3) the " Ae«:aKv»j/uoc Jtmbriata Barrelieri " 3 , which was named by Liuck barbata, as
he supposed it to be different from the other two. All three, however, are really identical,
being simply local -varieties of one and the same species, viz. the British Antedon rosacea,
or the Comatula mediterranea of Lamarck. Thus, Fabius Columna described the
Neapolitan variety, and Barrelier another obtained at the mouth of the Tiber, while Llliuyd
based his description upon specimens found upon the coast of Cornwall near Penzance.

Linck's second genus, the T pianaileKUKv^oc, was based upon a specimen with thirteen
arms, previously described by Petiver 4 as " Stella chinensis ;" this specimen, however,
was suspected by Linck to have been mutilated. His third genus he called " Caput-
Medusce," and described it as including those specimens which " ex centro corporis parvi
umbonatique per quiuque truncos primum bifidi, mox nullo constanti nurnero multifidi,
in 60 et plurcs surculos geniculatos rectos simpliccs abeunt, quos gracilescentes fibrilkc
arise pilorum instar vestiunt."

Linck referred two species to this genus, viz. Caput Medusae cinereum and C. brnnnum;
and he gave good figures of both (tab. xxi. n. 33, and tab. xxii. n. 31), from which it
may be determined with tolerable certainty that they represent species now known to
belong to two different types among the Coinatuke — namely, to the genera Antedon
and Actinometra respectively.

(§2) Although Llliuyd 5 , and after him Bosinus 1 ', had explicitly pointed out the
relationship between the recent Comatulae and the fossil Crinoidea, and although Linck,
while supporting and repeating Llhuyd's views, had clearly differentiated the former
from the Asteroidca and Ophiuroidea, yet Linnaeus 7 , instead of adopting the more
correct views of some of his predecessors as to the true relations of the Crinoidea, was
so misled by the jointed structure of their stems as to rank them among zoophytes in
his genus Isis, whilst he grouped the Comatulidas, together with all the other Starfish,
under one common name Asterias. Linck's three species of Decacnemus were rightly
regarded by him as identical ; and he placed them, together with Petiver's Stella chinensis,
in one species, Asterias peclinata, to which he also referred a specimen previously
described by Betzius 8 . We now know, however, that this last is an Actinometra, dif-

1 Phytobasanus, sivc Planfcarum aliquot Historia. Neapoli, 1592.

- Eduaedi Luidi ' Lithophylacii Britanniei Icknographia ' p. 149. Londini, 1699.

'■' Jac ib] Bakeelieei ' Plantae per Galliam, Hispaniain ct Italiam observatrc ' Paris, 1714, p. 131.

4 ' Gazophylacium Natura: et Artis,' Londini, 1711 ; and also ' Aquatilium animalium Aniboincnsiurn Ioones ct
Nomina,' 1713.

5 Prsclcctio de Stellis marinis Oceani Brit, nee non de Asteriarum, Entrocliorum, et Encrinorum Origine, pp. 149-
155, Oxford, 1733.

c Tontaminis do Litliozois ae Lithopbytis olini marinis, jam vero subterraneis, prodromus ; sive de stellis marinis
quondam, nunc fossilibus, disquisitio. Hamburg, 1719.

7 'Cystoma Natural,' editio decima tertia (Lipsire, 1783), pars vi. p. 3100.
s Nova Acta, Stockholm, 17S3, p. 234, n. 12.

ME. P. II. CAEPENTEE ON THE GENUS ACTLXOMETEA. 3

fering very considerably from the type represented by Decacnemus (Antedon) rosacea,
although resembling it in only having ten arms.

In like manner both of Linck's species of his genus Cajiut-HIedusa?, the one an
Antedon and the other an Actinometra, were united by Linnaeus, together with a many-
armed specimen described by Retzius, into one species, Asterias maltiradiata.

(§3) For some years after the commencement of the present century the Linnean
nomenclature held its own, and the few species of recent Comatukc with which the
naturalists of that time were acquainted were described as different species of the
Linnean genus Asterias.

The first among the post-Linnean zoologists who recognized the claims of this form of
Sea Star to a distinct generic rank was De Ereminville \ who in 1811 presented to the
Societe Philomatique de Paris a " Memoire sur un nouveau genre de Zoophytes de
l'Ordre des lladiaires," to which he gave the name of Antedon. His definition of the
genus was as follows : — " Animal libre, a corps disco'ide calcaire en dessus, gelatineux
en dessous, environne de deux ran gees de rayons articules, pierreux, perces dans leur
largeur d'un trou central ; ceux du rang supcrieur plus courts, simples, et d'egale
grosseur dans toute leur longueur ; ceux du rang inferieur plus longs, allant en dimi-
nuant de la base a la pointe, et garnis dans toute leur longueur d'appendices alternes
egalement articules; bouche inferieure et centrale."

It is not very clear which of the two apertures on the ventral (or, as he called it,
inferior) side of the disk was regarded by De Freminville as the mouth ; it is very pro-
bable that, as he was only able to examine a spirit-specimen, he failed to recognize more
than one — that namely, which, placed at the extremity of a long tube projecting from
a point near the centre of the disk, we now know to be the anus.

Adams 3 , who had studied living specimens of Linck's Decacnemus rosacea, had, how-
ever, pointed out some years previously the existence of two orifices to the digestive
cavity ; but his observations seem to have escaped notice ; for Lamarck 3 , Miller i , and
many other naturalists, all regarded the aperture at the end of the anal tube either as the
mouth alone or as a combined mouth and anus ; and it was not till 1823 that the exist-
ence of distinct oral and anal orifices was fully recognized.

It was announced as a new discovery by Leuckart 5 , in a letter to Von Schlotheim ;
and he was followed shortly afterwards by Meckel 9 , Gray 7 , and Heusinger s .

1 Nouv. Bull. d. Scien. par la Soc. Philomat. torn. ii. p. 340. Paris, 1811.

2 " Description of some Marine Animals found on the Coast of Wales," Trans. Linn. Soc. vol. v. p. 7 (1800).

3 Systeme dAnimaux sans Verfcebres, 2 me e'd. (Paris, 1816), torn. ii. p. 532.

4 A Natural History of the Crinoidea (Bristol, 1821), p. 128.

5 Von Schlotheim, Naeht. z. Petrcfact. ALtli. ii. p. 48 (Gotha, 1S23); and Leuckart, "Einiges iiber Asteriden
Gesehlecht Comatula Lam. uberhaupt, und iiber Coi . a insbesondere," Zeitsch. fur organ. Physik, iii.
p. 385 (1833).

6 " Leber die Ocffnungen des Speisekanals bei den Comatulen," Meckel's Archiv fur Physiol. Band iii. p. 470
(1823).

7 " Notice on the Digestive Organs of the Genus Comatula and on the Crinoidea of Miller,'' Ann. of Philos. n. s.
vol. xii. p. 392 (1826).

8 " Bcmerk. iiber d. Verdauungskanal dcr Comatulen, Meckel's Archiv" fur Physiol. 182(5, p. 317; and "Anat. Uu-
tersuch. d. Comatula mediterranea," Zeitsch. fur organ. Physik, iii. p. 300 (1S:j3).

1*

4 ME. P. H. CAEPENTER ON THE GENUS ACTINOMETRA.

Dc Preniinville's specimen was found on the keel of a vessel which had come from a
warm climate ; it had ten arms and twenty cirrhi, and was named by him Antedon gor-
gonia. He gave no further description of it, but simply referred to the figure of Stella
decacnemus rosacea, Linck, in the ' Encyclopedie Methodique ' \ which represents the
ordinary European Comatula rosacea, as it is now called.

This species, however, is not identical with De Premiuville's Antedon gorgonia, which
was referred by Lamarck 2 to his Comatula carmata, under which name he described
some specimens brought by Peron from the Isle de Prance. Nevertheless the two
species resemble one another in some important points, viz. the presence of ten arms,
of a central or subcentral mouth, and of an excentric anal tube.

In 1815 Leach 3 rescued the three genera contained in Linck's classis CrinitcB
from the confusion of the Liunean genus Astcrias, and united them into one genus,
Alecto, comprising three species, viz. Alccto europcea (= Decacnemus rosacea, Linck),
Alecto horrida (= Caput-Medusoe, Linck, or Asterias multiradiata, Linn.), and Alecto
carinata (which seems to have been the same as De Premiuville's Antedon gorgonia).

Leach defined Alecto as having the " os inferius, irregulare," a description which would
suit equally well either for the true mouth or for the anal opening, though perhaps
more applicable to the former. He seems, however, like his predecessor De Preminville,
to have regarded the mouth as situated at the extremity of the anal tube ; for in the
explanation to Schweigger's figure 1 of Leach's specimen of Alecto horrida the latter is
described as the " rohrenformig hervorstehender Mund." It is obvious, therefore, that
we cannot make any use for systematic purposes of the definitions of Antedon and Alecto
as given by Leach and De Preminville respectively, as far as the position of the mouth
is concerned.

Schweigger's figure of the disk of Alecto horrida shows clearly enough that the five
trunks of the ambulacral grooves converge towards the centre of the disk, as in Antedon
rosacea (Alecto europcea, Leach), Plate I. fig. 1. Leach's Alecto horrida was therefore
a true Antedon in the modern sense of the term, although belonging to that division of
the o-enus in which the repetition of the bifurcation of the ten primary arms is carried
to a great extent.

(§4) Leach was apparently unacquainted with the memoir of Preminville; but the
same was evidently not the case with Lamarck (1S1G), who, like Leach in the previous
year, united Linck's three genera into one, to which he gave the very appropriate name
Comatula 5 . His definition of the genus differs but little from that given for Antedon
five years previously by De Preminville, whose original specimen Lamarck seems to have
examined ; and it is difficult to see why he did not adopt the name Antedon to designate
the genus, which, like Leach and De Preminville, he clearly distinguished as belonging
to a different type from the Asteridse, Ophiuridoe, and Eur y ale.

1 Partie des Vers, pi. 42. fig. G. 3 Tom at. p. 534.

3 Zoological Miscellauies, vol. ii. p. 61 : London, 1815.

4 Beobacht. auf naturhistorischen Reiseu, p. GO, and Taf. iv. fig. 42 : Berlin, 1819.

5 Turn. cit. p. 530.

ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 5

Laniarck included eight species in bis new genus Comatula. In only three of these is
the mouth at or near the centre of the disk, viz. C. mediterranea (= Stella decacnemus
rosacea, Linck), C. carlnata(= Antedon gorgonia, T?rem. ?), and C. adeonce.

In all the other five species described by Lamarck, viz. C. Solaris, C. brachiolata,
C. rotalaria, C. fimbrlata (= Stella chinensis, Petiver), and C. multiradiata (= Caput-
Medusce brunnum, Linck), the mouth (as I know from examination of the collection of
Comatula? in the Paris Museum, which still contains many of Lamarck's original spe-
cimens) is nowhere near the centre of the disk, which is occupied by the anal tube, but
is excentric, or even marginal. At the same time the five primary groove-trunks do not
converge towards the centre of the disk, as in Antedon rosacea (Com. mediterranea, Lam.)
(PL I. fig. 1), but they unite more or less completely into a horseshoe-shaped furrow, at
one point of which is situated the excentric mouth (PL I. figs. 2-5).

This type will be described further on, under the name of Actinomelra. Lamarck, who
found it in more than half of the species constituted by him, seems to have regarded
it as common to all ComatuLw His description of it is worth quotiug, as it is the first
notice of a true Actinometra that I have been able to find. lie says 1 : — "Le disque
iuferieur ou ventral offre un plateau orbiculahv plus large que le dorsal, entoure de rayons
simples, cirreux. Pres de la circonference de ce plateau, on apercoit un sillon irregu-
lierement circulaire qui s'ouvre sur la base des rayons pinnes, et se propage le long de
leur face inferieure, aussi que de celle des pinnules. Cc sillon neanmoins, ne s'approche
point de la bouche [i. e. the anal tube] et ne vient point s'y reunir, comme cela a lieu
pour la gouttiere des rayons dans les Asteries. Au centre du disque iuferieur ou ventral
des Comatules la bouche, membraneuse, tubuleuse, ou en forme de sac, fait une saillie
plus ou moins considerable suivant les especes."

Although Antedon and Alecto were both constituted previously to Comatula, yet
Lamarck's authority was sufficient to establish the latter name, and to bring it into
general use, though Cuvier adopted Leach's genus Alecto, and used it in preference to
Comatula. The latter, however, was more generally employed by all subsequent
observers (who pointed out Lamarck's error respecting the position of the mouth)
thenceforward till the time of Johannes Miiller.

(§ 5) During this period the skeleton both of the recent and of the fossil Crinoids was
made the subject of careful investigations by Miller 3 and Goldfuss 3 . The latter
author divided his class Stellerites into two Orders : — (a) Stilasteritse, or Gestielte
Seesterne ; and (b) Asterites liberi, or Preie Seesterne. The former he again divided into
Articulata and Inarticulata, placing in the first group a number of fossil Mesozoic
Crinoids, viz. Eugeniacvinites, Miller, Solanocrinites, Goldf., Pentacrinites, Encrinites,
and Ajuocrinites.

In his descriptions of these genera he adopted and considerably improved the some-
what inapplicable system of nomenclature, introduced by Miller for the parts of the
skeleton of the fossil Crinoids and of Comatula.

Portions of his diagnoses of Eugeniacrinites and Solanocrinites are of considerable
interest, both zoologically and morphologically. Of the former, he says 4 : —

1 Tom. tit. p. 532. ' ; Loc. cit. 3 Petrefacta Germanic, i. (Dusseldorf, 1826-35). 4 Turn. tit. p. 102.

6 ME. P. H. CAEPENTEE ON THE GENUS ACTIXOITETEA.

" Die kurze runde mit eiuem runden Kanale durchbohrte Saule besteht aus weuigen
walzigen verlangerten Gliedern, und nininit am obern Ende allmalig zu. Hire Gliede-
rung wird oft nur durcb Binge angedeutet, unten endet sie in starken Wurzelu. Das
letzte verdickte Saulenglied vertritt die Stellc des Beckens und articulirt durcb eine
Gelenkflache mit den Bippengliedern."

Tbe term " Becken " is bere meant to signify the circlet, of basals which, in Penta-
crinus and other stalked forms, intervenes between tbe stem and tbe circlet of first
radials, tbe " Bippenglieder " of Goldfuss.

These basals are also present in Solanocrinites, which genus, as Goldfuss well remarks,
" hat in verscbiedener Hinsicbt Aehnlichkeit mit den Pentacriniten, und bildet zugleich
einen Uebergang zur Gattung Comatula."

Its most important characters are as follows ! : —

" Die Saule ist sehr kurz, beiuabe so dick als der Kelcb, funfseitig und an ihrer
Basis nicbt mit Wurzelsprossen, sondern mit ausstrablenden B/unzeln versehen. Hire
Glieder sind mit einander verwachsen und haben an den Seitenfiacben Gelenkhohlungen
fur den Ansatz zahlreicher dicker Hiilfsarme. Die oberste Gelenkfliicbe der Saule zeigt
fiinf strablenformige Erhabenheiteii, auf welchen das Becken articulirt. Das Becken
wird nicht durcb das oberste Saulenglied vertreten, sondcrn es besteht aus fiinf Gliedern,
welche zwischen die Nahte der fiiuf Bippenglieder eingefiigt sind, oder sie bedecken.
Die fiinf Glieder des Beckens bilden entweder nur schmale Strahlen, die zwischen die
Nahte der Bippenglieder einsenken (S. costatus, S. scrobiculatus), oder sie sind breiter,
stossen seitlicb an einander, und stellen eine tiefe mit 5 Strahlenfurclien ausgehohlte
Gelenkflache dar (S.jcegeri)."

The first genus in Goldfuss's order " Asterites liberi" is tbe Comatula of Lamarck,
of which he says 2 , ' ; Diese Gattung bildet den Uebergang von den Stilasteriten zu den
freien Seesternen und steht zunachst mit den Solanocriniten in nachster Verwandtschaft."

Besides describing five fossil species, he gives some account of two recent ones,
dwelling more especially upon tbe structure and composition of the skeleton.

Thus "Bei der G. meditcrranea besteht die Saule aus drei Gliedern : das Becken fehlt,
und die Bippenglieder sitzen unmittelbar auf dem letzten Saulenglicde." "Bei der in
den Ostindiscben Meeren lebenclen C. multiradiata Lam. bingegen, finden sich Becken-
glieder, so dass man berechtigt sein konnte sie als eine eigene Gattung zu betracbten.
Ihr Saulenrudiment besteht aus einem einzigen schiisselformigen Gliede, an dessen
Bande fiinf schmale dreieckige Beckenglieder ansitzen und mit ihm verwachsen sind.
Diese stossen mit ihren Seitenfiachen nicbt an einander, sondern stehen so weit entfernt
dass die ersten Bippenglieder unmittelbar zwischen ibnen auf dem Siiulengliede ansitzen,
und sie durch einen Abscbnitt der unteren Ecke zwischen sich aufnehmen. In der
Mitte des innern unteren Ptandes jedes Beckengliedes entspringt ein zahnformiger
Eortsatz, der sich als knospeliger schmaler Streifen bis zum Mittelpunckte des Sau-
lengliedes verlangert, in welcbem er durch eine Binne aufgenommen ist. Diese Becken-
glieder sind also denen der Solanocriniten vollkommen analog."

(§ 6) Goldfuss, who, though acquainted with the name of Alecto, yet used Comatula

1 Tom. tit. p. 166. - Tom. at. p. 202.

ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 7

in preference to it, paid very little attention to the soft parts of either of the Comatula?
which he dissected.

In the following 1 year, however, De Blainville ' described the visceral mass at some
length. Like his predecessors, he adopted Lamarck's genus Comatula, making it the
only representative of his section, " Les Asterencrinides libres," while at the same time
he acknowledged the prior claims of de Ereminville's name Antedon. He was, of course,
acquainted with Lamarck's error respecting the position of the mouth, which he
described as " assez anterieure, isolee, membraneuse, au fond d'une etoile formee par cinq
sillons bil'urquc's." The species which lie dissected was a foreign one preserved in
spirit ; it had a large number of arms ; and from the not very clear description which he
gives of its ventral surface it would seem to have been a true Actinometra.

After speaking of the tentacular furrows on the ventral surface of the arms, he says 2 ,
" En suivant ces especes de sillons dont lc nombre est proportionnel a celui des digitations
du rayon, on arrive par un sillon unique pour chacun d'eux et qui en occupe la base, au
centre d'une sorte d'etoile a bords epais, franges, et par suite a la bouche qui est au fond.
L'etoile formee par la reunion des sillons des rayons n'est pas symetrique, c'est a, dire que
ses branches sont tres-inegales : les unes que j'appellerai les anterieures, etant bien plus
courtes que les autres, ou posterieures. II en est resulte que la bouche n'est pas au
centre de l'etoile, mais bien plus proched'un cote quede 1' autre : elle est assez difficile a.
voir au contraire d'un autre orifice, dont il va etre question, et que M. de Lamarck
paroit avoir pris pour elle. Elle est profondement enfoncee dans l'etoile des sillons : elle
est ronde, sans aucune armature et conduit immediatement dans l'estomac."

The above description implies, if I rightly understand it, that the mouth of De Blain-
ville's specimen was nearer to one side of the disk than to the other, so that the primary
trunks of the ambulacral grooves were of unequal lengths. This will be subsequently
seen (section 14) to be the principal distinctive character of the genus Actinometra.

De Blainville evidently attached no importance to the position of the mouth as a
character of systematic value in the determination of the species of recent Comatulce ;
and from his definition of it as " assez anterieure," it would almost appear as if he
supposed the other species to agree in this respect with the one dissected by him.

This is, in fact, the case in five out of the eight species described by Lamarck, with
which De Blainville was probably acquainted, and to which he added no new ones,
except that he gave the name of Comatula barbata to Linck's third species of Decaciiemus,
the fimbriata of Barrelier, or " barbata " of Linck. Lamarck had been uncertain to which
of his species he should refer it, although, as we have seen above (section 1), it is really
only a local variety of his C. medilerranea.

Like the other naturalists of his time (1S38), Agassiz 3 also adopted Comatula in preference
to the other generic names of this type, but defined it as having the " bouche centrale en-
foncee," and with the five " rayons du disque bifurques," thus limiting the number of
arms in the genus Comatula to ten only. At the same time he erected Lamarck's species
C. multiradiata, with sixty or more arms, into a new genus, Comaster, which he defined as

Manuel d'Aotinologie, (Paris, 1834) p. 2 19. : Op. cit, p. 251.

" Prodrome d'une Monographic des Eadiaires ou Echinodernies," Ann. des Scicu. Nat. 2 e serie, Zool. vii. p. -~>7.

8 ME. P. H. CAEPENTER ON THE GENUS ACTINOMETEA.

follows : — " Ce genre a la meme organisation que le precedent ; mais les especes ont les
bras ramifies au lieu de les avoir simplement fourchus." Agassiz consequently used
Comatula simply as equivalent to the Decacnemus of Linck, while his new genus Comaster
was Linck's Caput-Mediisce, or the Comatula multiradiata of Lamarck. Of the seven
other species constituted by the last-mentioned naturalist, only two, C. rotatoria and
C. fimbriata, have more than ten arms ; in both of these the number of arms is usually
twenty, though it may reach twenty-four, or possibly even more. Strictly speaking,
therefore, these two species, according to the above definition, should be referred to
Comaster, and not to Comatula.

This character, the number of arms, upon which Agassiz founded a generic distinction,
is, in fact, extremely variable, and by no means of generic importance ; in fact, as
Goldfuss 1 remarked a little later, " Wollte man mit Agassiz die Theilung der Arme
als hinreichendes Gattungsmerkmal ansehen, so wiirde man folgerecht gezwungen sein
fast jede Art der Crinoideen als Gattung aufzustellen."

Leach and Lamarck had already recognized this fact in uniting Linck's three genera
under a common name ; and it is not a little strange that Agassiz should have seen fit
to separate them again. His doing so, however, led to somewhat important consequences
from a systematic point of view. Turning to the fossil Comatula;, we find that Agassiz
erected the C. pinnata of Goldfuss into a new genus, Pterocoma, and grouped together
his other three species, C. tenella, C. pectinata, and C. filiformis, under the generic
name Saccoma ; while he expressed his belief thaf Solanocrinus was really related to
the Comatula;, and more especially to the problematical fossil described by Goldfuss
under the name of Glenotremites, which he rightly recognized as the centrodorsal piece
of a free-living Crinoid.

In the year 1810 a new fossil Comatula was described by Hagenow 2 under the name
of Hertha mijstica. The specimen, consisting of the united first radials and hemispherical
centrodorsal piece, was somewhat worn ; but Hagenow was able to recognize the resem-
blance between it and the remains of Solanocrinus, and the corresponding parts of
Goldfuss's Comatula multiradiata, except that he was unable to find any trace of the
external basals which Goldfuss had described in both the above cases ; and though he
seems, and (as we now know) correctly, to have suspected "das Vorhandensein etwa
verdeckt-liegender Beckenglieder," he was, of course, unable to come to any satisfactory
conclusion upon the point.

(§7) The year 18-10 is a noteworthy one in the history of our knowledge of the Crinoidea J
for it marked the appearance of the first of a series of classical memoirs by Johannes
Miiller, who laid the foundation of nearly all our knowledge of the zoology and morpho-
logy of the group : the first 3 of these was devoted to an anatomical account of the recent
and very rare genus Tentacrinus, together with many observations upon Comatula.

1 "Beitr. z. rctret'actenkunde," N. Acta Acad. Leop.-Carol. Nat.-Cur. xix.A. p. 348.

2 " Monogr. d. Riigen'schen Kreide-Versteinerungen, II. Abtheil. Eadiarien und Annulaten," Ncues Jahrb. Minc-
ralogie, 1840, p. G64.

3 " Uebcr den Bau des Pentacrinus ccvput-Medusce," Abhandl. d. Berlin. Akad. 1S43 ; Abstract in Monatsb. der-
selben, 1840 ; also in Wiegmann's Archiv f. Xaturgescb. 1840, i. p. 307.

ME. P. H. CAEPENTEE ON THE GENES ACTINOMETEA. 9

Miiller seems at first, not unnaturally, to have supposed that Goldfuss was right in
referring the rnany-armed specimen dissected by him to the Comatula multiradiata of
Lamarck, for which species he adopted Agassiz's name Comaster '. But he did not use it
precisely in the same sense as Agassiz, who, in his definition of the genus, makes no
mention of the external basals, the presence of wbicli was regarded by Goldfuss as the
principal character distinguishing Comaster from Comatula.

Miiller adopted Comaster 2 in the sense in which Goldfuss used the name ; and when
he subsequently discovered 3 that the Comatula multiradiata or Comaster of Goldfuss was
not specifically identical with the specimen described asComatitla multiradiata by Lamarck,
he retained the name Comaster for Goldfuss's specimen only, which, like Solanocrinus, is
remarkable for having " kleine basalia zwischen den Insertionen der Kelchradien, oder
sogennanten Beckenstucke welche den eigentlichen Comatulen ganzlich fehlen " 4 .

At the same time he gave a careful description 5 of Lamarck's original specimen of
Comatula multiradiata, based upon an examination of it by Troschel ; but as he regarded
Comaster and Solanocrinus only as one subgenus of Comatula, he gave it a new specific
name " multifula," on the ground that " die Comatula multiradiata Goldfuss, als die
zuerst genau beschriebene, den Speciesnamen multiradiata behalten muss." Lamarck's
specimen Avas thus restored by Miiller to its previous position among the " Comatulen
im engern Sinne, nainlich Gattung Alecto, Leach (Comatula, Lamarck)," which he
grouped together with Comaster into one family, Comatulinm. The fossil Solanocrinus
was regarded by him as identical with the latter form, while he referred the Hertha
mystica of Hagenow, and Pterocoma, Ag. (C. pinnata, Goldf.), to Comatula or Alecto;
for at that time (1841) he used the two names indifferently, considering them (as, indeed,
they originally were) equivalent to one another.

Goldfuss put forward about the same time a somewhat similar classification. 6 In a
subsequent abstract (with additions) of his ' Beitrage zur Petrefactenkunde,' [which had
been published two years previously (1839)] he speaks of the two species dissected by
him as "die Typen der zwei nachst bezeichneten Genera (Comatula, Comaster), welche
daher nebst den zwei zuletzt folgenden (Solanocrimtes, Gasterocoma) als Verzweigungen
des Lamarck'schen weiten Geschlechts Comatula zu betrachten sind." He did not, how-
ever, agree with Miiller in regarding Solanocrinites and Comaster as identical, partly,
apparently, because nothing was known of the arms of the former, and partly because of
the differences in the form of the " Knopf," or centrodorsal piece, which he called a
short stem — although, as Miiller showed, this is not a character of any generic value.

Although Goldfuss had at first supposed 7 that the basals were really absent in Coma-
tula mediterranea, and that the first radials therefore rested directly upon the top of the
centrodorsal piece, or, as he expressed it, on the last stem-segment, he seems subsequently
to have changed his opinion ; for in his definition 8 of the genus Comatula, given in 1839,

1 " Beitrage zur Petrefactenkunde," ho. cit. p. 349. " Wiegmann's Archiv, 1840, i. p. 309.

3 " Ueber die Gattungen und Arten der Comatulen," Wiegmann's Archiv, 1841, i. pp. 14o, 147.

4 "Bau des Pentacrinus," loc. cit. p. L'7. 5 Wiegmann's Archiv, 1841, i. p. 147.

6 Neues Jahrbuch fiir Mineralogie, 1841, p. 818.

7 Petrefacta Germanise, torn. cit. pp. 203, 204. 8 Beitrage, &c. loc. cit. p. 349.
SF.COND SERIES. — ZOOLOGY, VOL. II. 2

10 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

he says, " Auf deniletzten Saiilengliede ruhen fiinf Beckeuglietler, und auf jedem derselben
em Bippen- ( = second radial) undein Scliulterglied (third or axillary radial), auf welchem
zwei einfache Arme eingelenkt sind," from which it is evident that he was wrongly led
to regard the first radials as representing the basals of Comaster and JPentacrinus. This
mistake is hardly a surprising one when we consider the remarkable metamorphosis
undergone by the embryonic or primitive basals, and their concealed condition in the
adult Coinatula mediterrcmea.

Midler, who examined a very large number of species of Comatula, never found one in
which the basals appeared externally, as described and figured by Goldfuss in Comaster,
and remarked ' : — " Daraus geht hervor, dass die Gegenwart wirklicher Basalia ohne Zer-
legung bei einer lebenden Comatule, auch dann, wenn sic wirklich solche besitzt, schwer
zu erkennen sein muss. Die Unterscheidung der Comaster uud Comatula wird daher
bei der Ordnung der lebenden Comatulen unpractisch." In fact he appears to have
given up the genus Comaster altogether ; for he adds in a note : — " Kurzlich habe ich
die einzige im Museum zu Bonn befindliche Comatula mmltiradiata (nicht das von Goldfuss
zerlegto Exemplar, wovon ich nichts mehr vorfand) untersucht. Ich habe daran nichts
von Beckenstiicken erkennen konnen. Die Gattung Comaster ist daher wohl zu un-
terdriicken." He seems finally 2 to have thought that it might possibly be identical
with the C. Bennett I of the Ley den Museum. As, however, Comaster has not been seen
by any naturalist since the time of Goldfuss, its position must still remain in doubt.

(§8) Up to the time of Midler no one paid any attention, from a systematic point of
view, to the arrangement of the tentacular furrows on the ventral perisome of the disk of
Comatula ; but Lamarck and De Blainville had, as we have already seen, examined and
described, with more or less accuracy, a condition which we now know to differ very
considerably from that presented by the Decacnemus of Linck, or the Antedon of De
Freminville. Both these observers seem to have regarded the former condition as the
normal one, and as common to all Comatuhe. Midler, who does not seem to have been
acquainted with their descriptions (for he makes no mention of them), took up the subject
systematically, and soon discovered that, using the distribution of the tentacular furrows
as a basis of classification, he could distinguish two, as he thought, very distinct types
of the genus Comatula, which he named Alecto and Actinometra respectively. In his
earlier communications 3 on the subject he described the ordinary Comatula and Penta-
crinus as having a central mouth and symmetrically distributed tentacular furrows;
i. e. the five main trunks formed by the union of the furrows of the five groups of arms
converge directly towards the centre of the disk, being separated by five "interpalmar "
areas, one of which, slightly larger tban the rest, is occupied by the anal tube, which is
therefore excentric in its position (PI. I. fig. 5, An.).

During his visit to Vienna in 1810 Midler had an opportunity of examining an un-

1 'Uebcrdie Gattung Comatula, Lam., und ihre Arten,' Separatabdruck aus den Abhandl. Berlin. Akad. lS49,p. 8.

a Ibid. p. 29.

3 " Ban des Pentacrinus" loe. cit. p. 47, and Wiegm. Arcbiv, 1S40, i. p. 311.

ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 11

usually large specimen of the Comatula Solaris of Lamarck — unfortunately, however,
only a dry one, which he found to differ so greatly from the other Comatulce then known
to him, that he described it under the name of Actlnometra imperialis \ " welche gene-
risch von andern durch die Bildung ihres Scheitels verschieden zu sein schien. Auf
dem Scheitel der mit hlumenartigen Kalkhlattchen bedeckt ist, ist keine Spur von den
Furcken zu sehen, die bei den Comatulen von den Arm en zum Munde fiihren. Auch ist
dort nichts vom Munde zu sehen. Die Mitte der Bauchseite nimmt eine Ptohre ein.
Die Arme haben die ventrale Furche der Comatulen, die Furchen der zehn Arme miinden
aber in gleichen Abstanden in eine die Scheibe am Rande umziehende Cirkelfurche.
Diese eigenthumliche Bildung Hesse sich durch eine unsymmetrischc Vergrosserung
desjenigen Intertentacularfeldes, worin die Afterrohre steht, iiber den ganzen Scheitel
unci auf Kosten der anderen Intertcntacularfelder erklaren, so dass der Mund aus der
Mitte des Scheitels ganz an die Seite zwischen je zwei Armen gerath ; es ist mir aber
nicht gelungen den Mund hier zu finden." (PL I. fig. 2.)

In a subsequent visit to Lund, Muller examined two dry specimens of Comatulce, which
had been described by Eetzius 2 many years previously under the names of Asterias pec-
tlnata 3 and Asterias multlradlata. These he found to belong to the same type as the
Vienna specimen, which he had already designated Actlnometra imperialis, and which
he supposed to be distinct from the true Comatula Solaris of Lamarck. When he visited
Paris, however, in 1844, he examined Lamarck's original specimen of this species, and
convinced himself of its specific identity with his Vienna Actlnometra. Consequently he
withdrew the specific name "imperialis," and described the type simply as Actlnometra
solaris\ Muller was unable to determine the position of the mouth in the dry specimens
of the Lund aud Vienna Museums on which he founded his new genus Actlnometra; but
subsequently he was able to examine many spirit-specimens both of his typical species,
Act. Solaris, and also of other " Comatulen von jener Anordnung der Purchen, sowohl

zehnarmige als vielarmige Siehe die beistehende Pigur von Comatula JFahl-

berghil 6 " (PL I. fig. 3). This last species he describes a few pages further on as Coma-
tula (Actlnometra) JVahlberghll.

He did not, however, appear to regard the position of the mouth as of any systematic
importance ; for he goes on to say : — " Der Mund ist bei der in Frage stehenden Ab-

1 Wiegm. Archiv, 1841, i. p. 141 ; and " Neuo Beitr. z. [Eenntniss der Arten der G'omutalen," Wiegm. Archiv,
1843, i. p. 132.

2 Dissertatio sistens species cognitas Astcriarum. Lundas, 1805.

3 The specific name " peetinata" Linnasus, included both Retzius's specimens and the Decacnemus of Linck.
These belong, however, to two very different types of the genus Comatula, and must be carefully distinguished from
one another. The former is, as above mentioned, an Actinometra, while the latter was called Ahcto by Jfuller,
being simply the common Comatula mediterranea, Lam. Pennant, Adams, and others naturally employed the
Linnean name for this last species ; but Dujardin, following Eetzius and Muller, applied it also to Retzius's original
specimen, which is really an Actinometra, and not an Antedon, like C. mediterranea. This has given rise to much
confusion in the synonymy of these two species.

4 Wiegiminn's Archiv, 1843, i. p. 133. ° ' Gattung Comatula,' pp. 12, 13.
6 Ibid. p. 9.

12 ME. P. H. CARPENTER ON THE GENUS ACTINOMETEA.

weichung allerdings vorhanden, cr liegt ganz zur Seite, cloch ist dies niclit die Ursac h
des Unterschiedes, es gibt vielmehr auch Comatulen von der gewblmlichen Anordnung
der Purcben, Lei denen gleichwohl der Mund seitlich, die Afterrohre central steht. Pig.
von C. multiradiata (PL I. fig. 4). Die fragliche Abweichung beruhrt vielmebr darauf,
dass die fiinf Purcben nicbt symnietriscb fur die fiinf Gruppen der Arme vertbeilt
werden, sondern dass von den fiinf Purcben einzelne herrschend werden und Aeste an
die meisten Arme abgeben. Indein diese Hauptfurcben, nachdem sie die Scbeibe
umzogen, sicb wieder annahern, so entstebt der Scbein eines Cirkels. An in Weingeist
aufbewahrten Exemplaren siebt man indcss, dass es kein gescblossener Cirkel ist;" and
further on (p. 10) be says, " Icb werde daber bei den Arten wo fiinf centripetale Purchen
beobachtet sind, den NamenJfec/o in Klammer clem Gattungsnamen Comatula beifiigen,
wo aber weniger Purcbcnstiimme den excentrischen Mund erreichen, den Namen Acti-
nometra demselben Gattungsnamen Comatula folgen lassen. Also z. B. Comatula
(Alecto) europcca; Comatula (Actinometra) Solaris"

Miiller does not appear to bave been acquainted with De Prcminville's name of An-
teclon ; but be distinctly states tliat Leacb's genus Alecto was constituted a year earlier
than Lamarck's Comatula. He bad in bis earlier communications employed the two
indifferently and as equivalent to one anotber ; but wben be became acquainted with tbe
type represented by Comatula Solaris and elevated it into a new genus, or rather sub-
genus, Actinometra, in contradistinction to Alecto, be retained Lamarck's name Coma-
tula, probably on account of its being so well known, and employed it to designate the
genus in which he included the subgenera Alecto, Actinometra, and Comaster.

Thus the sole cbaracter by which Midler distinguished the first two of tbese subgenera
from one anotber was tbe number of groove-trunks reacbing tbe peristome, irrespective of
the position of tbe moutb. It is tberefore easy to understand that, as many of the spe-
cimens which he examined were dry, and as in otbers, although preserved in spirits, tbe
arms were contracted over tbe disk so as completely to conceal it, be was unable satis-
factorily to determine more than three species of Actinometra. Two of these, Act. Solaris
and Act. Wahlberghii, have been already mentioned; tbe tbird was tbe small Comatula
rotalaria of Lamarck.

(§9) The Aster ias pectinata of Retzius, which presented the same " Bildung des
Schcitels ' as Act. Solaris, resembled it so greatly in other respects, that Miiller
regarded the two as almost identical, or, at any rate, as presenting only varietal
differences 1 .

lie seems also to have come to the conclusion that the other species described by
Betzius, the Asterias multiradiata, Linn., had a prior claim to this specific name over
either of the similarly named types described by Goldfuss (Comaster) or Lamarck (Co-
matula multiradiata) ; for be described it as " Comatula (Alecto) multiradiata nob." 3 .
It is difficult to understand why he called it Alecto; for he had already 3 described this

1 'Gattung Comatula,' p. 52. - Ibid. p. 25.

3 Wiegniann's Arehiv, 18-13, i. p. 133, and ' Gattung Comatula,'' pp. 9, 10.

ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 13

Lund specimen as belonging to the type of his new genus Actinometra, in which not five,
as in Alecto, hut " weniger Furchenstamme den excentrischen Mund erreichen."

Midler further examined a specimen of the " so-called " C. multiradiata in the Bonn
Museum ; and although he did not actually include it in his type under that name, yet
he seems to have been inclined to do so ; for he says 1 that " es stimmt durch den Besitz
der Syzygien an den Axillaria der Anne mit Comatula multiradiata Betz., " hut adds,
" Maul excentrisch, 5 Burchen der Scheibe sammeln die Burchen der respectiven
Arme und kommenam Mund zusammen." (See Bl. I. fig. 4.) Here, again, it is evident
that Midler's description of Alecto will not hold good ; for according to his own descrip-
tions, the Lund and Bonn specimens of Comatula multiradiata, Mull., however much
alike in other respects, differ so greatly in the distribution of the ambulacra on the disk
that one is Actinometra and the other Alecto.

Miiller also referred three specimens contained in'the Baris collection to this type ; and
he was perhaps thinking of the condition of the ambulacra presented by them when he
added the following sentence to his previous description of the Lund specimen, and
named the type Alecto 2 . " Mund excentrisch, aber an Weingeistexemplaren ergibt sich
dass die fiinf zum Munde fiibrenden Burchen sich ganz symmetrisch fur die fiinf Gruppen
der Arme vertheilen." This arrangement, which he called the "gewohnliche Anordnung
der Burchen," had been already 3 figured by him as occurring in C. 'multiradiata, which,
as he says, differs from the ordinary C. mediterranea in the excentric position of the
mouth (PL I. figs. 1, 1).

It is thus evident that, according to Midler's own nomenclature, two types, differing
only in the "Bilduug des Scheitels," but almost precisely similar in every other respect,
viz. the Lund specimen, on the one hand, and the Baris specimens, on the other, were
referred by him to the same species, Alecto multiradiata, Miiller. It will, however,
be shown further on that the distinction drawn by Miiller between Alecto and Actino-
metra is not a real one, and that the Lund and one of the Baris specimens, both of which
have an excentric mouth and a central or subcentral anal tube, really belong to one and
the same species, Actinometra multiradiata.

(§ 10) Bor a short time after the publication of Midler's Comatula-memoirs the genera
Alecto and Actinometra remained as he left them, both being regarded as subordinate
types of Lamarck's genus Comatula.

A singularly minute fossil species, discovered by Bhilippi 4 between the valves of an
Isocardia cor from the Sicilian Tertiaries, was named by him Alecto alticeps because of
the height of its " Kelchstiick," a character found both in Alecto Eschrichtii and in
A. phalangium, as Midler had already pointed out. A few new fossil species of a more
or less doubtful nature have been since described, and variously referred cither to Mid-
ler's family Comatulinae or to new and distinct types.

The typical genus of this family, Comatula, Lam., has undergone numerous changes in
its definition. Bcemer, who at first revived Linck's name Decacuemus, subsequently

1 'Gattung Comatula,' p. 29. : Ibid. p. 26. 3 Ibid. p. 9.

4 " Alecto alticeps, n. sp., eine tertiiire Comatula-Art von Palermo," Neues Jahrb. fur Mineral. 1S44, p. 540.

14 MR. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

withdrew it in favour of Comattda 1 , the" Knopf," or centrodorsal piece of which was
described by him as an " Ueberrest der verkiimmerten Saule," while, like his prede-
cessors, he mentioned the absence of basals. The existence, however, of external basals,
both in Solanocrmus and in Comaster, led him to regard them, like Midler, as generically
identical ; and he used the name Comaster for this type in preference to Solanocrinus,
as it " bezieht sich nicht nur auf einen lebend und vollstandig bekannten Typus, sondern
druckt auch die Veiwandtschaft richtig, wie Solanocrinus unrichtig, aus." About the
same time D'Orbigny 2 , and, a few years later Pictet 3 , transferred the name Comatula to
this last-mentioned type, in which the basals appear externally ; while they revived
Linck's name Decacncmus (or, as they named it, Decameros) for the Antedon of De Fre-
minville and the Alecto of Leach. They characterized the genus as only differing from
Comatula, in their sense, in the total absence of the five basals, so that the radials
rest directly upon the centrodorsal piece. Fortunately, however, this peculiar inversion
of the nomenclature employed by Midler was not destined to last ; for in Bronn's
' Klassen und Ordnungen des Thierreichs ' 4 , all the above genera are united into one,
Comatula, which with Glenotremites and another doubtful fossil constitute the family
Comatulkke. Saccocoma and MarsupUes are restored to the places originally assigned to
them by Midler, in special groups, Costata and Tessellata respectively, among the
unstalked Crinoids ; Avkile Eugeniacrinus, which Goldfnss regarded as nearly related to
Solanocrinus, is placed with a few similar forms in a family Eugeniacrinidae, which, toge-
ther with the Pentacrinidae and Apiocrinida?, make up the group Articulata of Midler.

(§ 11) The family Cornatulidse was considerably enlarged a few years later by Du-
jardin and Hup6 s , who included in it, as D'Orbigny and Pictet had already done, not only
the tribes Comatuliens and Saccocomiens, but also the Eugeniacriniens, which both the
above authors had ranked among the stalked Crinoids, while MarsupUes, which they
referred to the Cornatulidse on account of its calyx being free, was transferred to the
Cyathocrinidae by Dnjardin, who could " ne voir qu'un caractere secondaire dans
l'absence d'une tige chez plusieurs de ces Crinoides." He distinguishes the three tribes
as follows : — "Nos trois tribus seront suffisamment caracterisees : la premiere, cede des
Eugeniacriniens, par son calice adherent ou pedoncule, jamais libre ; les deux autres,
dont le calice est libre a l'etat adulte, se distinguent parce que celle des Comatubens
porte cirrhes ou rayons dorsaux, dont la derniere, celle des Saccosomiens, est censee
depourvue."

The position of Eugeniacrinus does not concern us at present. Let us now investigate
the species included by Dujardin in the tribe Comatnliens. Under this head he ranks
three genera, viz. Comatula, Lam., Actinometra, Midi., and Comaster, Ag., using the
latter name in the sense in which it was employed by Midler and Rcenier, namely as
equivalent to Solanocrinus. Dnjardin's genus Comatula, however, is not precisely
equivalent to that of Midler, who included in it the two genera or subgenera Alecto and

1 Lethaea geognostica, iii ,c Auflage, 1851, Theil iv. p. 133, and Thcil v. p. 177.

- Cours element, de Paleontol. et de Geol. stratigraph. 1S50-1852, vol. '2, i. pp. 138, 139.

s Traite de Paleontol, (Paris, 1857) vol. iv. p. 288. 4 Band ii. Aktinozoen (1SG0), p. 233.

5 Hist. Nat, des Zoophytes, Echinodermes, (Paris, 1SG2) p. 186.

ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 15

Aclinometra. The latter was erected into a separate genus by Dujardin, who limited the
application of the name Comatula to those forms only which had been described as
Alecto by Midler — those, namely, in which five main groove-trunks reach the mouth,
irrespective of its position, to which Midler seems to have attached no importance as a
character of any systematic value ; so that Dujardin, following him, says of the mouth
of Comatula (i. e. Alecto) that it is only " ordinairement au centre " \

Further, Dujardin, though really employing the name Actinometra in the same sense
as Midler did, does not describe it in the same way ; he takes no account of the number
of groove-trunks reaching the peristome, to which Midler attached so much importance,
but simply says 2 , " Ce genre ne differe guere des vraies Comatules que par la position de
l'anus au centre et de la bouche au bord du disque. II en resulte que les gouttieres
ambulacraires, au lieu de se rendre a la bouche en suivant la direction des bras comme
chez les Comatules, s'inflcchissent et suivent le contour du disque." Dujardin adds,
with perfect truth, that the distinctive characters of Actinometra are hardly yet suf-
ficiently established. It will be shown, further on, that his definition of the genus is
really the correct one, and that we must refer to it all those forms of Alecto (Comatula,
Dujardin) in which, as described by Midler 3 , the anal tube occupies the middle of the
disk, " so dass der Mund seitlich gegen den Rand der Scheibe riickt, ohne dass die
Ambulacra ihre symmetrische Vertheilung auf die 5 Armstamme einbiissen " (PL I. fig. 4).
As Midler had only employed the names Alecto and Actinometra to designate sub-
ordinate types of the Comatula of Lamarck, it is rather unfortunate that Dujardin
should have erected the latter into a separate genus, in contradistinction to Comatula,
and restricted this name to the Alecto of Midler ; for we now know, as mentioned in
section 5, that most of the species described as Comatula by Lamarck belong really to
Actinometra, not only in the somewhat limited sense in which this name was used by
Midler, but also in its wider application as employed in this memoir.

Thus, for example, Midler stated expressly i that Lamarck's original specimen of
C. Solaris in the Paris Museum is identical with the large Vienna specimen, also bearing
the name of C. Solaris, Lam., which he made the type of his new genus Actinometra.

Dujardin, however, paid no attention to this identification of Midler's, and described
the two specimens as C. Solaris and Act. hmperialis respectively, simply on the basis of
Mailer's original diagnosis, published before his visit to Paris. Dujardin thus made
not only two different species, but also two different genera, out of the same type, while
he makes a third species out of the Asterias pectiuafa of the Lund Museum, which
Midler regarded as a variety of Actinometra Solaris.

I have examined a considerable number of specimens of this type, and find it to
exhibit an enormous range of variation in minor points, such as the number and relative
proportions of the cirrhus-segments, the colouring, the presence or absence of a faint
keel on the dorsal side of the arms, &c, and am convinced that none of these can be
regarded as of specific value. A number of such varieties group themselves around a

1 Op. tit. p. 104. 2 Op. tit. p. 208.

3 " Nachtrag zu der Abhandl. iiber die Comatulen," Monatsb. der Berlin. Akad. 1S4G, p. 177.

4 'Gattuug Comatula,' p. 13.

16 ME. P. H. CARPENTER ON THE GENUS ACTINOMETEA.

type possessing certain definite characters, by which it may he distinguished from other
types forming the centres of similar groups of varieties ; but the characters above men-
tioned are usually so excessively variable within each group, that it becomes utterly
impossible to make any use of them as specific distinctions, as Dujardin has done.

Dujardin seems to have detected Muller's oversight in classing the Asterias multi-
radiata of Retzius as an Alecto after previously describing it as an Actlnometra ; for he
transferred it to this genus under the name of Actlnometra multiracUata, and adopted
Muller's specific designation multifield for the original specimens described as Comatula
multiracUata by Lamarck. The third form to which this name has been applied, viz. the
C. multiracUata of Goldfuss, was regarded by Dujardin as a separate genus on account
of its possessing external basals, or, as he called them, " interradials ;" and he restored to
it the old name of Comaster, which had been given up by Midler, including in this genus,
as Roerner had previously done, all the species of Solanocrinus.

(§ 12) Muller had, as we have seen above, referred hoth Alecto and Actlnometra to the
one genus Comatula, while Dujardin limited the application of the latter name to the
species of Alecto only, and gave up the name of Alecto altogether, as had been previously
done by D'Orbigny and Pictet. This was a step in the right direction ; for, as Muller
had already pointed out, this name had been employed since 1821 to designate a section
of the Polyzoa established by Lamouroux. It is a pity, however, that Dujardin, instead
of limiting the application of Lamarck's name Comatula to the species of Midler's sub-
genus A lecto, did not revert to the old name of Anteclon, which was proposed by his
countryman De Ereminville in 1811, and had since received but little notice. This step
was taken by Mr. Norman l a few years later. He did not, however, use Anteclon as
precisely equivalent to Alecto, but applied the name to those forms only in which the
mouth is central and the anus lateral ; and he has been followed by nearly all the sub-
sequent writers upon the Crinoids.

The etymology of Anteclon is somewhat obscure. De Preminville described his typical
species as Anteclon gorgonia, which gives no information as to the gender of the name.
Mr. Norman, however, arrived at the conclusion that it is masculine, and hence
described the common British species as Anteclon rosaceus. In this respect all the later
writers have agreed with him with the exception of Pourtales 2 , who employs Anteclon
as a feminine name; and in this step he has since been justified by the result of the
recent etymological researches of Mr. Spedding 3 .

It will be used in the same manner in the following pages, both because this seems to
be ctymologically correct, and for the sake of convenience ; since, as long as Muller's
system of trinomial nomenclature is employed for the Comatula, it is far simpler to
write Comatula (Anteclon) rosacea than Anteclon ?°osaceus= Comatula rosacea. In any
case, we are now acquainted with so many different types, Anteclon, Actlnometra, Co-
master, Phanogenia, and 02)hiocrinus, to all of which Lamarck's designation Comatula

1 " On the Genera and Species of the British Echinodermata," Ann. and Mag. Nat. Hist. ser. 3, vol. xv. p. 98.

2 Bull, of the Mus. of Comp. Zool. vol. i. No. 6, " Contributions to the Fauna of the Gulf Stream at Great Depths,"
p. 11 1 ; and Xo. 1 1, " List of the Crinoids obtained on the Coasts of Florida and Cuba in 1867, 18G8, 18GSJ," p. 355.

J ' Nature,' vol. xv. p. 366.

ME. P. II. CARPENTER ON THE GENUS ACTINOMETRA. 17

is equally applicable, that this last can only he used to designate the family, while one
of the two names, Antedon and Alecto, which have precedence over it in point of time,
has gradually become more limited in its meaning, and the other has ceased altogether
to be applied to the Crinoids.

II. On the Characters of the Genus Actinometra.

(§ 13) We have seen that while the distinction drawn by Midler between Alecto and
Actinometra depended upon the number of groove-trunks reaching the peristome, irre-
spective of the position of the mouth, the genus Antedon, as defined by Mr. Norman,
and as subsequently used, is distinguished by having the mouth central and the anus
lateral.

There are, however, numerous species of Alecto in which, according to Miiller 1 , " die
Afterrohre nimmt die Mitte der Scheibe ein, so dass der Mund seitlich sesen den Rand
der Scheibe riickt ohne dass die Ambulacra ihre symmetrische Vertheilung auf die fiinf
Avmstamme einbussen." These forms have obviously no place in the genus Antedon,
while they were excluded from Actinometra by Miiller, who goes on to say, " In
andcrn abweichenden Arten geht die gleiche Vertheilung verloren, indem der excentriseh
liegende Mund weniger als fiinf Furchen der ambulacra aufnimmt, dann werden
einzelne dieser Furchen herrschend unci verasteln sich, indem sie einen grossen oder den
grossten Theil der Scheibe umziehen, auf mehreren Armstammen zugleich, so dass die
Scheibe von einem Furchenkreis umgeben ist, der jedoch an einer Stclle nicht gcschlossen
ist (Actinometra) (PI. I. figs. 2, 3, 5).

In PL I. figs. 6-16 is represented the distribution of the groove-trunks or ambulacra
on the disks of the eleven specimens of Act. polymorpha which I have been able to
examine. A glance at these, no two of which arc alike, will suffice to show that within
the limits of one and the same species there may occur individuals, some of which would
have been referred by Miiller to Alecto, some to Actinometra, and some which, strictly
speaking, have no place in either of these genera.

Thus, for example, the specimen represented in fig. 16 would probably have been
classed as Alecto by Miiller ; but although five groove-trunks leave the peristome, their
branches are by no means equally and symmetrically distributed to the different arms.
On the contrary, one of them gives off far more branches than any of the others, sup-
plying all the arms borne by two radii (I^-Eo), together with half of those of another
radius (AJ ; while another trunk running straight, out from the peristome bifurcates but
once, and only supplies two of the arms of one radius (Bo). Again, in fig. 15, only four
groove-trunks leave the peristome, one of which gives off a large number of branches, as
in fig. 16, also supplying all the arms of two radii (D^Ea). According to Midler's system,
therefore, this individual is an Actinometra .

In all the other nine specimens of this species which I have examined, however, there
are invariably more than five groove-trunks running out from the excentric peristome
(PL I. figs. 6-11). Even in the small specimen with thirteen arms, represented in fig. 6,
there are six groove-trunks, while in fig. 11 there are eight, and in all the other figures

1 Berlin Monatsberichte, 18iG, p. 177.

SECOND SERIES. — ZOOLOGY, VOL, II. 3

18 MB. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

either six or seven. In no case are there only five with their branches so regularly
distributed as Midler figured them in Alecto multiradiata (PL I. fig. 4) ; nor in the single
individual with only four primary trunks (fig. 15) is the distribution so regular and
symmetrical as in Midler's figure of Actmometra WahlbergMi (fig. 3). Farther, the dis-
tribution of the ambulacra on the disk of the specimen of Act. Solaris, represented in
PI. I. fig. 5, is by no means so symmetrical as Miiller found it to be in the large Vienna
specimen wbich he made the type of his new genus Actmometra (fig. 2). It can hardly
be said of fig. 5 that the " Furchen der zehn Arms miinden in gleichen Abstiinden in
eine die Scbeibe umziehende Cirkelt'urche."

The above instances, which could be multiplied indefinitely, suffice to show the
impossibility of classifying the Comatulce according to the distribution of the ambulacra
on the disk. We have already seen (sect. 9) that Miiller found the Lund and Paris
specimens of his species Com. multiradiata to agree in every respect but this ; so that, had
he adhered strictly to his own system of classification, he would have had to refer the
former to Actinometra and the latter to Alecto. In this case, however, as in all the
specimens represented (PL I. figs. 2-16), there is one point of agreement, viz. the relative
positions of the mouth and anal tube. In the Paris, Bonn (fig. 4), and Lund specimens
of C. multiradiata, Mull., in both the specimens of Act. Solaris, represented in PL I.
figs. 2, 5, in A. Walhberghii (fig. 3), and, lastly, in all the eleven specimens of A.polymorpha
(figs. 0-16), the centre of the disk is occupied by the anal tube, and the mouth is situated
excentrically, either close to the margin of the disk (fig. 11), or at some point rather
nearer to the centre.

(§ 14) After arriving at the conclusion that in this character, the central or excentric
position of the mouth, lies the real distinction between Antedon and Actinometra, and
that the number of groove-trunks reaching the peristome is a character of very minor
importance, I wrote to Dr. Liitken, of the University Museum, Copenhagen, upon the
subject, and was not surprised to learn that he had held this opinion for some time past.
With his usual kindness he has permitted me to make use of the following extract from
an unpublished MS. of his, containing descriptions of some new species of recent
Comatulce : —

" One of the reasons why it is so difficult to identify Midler's species is, that he does
not always mention the positions of the mouth and anal tube, and the direction of the
ambulacra on the disk, but has evidently established a somewhat unnatural distinction
between the differences which may occur in these characters. Two cases may occur : in
the one the mouth is subcentral (' quite central ' probably never occurs), and the ambu-
lacral farrows convennni? from the arms unite into five trunks, which all run directlv
towards the mouth along the shortest line ; they differ, therefore, but slightly in length ;
and the ' interpalmar ' areas defined by them are of almost ecmal size, that containing
the anal tube having sometimes, however, a slight preponderance in size, especially when
the anal tube is placed close to the mouth, almost centrally. In the other case the
mouth is removed towards the margin of the disk ; and of the ambulacra, those only which
come from the arms nearest to the mouth run directly towards that orifice, while the
others, and especially the two enclosing the anal area, are obliged to make a large

MB. P. H. CAEPEXTEE ON THE GENUS ACTLXOATETEA. 19

deviation, and reach the mouth, after a circuitous course, parallel to the margin of the
disk. It is often difficult to state the number of amhulacral furrows abutting ou the
mouth, as they frequently unite immediately before reaching it ; in different species, and
in different specimens of the same species, I have counted 4, 5, 6, 7, 10 stems ori-
ginating from the mouth ; this difference, therefore, is of no importance at all. The
anal tube in all these species [Actinometra, mihi] has a central or subcentral position,
and the anal area occupies the larger portion of the disk. Using this difference as
* fundamentum divisionis,' I have never encountered a doubt whether any type should
be referred to Antedon or to Actinometra, although I have examined a great number of
specimens and species. Moreover the lower or oral pinnules of Actinometra arc always
very different from the others, being flagelliforrn and presenting a more or less distinct
serrature or comb (pinnuloe oralcs prehensiles) ; while in Antedon they are only slightly
differentiated from the others, or are transformed into strong rigid spines, forming a
protective covering over the disk \_A.protectus, mihi]. It will, perhaps, be thought
improper to elevate these sections into genera, as the fossil Antedons would usually not
be generically determinable 1 ; but they are at least very good subgenera for the distri-
bution of the numerous species. The mode of classification here proposed is concerned
with the main point of that established by Midler, but is evidently an amelioration of
it. In Actinometra he describes the five ambulacra as partially uniting before reaching
the mouth, so that their number becomes reduced to three or four [C. Solaris, PI. I.
figs. 2, 5, and A. Wahlberghii, PI. I. fig. 3], while he refers to Alecto all those specimens
in which five ambulacra separately reach the mouth, even though this orifice be quite
excentric and marginal, and the length of the ambulacra therefore esceedin^lv different,
as in Alecto multiradiata [PI. I. fig. 4]. This mode of distinction used by Midler is,
however, very unnatural, and often quite arbitrary or illusory. It is the marginal or
subcentral position of the mouth that is of importance; and this character is never
ambiguous. C. multiradiata [PL I. fig. 4] is not less a true Actinometra than A. soh
[PL I. figs. 2, 5] and A. Wahlberghii [PL I. fig. 3]."

It will be seen from the above note that Dr. Liitken considers Antedon and Actino-
metra as two subgenera of Comatula, Antedon having a subcentral mouth and but slightly
differentiated oral pinnules, while in Actinometra the mouth is excentric and the oral
pinnules bear a terminal comb (PL III. figs. 1-3). At the time I received Dr. Liitken's
note I had had no opportunity of examining any large collection of Comatula?; and his
statement that the oral pinnules of all Actinometra were marked by a terminal comb
was therefore new to me. I have since been able to examine a considerable number of
Comatula?, and, like Dr. Liitken, have never had the least difficulty in determining to
which type any given specimen shoidd be referred, while at the same time I have always -
found that in Actinometra, or Comatula? with an excentric mouth, the oral pinnules bear

1 It -will be shewn further on that there are very decided differences in the shape of the calyx in the two genera
Aatedon and Actinometra. These render the determination of fossil Comatulai (and also of recent specimens from
which the disk is lost) less impracticable than Dr. Liitken supposes.

2 The above passage was written early in 1877. Since then I have examined the large collection of I
brought home by the Challenger.' Out of nearly fifty species with an eccentric mouth, all but two have a terminal
comb on the oral pinnules.

3*

20 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETBA.

a terminal comb. These two characters, however, do not always coexist; forPonrtales 1
describes Antedon meridionalis, A. Ag., as having an excentric mouth, while he says
nothing about the first pinnule, except that it is "rather long, the first five or six joints
webbed by the perisome."

Again, in many of the Comatula with an excentric mouth which I have examined,
the terminal comb is not limited to the oral pinnules only, but may occur at intervals on
different pinnules till near the end of the arms, although it is never so well developed as
it is on their basal or oral pinnules, fewer of the terminal segments bearing the processes
which go to make up the comb.

Loven 2 has found the same to be the case in the new Comatula which he has described
under the name of Phanogenia typica. Speaking of the pinnules, he says, on p. 232 : —
" In nonnullis (omnibus ?) articuli 8 1. 9 ultimi convoluti, pectiuati, margine cujusvis ex-
terno in laminam lanceolatam niagnam erectam producto ; " but he goes on to say (p. 233).
" Os centrale. Tubus analis crassus in media area interradiali. Sulci tentaculiferi fere
quales in Antedone."

Here, therefore, a terminal comb on the pinnules coexists with a central mouth ; so
that all the four possible variations may occur of these two characters, viz. the position of
the mouth aud the condition of the terminal segments of the pinnules.

Thus Antedon rosacea Sac. have a central mouth and no comb.
„ Phanogenia tijpica lias a central „ and comb.

„ Actinonielra Solaris kcAmvean excentric „ and comb.
„ Comatula meridionalis lias an excentric ,, and no comb.

Leaving Phanogenia out of consideration for the present, as it was unknown to
Johannes Midler, the following scheme will represent the relations of Antedon, and Act i-
nometra as used by Dr. Liitkeu and myself, to Alecto and Actmometra as used by
Midler :—

Alecto. Ambulacra symmetrically | mouth central. Antedon. Oral pinnules not specially

distributed on the disk. J mouth excentric. "^ distinguished.

Actinomctra. Ambulacra unsymmetri- ] fActinometra. Oral pinnules nearly

cally distributed on the l mouth excentric. | always have a ter-

disk. J minal comb.

(§ 15) "We are now in a position to investigate which species of the numerous Coma-
tulcB described by Midler can be referred to Actinometra under its new definition, and
what further subdivisions of the genus are possible according to the principles of classi-
fication introduced by Midler. Before doing so, however, it will be advisable to devote
a little time to a consideration of the descriptive terminology which he employed, and
of the manner in Avhich it has been modified by later writers.

In Pentaerinus and Comatula Miiller regarded the arms as starting directly from the
five radial axillaries. The two primary arms borne by each of these might either remain

1 hoc. cit. No. 11, p. 355. Pourtales bcre evidently uses Antedon as equivalent to Alecto, and not in the sense in
■which it was employed by Mr. Norman, viz. to designate those forms only in -which the mouth is central (or nearly
so) and the anus lateral. To avoid confusion, I shall speak of this species simply as Comatula meridionalis.

■ " Phanogenia, ett hittills okiindt slagte af fria Crinoideer," Ofver. af Kongl. Yetensk.-Akad. Forhandl. 1SGG,
No. 0, p. 223.

ME. P. II. CAEPENTEE ON THE GENUS ACTINOMETEA. 21

simple, as in Ant. rosacea, or divide more or less frequently into secondary, tertiary,
&c. arms, as in Act. multiradiata and in Pentacrinus; and every segment, like the
radial axillary, preceding- a bifurcation, was called by Muller a "bracbial axillary."

In some of the Tessellate Crinoids, however, the arms do not become free at the radial
axillary, but " der Kelch setzt sich noch weiter fort ; die lladien zerfallen dann in zwei
Distichalradien mit radialia distichalia, die jedes mit einem distichale axillare endeu,
wie bei Actinocrinus moniliformis und Eucalyptocrinus " \ In this case the distichal radii
represent the primary arms of Comatula and Pentacrinus, though Muller never used
the name " distichals " in his descriptions of the species of Comatula; for, as in the
Tessellata the segments composing two adjacent distichal radii are united laterally with
one another by intermediate plates, he regarded them as forming a part of the calyx,
and considered the arms of this group as starting from the distichal axillary, and not
from the radial axillary, as in the Articulate Crinoids.

The two primary arms, or distichal radii, borne upon a single radial axillary, were
called by Muller a " distichium ; " and the interval between two successive distichia
dorsally between the calcareous segments, or ventrally between the corresponding
grooves on the disk, was spoken of by him as " mfcrpalmar,' 5 while the interval between
the two primary arms or distichal radii borne by the same radial axillary, or, as Muller
called it, "die Kluft eines Distichiums," was " interbrachial " or " intrapalmar."

The words " interambulacral," " interradial," and " intertentacular," have been also
used by Muller and others to designate the interpalmar areas on the disk of Comatula.
Either of these is preferable to " interpalmar," for reasons which will presently appear,
though " intertentacular " is not universally applicable, as in certain Actinometrce the
posterior ambulacral grooves bounding the large area in which the anal tube is situated
are not provided with tentacles at their sides.

The term "interbrachial" is decidedly preferable to "intrapalmar," which was used
by Muller to designate the small areas on the disk, bounded by the two branches of
each of the five primary groove-trunks. " Intrapalmar " does not convey any clear idea
of the relation of these areas to the divisions of the skeleton, while "interbrachial" dis-
tinctly indicates that they correspond to the intervals between the two primary arms
borne by every axillary radial.

Reenter 2 adopted Muller' s nomenclature for the fossil Crinoids, and, like him, con-
sidered that the distichal radii, when present and united laterally to one another, formed
a part of the calyx ; so that the arms were regarded by him as commencing from the
axillary distichals, while he distinguished their different divisions simply as rami of the
first, second, and third order. De Koninck and Le Hon 3 , however, regarded the arms
as commencing from the first bifurcation, i. e. from the axillary radial, whether they
become free at once or whether they remain united with the calyx for a longer or shorter
distance. Nevertheless they distinguished the arm-segments by different names in the
two cases, using the expression " pieces brachiales " for the distichals of Muller, i. e. for
those segments which are immovably united with the calyx, while they gave the name
" articles brachiaux " to the movable segments, the brachials of Muller.

1 Baudes Pentacrinus, p. 31. " Lethoea geognostica, Band i. Theil 2, pp. 210, 215.

3 Recherches sur les Crinoides du Terrain Carbonifere de la Belgique, (Bruxelles, 1854) p. 69.

22 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

This view, although unquestionably correct in the case of the Articulate Crinoids
( Comatula, Pentacrinus, &c.), is, as Roenier has pointed out, beset with some difficulties
in its application to the fossil Tessellata ; and Schultze x accordingly reverted to the
original view of Miiller, saying, " Die Arme (brachia) beginuen unveranderlich da, wo
eine dcutliche Gelenkfacette eines festen Kelckstiickes, ihren TJrsprung anzeigt." In
describing the divisions of the arms, he speaks of the brachial axillaries of the first,
second, and third order, without giving them any special names. These are perhaps
scarcely necessary when the number of segments between each division varies so much
in different specimens and in different arms of the same specimen as it does in many
fossil Crinoids, and in Pentacrhtus. Among the Comatulce, however, the number and
character of the segments between the successive divisions of the arms exhibit variations
which are to a certain extent constant in different species, and thus give us the means of
classifying them into larger or smaller groups.

Miiller has availed himself of this character to a certain extent in the scheme which
he gives 3 of a classification of the Comatulce according to the presence or absence of
syzygia in the various brachial axillaries ; but though, in his descriptions of the different
species, he furnishes the material for carrying this classification much further, and for
separating species which, in his scheme, stand very near to one another, he never made
anv use of it, simply classifying the Comatulce in the groups which he had constituted,
according to the number of their arms — 10, 20, 40, or more. Under these circumstances
he woidd have been puzzled where to place Act. polymorphs, in which I have found the
number of arms to vary from 13 to 39.

(§ 16) It has been already stated that the arms proper of Comatula begin from the
radial axillaries. In many cases they are united by perisome as far as their second or
third division; and in Act. multifida this perisome contains numerous small calcareous
plates, which render the union of the arms with one another and with the calyx some-
what firmer than usual ; but they are never so united as to be immovable, as their
various segments are connected with one another, except, of course, at the syzygia, by
muscles and ligaments. There is one point about the nature of this union which has
not, I think, received sufficient attention ; and as it shows clearly that the arms of
Comatula and Pentacrinus begin from the radial axillaries, it is worth considering here.
It is this : the first and second segments beyond every axillary, whether radial or
brachial, are nearly always united together in the same manner as the second and third
(axillary) radials.

Thus, for example, in Act. Solaris, and in the forms allied to it, the second and third
radials are united by a syzygy. The same is the case with the first and second brachials.
In Ant. rosacea, and in the various species which are closely allied to it, there are no
muscles between the second and third radials ; but their opposed articular faces
present a vertical and not a transverse ridge, and are so united by ligament that the
two segments are only capable of a lateral movement upon one another, and cannot take
part in any movements of flexion or extension, in which they act as a single segment
only. The first and second brachials are united in precisely the same manner.

1 Monographie der Eclrinodermen des ELflerkalks, (Wien, lSU(i) p. 5.

2 Gattung Comatula, p. 11.

ME. P. H. CAEPENTEE ON THE GENES ACTINOMETEA. 23

In both these groups the primary arms do not subdivide ; so that the total number of
arms is limited to ten ; and we are as yet unacquainted with any Comatula in which the
second and third radials are united by a syzygy and there are more tban ten arms 1 . This
is, however, the case in Pentacrinus Millleri, in which, in like manner, the first and
second segments beyond every brachial axillary are also united by a syzygy 2 . On the other
hand, Pentacrinus asteria, L. (=P. caput-Medusce, Midler), is remarkable for having
muscles between the second and third radials as well as between the first and second 3 .
In the same manner the first and second segments beyond every axillary are united
by muscles, and the syzygium is between the second and third segments '.

In nearly all the Comatulce with which we are acquainted, with the exception of Act.
Solaris and the species most nearly allied to it, the second and third radials are united
by ligament only, as in Ant. rosacea, their opposed faces being marked by a vertical
articular ridge (PL VII. figs. 2 b, 3 a, 5 b, 6 a, i). In almost all of these species which have
more than ten arms, the first and second segments beyond every axillary are united by
ligament only, just like the second and third radials 5 . Thus in Ant. Savignii every third
segment, so long as the division lasts, is an axillary, and the first and second segments
beyond each axillary are united by ligament only. But in Ant.palmata only two segments
follow each bifurcation, the second of which is again axillary ; it is nevertheless united
to the first one by ligaments only. I have found these same two conditions to occur
together in Act. polymorpha (PI. II. fig. 8), in which the normal number of segments
between every two points of division is three (PI. II. figs. 7, 9, 10), of which the third
is axillary Avith a syzygiuin, as in Ant. Savignii, while the second is united to the first by
ligament only. In exceptional instances, however (PL II. figs. 8, 11), the second segment
may be axillary, and united to the first by hgaments only, as in Ant. palmata.

In every case, after the division has ceased, the union of the first and second brachials

1 There are three Comatula in the ' Challenger ' collection which answer to this description. In two of them the
first and second distichals and the first and second brachials are united by syzygies, like the second and third radials.
But in the third species there is a curious exception to the rule. The rays may divide eight times : and in the primary
divisions there are three distichal joints, the first two of which are united by ligaments and not by syzygy. But in
all the subsequent divisions the first two joints beyond each axillary form a syz3'gy, like the second and third radials.

" This agreement between the mode of union of the second and third radials, and of the first and second brachials
respectively, is seen also in Encrinus moniliformis, in which these segments are united by syzygia as in P. MiiUeri.
See ' Fetref. Germ.' Taf. liv. figs. F, G. The same is the case in Rliizocrinus (Sars, ' Crinoides vivants," pp. 1-3, 22).

3 P>au des Pentaermus, p. 30, and Taf. ii. fig. 8.

4 This does not appear, however, to be always the case ; for iliiller described the syzygium as uniting the first and
second arm-segments in the specimen examined by him, while I have found the same to be tho case in a specimen of
this species contained in the Zoological Museum of the University of Wiirzburg, in which there is certainly no
syzygial union between the second and third radials. In Pentacrinus the opposed faces of two elements, which are
united by a syzygium, are simple, and not radially striated as in Antedon. Sars has found this to be the case in
Bhizoorinus also ; but in its predecessor, Apioerinus obeonicus, Goldfuss, the radial striation of the syzygial surfaces
is very distinct (Petref. Germ. Taf. lvii. fig. 5).

5 The ' Challenger ' collection includes two very abnormal species which present a singular exception to the rule.
The rays divide three times ; and the first two segments (distichals) of each of the ten primary arms are united by
ligaments only, like the second and third radials. So far the rule holds good ; but with the next arm-division there
is a new point of departure. The third or axillary distichal bears the secondary arms, which consist of one axillary
segment only. But this segment is itself primitively double, i. c. it consists of two parts united by a syzygy ; and
the first joints of each of the ultimate arms borne by tis axillary a gree with it in being syzygial segments.

24 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

of the free and undivided secondary or tertiary arms is of precisely the same nature as
the union of the first and second segments of the primary arms home hy the radial
axillary. In fact, it is not at all uncommon for one of the primary arms to remain
simple and the other to divide, as in PL II. fig. 9, c, d, e, which shows that the arms
taken in the strict sense of the word, cannot he regarded as commencing from any point
but the axillary radials.

(§ 17) In practice, however, it is more convenient to regard the arms of Comatula as
beginning from the last bifurcation, i. e. from that axillary the two branches borne by
which do not further divide, but remain composed of a series of simple brachial seg-
ments (b l} b 2 ), &c. In the ten-armed Coma tit lee the brachials are, of course, borne directly
by the radial axillary. But in those forms, such as Act. mvltiradiata, in which the sub-
division of the ten primary arms is carried to a very great extent, it is most convenient
to regard as brachials only the segments of the ultimate branches borne by the last
axillaries, and to give special names to the segments composing the primary and
secondary arms ; for we have already seen that the number of the segments composing
these arms, i. c. between every two successive axillaries, varies in different species, and
it consequently becomes desirable to have some system of nomenclature by which
these differences can be briefly indicated. Under these circumstances, therefore, the
term distichals may be applied to the segments composing the ramified primary arms of

e Articulate as well as of the Tessellate 1 Crinoids, but onlv on the distinct under-
standing that they are really arm-segments and do not enter into the formation of the
calyx, as in the Tessellata ; so that the name is purely a conventional one, employed for
greater convenience in the description of species.

Supposing the secondary arms borne by the distichal axillaries to divide again, we
may consider them as composed of two, three, or four palmar 2 segments, of which the last
is a " palmar axillary " (figs. 10, 11, p.a.), and bears two tertiary arms. These may either
remain simple and composed of brachial segments, or they may continue to divide more
or less frequently. The latter case, however, is somewhat rare ; for if complete series of
distichals and palmars be developed on each radius, the total number of arms rises from
10 to 40 ; and there are not many Comatula? in which this number is exceeded.

If we apply this nomenclature to the species of which mention has already been made,
we should describe Antedon Savignii with 20 arms, as characterized by the presence of
three distichals composing each primary arm and bearing the brachials directly, while
in Antedon palmata, with 30 or 40 arms, there are only two distichals, which are fol-
lowed by two palmars in the secondary arms. Act. polymorpha, again, normally has
three distichals and three palmars (PI. II. figs. 7, 10), while Act. multlrad'iata, Mull., has

1 /. e. As understood by MM. do Koninck and Le Hon.

- I have been accustomed for some years past to use the term " palmar " to designato the secondary arms of the
Crinoids. Professor Huxley, in whose lectures I first heard it used in this sense, informs me that he believes it to
have been so employed by Miillcr ; but I have searched in vain through Midler's works for any definition of the term.

In his description of tho Tessellate Crinoids, however, he describes the plates which continue the series of inter-
radials and interaxillaries in a peripheral direction as " interpalmaria;" and as these partially correspond to the
intervals between the secondary arms, when such arc developed, the latter may perhaps be not incorrectly regarded as
composed of " palmar" segments. It will now be apparent why " interpalmar " is not a very suitable designation
for the strictly " interradial " areas on the disk of Comatula, as was remarked in section 15.

ME. P. H. CABPENTEE ON THE GENUS ACTINOMETEA. 25

only two palmars in each secondary arm, although the number of distichal segments in
the primary arms is usually three. Another proof, if proof were wanted, that the arms
proper of the articulate Criuoids begin from the axillary radial is seen in the fact that
whenever there are three segments in a distichal or palmar series, the second of these,
which is united to the first by ligament only, always bears a pinnule, while the third, or
axillary, is a double segment, i. e. it consists of two primitive segments united by a
syzygium. This is in precise accordance with what we find in all the ten-armed
Comatulce, even in those forms in which, as in. Act. Solaris, the first and second brachials
are united by a syzygium like the second and third radials. In these the second brachial
or the epizygial element of the syzygium bears a pinnule, while the next segment is also
a double one, and corresponds with the compound third brachial of Ant. rosacea and of
the ordinary ten-armed Comatulce. In these last the second brachial is laterally movable
upon the first, and bears a pinnule as in Act. Solaris, while the third has a syzygium (/. e.
is a double segment). This is exactly what Ave find to be the case in those primary and
secondary arms of the multiradiate Comatulce which consist of more than two segments.

(§ 18) The principal character of the genus Actinometra is, as we have seen in sect. 14,
that the mouth is situated excentrically, while the centre of the disk is occupied by the
anal tube. The position of the mouth relatively to the radii or ambulacra, however, is
not the same in all Actinometrce ; thus in Act. Solaris (PL I. figs. 2, 5) the mouth lies
in a radial or ambulacra! plaue, while in Act. WahlbergMi and many other species (PI. I.
figs. 3, 4, 6-16) it is interradial or interambulacral. If we place the disk of an ordinary
Antedon in such a position that the interradial area containing the anal tube is nearest
to us (PI. I. fig. 1), the odd ambulacrum lies in front of the mouth. Let us designate
this as ambulacrum or radius A, and the two branches of its groove-trunk corresponding
to the two primary arms as A L and A 2 respectively, A,, being that on the left of the
mouth. Proceeding round the disk in the direction of the hands of a watch, we may call
the other four ambulacra B, C, D, E respectively, and their primary divisions L\ B..
. . . . E x E 2 . The anal area is then bounded by the two postero-lateral ambulacra C, D ;
and a plane passing through the mouth and anus, so as to divide the disk into two
symmetrical halves, passes along the odd ambulacrum or radius A, in front of the mouth,
which may therefore be regarded as radial in position.

In Act. Solaris the same is the case, as may be seen in Midler's somewhat diagrammatic
figure (PL I. fig. 2) 1 , and still better in PL I. fig. 5, which was drawn from a spirit-
specimen, and not from a dry one like Midler's figure. Here, as in Antedon, the odd
ambulacrum is in front of the mouth, which, although excentric in position, lies in the
radial half of a plane which passes through the mouth and anus, so as to divide the disk
into two symmetrical halves. The same is the case in a new Actinometra from the

1 It is not usual to meet with specimens of Actinometra in which the branches of the ambulacral grooves are
distributed with such symmetry as is represented in Muller's diagrams (PL I. figs. 2-4) and in PI. II. fig. 1. Thus,
for example, Midler's figure of Act. Solaris (PL 1. fig. 2) is remarkably regular, much more so than that repre: en1 !
in fig. 5; and I have examined other specimens with more than 20 arms and a radial mouth, in which the regularity
is by no means so distinct as in PI. II. fig. 1. A great range of variation in this respect is seen in PL I. figs. 6-16.
which represent the disks of eleven different individuals of Act. jpolymorpha, no two of which are alike ; the position
of the mouth, however, is constant in all individuals of the same species.

SECOND SERIES. — ZOOLOGY, VOL. II. 4

26 ME, P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

Philippines (PI. II. fig. 1) with 23 arms. In both these species the dividing plane passes
in front of the mouth between the two primary divisions, A 1 A 2 of the odd anterior
ambulacrum A, while behind the mouth it is interradial, and separates the two postero-
lateral ambulacra C, B*.

In Act. WahlberghU (PI. I. fig. 3), Act. multiradiata (fig. 4), and Act. polymorpha
(figs. 6-16) the case is different. If, as in Antedon, we designate the two ambulacra
bounding the anal area as C and D respectively, we find that the latter is the odd
ambulacrum, and that a plane cutting the mouth and anus is radial behind the mouth,
in front of which it passes between the two ambulacra A and B ; so that if the centre of
the disk be regarded as the centre of radiation, the mouth lies in an interradial or inter-
ambulacral plane. This is clearly seen when we turn to the dorsal side of the disk, in
which the radii converge to a central point, and not to an excentric one, like the ambu-
lacra of the ventral side. Thus in PI. II. figs. 9-11, the position of the mouth relatively
to the radii is indicated by a X , which in each case is between the two anterior radii
A and B, or interradial.

So far as my experience goes, this type of Actinometra, in which the mouth is inter-
radial and the odd ambulacrum lies behind it, is slightly more common than the simpler
type, in which the mouth is radial and the odd ambulacrum anterior, as in Antedon 1 .

(§ 19) We are now in a position to investigate which of the numerous species of
Comatulce described by Miiller belong to the type of Antedon, and which to Actinometra,
and into what groups the latter may be divided according to the principles of classifica-
tion discussed in the last four sections. Midler's specific descriptions afford very little
information in this respect ; for though he says that the mouth, in some instances, is
excentric, and that in others the oral pinnules have a terminal comb, he does not always
do so ; and he makes no use whatever of these two characters in his distinction between
Alecto and Actinometra. In some cases he simply designates a species as Comatula,
without attemping to name it more exactly. This is often, no doubt, simply due to the
fact that, in the specimens which he examined, the arms were so closed over the disk that
he was unable to investigate the distribution of the ambulacra. This, however, is not the
case in the C. trichoptera of the Paris Museum, the disk of which can be readily ex-
amined ; and I believe that Midler did not define this species more precisely because he
was unable to decide whether it should be referred to Alecto or to Actinometra ; for in
one of the two Paris specimens five groove-trunks start from the excentric peristome,
while in the other there are only four. This example alone suffices to show the unsatis-
factory nature of the only distinctive character established by Miiller between his genera
Aleclo and Actinometra, The Paris Museum- contains a very large majority of the

1 Since the above was written, I have examined three large C'omatula-colhctions : — (1) that of the 'Challenger;'
(2) that made by Prof. Semper in the Philippines ; and (3), thanks to the kindness of Dr. (jiinther, that in the
British Museum.

I have, been able to determine the position of the mouth in 80 species of Actinometra. In 4-j of these it is inter-
radial, as in Act. polymorpha : while in the rcmainig 35 it is radial, as in Antedon and in Act. Solaris.

- I would here express my most hearty thanks to Mons. Edmond Perrier, Assistant-Naturalist at the Museum of
Natural History, Jar:! in des Plantes, who lias charge of the Echinoderm collection, and also to his two Assistants, for
the kindness which they showed me during my stay in Taris, and for the readiness with which they afforded me
every possible facility in the prosecution of my work.

ME. P. H. CAEPEXTEE OX THE GENUS ACTIXOMETEA.

27

species of Comatula described by Midler, who personally examined this collection. Last
autumn (1876) I was also able to examine it for myself, and thus to determine which
species should be removed from Midler's genus Alecto and transferred to Actinometra, in
the sense in which this name is understood by Dr. Liitken and myself.

In the following scheme all those species to which no note of interrogation and the
name of no authority is attached, have been determined by myself as Actinometra^, i. e.
as having an excentric mouth and a terminal comb on tbe oral pinnules.

With 10 arms.

Actinometra.

Second and third radial*;
united by a syzygy. <(

Mouth radial.

''Act. Solaris, Miill.
Act. pectinata, Miill.
Act. brachiolata.
Com. purpur a ? 1.

I Com. rosea '.'

y_Act. robusta, Liitken.

Second and third radials f Com. eehinoptera ? -.
united hy ligaments -j
only. Mouth ? (_ Com. meritUonalis'? 3.

With more

than 10

arms.

Second and

third radials

united hy

ligaments

only.

" 2 Distichals.
Axillary has no

syzygy.

<; 3 Distichals.
Axillary has a -^
syzygy.

3 palmars.
Axillary has a V Act. rotatoria, Midi. 4.

syzygy-

/(Act. polymorpha.
Mouth interradial. -< Act. trichovtera.
_ . 1 .

2 palmars. ^|
Axillary has
no syzygy.

^ Mouth radial

'2 palmars.

Axillary has

a syzygy.

3 palmars.

axillary has a

syzygy.

[Act. Wahlberghii, Miill.
Act. finihriatn. 5.

Act. multifi&a. <>.

> Mouth interradial . Act. multiradiata.

C Act. parvicirra. 7.
■i Act. polymorpka. S.
[Act. Bennettii, Bohlsche. 9.

Remarks.

1. I have not personally examined either C. purpurea or C. rosea. Muller seemed to
think that the former might be a young condition of Act. Solaris 1 ; so that it is most pro-
bably a true Actinometra. C. rosea, however, presents a difficulty; for Muller says
expressly 2 that the first pinnules are uot specially distinguished; but, except in this
point, he regarded it as very closely related to C. brachiolata, which is a true Acti-
nometra.

: Gattimg Comatula? p. 13.

2 Ibid. p. 14.

! :;

28 ME. P. II. CARPENTER ON THE GENUS ACTINOMETRA.

2. O. echinoptera, on the other hand, has, according to Midler 1 , a very marked comh
on the oral pinnules. He says nothing, however, about the position of the month ; and I
have unfortunately not been able to examine the species for myself.

3. According to Pourtales - the mouth is excentric in Com. meridionalis ; but he makes
no mention of a comb on the oral pinnules. If it should be absent in this species, and
also iu Com. rosea, while it is present in Phanogenia, in which the mouth is central, it
becomes obvious that the only external character, besides the shape of the calyx, on which
we can rely with any certainty in the determination of the generic position of any
Comatula, is the nearly central or the excentric position of the mouth.

4. According to Midler ;1 there are only two radials in Act. rotalaria which are united
by a syzygium, while they bear the distichal axillaries directly ; and these are also syzy-
gial segments. Although, like Miiller, I examined Lamarck's original specimen of this
species, I cannot confirm the above statement. It is true that only two radials are visible
externally; but this is often the case in Comatula?, with a wide centrodorsal piece; and
I A\as quite unable to satisfy myself that they are united by a syzygium, as Miiller
describes, and as is the case with the second and third radials of Act. Solaris, Avhile I was
equally unable to determine a syzygial union between the two segments of which the
distichal series is composed. Lamarck's original specimen of this species in the Paris
Museum is wrongly labelled C. brevicirra, Troschel.

5. The dry specimen of Act. fimbriate in the Paris Museum, from the voyage of
Peron and Lesueur, is labelled C. multiradiata, Lam. ; while Peynaud's original speci-
mens from the Strait of Sonde are labelled C. brevicirra, Trosch., together with Act.
rotalaria and the Vavas variety of Miiller' s C. parvicirra.

6. In Jet. multijida the tertiary arms borne by the palmar axillaries may divide again
several times. In every case there are only two segments between each division, and all
the successive axillaries, like the palmar axillaries, have no syzygia. Although Miiller
mentions this in his description of the species, it is placed in his scheme in a group in
which the axillaries of the arms have syzygia. In reality, however, this is the case with
the distichal axillaries only.

7. The type specimen of this last species does not exist in the Paris Museum under
that name, nor even under C. brevicirra, Troschel, which seems to have been used as
an equivalent for it ; but I believe that three small spirit-specimens from the voyage of
Peron and Lesueur in 1803, which arc classed, with two specimens of Act. pectinata,
under the name of C. simplex, are really those which were described by Miiller as
C. parvicirra.

8. It will be noticed that Act. polymorpha has already appeared higher in the list as
a species in which palmars are not developed. In some individuals but few of the 10
primary arms bear axillaries ; so that the total number of arms is less than 20 (PI. II.
fig. 9) ; while in others all the primary arms divide again, and so do the resulting secon-
dary arms, so that the total number is little short of 40 (PI. II. figs. 10, 11). This is
very unusual ; for I only know of two other species which present the same kind of

1 '(juttuug Comatula,' p. 14. " Loc. cit. No. 11, p. 355.

3 ' Gattung Comatula, ' p. 20.

MR. P. H. CARPENTER OX THE GENUS ACTINOMETRA. 29

irregularity. As a general rule all the individuals of a species agree in the presence or
absence of distichals and palmars.

9. In Act. Bennettii there are more than 70 arms; but all the axillaries are like the
first one (distichal), and not different from it, as in Act. multifield. According to
Miiller 1 , every fourth segment is an axillary without a syzygium ; but Bohlsche 2 has found
this to be incorrect. There are, indeed, four segments between every two points of division ;
but the last two are united by a syzygium ; so that the formula becomes three distichals,
palmars, &c, of which the axillary has a syzygium. Bohlsche's figure of the disk of his
specimen is noteworthy; for though five groove-trunks leave the excentric peristome, as in
Alecto, yet their distribution to the arms is not by any means symmetrical, so that he
seems to have decided upon calling it Actinometra. Midler named it simply Comatula.

(§ 20) In the above scheme are included all the species of Comatula which have been
determined by myself or by others 3 , as far as I know, to belong to the type Actiuometra.
Fourteen of these were known to Midler ; and of the remaining 23 species described by
him I have been able to refer 16 to Antedon, viz. : —

Ant. adeonte. Ant. macrocnema. Ant. phalangium.

Ant. articulata. Ant. Milberti. Ant. Philiberti.

Ant. carinata. Ant. Milleri. Ant. Reynaudii.

Ant . Eschrichtii. Ant. palmata. Ant. rosacea.

Ant. Jacquinoti. Ant. petasus. Ant. Sarsii.

Ant. Sarignii.

To these must be added

Ant. armata, Pourt. Ant. cubensis, Pcmrt. Ant. Hagenii, Pourt.

Ant. bicolor, Mus. Paris. Ant. dividua, Mus. Paris. Ant. polyactinis, Mus. Paris.

Ant. celtica, Barrett. Ant. Dubenii, Bolsche. Ant. rubiginosa, Pourt.

The following list contains the seven remaining species of Comatula described by
Miiller which I have not been able to examine, and of which I know no descriptions from
which it is possible to obtain any information as to the position of the mouth or
the character of the oral pinnules.

C. Cumingii. C. elongata. C. flagetlata.

C.japonica. C. nova Guinea?. C. tessellata,

and C. timorensis. To which must be added C. brevipinna, Pourt.

III. External Characters of Act. poltmorpha, and Specific Diagnosis of the Type.
(§ 21) In Act. yolymorpha, as in all Actiuometrce, the mouth (PI. I. figs. 6-16, m)

1 ' Gattuug Comatula' p. 28.

: '• Uebar Actinometra Bennettii und erne neue Comatida-Aib {Antedon Dubenii) " Wiegm. Archiv, l^GG, i. p. 90.

3 Dr. Liitken has named several new species of Actinometra besides Aet. robusta — for example, Act. tenant; and
Act. trachygaster. But his descriptions have not, as far as I know, been published ; and I have had no opportunity of
examining any specimens of his new species except Act. robusta ; so that I am unable to llace them in the classi-
fication given in the previous section.

Grube has described a new Actinometra from Borneo, and two new species of what be calls Comatula. His de-
scriptions (Jahresber. d. Schlesisch. (Jesellsch. 1375, Nat.-Hist. Sect. pp. 54, 55) are, unfortunately for me, not to be
obtained in this country.

30 MR. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

does not occupy a central or subcentral position on the ventral surface of the visceral
mass as it does in Antedon, but is placed more or less excentrically, and may be some-
times almost marginal (PL I. fig. 11). It occupies the centre of the peristome, P, and is
bounded by two lips, a large anterior and a smaller posterior one ; so that its opening is
very inconspicuous, and usually so much extended in a direction transverse to the
antero-posterior diameter of the disk, that it presents the appearance of a simple slit, as
is well seen in PL II. fig. 2.

The circumoral portion of the peristome, or the peristome proper, is a more or less oval
depression in the ventral perisome of the disk, which completely surrounds the oral
opening, and gives origin to the ambulacral grooves or, more shortly, the ambulacra.
Beneath this depression lies the water- vascular ring which gives off a trunk imder
each of the ambulacra radiating from it. The number and distribution of these are
very variable, as is seen in PL I. figs. 0-16. This principally depends upon the
way in which the ambulacra divide, so as to give rise to the groove-trunks corresponding
to the ten primary arms. As a general rule, the two ambulacra corresponding to the
radii T) and E unite into one large posterior trunk, from which the branches are dis-
tributed to the various arms into which these radii divide (PL I. figs. 8-10, 12-10). In
other cases the left lateral ambulacrum, E, leaves the peristome alone (figs. 6, 11) ; while
in others it is partially united with the posterior ambulacrum. D, its anterior division,
E 2 , leaving the peristome by a separate trunk, while its posterior division, E„ unites with
the posterior ambulacrum (fig. 7).

As a general rule, the right lateral ambulacrum, C, leaves the peristome alone, and
supplies the arms of the corresponding radius ; but in figs. 9 & 15 it is seen to unite with
the posterior division, B 2 , of the right anterior ambulacrum, B.

The mode of division of the two anterior ambulacra is excessively variable. As a
general rule there are no principal trunks corresponding to the two radii A and B, and
the primary divisions, A 13 A,, B^ B,, start directly from the peristome. In the specimens
with but few arms, however, each pair may be united for a longer or shorter distance
(PL I. figs. 0, 7), as in Antedon (fig. 1). Not unfrequently the posterior divisions
A l5 B 2 , of these two anterior ambulacra unite for a longer or shorter distance with the
two large ahoral groove-trunks, to form an open horseshoe-shaped curve bounding the
anal area (figs. 12, 15, 10). The position of the anal tube in this ai'ea, and also with
regard to the whole surface of the disk, varies somewhat with the position of the mouth ;
it is rarely, if ever, absolutely central. Its appearance differs very much according
as it is full or empty : sometimes its aperture is so completely closed as to be scarcely
discernible, though the tube below is widely distended ; and sometimes the aperture is
patent with its edges everted and crenate, and the tube leading to it quite shrunk and
flaccid (PL II. fig. 2).

(§ 22) In Antedon the median line of the ventral perisome of all the arms is occupied
by an ambulacral groove, with a floor of ciliated epithelium. This groove extends also
on to all the pinnules, wdth the exception of those borne by the second distichals and
second palmars, &c. (when present), and by the lowest brachial segments. Beneath it
lie the radial water-vascular and blood-vascular trunks, between which last and the

ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 31

ciliated epithelium of the floor of the groove lies a fibrillar structure, to which I have
given the name of the " subepithelial band " l , and to which a nervous character has
been attributed by myself and by all the other observers who have described it. Each
side of the ambulacral groove is bounded by an elevated fold of perisome, the edge of
which is not straight, but cut out into a series of minute valvules, the cresceutic or
respiratory leaves (Wyv. Thomson), or " Saumlappchen " of the German authors.

At the base of each leaf, and to some extent protected by it, is a group of three
tentacles, one of which, the more distal one, is larger than the other two. This trind
group of tentacles and the cavity of the respiratory leaf adjacent to them receive a
common branch from the radial water-vessel. These groups of tentacles alternate on
the opposite sides of the ambulacral groove from the base to the tip of each arm, and are
distributed in the same manner at the sides of the ambulacra of the disk, though they
are not so markedly developed, especially near the peristome, where every lateral branch
of the water-vessel supplies only one tentacle. The crescentic leaves at the sides of the
groove are also far less distinct than in the arms, the edges of the folds of perisome
bounding the groove being only marked by a faint wavy line, and not distinctly cut out
into " Saumlappchen."

In many Actinometrce, however, the above description only applies to the arms of the
two anterior radii, A, B (PI. II. figs. 2, A), and to more or fewer of the antero-lateral
arms, C^ and E^. The arms of the posterior radius, D, and of the posterior divisions of
the lateral radii, C 2 and E l5 are often entirety devoid of tentacles ; and in many of them
the ventral perisome not only exhibits no ambulacral groove, but is, on the contrary,
convex, as in the oral pinnules of Antedon (PL II. figs. 5, 6).

We have just seen that in Act. pohjmorplia, as in all Actinometrce with an interradial
mouth, the anal area is bounded by two large aboral groove-trunks, which start from the
posterior angles of the peristome, and form a horseshoe-shaped curve, tbe limbs of which
are unequal in size (PL II. fig. 2). The smaller right limb is formed by the right lateral
ambulacrum, C ; while the larger left limb represents the posterior ambulacrum, D, com-
bined with part or the whole of the left lateral ambulacrum, E. In neither of these limbs
are the tentacular groups and crescentic leaves so well developed as they are in the
two anterior ambulacra. After the branches to the two antero-lateral primary arms,
C^ and E 2 , have been given off, or sometimes even sooner (PL I. figs. 13, 15), the ten-
tacles at the sides of the two aboral groove-trunks become more and more insignificant,
and finally disappear altogether, while the position of the crescentic leaves is only
indicated by a very faint wavy line at the edge of each groove.

In small specimens with but few arms (PL I. figs. 6, 9) the grooves of the posterior
(D) and postero-lateral arms (C 2 , E,) may remain in this condition; but in larger speci-
mens with many arms all trace of the crescentic leaves disappears, and the two edges of
the groove meet and unite so as to produce the condition represented in PL II. figs. 5 & 6,
where the ventral surface of the arms and pinnules is convex, and does not show the
least trace of a groove of any description.

" Kcmarks on the Anatomy of the Arms of the Crinoids. Part I.,'' Journ. of Anat. and Physiol, vol. x. p. 57!'.

32 ME. P. H. CARPENTER ON THE GENUS ACTINOMETEA.

The position of the point at which the two folds of perisonie hounding the sides of the
original ainbulacral groove meet and unite, varies extremely. The fusion may, though
rarely, take place on the disk ; sometimes it is at the hase of the arms, and sometimes
not till near their middle or terminal portions. In any case, however, the fusion, when-
ever it occurs, is so complete that all trace of the original ainbulacral groove is entirely
obliterated.

(§ 23) The bearings of this fact upon the different views advanced by Greeff l and
Ludwig 2 respecting the homologies of the ainbulacral grooves of the Crinoids will he
best discussed at a later period, when the changes undergone by the various structures
underlying the grooves are described and illustrated. One point, however, must he
noticed here on account of its importance with respect to the two views now entertained
regarding the nervous system of Comatula.

As long ago as 1865 it was stated by Dr. Carpenter 3 that the cord which traverses the
length of the arms between the subtentacular and cceliac canals, " and which was regarded
by Professor Miillcr as a nerve, really belongs to the reproductive apparatus. But it
will also be shown that a regular system of branching fibres proceeding from the solid
cord (described by Professor Miillcr as a vessel) that traverses the axial canal of each
calcareous segment of the rays and arms, is traceable on the extremities of the mus-
cular bundles ; and reasons will be given for regarding these fibres as probably having
the function of nerves, though not exhibiting their characteristic structure." During
his residence in the Philippine Islands, Professor Semper had also discovered that the
arm-nerve of Miillcr is really a part of the generative system ; and in a short paper *
published some time after his return he announced this fact, and suggested at the same
time, " dass der lusher immer als Gefass aufgefasste Strang im innern des Kalkskelettes
ein Nervenstrang sei, und dann ware wohl das im Kelch liegende sogenannte Herz als
ein Ganglion anzusehen."

These observations of Dr. Carpenter's and Professor Semper's were unfortunately
overlooked for many years, so that even as late as 1871 Midler's erroneous statements
with regard to the nervous system of Comatula were repeated in the valuable text-book
of Gegenbaur 5 and in many smaller works. At the commencement of 1876, however,
two very different views respecting the nervous system were put forward nearly simul-
taneously by Greeff and by Dr. Carpenter. The former 6 described the whole floor of the
ambulacral grooves on the arms and disk of Ant. rosacea as constituting a radial
nervous system, starting from an oral nervous ring in the peristome, and corresponding

" Peber den Bail der Echinodermen. III. Mittheilung," Sitzungsb. der Gesell. z. Beforder. d. gesamm. Naturwiss.
zu Marburg, 1S72, No. 11, p. 155.

- - Beitr. z. Anat. der Crinoideen," Nachrichten von der Konigl. Gesells. der "Wissens. u. der G. A. Universitat
zu Giifctingen, 1876, No. 5, pp. 107, 10S.

3 " Researches on the Structure, Physiology, and Development of Antedon rosaceus. Tart I.," Philos. Trans,
vol. clvi. p. 705.

J " Kurze anatom. Bemerk. iiber Comatula," Arbeit, aus d. zool.-zootom. Inst. zuWurzburg, Band i. (lS74),p. 262.

5 Grundriss der vergleich. Anat. p. 222.

6 '• Peber den Bau der Crinoideen," Marburg. Sitzungsb. 1»70, No. 1, Jan. ^^6, p. 21.

MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 33

in position and histological structure with the typical Echinoderm nerves. At the same
time he denied the nervous nature of both the structures described as nerves by Miiller
and Dr. Carpenter respectively, viz. the genital cord, the so-called " rachis," on the one
hand, and the axial cords in the centre of the calcareous segments on the other.

A week after the publication of Green's views, Dr. Carpenter 1 announced his belief
that a complicated apparatus, " consisting of the outer cylinder of the Crinoidal stem, of
the five-chambered central organ formed by the dilatation of that axis within the centro-
dorsal basin, and of the cords proceeding from it to the arms and eirrhi," should be
regarded as the central portion of a nervous system. This view was based both upon
anatomical and upon physiological considerations : —

(a) That while a siugle arm may be made to coil up by irritating one of its pinnules,
the whole circlet of arms closes together when an irritation is applied to the pinnules,
which arch over the mouth — an act which affords a strong indication of the "inter-
nuncial" action of a definite nervous system.

(b) That stimulation of the central quinquelocular organ ("heart" of Miiller and
Grceff) contained in the calyx, with which the axial cords of the arms are in connexion,
is followed by sudden and simultaneous flexion of all the arms.

(c) That these axial cords give off successive pairs of branches, which ramify upon the
muscles connecting the arm-segments.

Shortly after the announcement of these views on the part of Dr. Carpenter, Ludwig 2
described a ventral nervous system as existing in Comatula in common with all the
other Echinoderms. He attributed a nervous character, not to the whole epithelial floor
of the ambulacral grooves, as was done by Greeff, with whose researches he was unac-
quainted, but to a fibrillar layer beneath it, and more or less distinctly separated from
it. This layer, which was also discovered independently by myself 3 and Teuscher 4 ,
and was regarded by us both as of a nervous nature, is the " subepithelial band " men-
tioned in sect. 22. Ludwig, like Greeff before him, denied the nervous character of the
dorsal axial cords of the arms ; Teuscher discussed it as possible, but hesitated to accept
it on account of the morphological difficulties involved in such a view.

Baudelot 5 , who seems to have been unacquainted Avith Dr. Carpenter's earlier state-
ments, was apparently struck with the nature of these cords, though he could not regard
them as nervous. After describing; their structure and their union in the calvx to form
the pentagonal commissure, he adds, "Ainsi done chez les Comatules il existe des parties
qui eciclemment rC appartiennent point au systeme nerveux (!), et qui dans leur disposition
aussibien que dans leur structure offrent une analogie presque complete avec les cordons
nerveux des autres Echinodermes." It must be remembered that Baudelot wrote before
the discovery of the so-called " ventral nerve " of Comatula ; but, in any case, I do not
quite see the force of his " evidemment."

1 " On the Structure, Physiology, and Development of Antedon rosacew" Troc. Roy. Soc. no. 160, Jan. 20th,
1876, pp. 219-226.

- Gottingen Nachriehten, no. 5, Feb. 23rd, 1S70, p. 106.

3 Journ. Anat. Phys. x. p. 578.

4 " Beitr. z. Anat. der Echinodermen, I. Comatula mediterrawa," Jenais. Zeitseh. B. x. p. 253.

b " Contribution a l'histoire du systeme nerveux des Echinodermes," Arch, de Zool. Exp. et Gen. i. p. 211.
SECOND SERIES. — ZOOLOGY, VOL. II. 5

3 ± ME. P. II. CABPEXTEE ON THE GENUS ACTIXOMETEA.

Ill the centre of every segment of the skeleton of Act. polymorpha and of all the other
Comatulce which I have examined, from the first radials to the ends of the arms and
pinnules, and also in the cirrhus-segments, these axial cords increase considerahly in
size, and give off four principal branches. Two of these run towards the ventral side, and
in the calyx disappear in the neighbourhood of the muscles connecting the segments,
though I must confess that I have never been able to trace them any farther (PL VIII.
fig. 3, »'). In the arm-segments, however, they continue their course towards the ventral
surface and break up into numerous branches, some of which, as I have already described 1 ,
extend to the tips of the cresceiitic leaflets at the sides of the tentacular furrow. The
two inferior or dorsal trunks run towards the surface of the skeleton ; and while some
of their branches are lost in the plexus of tissue forming its organic basis, others seem to
become connected with epidermic structures in a manner which will be described at
length further on.

Not one of the German observers makes any mention of these branches, although two
of them at least have examined Antedon Eschrichtii, while they have all cut sections of the
arms of species of Actinometra, in which genus I find them to be particularly distinct.
It is obvious that the facts above stated strongly support the view expressed by Dr. Car-
penter and by myself, that the axial cords of the arms are of a nervous nature ; and the
experiments made by Dr. Carpenter 2 at Naples have shown conclusively : —

1. That the quinquelocular organ is the instrument of the perfect coordination of the
swimming movements of the arms, which involve the conjoint contraction of several
hundred pairs of muscles.

2. That nothing contained in the visceral mass is essential to the perfect coordination
of the swimming-movements, and that therefore the subepithelial band or arnbulacral
nerve of the German authors has no immediate relation to those movements, even if it
be a nerve at all.

3. That section of the subepithelial band in an arm does not prevent its playing its
usual part in the regular swimming-movenuiiLS.

4. That destruction of the axial cord of an arm by the application of acid causes the
arm to become rigidly stretched out, while all the others work as usual.

Since the publication of these experiments Greeff seems to admit the nervous nature
of the axial cords, and of the yellowish fibrillar envelope (PI. VIII. figs. 1-3, N) of the
quinquelocular organ from which they proceed. Ludwig 3 , however, while allowing
their force, cannot admit the existence in the Crinoids of an antiambulacral nervous
system, of which Ave know as yet no homologue in the other Echinoderms, but sees
no difficulty in regarding the quinquelocular organ, its fibrillar envelope, and the
axial cords proceeding from it, as parts of a blood-vascular system, like that of the other
Echinoderms, although he admits (p. 87) that " ihncn vergleichbare Gebilde sind bis
jetzt bei anderen Echinodermen nicht bekannt gewordert." The axial cords of the

1 Journ. . x. p. 584.

• "Su [omenta] on the Structure, Physiol, and Develop, of Aniedm rosacius," Proc. Eoy. Soc.
no. 169, 1876.

3 "Bcitr. zur Anat. der I ." Sep— Abdruck aus der Zeitsch. f. wissensch. Zool. B. sxviii. Heft 3, p. 81.

ME. P. H. CARPENTER OX THE GENUS ACTIXOITETEA. 35

calcareous segments are regarded by him (pp. 80, 80) as " unverkalkt gebliebene Tbeile
der bindegewebigen Gruudlage der Kalkglieder, deren Aufgabe es ist, aus dem Blutge-
fasssysteni, genauer aus den fiinf Kammern die kemniihrende Eliissigkeit aufzunebmen
unci den Arm- und Pinnulagliedern zuzufiibren."

"Without going into the question as to how far the organic basis of the calcareous
skeleton can be regarded as of a connective-tissue nature, I would only remark that it
is difficult to see why the axial cords, which Ludwig supposes to consist of uncalcified
connective-tissue fibres, should give off branches the terminations of which are entirely
outside tbe skeleton, as is the case with those which reach the erescentic leaves at the
sides of the tentacular groove, and which therefore cannot take any part in the nutrition
of the tissue forming the organic basis of the skeleton.

(§ 24) This is not the place for a full discussion of Ludwig's views on the nervous
system of Comatula ; but one point must be briefly referred to. I have already ' stated
that in some arms, and in most of the pinnules, of many Actinometrce, the subepithelial
band or nerve of Ludwig is entirely absent, and also that "if the axial cords are not
nerves, and if the subepithelial bands are to be regarded as the only nervous structures
in the whole Crinoid organization, the difficulty presents itself that the oral pinnules of
the European Crinoids, and more than half the arms, with the majority of the pinnules
of some forms of Actinometra, are entirely devoid of a nervous supply.

"The oral pinnules of Antedon have been shown by Dr. Carpenter 2 to be extremely
susceptible of irritation; when they are touched in the living animal, the whole circlet
of arms is suddenly and simultaneously coiled up over the disk, while irritation of one of
the ordinary pinnules is simply followed by flexion of the arm which bears it.

" The structure of tbese oral pinnules, which are borne in Antedon rosacea by the
second brachials, differs very considerably from that of the pinnules borne by the other
brachial segments ; for not only are they sterile, but they have neither tentacular appa-
ratus nor ambulacral groove, their ventral surface being slightly convex instead of being
concave as in tbe ordinary arms and pinnules. Tbis has been mentioned by Teuseher 3 ;
but he has omitted to state that the ordhiary ciliated epithelium of the ambulacral
groove, with its subjacent nervous layer and nerve-vessel, are also absent."' Ludwig
enthely ignores tbis argument, althongh he confirms the above statement concerning the
oral pinnules of Antedon; in the text he is obliged to confess that "Fraglich est mir
geblieben ob die oralen Pinnuhc einen Zweig des radiaren Xerven besitzen oder nicht "
(p. 75) ; while his figure of a section of an oral pinnule (pi. xvii. fig. 55) entirely con-
firms the statement quoted above, to which, however, he makes no reference.

This condition, which is limited in Ant. rosacea to the oral pinnules, sometimes exists
in whole arms and in all the pinnides borne by them in many species of Actinometra.
Even in the arms which come off from the anterior or oral side of the disk the ambulacral
groove does not give off regular branches to the pinnules borne by the third and succes-
sive brachial segments ; but a variable number of these first pinnules (sometimes only

1 " Remarks on the Anatomy of the Arms of the Crinoids. part ii.,"' Journ. of Anat. and Physiol, vol. xi. October,
1876, p. 89.

■ Proc, Roy. Soc. no. 1G6, p. 226. 3 Jenaische Zeitschrift, x.p. 249.

36 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

three or four, sometimes as many as forty) resemble in this respect the oral pinnules,
their ventral surface being convex, and devoid of any ciliated epithelium or subepithelial
baud ; while their water-vessel is simple, without any lateral extensions to respiratory
leaves and tentacles. In these oral arms, however, branches of the ambulacral groove
enter the pinnules sooner or later, so that the terminal ones are always provided with a
distinct tentacular apparatus, while the floor of their median groove is of the usual cha-
racter, consisting of a ciliated epithelium and a subepithelial fibrillar baud.

We have seen in sect. 22 that in many cases the ambulacral grooves going to the
aboral arms become less and less distinct as they get further and further from the peri-
stome, and that their tentacles diminish and finally disappear. At the same time the floor
of the groove becomes very much reduced in extent, its epithelial layer thinner and
thinner, and the subepithelial band almost invisible, until, in those cases in which the
two sides of the groove meet and unite, the ciliated epithelium and subepithelial band
disappear altogether. Consequently, when this union takes place on the disk, whole arms
are entirely devoid of any nervous supply, if we suppose, with Ludwig, that the anti-
ambulacral axial cords are not of a nervous nature, and that the " subepithelial bands "
are the only nervous structures in the arms. In such cases it would naturally be
expected that these arms would be incapable of performing the regular swimming-move-
ments like those in which there is an open ambulacral groove and a subjacent " ambu-
lacral nerve;" but Professor Semper, who has kept Actinometrae in his aquaria for
weeks together, informs me that he never saw the least trace of any irregularity in the
alternating movement of their arms while swimming.

The gradual obliteration of the ambulacral grooves by the approximation and fusion of
the elevated folds of perisome at their sides, which may occur to so great an extent in
Actinometra, is found also at the ends of the arms and pinnules of Aiitedon Eschrichtii.
Ludwig states (p. 75) that their terminal segments have no ambulacral groove or
tentacles ; and he gives a figure of a section through the end of a pinnule (pi. xiii.
fig. 12), the ventral surface of which is convex, while there is no ciliated epithelium
or subepithelial band (ambulacral nerve), although in the text Ludwig makes no
mention of their absence. I have found the gradual obliteration of the groove in
these cases to take place in precisely the same manner as in Actinometra, the only
difference being that the point at which the sides of the groove meet and fuse is much
further from the disk in the one case than in the other.

If we suppose, with Ludwig, that the subepithelial band is the sole structure of a
nervous nature in the whole Criuoid organization, it is difficult to understand the fact,
which Ludwig himself admits (p. 10), that it gives off no branches except those which
go to the tentacles. It is true that in the Ophiuridea the ambulacral nerve does give
off branches which go to the muscles, besides those proceeding to the tentacles, as
described by Lange \ Teuscher -, and Simroth - ; but the researches of the first-mentioned
observer render it very doubtful whether the representative in the Ophiuridea of the

1 " Beitr. z. Auat. und Histiol. d. Astericn und Opkiuren," Morphol. Jahrb. ii. Heft 2, p. 241.

2 " Bjitr. (fee, II. Opliiurida?," Jenais. Zeitsc.h. x. p. 274.

8 " Ar.at. und Scbizogonie der Ophiactis virens, Rars," Zcitsch. f. wissensek. Zool. xxvii. p. 473

ME, P. H. CARPENTER ON THE GENUS ACTINOMETRA. 37

subepithelial baud of Comatula takes any part in the formation of these branches.
Ludwig further admits that he has been quite unable to find any sense-organs at the
ends of the arms or pinnules of Comatula like those which exist in the Asteridea, and,
in discussing the views of Greeff, expresses it as his opinion (p. 78) that " die subepi-
theliale Faserlage, welche durchsetzt wird von fadenformigen Verliingerungen des dariibcr
gelegenen Epithels allein den Nerven darstellt." There can, I think, be little doubt
that this subepithelial band is of the same nature in the Crinoids and Asterids ; and
it is therefore very interesting that the nervous nature of this structure in the Asterids
has recently been disputed by Lange \ who regards as nervous only some cellular
masses separated from the subepithelial band by a lamella of connective tissue, and
projecting into the lumen of the two nerve-canals. He believes these cell masses to swell
into a large ganglionic mass beneath the pigment-spot ; while, in his opiniou, the sub-
epithelial band, together with the ciliated epithelium and the cuticula, constitutes a pro-
tecting integumentary layer. Lange finds a corresponding condition in Ophiura te.vlu-
rata, in which the radial nervous system is better developed than in the Asterids, and
consists of a series of paired ganglionic masses, connected with one another by transverse
and longitudinal commissures. On the ventral side of this ganglionated cord is a longi-
tudinal band, which Lange regards as the homologue of the protecting integumentary
layer formiug the floor of the ambulacral groove of the Asterids, and which, as is uni-
versally admitted, corresponds to the subepithelial band, epithelium, and cuticula of the
ambulacral grooves of the Crinoids.

Langc's views have been partially accepted by Simroth 2 ; but the correctness of them
is altogether denied by Teuscher 3 , who regards Lange's nervous cell-masses in the
Asterids simply as the " geschichtetes Epithel " on the wall of the nerve-canals; while
the terminal ganglionic mass under the eye-spot described by Lange is represented by
Teuscher (pi. xix. fig. 22) simply as a " bindegewebiges Polster." Ludwig 1 , too, speaks
of the nervous cell masses as local thickenings of the epithelium of the nerve-canals, which
are not present in every species. This is naturally a very strong argument against
Lange's views ; but Ludwig omits to apply similar reasoning to his own opinions regarding
the Crinoid nerves. The subepithelial bands {his nerves) are not constant in every
arm of many species of Actimomeira. Still less do Teuscher and Lange agree about the
nervous system of the Ophiurids ; Lange's ganglionic masses are described as artificial by
Teuscher, who, as in the Asterids, regards as the nerve only the fibrillar structure repre-
senting the subepithelial band of Comatula.

The question is still an open one ; and it is therefore of no slight interest to learn that
the supposed ambulacral nerve, or subepithelial fibrillar band, is not always present in
the arms of Comatula, and that even when it exists it is certainly not motor in function,
even if it be a nerve at all 5 -,

1 Morph. Jalirb. ii. 274. - Zeitseh. f. wiss. Zool. xxvii. pp. 550-500.

3 " Beitr. &c, III. Astcriden," Jen. eitsch. x. p. 513.

* "Beitrage zur Anatoinie der Aetoriden," Zeitschr. fur wiss. Zool. xxx. p. 191.

s It is worth noticing here that the " ambulacral nerve " of Comatula must bo derived either from the niesoblast
or from the hypoblast of the embryo. It is developed immediately beneath the tentacular atrium of the pentacrinoid
larva, which Gotte has shown to be the most anterior portion of the left peritoneal sac. This is lined by hypoblast,

38 MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

(§ 25) We have seen in sect. 22 that in certain of the arms of Actinometra the water-
vessels are simple tubes, like the integumentary water-vessels of the Jfolpadhlce, and are
not in connexion with any tentacular appar/atus. "Whether the mouth be radial or
interradial, the non-tentaculiferous arms are invariably the aboral ones ; so that in the
latter case they belong to the trivium (PI. I. figs. 0-15), and in the former to the bivium
(PL I. fig. 5) \ This last, however, is not always the case ; for I have a specimen of
Act. solaris in which an anterior arm (d) of one of the two ambulacra of the bivium is
tentaculiferous, while a posterior arm (EJ in the trivium has no tentacles ; it is never-
theless aboral in position, as maybe seen from PI. I. figs. 2, 5.

In only one individual of Act. polymorpha (PI. I. fig. 15) have I found a non-tenta-
culiferous arm on one of the two anterior radii (A, B.) ; but this was a very remarkable
case. Out of 31 arms, 19 were entirely devoid of a tentacular apparatus ; and in
15 of these the fusion of the two sides of the ambulacral grooves had taken place cither
on the disk or in the basal arm-segments, so that an " ambulacral nerve " was wanting
in nearly half the total number of arms. In the other four non-tentaculiferous arms the
groove remained open for a short distance, and then closed in the manner above
described. Three of these four arms constituted the anterior division (E 2 ) of the left
lateral ambulacrum ; but the fourth was the first arm of the left anterior ambulacrum (A x ),
and was borne upon the same palmar axillary as a well-developed ordinary tentaculiferous
arm. Pieces of the middle portions of these two arms are represented in PI. II. figs. 3 and 5,
and their terminations in figs. 4 and G. With this exception, I have invariably found the
non-tentaculiferous arms on the aboral side of the disk ; their number and distribution,
however, vary extremely, not only in different species but in different individuals of
the same species.

Thus in Act. poly morpha, in Plate II. fig. 8, the former is as low as -^ of the total
number of arms, while in fig. 15 it reaches Jf. Even in two individuals with the same
number of arms it may be different ; thus in figs. 8 and 9 it is -£q and ^ respectively,
and in tigs. 12, 13 it is ^f and |f . The individual represented in fig. 12 was also
remarkable for the fact that one of its aboral arms belonging to the posterior division
of the left lateral ambulacrum (E,) was tentaculiferous, while those on either side of it
were not so.

In all the specimens of the type of Act. polymorpM which I have examined, and in
three of its varieties, of which I have, unfortunately, only single specimens, more or

and appears to be separated from the hypoplastic epithelium lining the water-vascular ring by a remnant of the meso-
blastic tissue which occupied the blastoccel of the Echinopoedium. One or other of these two layers, the hypoblast lining
the atrium, or the mesoblast between it and the epithelium of the water-vascular ring, must give rise to the " ambu-
lacral nerve,'' which cannot be in any way derived from the epiblast. I am inclined to believe that the " nerve " is
most probably of mesoblastic origin, and that the remainder of the mesoblast (in this position) is converted into the
muscular layer of the ventral wall of the water-vascular ring; whilo the blood-vascular ring is a remnant of the
primitive blastoccel. Huxley ('Anatomy of Invertebrate,' p. 559) has suggested a similar origin for the nerve-canals
(perihamral canals, Ludwig) of the Asterids.

1 In all these figures (PI. I. figs. 5-16) the tentaculiferous ambulacra are indicated by dark lines, and the non-
tentaculiferous grooves by fainter lines.

ME P. H. CARPENTER OX THE GENUS ACTINOMETRA. 39

fewer of tlie arms have no tentacular apparatus ; but in the fourth variety (PI. I. fig. 1G)
all the arms are of the usual character, with open ambulacral grooves fringed with
crescentic leaves and groups of tentacles. I have found the same variation to occur also
in Act. Solaris. In this case the number of arms is limited to ten, which may be all ten-
taculiferous, or from one to four of the aboral arms may have no tentacular apparatus.
[Note. February 1879. — No less than 23 out of 48 species of Actinometra brought
home by the ' Challenger ' have more or fewer grooveless arms. I have cut sections of
these arms in two species, and have obtained the same results as with Act. polymorpha
and Act. Solaris. The "ventral nerve" and ambulacral epithelium are conspicuous by
their absence, while the axial cords in the skeleton give off branches freely in the centre
of each arm-joint, as I have already described for other species, both of Antedon and of
Actinometra. Two points are noteworthy. In one species one of the posterior ambu-
lacra stops quite abruptly on the disk, and the two arms to which it would naturally
have gone, with its " nerve," tentacles, &c, receive no branches from any of the adjacent
grooves to supply the deficiency. Again, in one of the largest Comatulce 1 have ever
seen (a ' Challenger ' specimen from the Philippines) there are more than 100 arms,
manv of which are both grooveless and nerveless, as I have found from sections. But
these abnormal arms are not limited to the hinder part of the disk as is usually the case ;
for there are several on each radius.]

The distribution of the non-tentaculiferous arms in Act. polymorpha varies, like their
number, to a very great extent. In any case they always occur on the odd posterior radius,
D (PI. I. fig. 8) ; when more are developed they may occur on the posterior divisons,
G, and E l5 of the two lateral radii, C, E ; and they may then be called postero-lateral
(PI. I. figs. G, 12-11) ; and when the proportion of non-tentaculiferous to tentaculiferous
arms hecomes very great, more or fewer of the antero-lateral arms, C 13 E 2 , belong to the
former class (PL I. figs. 7, 9-11, 13), while in exceptional cases non-tentaculiferous arms
may occur even on the anterior radii (fig. 15, A).

(§ 20) The condition of the ambulacral groove and of the tentacular apparatus is not
the only point in which the anterior or oral may differ from the posterior or aboral
arms. The former taper very slowly, contain far more segments, and are much longer
than the latter, while the form of their terminal portions and of the pinnules which
these bear is altogether different (PL I. figs. 4, C). When viewed from the dorsal side
(PL II. fig. 7) the basal portions of the two kinds of arms are precisely similar ; they
widen slightly between the first and second syzygia, *. e. from the third to the tenth
brachial, remaining uniform till the third syzygium on the fourteenth brachial, after
which they hegin to taper. Up to about the twenty-fifth or thirtieth segment the oral and
the aboral arms decrease in width at about the same rate ; but from this point onwards
there is a great difference between them. The arms borne by the two anterior radii,
A and B, taper very slowly, the length of their segments increasing considerably, while
the breadth only diminishes very gradually ; at the same time the middle and terminal
pinnules, in which no genital glands arc developed, become very long and filiform, and
remain so until the last few segments, when their length suddenly diminishes very
considcrahly (PL II. fig. 1).

40 ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

I have never been able to ascertain what is the precise mode of termination of these
anterior arms ; even when the arm ends in such a manner that there is no reason to
suppose that its terminal segments have been broken off, its few last pinnules appear
simply as immature, and the last pair are separated by a delicate prolongation of the
arm-stem, on which no pinnules have been as yet developed. Dr. Carpenter : has found
the same " growing-points " at the ends of the arms of Ant. rosacea, all of which are of
the same character as the oral tentaculiferous arms of Actinometra ; and he was never
able satisfactorily to determine the normal mode of termination of the arms.

With the posterior arms of Actinometra, however, the case is different. From the
twenty-fifth segment onwards they taper very rapidly, and instead of reaching a length
of 145 millims., as the anterior arms with some 150 segments may do, they have only
some 80 segments, and rarely attain a greater length than G0-70 millims.

At the same time their terminal pinnules are little, if at all, longer than those of the
middle portion of the arm (PI. II. figs. 5, 6) ; and the centre of the dorsal half of each
of- their segments is occupied by a dark-brown egg-shaped body, of a peculiar cellular
nature, which I have reasons for believing to be a sense-organ 2 (PI. II. fig. 6, o.b).
These bodies commence to appear in the pinnules at about the beginning of the second
third of the length of the posterior arms, and are continued to their extremities. The
pinnules of the last few segments decrease very slowly in size ; and the arm ends in an
axillary segment which bears two pinnules of the ordinary character, each provided with
the brown ovoid bodies or " sense-organs" (PL II. fig. 6, o.b).

These bodies, which may occur, though but rarely, on one or more of the anterior
tentaculiferous arms, do not exist in all the specimens of Act. polymorpha which I have
examined. In three out of my eight specimens of the type they are entirely wanting ;
and they are also absent in all the single specimens of the four varietal forms which
I have investigated. I have also failed to find them in the non-tentaculiferous arms of
Act. Solaris*.

Between these two kinds of arms, the long anterior ones on the radii A, B, with a
wide ambulacral groove and a well-developed respiratory apparatus, and the short
posterior ones of the radius D with a closed groove and no external respiratory appa-
ratus, all possible forms of transition may occur. As a general rule, more or fewer of
the antero-lateral arms, d and E 2 , are tentaculiferous ; but they never reach such a great
length as the anterior arms, and their terminal pinnules are by no means so long and
slender. At the same time the postero-lateral arms, C 3 and E l9 although generally non-
tentaculiferous, have, except in rare cases, a more or less open groove for the greater part
of their length, which, while greater than that of the posterior arms of the radius D,
is less than that of the antero-lateral arms of C L and E 2 ; and their pinnules increase
slightly in length from the middle till near the end of the arm.

1 Phil. Trans. I860, p. 723, plate xxxviii. fig. 4. = Journ. Anat. & Phys. vols. x. xi. Incc. citt.

3 Sense organs occur in two of the ' Challenger' species — one from Banda (which is probably the young of Act.
jydlymorpha), and one (a new species) from the Admiralty Islands. In both cases they are limited to the hinder arms,
some of which are grooved and others not.

MR. P. H. CARPENTER ON THE GENUS ACTINOMETEA. 41

The arms of Act. polymorpha may thus be roughly classified as follows : —

Anterior, on radii A and B, 120-150 segments. Pinnules increasing in length to the terminal ones,
which are very long and slender. Tentacnliferous.

Anterolateral, on C t and E s , 100-120 segments. Terminal pinnules long and slender. Tenta-
cnliferous.

Posterolateral, on C 2 and E,, 80-100 segments. Terminal pinnules stout, and rather longer than the
median ones. Usually have " sense-organs " and narrow ambnlacral grooves, but are non-tentaculiferous.

Posterior, on radius D, 60-80 segments. Terminal pinnules stout, but shorter than median ones.
Sense-organs. Usually no grooves. Non-tentaculiferous.

Another difference between the anterior and posterior arms is that the genital glands
in the latter are far more developed than in the former. Not only is their number
greater, although the total number of pinnules on a posterior arm may not be much
more than half that of an anterior arm, but they also attain a very much greater size ;
the basal and median pinnules of an anterior arm being very much less swollen than
the corresponding pinnules of a posterior arm. A similar inequality in the development
of the genital glands has been noticed by Alex. Agassiz 1 as occurring in the Echini. This
difference in length in the anterior and posterior arms of. Act. polymorphs, and in the
character of their terminal pinnules, seems to be to a certain extent dependent upon
the condition of the respiratory apparatus occupying their ventral surface. When this
is well developed the arm seems to have the power of indefinite growth ; for in the
single specimen (PL I. tig. 16) in which all the thirty-three arms were normal and ten-
taculiferous as in Antedon, there was no very appreciable difference in the lengths of
the anterior and posterior arms-. The shape of the terminal pinnules, however, was of
a slightly different character in the two cases, though the development of the genital
glands was about the same ; and we have just seen that those arms are the shortest in
which the ambulacral groove entirely closes, and the water-vessel is reduced to a simple
tube without any lateral tentacular branches, while it is in these arms only that any
definite mode of termination is known. This may occur before half the number
of segments have been developed which are commonly met with in an anterior ten-
tacnliferous arm.

(§ 27) The ventral surface of some of my specimens of Act. polymorpha is marked by
small calcareous concretions, somewhat resembling the " blumenartige Knotchen mit
mehreren blattartigen Fortsatzen " described by Midler 3 in the Vienna specimen of
Act. Solaris. When present, they are usually scattered around the peristome, and

1 ' Revision of the Echini," part iv. pp. 680, 681.

2 Not only are the arms of different, lengths in the ' Challenger ' species of Actinometra, which have ungrooved
hinder arms, but there are three species in which the anterior arms are longest, although all, anterior and posterior
alike, are grooved and bear tentacles. In another species the arms are all grooved and all equal in length, but the
distribution of the syzygia is quite different in the anterior and posterior arms.

:i • Gattung C'omatula,' p. !"_'.

SKCON'U SERIES. — ZOOLOGY, VOL. LI. b"

42 ME. P. H. t'AKPENTEK ON THE GENUS ACTINOMETRA.

disposed along the edges of the primary groove-trunks proceeding from it, and there
are generally some upon the sides of the anal tube. They are particularly well de-
veloped in the dark variety from Ubay, in which all the arms are tentaculiferous.

(§ 28) The "oral pinnules" of Act. polymorpha, those, namely, which arch over the disk
so as to protect it, are borne by the second distichals and second palmars when these are
present, but in any case upon the second brachials, those of the distichals and palmars
being the longest. They are all very long and slender, consisting of some 30 or 40
segments ; and their terminal portions exhibit the peculiar characteristic comb made up
of processes which rise from the outer margin of the ventral surface of each calcareous
segment (PI. III. fig. 2), just as in Act. Solaris and Act. pectmata (PI. III. fig. 1). The
number of segments on which these processes may be developed varies from 10-12 on
a distichal pinnule, to 6-8 on a brachial pinnule ; but in cases in which no second
distichals or palmars are developed, so that the pinnule on the second brachial is the
first of the series, it is much longer than usual, and more of its terminal segments bear
the comb-like processes.

The oral pinnules of the dark Ubay variety of Act. polymorpha differ considerably
from those of the type and of other Actmometrce ; not only are they much stouter, but
their terminal comb is differently constituted (PL III. fig. 3). As is usually the case, the
lower processes gradually develope themselves from the outer margin of the ventral
surface of each calcareous segment ; but towards the end of the pinnule they gradually
come to rise less and less from the outer margin, and more and more from the median
portion of the ventral surface of each segment, until finally, on the last two or three
segments, they are developed from the inner margin. Consequently the comb, when
viewed from above, is seen not to lie altogether on the outer side of the pinnule, as is
usually the case, but to start from the outer side, cross its ventral surface, and finally
come to lie on the inner side of each pinnule, i. e. on the one nearest the arm.

Both in the type of Act. polymorpha and in all the four varieties, the pinnules
diminish in length from that of the second distichal (when present) to those borne by
the fourth and fifth brachials; that of the sixth brachial is longer, and usually contains

well-developed genital gland, so that it is slightly swollen. Prom this point onwards
the pinnules increase in length till about the thirtieth brachial, after which their length
and character vary according as the arm is tentaculiferous or non-tentaculiferous.

(§ 29) The dorsal aspect of Act. polymorpha differs from that of most Antcdons, and
especially from that of Ant. rosacea, in the fact that the plane of the second and third
radials, like that of the first, is parallel to the vertical axis of the calyx, and not inclined
to it, as in Antedon ; so that the dorsal surfaces of the whole of the pieces of the calyx
lie in one horizontal plane. The centrodorsal piece is circular (PI. II. figs. 9, 10, c d), or
pentagonal (fig. 11), and conceals a large portion of the pentagon formed by the first
radials, less in young specimens with but a few arms (fig. 9) than in large and full-grown
specimens with many arms (PI. II. figs. 10, ] I & PI. VI. fig. 2). It is usually a flattened
plate with a slight concavity in the centre of its outer surface; and around its margins

ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 43

are disposed some 20 or 25 cirrhi in one row, but with occasional traces of a second,
in which the cirrhi alternate in position with those of the first row. The number of
segments in each cirrhus is normally from 11 to 14, of which the last forms a recurved
claw, while a more or less distinct spine is usually visihle upon the dorsal edge of each of
the three or four penultimate segments (PI. III. figs. 8-11).

In PI. II. fig. 8, is seen an abnormal condition of the centrodorsal piece, which is of
an irregular oval form, and so extended as to conceal large portions even of the second
radials. These last are usually more or less completely united with one another late-
rally. The amount of their union is to a certain extent dependent upon the number
of arms developed. Thus in the small specimen with only 13 arms, represented in
PL II. fig. 9, the second radials are not united laterally for more than half their length;
in fig. 10 (2(3 arms) the union is somewhat more complete, and even more so in fig. 8
(28 arms), while in the variety with 39 arms, represented in fig. 11, the second radials
are completely and closely united with one another all round. This rule, however,
appears to he only a specific one, and not generally applicable to all Comaiulce ; for in
the SO-armed Phanogenia the second radials, as figured by Loven ', do not appear to
be united with one another any more closely than they are in the small 13-armed speci-
men of Act. polymorpha (PI. II. fig. 9).

In Act. polymorpha the two segments (first distichals, palmars, or brachials) borne by
any axillary are united to one another laterally to about very much the same extent as
the second radials are ; i. c. when the number of arms is small, their first segments,
whether primary, secondary, or tertiary, are not laterally united in pairs with such com-
pleteness as when the division of the ten primary arms is carried to any considerable
extent (PI. II. figs. 8-11, d lt p u &J.

When the arm-division is unequal it is generally carried further in the trivium or
posterior radii, C, D, E, than in the two anterior radii, A, B, which form the bivium.
This is well seen in PI. II. fig. 9, in which no distichals are developed on either of the
two anterior radii ; and again in fig. 10, in which, while distichals are developed all
round, the division is carried no further in one of the anterior radii, while in each of the
others from one to three palmar series may be developed. In only four normal cases
have I found an anterior radius to bear more arms than a posterior one. In each of
these the total number of arms was considerable, and one at least of the two posterior
radii bore the same number of arms as the abnormal anterior one. Thus, for example, in
PL II. fig. 11, each of the radii bears eight arms, with the exception of the posterior one
(D), on which only seven are developed. This, however, is an abnormal case of fracture
of the whole radius between its second and third segments. The new portion is con-
siderably smaller than the old, the proximal articular face of the new axillary being far
less wide than the corresponding distal face of the old second radial ; while both the
distichal series which it bears are imperfect and abnormal, so that the absence of a
further division in one of the secondary arms is not. particularly remarkable.

(§ 30) The number of arms that may be developed in Act. polymorpha is a character

1 " Phanogenia," loc. cit. p. 230.

SECOND SERIES. — ZOOLOGY, VOL. II. 7

41

3IB. P. H, CAEPENTEE OX THE GENTTS ACTLXCttlETEA.

of extreme variability. In the specimens I have examined it varies from 13 to 39 ;
so that, with one remarkable exception (PI. II. fig. 8), the ten primary arms do not, at
the most, divide more than twice, while in two specimens with 18 and 13 arms
respectively two and seven of the primary arms remain undivided. I believe, how-
ever, that, as a general rule, an axillary is developed on each primary arm, and that the
amount of further division is variable, but that a tertiary division is probably excep-
tional, so that the number of arms in this species will be found rarely to exceed 40.

It will have been already apparent from the position assigned to Act. polymorpha in
the classification given in sect. 20, that I consider the typical number of distichals and
palmars in this species to be three, of which the second (d,) bears a long pinnule, while
the third or axillary segment (d a) consists of two primitive segments united by a
syzygium. A typical specimen of this condition is seen in PI. II. fig. 10. Out of the
twelve specimens of this species which I have examined, but four others resemble this
one in having all then distichal and palmar series regularly developed. In each of the
other seven specimens one or more of the distichal or palmar series is irregular, con-
sisting onlv of two segments, the second of which is axillary without a svzvgium. In
one very remarkable case, represented in PI. II. fig. 8, one of the palmar series is
reduced to a single segment placed on the distichal axillary, being also itself an
axillary bearing the brachials directly on one of its articular surfaces, while on the other
are two segments which may be called suprapalmar, of which the second [sp.a) is an
axillary without a syzygium, and bears two arms.

Excluding this remarkable case, the comparative frequency of the usual variations in
the distichal and palmar series in the twelve specimens of Act. polymorpha examined by
me is seen in the accompanying Table. Prom this it appears that out of 111 distichal

Table I. — Showing the Variations in the Distichal and Palmar Series.

Specimen.

Total
number

of
Arms.

Bis-iehal Series. Palmar Series.

Total
n umber.

Of three

segments.

Of two
segments.

Total
number.

On two distichals. On three distichals.

Of two
segments.

Of three
segments.

Of~two

segments.

Of tl

segments.

I.

II.

III.

IV.

V.

VI.

VII.

YIII.

Yar. 1
2
3

4

Total

13

1-
20
25
26
28
28
31

20
29
39
33

3

8
10
10
10
10
10
10

10
10
10
10

3

4
5
10
10
7
10
10

10

10

8

9

4
5

'3

"2
1

5
6
5

8
11

ii
19
13

2

1
1

3
1

2
1

2
6
5

7
11

9
14
11

310

111

96

15

70

2

2

1

65

ME. P. H. CAEPEXTEE OX THE GEXTJS ACTIXOATETEA. 45

series 96 were normal, and that out of 70 palmar series 65 were normal, i. e. consisted
of three segments, of which the second bore a long pinnule, while the third was axillary,
with a syzygiiun. The three forms of variation exhibited by the abnormal palmar
series are of considerable interest, because some of them, at least, represent the normal
condition of the palmars in other groups of Actinometrce. Thus the most frequent one,
two palmars on three distichals, is typical for Act. multifida, while that of three palmars
on two distichals is typical for Act. rotalaria. The third variation, two palmars on two
distichals, occurs in Act. tenax, Ltk , and in a few new ' Challenger ' species ; and it is
typical in several species of Antedon — for example, in. Ant. palmata and Ant. articulata.
Specimens Xos. II. and III. are remarkable for the fact that the numbers of regular and
irregular distichal series are in each case equal to one another; so that a specific diagnosis
based upon either of these specimens alone, would, as is evident from the above Table,
have been entirely incorrect.

The amount of variation in these characters is so enormously great that only after
examination of a considerable number of specimens is it possible to draw conclusions
of any value respecting the use which may be made of these characters for systematic
purposes. The above Table, however, will, I think, show clearly that I am justified
in assuming the normal number of both distichal and palmar segments in this species
to be three, of which the second bears a long pinnule, and the third is axillary with a
syzygium.

(§31) The same variability occurs in the position and distribution of the syzygia
in the arms, but, as might be expected from the nature of the case, to an infinitely
greater extent. In most of his specific diagnoses Miiller gives the position of the first
svzv^iuin on the arm and the average number of segments which occur between everv
two successive syzygia throughout the rest of the arm. Only in a very few cases
does he make mention of the position of the second syzygium, which I believe to be
a character of nearly or quite as great systematic value as the position of the first ; and,
owing to its greater constancy, of considerably greater value than the number of seg-
ments between every two successive syzygia, which I will call the " syzygial interval."
It will be seen from Table I. that the total number of arms in the 12 specimens of
Act. polymorpha at my disposal reached 310 : 11 of these were broken below the third
segment ; but of the remaining 299, the first syzygium was on the third brachial in
283 cases ; and in 156 of these the second svzv«ium was on the tenth brachial. The
irregnlarities in the position of the first syzygium were limited to tbree specimens,
and, as will be seen from Table II., nearly all confined to one variety.

46

ME. P. H. CAEPENTEE ON THE GENES ACTINOMETEA.

Table II. — Showing Irregularities in the Position of the first Syzygium.

No.

Irregular series of Syzygia.

Type.

Variety.

III.

VI.

4.

Brachials.

1.

On 4, 9, 13.

. .

1

'2.

4, 10, 14.

2

3.

5, 13, 16.

1

4.

6, 10, 14.

1

5.

9, 13, 17.

2

6.

10, 14, 16.

1

7.

10, 14, 18.

1

4

2

8.

10, 14, 19.

1

Total

2

4

10=16

In nearly every case the irregularity appears to have been the result of regeneration,
the arm having been broken, either in the distichal or in the palmar series, or between
the third brachial and the preceding axillary, and a new one developed with an irregular
syzygial series ; although in many cases similarly regenerated arms of other specimens
exhibit perfectly normal series of syzygia. One of these unusual cases, in which there is
no syzygium on the third brachial (b 3 ) and the first syzygium occurs on the tenth seg-
ment (6 10 ), which is usually the position of the second syzygium, is seen in PI. II. fig. 8.

We have seen that when the first syzygium is on the third brachial, the position of the
second is in the great majority of cases on the tenth brachial ; that is to say, the first
syzygial interval is six simple segments, while the second and all the subsequent intervals
are, as a general rule, only three simple segments, though the range of variation on cither
side of this number is very great.

Table III. shows the variations in the positions of the second and third syzygia in all
those 283 arms in which the first syzygium is on the third brachial. From the last
column of this Table it is evident that in Act. polymorplia and its varieties the normal
position of the second syzygium is on the tenth brachial, and that in those cases in
which it does not occupy this position it is much oftener on the eleventh or twelfth
segment than on the eighth and ninth ; i. e. that variation, when it occurs, is in the direc-
tion of increase rather than of decrease in the length of the first interval. This is more
clearly seen in Table IV., which shows the number of segments intervening between the
first and second syzygia in all the above cases.

MR. P. H. CAEPENTEE OX THE GENUS ACTIXOMETEA.

47

Table III. — Showing the Variations in the Positions of the second and third
Syzygia in the Arms of twelve specimens of Act. polymorpha.

No.

Positions of the
first three Syzygia.

Type.

Vnr. 1.

Var. 2.

Var. 3.

Var. 4.

Total.

No. of Variations

in the

first interval.

Brachials.

1.

On 3, 4, 5.

1

1

2.

3, 4, 9.

1

1

3

3.

3, 4, 10.

••

•■

1

1

4.

3, 5, 10.

2

2

2

5.

3, 6, 11.

1

1

1

G.

3, 7, 11.

o

o

2

7 •

3, 8, 11.

1

1

8,
9.

3, 8, 12.
3, 8, 13.

1
1

2

o

5
1

8

10.

3, 8, 14.

1

1

11.

3, 9, 11.

1

1

12.

3, 9, 12.

1

1

23

13.

3, 9, 13.

10

h

3

20

14.

3, 9, 14.

1

1

15.

3, 10, 12.

1

1

16.

3, 10, 13.

3

1

4

s

17.

3, 10, 14.

76

10

23

16

13

138

18.

3, 10, 15.

1

1

150

19.

3, 10, 16.

1

1

1

3

20.

3, 10, 17.

3

1

. .

4

21.

3, 10, 19.

1

1

22.

3, 11, 14.

3

1

4

23.

3, 11, 15.

33

4

1

3

2

43

50

24.

3, 11, 16.

3

••

3

25.

3, 12, 14.

1

1

26.

3, 12, 16.

30

1

1

32

34

27.

3, 12, 18.

1

1

28.

3, 13, 16.

1

1

29.

3, 13, 17.

1

1

2

30.

3, 14, 18.

Total number of 1

variations . . J

1

1

2

2

174

20

28

38

23

283

283

48

ME. P. II. CAEPENTEE ON THE GENUS ACTINOMETEA.

Table IV. — Showing the Variation in the number of Segments in the

first Interval.

No. of
Segments.

0.

1.

2

o.

4.

5.

(5.

7.

8.

9.

10.

Total.

Type ....

Var. 1 . . .

2

3 ...

4 ...

1
2

2

1

2

4

2
o

13

i
3

81
14
25
22
14

39
4

I

Q

O

32
1
1

1
1

1
1

174

20
28
38
23

Total

3

')

1

2

8

23

156

50

34

2

2

283

It is also seen in Table III. that even when the second syzygium is abnormally placed,
it is usually the case that the interval between it and the third is the normal one of three
simple segments, so that scries like 3, 9, 13 ; 3, 11, 15 ; and 3, 12, 16, are very common.
This is well seen in Table V., which shows clearly that the length of the second interval
is normally three segments ; that, like the first, it tends to vary in the direction of excess
rather than of defect, and that the range of variation in both directions is greater in the
varieties than in the type of Act. polymorpha.

Table V. — Showing the Variation in the number of Segments in the

second Interval.

No. of

Segment 3.

0.

1.

2,

3.

4.

g

6.

8.

Total number
of Arms.

Typo

Var. 1 ....
2

3

4 ....

i

o

y
1

'4
1

151
16
25
30
21

8

3

1
1
1

3
1

1

174
20
28
38
23

Total

1

3

15

243

10

4

1

2S3

After the fourteenth brachial a syzygium usually occurs on every fourth segment ; so
that the number of segments composing the syzygial interval is normally three. It is, how-
ever, very unusual to meet with an arm in which this interval is constant throughout
its whole length and does not vary to a greater or less extent. In only seven arms out
of the whole number which I have examined have I found this to be the case, together
with normal first and second intervals, although twenty-three other arms were regular
from the second syzygium onwards. These thirty arms were distributed among five out

ME. P. H. CAEPENTEE ON THE GENES ACTIX03IETEA.

49

of the eight specimens of the type, while in none of the other three was the syzygial
interval constant throughout the length of any of the arms ; the same was the case with
the four varietal specimens.

Table VI. — Showing the Variations in the Syzygial interval (usually 3 segments)
in the Arms of twelve specimens of Act. poly morjpha.

3.

4.

5.

6.

7.

8.

9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
2-6.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.

Number of Segments in
each interval.

1.

1,2.
1,5.

0.

0, 0, 0.

0, 2.
1.
1,
1,
1,
1,

1, 2,2.
1, 4.

1, 11.
o

2, 0.
2,1.

oi l -> ■)

**) X 5 "*-> ~ ? — ■

—) x , ^j — •
2, 1'.

.) .. o

— , — , _.

., .; ., ..

— , — . — , — .

■) ■) -. •) g

O ■> .> o .) ., a

— > -> -. -- -- -. -•

2,4.

2, 4, 2.

2, 4, 5, 1.

2, 5.

2, 0.

4.

4 1 2 4

4,2.

4 2

4, 2, 2, 1, 2.

4,4.

4 4 12

4, 4, 4.
5.
5,2.

5, 2, 1.
6.
6,4.

7.
10.

Type.

:;

1

11

2

1
1

5
2

1

109

1

3
1
1

31
1

11
4
4
1
o

18

1

o

3

2

1
3

Var. 1.

Var. 2.

13

Var. 3.

Var. 4.

1
1

10

1

1

is

15

Total Cases.

11

1

24

8
3
1
12
2
1
1
5
2

3
1

146

2

3
1
1

32
1

11
4
4
1
3
1
1
1
1

82
1
3
1
1
1
1
1
8
o

1

5
1

2
1

50 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

As might be expected from the nature of the case, the range of variation of the
syzygial interval for the whole length of the arms is considerably greater than that of
the second interval alone. In the type and in variety 3 it differs also in a tendency to a
decrease rather than an increase in the length of the interval, which is more often two
segments than four, as is seen from Table VI. ; while in the other three varietal spe-
cimens the tendency of the variation is to increase in the length of the interval, four
segments occurring much more commonly than two. "With respect to these varieties,
however, it must be remembered that these conclusions arc all based upon an examina-
tion of single specimens, which, as already mentioned, may in some instances be very
misleading.

(§ 32) The colour of Act. polymorpha is usually (in spirit-specimens) a yellowish
brown, which is much darker in the soft parts of the body than in the elements of the
skeleton. In variety 4, from Ubay, the colour is the same as that of the type, but con-
siderably darker, so that the disk appears almost black. In varieties 1 aud 2 the colour
is rather a greyish brown, which is considerably lighter on the ventral surface of the
disk and arms than on the dorsal skeleton ; and in variety 3 it is a somewhat reddish
brown.

In varieties 2, 3, and 4 the dorsal surface of the skeleton is marked by a median
white line, with more or less defined dark borders, which commences on the radials, and
extends for some distance onto the arms. Its distinctness varies in different specimens
and in different arms of the same specimen ; but it is especially well marked in the
darkly coloured var. 4, as is seen in PL II. fig. 7.

In an adult specimen of Act. polymorpha the total diametex*, including the arms, is
about 200 or 220 millimetres, of which about 20 millims. represents the diameter of the
disk alone ; but in one young specimen I found these two diameters to measure only
105 and 7 millims. respectively. The three specimens of varieties 2, 3, and 4 were of
about the same size as the type, but the single specimen of variety 1 was considerably
smaller, its longer diameter being only about 100 millims., and its shorter (that of the
disk) about 8 millims. This specimen, however, was, I believe, full-grown ; for it had
very large and well-developed genital glands ; while in the young and small specimen of
the type mentioned above, the size of which was about the same as that of variety 1,
the genital glands were scarcely developed at all.

(§ 33) The manuscript name of Act. armata has been given to the type of Actino-
metra here described, by Professor Semper, on account of the small spines with which
the segments of the arms and pinnules are fringed, more especially upon their dorsal
and aboral margins. As, however, this character is a very general one among the Co-
matulce, and as it is by no means so well developed in this type as in many others, I have
thought it advisable not to adopt Professor Semper's specific name, "armata," more
especially as it has been already employed by Pourtales to designate a new American
Anictlon. Under these circumstances, I propose to designate this type as Act. poly-
morpha, having regard to the enormous amount of variation which I have found to exist
in nearly all its characters.

I believe it to be very closely allied to, if not actually identical with, the type described

ME. P. H. CAEPEXTEE ON THE GENUS ACTLNOMETEA. 51

as Alecto parvicirra by Midler 1 , who gave this name to three small spirit-specimens in
the Paris Museum, from the voyage of Peron and Lesueur in 1803, which I recently
found there under the name of Comatula simplex, Mus. Midler's diagnosis of Alecto
parvicirra was based upon his examination of the three Paris specimens, which all
have an excentric mouth and a terminal comb on the oral pinnules, and is exactly appli-
cable to Act. polymorpha, except that he describes the pinnules as " ziemlich gleichformig."
In their yellow colour and smaller size (about 100 millims.) these also differ slightly
from the type of Act. polymorphic, but without a very much closer examination of them
than I was able to make, it would be impossible to arrive at a definite conclusion as to
the identity or difference of these two species.

The Vavao variety of Alecto parvicirra described by Miiller occurs in the Paris
Museum under the name of Comatula brevicirra, Troschel. This specimen differs from
Act. polymorpha in many subordinate characters, and is not absolutely identical either
with the type or with either of the four varietal specimens which I have examined ;
and I cannot but regard it as representing another of the slight and probably very
numerous modifications of this type, of which I think it most likely that Midler's
original species, Alecto parvicirra, is also a varietal form.

(§ 34) The following diagnosis will, I believe, be found sufficient for the future iden-
tification of Act. polymorpha and of tbe four varieties here referred to.

Actinometra polymorpha, n. sp.

Centrodorsal piece. A circular or irregularly pentagonal disk almost completely concealing the first
radials. Surface flattened, and slightly concave in the centre.

Cirrhi marginal, 15-25, of 11-1-1 segments, of which the fifth and sixth are the longest; hasal ones
thick, and wider than long; remainder taper gradually, and terminal ones are laterally compressed; the
last 5 or 6 segments have a small dorsal spine, increasing in distinctness up to the penultimate segment,
which hears the terminal claw.

Radials 3, of which the first are barely visible ; the second are short, and in the middle of the same
height as the first, but somewhat lower at the sides, for nearly the whole length of which they are united
to one another in pairs.

Axillary radial pentagonal, about twice as wide as the second, to which it is united by ligaments only.

Arms from 13-40 ; rays may divide three times. First segments borne by each axillary in contact
for nearly their whole side.

Distichals and Palmars. When present, 3 ; second bears a long pinnule, and is united to the first
by ligamentous articulation only. Axillary has a syzygium.

Syzygia. First on third brachial, then an interval of 6 segments to the next, and then a general
interval of 3 throughout the arm, variable from 0-6, but usually varying to < 3.

Arms. Anterior arms much longer than the posterior, which are usually non-tentaculiferous. Width
increases from 3rd to 10th segment, remains uniform till about the 14th, and then decreases, slowly
in the long anterior arms, and rapidly in the short posterior ones. Arm-segments wedge-shaped, slightly
overlapping one another, and fringed at the borders with short spines.

Pinnules. The second distichal and the second palmar, when present, bear long pinnules, of which
the palmar bears the shorter one; the next is on the second brachial, and still shorter, and the length

1 ' Gattung Comatula,'' p. 24.
SECOND SERIES. — ZOOLOGY, VOL. II. 8

52 ME. P. H. CABPENTEB ON THE GENUS ACTINOMETBA.

gradually diminishes to the pinnules of the fourth and fifth brachials, -which are the shortest on the
whole arm. From the sixth brachial onwards the pinnules are long and stout, gradually increasing in
length and thickness to near the middle of the arm ; the thickness is greatest in the short posterior arms,
in which both length and thickness rapidly decrease from the middle to the end of the arm, while in the
long anterior arms the thickness slowly diminishes and the length slightly increases, so that the terminal
pinnules are long and slender.

Comb. The last six or eight segments of the distichal, palmar, and first eight or ten brachial pinnules
have the outer ventral margin of each calcareous segment produced into a small lancet-shaped process
which bends over towards the ventral side, so that the end of the pinnule has a comb-like appearance.
Many of the other pinnules till near the end of the arm have similar processes upon their four or five
terminal segments.

Disk. Mouth excentric and interradial ; posterior ambulacra very indistinct, and often nearly obli-
terated. Small calcareous concretions occasionally present in the neighbourhood of the peristome and
anal tube.

Colour. Yellowish brown to dark brown.

Diameter. About 20 centimetres.

Locality. Bohol.

The following are the points in which the varieties differ from the type as described above : —

Variety 1.

Cirrhi. 25, of 13-15 segments, with terminal claw ; spines on dorsal face of terminal segments not very
distinct, but the segments are laterally compressed.

Radials. Second radials completely united with one another in pairs.

Arms. 20.

Syzygial interval. Usually 3, but varying from 1-10 segments ; generally to >3.

Comb. From 2nd distichal to 6th brachial pinnules, and then, at intervals, to about 20th brachial
but no further.

The basal pinnules of the arms have a faint dorsal keel, and the distal ends of their segments are
rather wider than the proximal ends.

Diameter. 105 millims.

Colour. Greyish brown.

Locality. Ubay.

Variety 2.

Centrodorsal piece. Small, but rather thick.

Cirrhi. 10, of 11 or 12 segments, with a terminal claw; the fourth and fifth are longest; the spines
on the dorsal border begin from the middle segments, and the opposing process on the penultimate
segment is well marked.

Radials. Second radials only incompletely united ; second and third very convex, and much higher
than the first ; median dorsal line of skeleton marked by a white line with dark borders, which is lost
about the middle of the arms.

Arms. 29.

Syzygial interval. Usually 3, but varying from 1-10 segments; generally to >3.

Comb. Limited to distichal, palmar, and first five brachial pinnules ; those of the Gth and next
succeeding brachials have a dorsal keel, and the distal ends of their segments are much wider than the
proximal ones.

Colour. Greyish brown.

Locality. Cabulan.

ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 53

Variety 3.

Centrodorsal piece large and thick, with only 3 cirrhus-scars.

Radidls. Second radials completely united all round. Centre of dorsal surface of the skeleton, from
the centrodorsal till near the end of the arms, marked by a faint white line with dark borders.

Arms. 39.

Syzygial interval. Usually 3, but varying from 1-7 segments; generally to <3.

Comb. On second distiehal, palmar, and brachial pinnules, and occasionally also on those of the
3rd— 5th brachials, but on no others.

Colour. Reddish brown.

Locality. Bohol.

Variety 4.

Centrodorsal piece large and thick, with only 3 cirrhus-scars.

Radials. Second radials closely united all round. Median white line on dorsal surface of skeleton
very marked.

Arms. 33, all tentaculiferous, and tolerably uniform in length and in the character of their pinnules.

Syzygial interval. Usually 3, but varying from 0-6 segments; generally to >3.

Pinnules. Oral pinnules much stouter than in the type ; that of third brachial but little shorter than
that of second. Comb limited to these and to the distiehal and palmar pinnules, and the processes forming
it gradually come to rise from the ventral surfaces of the calcareous segments instead of from their outer
margins.

Colour. Blackish brown.

Locality. Ubay.

IV. Tiie Skeleton.
(i.) The Skeleton generally with its Ligaments and Muscles.

(§ 35) The general structure of the skeleton of Actinometra, and of the ligaments
and muscles which connect its component pieces, is precisely the same as in Antedon ;
and as this has been already described by Dr. Carpenter 1 , there is no need to repeat it
here : a few points, however, must be treated somewhat more in detail. The component
pieces of the skeleton of Actinometra, as of all the other Echinoderms, consist of a
calcareous reticulation formed by the calcification of an organic basis of a protoplasmic
nature, in which numerous nuclei and pigment-granules are imbedded. This " nuclear
tissue," as Simroth 2 has called it, is in the form of a network, around the meshes of
which the calcareous material is deposited.

The character of the calcareous reticulation varies greatly in different parts of the
skeleton, being much closer at the synostoses and syzygia and at the articular surfaces
than in the interior of the segments ; and in correspondence with this greater compact-
ness of the calcareous tissue, the organic plexus which forms its basis becomes remark-
ably modified at these points, as will be seen further on. The various modes of union
of the different pieces of the Crinoid skeleton have been closely investigated by Miiller
and by Dr. Carpenter. The former 3 described the stein-segments of Pentacrinus as united
to one another in two different ways — (1) by the tendons which traverse the whole

1 Phil. Trans, loc. cit. p. 702. : Op. cii.p. 433. 3 Bau des Pentaerinus, pp. 17-20.

54 ME, P. H. CARPENTER OX THE GENUS ACTIXOMETRA.

length of the stem, passing through the substance of its various segments, and (2) by the
" elastic interarticular substance" between the individual segments.

The substance of these tendons consists of a white fibrillar tissue very like the ten-
dinous tissue of the higher animals ; but Midler supposed the elastic interarticular sub-
stance to be of a totally different nature, consisting of " lauter senkreckt stehenden
Fasersaulchen, die durch Reihen bogenformiger Schlingen einfacher Fasern verbunden
sind," and " diese Scblingen gehen mit den regelmiissigsten Arkaden in ganz gleichen
Abstiinden aus einem Fasersaulchen in das andere liber." This substance fills up
the whole space between the successive stem-segments which is not occupied by the
tendons, and is connected in the closest possible manner with the opposed surfaces of
every pair of segments, even extending for a short distance into their superficial cal-
careous tissue. Each of the arcades above mentioned consists of a single primitive
fibre, the terminations of which are lost in the " Fasersaulchen ;" and the passage of
these fibres in loops from one fibrous column to another gives an elasticity to the whole
tissue, viz. a power of contraction after lateral displacement, and of extension after
vertical compression, although the individual fibres are not elastic in the ordinary
sense of the word.

Miiller described the basals of Pentacrinus as united with the top stem-segments in
the same manner as the successive stem-segments Avith one another, namely, by this
elastic interarticular substance, while their sides simply " stossen an einander " (p. 25).
This mode of union between the stem-segments was generally called by him a "Nath,"
or suture ; but he sometimes spoke of it as an articulation, though he usually employed
this last term only in those cases in which two segments are movable on one another
through the intervention of muscles and ligaments which pass between them.

He further described the union between the first radials and the basals of Pentacrinus
and between the first radials and the centrodorsal piece of Comatula as a suture, which
name he also gave to the lateral union of the five first radials with one another (pp. 28, 29) ;
but he does not seem to have supposed that in these cases the various elements were
connected by the elastic interarticular substance which he found between the likewise
suturally united stem-segments. In fact, in speaking of the syzygia, which he called an
immovable sutural union of two segments, he said expressly that not only the muscles
but also the clastic interarticular substance was absent, On the other hand, the latter is
to be found between the segments which are capable of motion upon one another, whether
ligaments and muscles be present, as between the first and second radials and between
most of the brachials, or ligaments onlv, as between the first and second brachials and the
second and third radials ; for Miiller described the ligaments connecting two mutually
movable segments as having essentially the same structure as the elastic interarticular
substance of the stem, except that their surface is plain and not " krausenartig gefaltet"
(pp. 30-38).

In those more common cases in which there is a muscular union between two segments,
such as the first and second radials, in contact by transverse articular ridges upon their
opposed faces (PL VII. figs. 1 b, 2 a, I b, 5 a ; I), Miiller drew no distinction between the
pair of ligamentous bundles on the ventral side of the articular ridge, and the single

MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 55

mass which occupies the whole space between the opposed faces on the dorsal side of
this ridge, describing them as alike consisting of elastic interarticular substance, the
function of which is extensor and antagonistic to the flexor action of the muscles.
Dr. Carpenter ! , however, regards the former as interarticular, with the special function
of holding the pieces together, but allowing a certain amount of movement between
them, while he describes the single dorsal mass as elastic, and as antagonizing by its
extensile powers the action of the flexor muscles. Histologically he finds no differ-
ence between them, both consisting of minute, straight, and nearly parallel fibres, very
much, in fact, like those which Miiller described as composing the tendons of the stem of
Pentacrinus. At their points of attachment to the pieces of the skeleton these fibres
pass into their basis substance and become incorjiorated with it.

The union of the first radials with one another and with the centrodorsal piece, which
had been spoken of as sutural by Miiller, is described by Dr. Carpenter as " an
adhesion of expanded surfaces closely fitted together, and held together by the con-
tinuity of their sarcodic basis substance " (p. 701) ; so that the different elements are
cemented together by a " thin layer of sarcodic substance, continuous with that which
occupies the meshwork of their own calcareous reticulation." This mode of union may
be conveniently described by the word " synostosis," which has been employed by
Simroth to designate the mode of union of two faces which " verkitten sich " in the
skeleton of the Ophiuridce. It is essentially the same as the syzygial union which
occurs between certain pairs of the primitive arm-segments, although differing from it in
points of detail.

(§ 3G) We have seen that Miiller regarded the tendinous tissue of the stem of 'Penta-
crinus, and the fibrous ligamentous bundles, or, as he called it, the elastic interarticular
substance uniting the movable elements of the skeleton both of Pentacrinus and of
Comatula, as distinct from one another. I believe, however, that they are fundamentally
identical, not only with one another, but also with the so-called "cement-substance"
between two segments which are united by synostosis. This last consists, in the Ophin-
ridce, according to Simroth 2 , of connective-tissue fibres which lose themselves in the
organic basis of the skeleton, and are of the same nature as the substance of the
masses of connective tissue uniting two articulating surfaces, both tissues staining
deeply with picro-carmine. I find the same to be the case in Comatula and Pentacrinus.
The tendons of the stem of the latter genus, the ligamentous bundles, composed, according
to Miiller, of elastic interarticular substance, which connect every pair of movable
arm-segments, and, lastly, the "cement-substance," uniting the first radials to one
another and to the centrodorsal piece, all stain very deeply with picro-carmine, and are
of essentially the same histological structure.

Dig. 1, on Plate III., represents a portion of a horizontal section through the suture,
or, as I prefer to call it, the synostosis of two of the first radials of Pentacrinus. In the
immediate neighbourhood of their apposed lateral faces there are none of the nuclei nor
pigment-granules which are imbedded so abundantly in the more internal portions of their
protoplasmic ground-substance, and the threads of the plexus of which it is composed
1 Phil. Trans, he. cit. pp. 703-714. - Op. cit. p. 435.

56 JIE, P. H. CAEPEXTER OX THE GENUS ACTIXO.METEA.

become excessively attenuated and disposed with great regularity almost parallel to one
another. At the same time the meshes of this organic plexus become greatly elongated
in the intervals between the parallel threads or fihres, which are connected with one
another by very delicate fibrils passing in the form of loops from one fibre to another.
These loops, which forcibly recall Midler's description of the arcades connecting the
fibrous columns of the elastic interarticular substance in the stem of Pentacrinus, are
simply the expression of the ends of elongated meshes of the protoplasmic plexus
forming the organic basis of the skeleton. In the neighbourhood of each of the two
opposed surfaces the fibrous elements of this plexus assume the character of closely
placed parallel connective-tissue fibres, with no pigment-granules nor nuclei imbed-
ded in them, but staining deeply with picro-carmine, while the normal protoplasmic
basis of the interior part of the calcareous segments is but little affected by this
reagent. These fibres pass from the organic basis of the one segment into that of the
other so that the two are firmly united, and the superficial denser layer of calcareous
tissue is deposited around their ends, which corresponds with Midler's description of the
elastic interarticular substance of the stem of Pentacrinus as extending for a short
distance into the calcareous substance of the opposed faces of the segments. The
superficial layer of calcareous reticulation which occupies the small intervals between
the ends of the fibres thus becomes extremely close and compact ; but the central
portion of the fibrous tissue (PI. III. fig. 4, L) does not calcify, remaining as a thin layer
of fibrous cement-substance between the two opposed surfaces, precisely like the inter-
articular substance in the stem of Pentacrinus, with which I believe it to be identical.
It is, at any rate, of the same nature as the substance of the ligaments connecting the
first and second radials, which Miiller described as identical with that connecting the
stem-segments ; for at the angles of the radial pentagon the fibres of the cement-sub-
stance connecting the adjacent first radials with one another in pairs pass directly into
the fibres of the ligamentous bundles between the first and second radials. These, which
are of precisely the same character as the ligamentous bundles between the successive
brachial segments, also stain deeply with picro-carmine, and only differ from the cement-
substance in the greater length of their fibrous element.

At the points of attachment of the ligaments to the pieces of the skeleton, the meshes
of the organic plexus become greatly elongated, and its fibrous bars regularly disposed
and connected with one another by loops, as above described. As, however, the distance
between the two articulating faces is very much greater than in a synostosis, several of
these minute primitive fibrils unite to form one of the larger fibres composing the liga-
mentous bundle, at the other attachment of which these primitive fibres again separate,
become connected with one another by transverse loops, and finally pass into the bars
of the protoplasmic plexus forming the ground-substance of the next segment.

(§ 37) I have found the fibres composing the ligamentous bundles between the
arm-segments of Antedon to terminate in the manner above described for Pentacrinus;
but in Act. poli/morpha they do not pass so directly into the organic basis of the seg-
ment. At the ends of the ligamentous bundles, where their component fibres begin to
break up into primitive fibrils, the latter cross one another in all directions, very much

ME, P. H. CAEPENTEE OX THE GENUS ACTINOMETEA. 57

as described by Simrotb 1 in Ophiactis virens, so as to form a network of delicate
threads without any imbedded nuclei, although it may contain pigment-granules ; and
this network passes very gradually into the nucleated protoplasmic plexus forming the
organic basis of the brachial segments (PI. III. fig. 7, L x ).

The tendons of the stem of Pentacrinus are, I believe, of precisely the same character
as the ligamentous bundles between the arm-segments, although, of course, enormously
longer. They stain deeply with picro-carmine, and are composed of parallel fibres, which
may be teased out into very much finer ones, and their upper ends pass into the organic
ground-substance of the five basals, precisely in the same manner as the fibres of the
arm-ligaments pass into the protoplasmic network composing the organic basis of the
successive segments.

In Pentacrinus Wyville-Thomsoni, in which the five basals are completely in contact
with one another in pairs, the two elements of every pair are united by a synostosis,
and the union of the basals with the radial pentagon above them is of the same cha-
racter. The first radials of Comatula are connected with one another and with the
centrodorsal piece in the same manner, as is seen in PI. III. figs. 5, 6, where L, I repre-
sent the tracts of fibrous tissue connecting the first radials with one another and with
the centrodorsal piece respectively. The terminal portions of this fibrous tissue become
calcified to form the compact superficial layers of calcareous substance on the apposed
faces, while the middle portion remains as the fibrous cement-substance uniting the two
calcareous segments, which is thus essentially of a connective-tissue nature.

The mode of union of the segments of the calyx of the Tesselate Crinoids, none of
which are connected with one another by a muscular articulation like the first and
second radials of Pentacrinus and Comatula, was most probably a synostosis of the
same nature as those just described. The immovable sutural unions between certain of
the brachial segments to which Miiller gave the name of " syzygia," arc, in Pentacrinus,
of precisely the same nature as the synostoses between the segments of the calyx, the
organic basis of the one segment being continuous with that of the other through the
fibrous cement-substance, which forms a thin layer between the whole of the two simple
opposed surfaces. This was described by Miiller 2 as a very delicate membrane, of a
different nature from the elastic interarticular substance between the likewise sutu-
rally united stem-segments.

In Comatula, however, the apposed surfaces of the two elements united by a syzygium
are not plain and simple, as in Pentacrinus and Bhizocrinus 3 , but marked by a series of
radiating ridges, as in Apiocrinus obconicus, Goldf. 4 The ridges of the two surfaces
correspond in position, and when the surfaces are in contact are closely applied to one
another, and united by fibrous cement-substance as in an ordinary synostosis. The fibrils
are very numerous and placed very close to one another, so that the calcareous reti-
culation forming the ridges is remarkably dense and compact, being formed around the
ends of these fibrils where they pass into the organic basis of the segments ; and these
ridges thus correspond to the whole of the syzygial surfaces in Pentacrinus and Rhizo-

1 Op. cit. p. 435. - Eau des Pentacrinus, p. 20. 3 Sars, he. cit. p. 22. " Petref. German. Taf. lvii. fig. 5.

58 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

crmus. In the intervals between them the organic basis of the one segment is directly
continuous with that of the other, little or no fibrous tissue being interposed.

The muscular fibres of Actinometra correspond very closely with those of Antedon, as
described by Dr. Carpenter l and Ludwig 2 ; their expanded terminations are simply
applied to the surfaces of the calcareous segments to which they are attached, not passing
into their substance as the ligamentous fibres do (PL III. fig. 7, b m s ), and there is no
trace either of a sarcolemma or of transverse striation.

(ii.) The Dorsal Cirrlii.

(§ 38) The Dorsal Cirrhi of Actinometra do not appear, so far, at least, as my observa-
tions have extended, to be developed over such a large surface of the centrodorsal piece
as is the casein A ntedon. In all the specimens which I have examined the cirrhi are
limited to its margin, while its central portion is entirely free from them and usually
slightly concave. There is generally only one row of these appendages ; but small and
rudimentary ones may occasionally be found interposed between the large and full-grown
ones at the extreme circumference of the plate, thus forming the commencement of a
second row. The number of cirrhi existing at any one time upon the plate-like centro-
dorsal piece of Act. polymorpha varies, I believe, between 15 and 20. Three or, in large
specimens (PL VI. figs. 1, 2), four are attached on each side of its more or less distinctly
pentagonal margin, while in var. 1 (PL VI. fig. 14), and in one specimen of the type
(fig. 7), the total number reached 25. In var. 2 (fig. 16) there are only 10; while
in vars. 3 (fig. 20) and 4 there is no evidence, in the single specimens which I have
examined, of the existence of more than three perfect cirrhi in the adult state, as there
are no sockets around the margins of the centrodorsal plate for the attachment of a
larger number ; though there may be minute openings here and there, which appear to
have corresponded with the central canals of lost cirrhi, the sockets of which have been
obliterated by a later calcareous deposit.

It is not a little singular that the dorsal cirrhi of Act. polymorpha, like the centro-
dorsal piece which bears them, should exhibit such a very slight range of variation, not
only in size but also in number (the three varieties just mentioned of course excepted) ;
for in nearly every other part of the skeleton the range of variation is very great. In
Antedon rosacea the reverse appears to be the case ; for the composition of the skeleton
is fairly constant in its simplicity, but the cirrhi vary considerably both in number
and in size.

In a fully developed cirrhus of Act. polymorpha (PL III. fig. 8 a) the number of
segments varies from 11-11, being usually 12 or 13, the last of which is in the form of
a stroug sharp claw. This is attached by simple suture to the penultimate segment,
which is prolonged at the base of the claw into a short opposing process on its concave
or aboral margin.

The diameter of the basal segment somewhat exceeds its length ; but in the second and
third segments this disproportion between the length and breadth is reduced, and in the
fourth it becomes reversed, the length of this segment being slightly greater than its

1 rhil. Trans, he. cit. p. 704. 2 Beitrage <fcc. he. cit. p. 40.

ME. P. H. CAEPEXTEE OX THE GEXTJS ACTIXOMETBA. 59

diameter. In the fifth and sixth the proportion between the length and breadth reaches
3:2; and a very slight degree of lateral compression is visible in the latter segment,
Avhile the fifth, like the four basal segments, is cylindrical, or nearly so. These are the
two longest segments of the cirrhus ; and from this point onwards the length of the
segments gradually decreases, until in the tenth and following segments ft becomes
again less than the dorso-ventral diameter. At the same time the transverse diameter,
which in the first five cylindrical segments is equal to the dorso-ventral diameter, under-
goes in the seventh and eighth segments a sudden decrease. A faint indication of this
is seen in the sixth segment, and it is continued on to the end of the cirrhus, so that its
terminal portion exhibits a considerable degree of lateral compression. In correspond-
ence with this, a small spine gradually developes itself on the dorsal margins of the sixth
and successive segments, on which it becomes progressively more and more marked,
until in the penultimate segment, i, e. the last one before the claw, it becomes the short
pointed opposing process above mentioned. This series of small spines, like the single
penultimate opposing process of the cirrhi of Ant. rosacea, seems to be characteristic of
those cirrhi only which have reached their full development ; for scarcely any trace of it
is visible in the still immature cirrhus represented in PI. III. fig. 8 c. Although its
penultimate segment shows a faint indication of an opposing process, the dorsal margins
of the segments immediately preceding it are almost, if not quite, even.

The segments of this cirrhus, although of the normal proportions, are not so large as
those of an adult cirrhus, and (counting the terminal claw) are II in number, the first
seven of which precisely resemble those of the cirrhus represented in PI. III. fig. 8 a,
with 12 segments and a chaw, so that the extra one in the former case would appear to
be interpolated between the middle and end of the cirrhus. As in Ant. rosacea, and for
precisely the same reason, greater facility of flexure, the ligamentous substance between
the terminal segments is thicker on the aboral side than on the oral (PI. III. ficr. 9 b),
while in the more cylindrical segments of the basal half of the cirrhus the ligamentous
substance is tolerably equally developed on both sides (fig. 9 a). In correspondence with
this, the canal {cc) which occupies the centre, or nearly so, of the circular opposed
faces of the basal segments (fig. 9 a), lies in the laterally compressed terminal segments
much nearer to the oral side of the oval articular faces, more than half of which is
occupied by the large fossa for the lodgment of the aboral interarticular ligament
(fig. 9 b). The opening of the central canal (c c) is surrounded in each case by a more or
less prominent articular surface.

(§ 39) Both the single specimens of Varieties 3 and I of Act. pohjnwrpha which
I have been able to examine had unfortunately lost all their few cirrhi ; but in Var. 1
twenty-five still remained attached to the centro-dorsal piece ; most of these are fully
developed, and present some slight differences from those of the type (PI. III. fig. 11).
Not only is the number of segments greater, varying usually from 13 to 15, besides the
terminal claw, but they also differ considerably in their relative proportions ; for in the type
the fifth and sixth segments are the longest, while in Var. 1 there is less difference between
them and the fourth and seventh in this respect. The lateral compression, which is not
visible till the eighth segment, becomes somewhat marked towards the end of the cirrhus,

SECOND SERIES. — ZOOLOGY, VOL. II. 9

60 ME. P. H. CARPENTER ON THE GENUS ACTINOMETEA.

which is more distinctly flattened than in the type, although the small opposing spines
on the dorsal margins do not appear at all until the three or four penultimate segments ;
and even on these they are but slightly developed.

In Var. 2 there are only ten cirrhus-sockets around the margin of the pentagonal
centrodorsal plate, two on each of its sides, and placed close to the angles (PL VI. fig. 16).
The number of segments in each cirrhus is eleven or twelve besides the terminal claw
(PI. III. fig. 10), and the width of the two basal segments somewhat exceeds their
length. In the third the length and breadth are nearly equal, but in the fourth and
fifth, which are the two longest segments of the cirrhus, the proportion between them
is as about 4 to 3. The sixth segment is slightly shorter than the fourth, and from this
point to the end of the cirrhus the length of the segments gradually decreases, while at
the same time they exhibit a slight degree of lateral compression. The dorsal spine,
the first indication of which is seen in the fifth and sixth segments, becomes very
marked indeed towards the end of the cirrhus, and develops in the penultimate segment
into a stout opposing process.

(§ 10) The development of the cirrhi of Act. polymorpha seems to take place somewhat
differently from their development in Aid. rosacea as described by Dr. Carpenter L . In
the latter species the individual segments usually present all the characters of maturity
from a very early date, viz. the relative proportions in the length and breadth of the
segments, the bevelling off of the opposed faces on the aboral side, and the develop-
ment of the opposing process on the penultimate segment. But in some rare cases, even
after the cirrhus has attained a considerable size and has the normal number of seg-
ments, the latter are of a very rudimentary character ; their basal segments are the
longest, and the following ones rapidly decrease in diameter, so that the whole cirrhus
tapers considerably from its base to its point. This condition gradually becomes less
and less marked as the segments increase in size, and their opposed faces become
bevelled off towards the aboral side, so that the cirrhus ultimately acquires all the
characters of maturity.

This mode of development, which is the exception in Ant. rosacea, seems to be the
rule in Act. j}olijmor})ha. All the very young cirrhi, both of the type and of Varieties
1 and 2, which I have met with, taper rapidly from the base to the apex ; and while
the four or five basal segments exhibit from a very early jieriod the same proportion
between length and diameter as is seen in a completely developed cirrhus, the following
ones are still in a very rudimentary condition. The sixth segment, instead of being as
long as the fifth, is much shorter ; the seventh is still shorter and more slender, while
the terminal segments are little more than a succession of small disks ending in a small
and very rudimentary claw (PI. III. fig. 8 b). They are thus not only of the smallest
dimensions, but have a much more immature appearance than the basal segments; and
it would therefore seem that the augmentation in the number of segments is effected
by the interpolation of new segments, not at the base, as is usually the case in Ant
rosacea, but between the middle segments and the terminal claw.

1 Phil. Trans, he. tit. p. 711.

ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 61

The fact already mentioned (sect. 38), that in two cirrhi in which the number of
segments is different, the character of the first six or seven is the same, would seem to
point to the same conclusion.

(iii.) The Centrodorsal Plate.

(§ 41) In all the Actluometrm with which I am acquainted, the external appearance
of the " Knopf," or centrodorsal piece, is very characteristic. Like the cirrhi which it
bears, it is far more constant throughout a considerable range of species from very
various localities, than it appears to be in the individual members of a single species,
both of the European and of some of the foreign Antedons, even when existing in the
same locality.

Thus, for example, the centrodorsal piece of Ant. celtiea may either be very shallow,
flattened, and bluntly rounded off at the base, where it was originally united to the
joint of the stem next beneath it, with only two rows of sockets ( V) for the attachment
of the cirrhi (PI. IV. fig. 1) ; or it may be deep and nearly hemispherical, with three or
four alternating rows of cirrhus-sockets, and terminating inferiorly in a flattened circular
or rudely pentagonal base (fig. 6) ; or, finally, it may have the form of a five-sided
pyramid, the apex of which is directed downwards and very slightly truncated, while
the sides bear three or four alternating rows of indistinct cirrhus-sockets (PI. IV. fig. 8).
This last character indicates that the animal had attained a considerable age ; and as all
the specimens of this kind which I have seen have been smaller than normal specimens
of Ant. celtiea, and exhibit slight differences from them in the characters of other parts
of the skeleton, I am disposed to regard them as dwarfed varietal forms rather than as
young and incompletely developed individuals of the ordinary type.

Again, of the two specimens of Ant. macrocnema, Val., in the Paris Museum, both of
which were brought from New Holland by Quoy and Gaimard in 1829, one has a large
and hemispherical centrodorsal piece, while in the other it is a short pentagonal or
nearly circular column, on which the cirrhi are disposed in three or four alternating
rows.

These instances, which might be greatly multiplied, suffice to show that the centro-
dorsal piece of Antedon may vary very considerably in its external appearance. In Acti-
nometra, so far as my experience goes, it is almost invariably a flattened circular, or
rudely pentagonal disk, somewhat hollowed in the centre of its dorsal surface, and with
low sloping sides marked out into distinct sockets for the articulation of the cirrhi
(PL V. figs. 1, 6, 15, and PL VI. figs. 1, 2, 7, 14, 16, 20). It generally conceals more or
less of the first radials which rest upon it (PL II. figs. 9-11, cd), but it may sometimes
be very irregularly extended so as to conceal a considerable part of one or more of the
second radials (PL II. fig. 8, c d).

As a general rule, only one row of cirrhus-sockets can be traced ; but in the large
Act. robusta (PL V. fig. 15) I have found two alternating rows of sockets to be distinctly
visible, and even traces of a third row, so that the dorsal surface of the plate becomes
dightly convex, though by no means to the same extent as in most Antedons. In fact,
the flattened plate-like condition of the centrodorsal piece, and the existence upon

9*

62 ME. P. H. CAEPEXTEE ON THE GENUS ACTINOMETEA.

it of only one or two whorls of marginal cirrhi, seem to be very characteristic of Actino-
metra ; and it would be almost possible to distinguish the Antedons from the Actino-
metrce among the Comatulce described in Midler's memoir, by simply referring to his
descriptions of the " Knopf."

When the centrodorsal is viewed in situ, with all the cirrhi attached around its sides,
its central flattened surface appears almost circular ; but when the cirrhi are removed,
so as to expose the low and more or less sloping sides to which they are attached, the
outer margin of the plate is seen to have a distinctly pentagonal form. This is well
seen in the large Act. robusta (PI. V. fig. 15), with its three rows of cirrhus-sockets ; but
in Act. Solaris (PL V. fig. 1) the angles of the pentagon are more rounded off, and there
is only one row, a complete one, however, of cirrhus-sockets, while the plate itself is of
a considerable diameter, so as entirely to conceal the first radials.

In a variety of this species, which I believe to be identical with the Act. pectinata of
Hetzius and Midler, the diameter of the centrodorsal (PL V. fig. G) is very slight, so that
the greater part of the superjacent radial pentagon is visible outside its pentagonal
margin ; and there are only ten distinct cirrhus-sockets, two at each angle, though one
of the angles (the upper one in the figure) is marked by the presence of a third socket,
which either indicates the commencement of a second row of cirrhi, or, and more pro-
bably, tbe incomplete obliteration of a pre-existing row.

In Act. polynwrpha the angles of the pentagonal centrodorsal plate, the diameter of
which varies from 3 to 5 millims., are sharp and distinct (PL VI. fig. 2) ; in varieties 3
(fig. 20) and 4, the dorsal and ventral surfaces almost meet at tbe edge of the plate,
which is very slightly truncated, and marked in three places by the large sockets ( U)
for the attachment of the few remaining cirrhi, while other small openings indicate the
former existence of others which have been since lost.

In var. 2 (PL VI. fig. 16) the diameter of the centrodorsal plate is very small, as in
Act. pectinata (PL V. fig. 6) ; and, also as in this species, it normally bears ten cirrhi, two
at each angle. The sockets for the cirrhi are, however, but slightly marked, so that the
edge of the plate is but little truncated, and it can hardly be said to have distinct sides
as in Act. Solaris (PL V. fig. 1).

In the type, on the other hand (PL VI. figs. 1, 2, 7), and more especially in var. 1
(fig. 14) j in which there are 25 cirrhi, the edge of the plate where the dorsal surface
passes over into the ventral is obliquely truncated, so that the plate has distinct sides,
which are marked by numerous cirrhus-sockets, as in Act. Solaris (PL V. fig. 1). Towards
the ventral surface the angles of the pentagonal margin of the plate are prolonged into
five short processes (PL VI. fig. 14, t), each situated between a pair of cirrhus-sockets (U).
Their distinctness varies in different individuals ; they are especially marked in var. 1
(fig. 14), and in that specimen of the type which resembles it in having 25 cirrhi (PL VI.
fig. 7), and they are best seen after removal of the centrodorsal plate from the radial
pentagon which rests upon it. They exist also in the other varieties of Act. polymorpha
(PL VI. figs. 16, 20), though they are not so distinctly visible externally, owing to the
greater extension of the dorsal surface of the plate than is found in the type and in

ME. P. H. CAEPENTEE ON THE GENUS ACTLVOMETEA. 63

var. 1, since it is not reduced in size by the truncation of its edges, as is the case in
these forms.

(§ 42) The meaning of these processes becomes apparent when we examine the
ventral or superior surface of the centrodorsal plate which bears the superjacent radial
pentagon. Its condition in Act. polymorpha is rather complicated, and will be better
understood after an examination of the simpler forms of this surface presented by other
species of Comatula.

In the large variety of Ant. celtica, the cavity of the centrodorsal piece (PL IV.
fig. 2, cd.c) is very deep, and its opening on the flattened ventral surface has a circular
or somewhat pentagonal form. Between the angles of this internal pentagon and those
of the outer more distinctly pentagonal margin of the piece run five slight ridges or
elevations (i.e). In the intervening radial areas (r.ar), between these ridges the surface
is somewhat depressed, so as to receive the convexities of the dorsal surfaces of the first
radials that rest upon it, while the five ridges correspond with five slight furrows
marking the lines of junction of these surfaces on the dorsal aspect of the radial pen-
tagon (PL IV. fig. 3), and are therefore interradial in position.

In Ant. celtica these ridges are tolerably uniform in width throughout their whole
course ; but in Ant. rosacea (PL IV. fig. 15) they are considerably wider at their internal
or central ends than they are towards the external pentagonal margin of the piece, so
that they have an elongated triangular form. Prom the base of each triangle a shallow
depression extends a little way towards its apex, cut out along the median line of the
ridge ; but it soon becomes obliterated by the gradual approximation of its two sides,
which meet and form a simple ridge, like that of Ant. celtica (PL IV. fig. 2), extending
outwards to the margin of the plate.

The central ends of the radial areas into which the ventral surface of the plate is
divided are marked in Ant. rosacea by five shallow depressions (PL IV. fig. 15, q),
placed close to the margin of the internal cavity, which bends somewhat inwards at
these points. These depressions correspond in position with five large radial openings
on the dorsal surface of the pentagonal base of the calyx (PL IV. fig. 16, Q), and receive
the blind euds of five diverticula of the body-cavity, which are enclosed between the five
radial spout-like processes of the rosette (PL IV. figs. 13, 16, />) and the internal faces
of the first radials. They are, however, occasionally absent in Ant. rosacea, whde, on the
other hand, traces of them may occur in Ant. celtica. In fact, the differences which
I have described above in the appearance of the ventral surface of the centrodorsal piece
of these two species must not be regarded as representing more than two extreme
variations of one and the same type.

(§ 13) We shall now be able to understand the meaning of the short processes (t)
above mentioned, which are seen projecting from the angles of the centrodorsal plate of
Act. polymorpha, when viewed from below. It will behest to begin with the examina-
tion of the ventral surface of the centrodorsal piece of variety 1 (PL VI. fig. 15), in
which they are more distinctly marked than in the type. This surface rises slightly
from the circumference towards the centre, which is occupied by the opening of a
shallow cavity, the centrodorsal ccelom (cd.c), the diameter of which is rather less than

64 ME. P. H. CAEPENTER OX THE GENUS ACTINOMETRA.

one third of the total diameter of the plate. The floor of this cavity is marked hy minute
punctations (u), which are the internal orifices of canals proceeding from it towards
the dorsal surface of the plate. They originally opened externally on the summits of
the small tubercles occupying the centres of the sockets for the articulation of the
first developed cirrhi in the young animal ; hut their openings have gradually become
obliterated by the deposit of new material upon the central portion of the external
surface of the plate, as described by Dr. Carpenter ' in Antedon rosacea. This is
accompanied by the continual removal of old material from the internal surface, so that
the minute openings (PL IV. fig. 15, u) seen on the central part of the floor of the
internal cavity of the centrodorsal piece are the original external openings of the first
developed canals, which have subsequently become closed externally by the new material
deposited upon the central part of the dorsal surface. The internal openings of the
canals proceeding to the last developed cirrhi are much larger, and placed more towards
the periphery of the floor of the cavity. Similarly in Actlnomelra polymorpha, the
internal openings of the canals proceeding to the existing marginal cirrhi on the plate-
like centrodorsal piece are placed under its projecting rim, so as not to be visible from
above. There are usually one or two large openings under the central margins of each
of the radial areas (PI. VI. figs. 3, 8, 10, 15, 17, 21, r.ar), and the canals which proceed
outwards from these internal openings break up into five branches, one of which reaches
the summit of each of the small tubercles occupying the centres of the five cirrkus-
sockets, which are placed along the outer or peripheral margin of each of the corre-
sponding radial areas (PL VI. figs. 1, 2, 7, 14, 16, 20, 77). These canals enclose the axial
cords of the cirrhi (PL VIII. figs. 1, 3, 4, 5, G, 8, n.c), which proceed from the fibrillar
envelope of the quinquelocular organ contained in the cavity of the centrodorsal plate
(PL VIII. figs. 1, 2, 3, 7, -V), and surround the cirrhus vessels arising from its chambers
(figs. 2, 3, 7, ch), from each of which there arises a single trunk ", dividing, sooner or
later, into branches for the individual cirrhi. In the specimen of var. 1 represented in
PL VI. fig. 15, the division is not completed within the cavity of the centrodorsal plate,
as two or, sometimes, even only one aperture can be seen under the inner margin of
each of the radial areas, so that the primary trunk enters the substance of the plate, and
there divides into the five branches for the cirrhi placed on the outer margin of each
radial area.

The rim of the cavity of the centrodorsal plate of Actinometra polymorpha, var. 1
(PL VI. fig. 15, cd.c), is ten-sided, or nearly circular, and is not marked by shallow
radial depressions, like those described above in Ant. rosacea (PL IV. fig. 15, q). The
radial areas rise very slightly from their peripheral to their central margins, and are
marked by various indistinct ridges and furrows. Their sides rise towards the five inter-
radial elevations, which, though not very much raised above the general surface of the
plate, are nevertheless very distinct ; for they are wide and marked by shallow grooves

1 Phil. Trans, lot. tit. pp. 742, 743.

2 This is in precise accordance with the origin of the vessels proceeding to the cirrhi which arc borne on tho stem
of Pen.tacrinus. At every nodal segment the five chambers which are placed radially around the central axis of the
stem enlarge slightly, and each gives off a single vessel to one of the five cirrhi.

ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 65

(PL VI. fig. 15, b.g), which occupy the greater part of their width, so that the simple
ridge, as seen in Ant. celtica (PI. IV. fig. 2, i.e), is here represented by the two sides of
the groove which is cut out along its median line. In Ant. rosacea (PL IV. fig. 15), as
we have already seen, these sides meet at a very short distance from the central end of
the groove so as to obliterate it. In this form, however, they approach one another very
gradually, and only just meet within the margin of the plate ; but the ridge formed by
their fusion does not end here as in Ant. rosacea, for it is continued a short distance
beyond the general surface of the plate, so as to appear as a short process (t) extending
outwards from the angle between two sides of its external pentagonal margin. Conse°-
quently these five short processes appear on the dorsal aspect of the plate, prolonging its
angles outwards, as we have seen in sect. 11 (PL VI. fig. 11, t).

(§ 14) The grooves (b.g) which are thus cut out along the median line of the inter-
radial elevations on the ventral surface of the centrodorsal plate of Actmometra are of
no little importance ; for there lie in them, as will be seen further on, the five rays of the
basal star (PL VI. fig. 13, S), which is in close connexion with the dorsal surface of the
radial pentagon; they may therefore be called the " basal grooves " (b.g).

As a general rule, these interradial elevations and the basal grooves are, like the
rays of the basal star, entirely devoid of pigment, which is, however, very abundant in
the organic basis of the calcareous reticulation composing the rest of the ventral surface
of the plate ; so that when this is first exposed by the removal of the centrodorsal from
the dorsal surface of the radial pentagon which rests upon it, five white rays are visible
on a dark background. Unless the plate is immediately removed from the alkaline
solution used to effect its separation, this distinction iu colour between the radial and
the interradial portions of its ventral surface rapidly disappears, owing to the destruc-
tion of the pigments contained in the former.

The development of these basal grooves is not only different in the type and in all the
varieties of Act. polymorphs but it differs in different individuals of the type, and even
to a certain extent in the same individual.

In the specimen of the type represented in PL VI. figs. 7, 8, which, like variety 1,
had 25 cirrhi, two only of the basal grooves are seen ; for the other three are occupied by
the rays of the basal star (fig. 8, S), which have become detached from the rest of the
star and from the first radials with which they were connected. But even these two
grooves do not resemble one another ; one extends almost to the margin of the plate,
beyond which the interradial ridge formed by the union of its sides is continued as a
short process (t), just as in var. 1 (fig. 15). The other open groove, however, terminates
very soon, as its sides, widely separated at its central end, bend sharply towards one
another, and meet some distance within the margin of the plate, to which the ridge
formed by their union does not extend, for it terminates abruptly in a blind and
rounded extremity.

In variety 1 (PL VI. fig. 15) the basal grooves (b.g) are narrow, and after increasing
a little at first, diminish gradually in width from their central to their peripheral ends°;
but in the specimen of the type (fig. 8) they are much wider in proportion to their
length, and the width increases slightly from their base to about the middle of their

66 iCE. P. H. CAEPZVTEE OX THE GESTS ACTIXOAIETKA.

com - - s to sive a leaf-like appearance to the rays of the non-pigmented interradial
- " on the ventral surface of the eentrodorsal plate.

In another specimen of the type, however PI. VI. fig. 3\ the sides of the narrow
hasal gi r es are almost parallel, and in every case meet at some distance within the
margin of the plate, while the interradial ridges resulting from their union scarcely
extend at all beyond the angles of the external pentagon.

-:ly, in the monstrosity represented in PI. VI. fig. 10, the dorsal aspect of which
is -een in PI. II. fie. S. both ridses and grooves are extremelv indistinct, and in no case
reach the outer margin of the plate ; while the margin of the internal cavity is markedly
pentagonal in form, and not ten-sided nor circular, as is the case where there are five
distinct interradial elevations alternating with the five radial areas (PI. VI. figs. 3, 8, 15) ;
and it does not project inwards so far as to conceal all the openings («) of the canals
leading to the marginal cirrhus-sockets. as is the case in the other two specimens of the
type (figs. 3j S and in var. 1 (fig. 15).

In all these three specimens of the type the ventral surface of the eentrodorsal plate
is not nearlv so flattened as in var. 1, but rises verv distinctly between its external and
its internal margins, while the radial areas are marked in the same wav bv various
indistinctly marked radiating ridges and furrows ; though as the floor of the central
cavity is also somewhat thicker, its depth is but little if at all greater. The same is the

se in the other three varieties, in each of which the basal grooves differ slishtlv in
form from one another and from the type. In every case they are widest about the
middle of their length, as in one of the specimens of the type (PL VI. fig. 8, b.g) ; this
is most marked in var. 3 (fig. 21), and least in var. 2 (fig. 17). They reach almost, if not
quite, to the margin of the plate, though the ridges formed by the union of their sides
extend but little if at all beyond it, except in var. 3 (figs. 20 & 21), in which two of the
angles of the external pentagon are marked by traces of the small processes (t) so
distinctly seen in var. 1 (figs. 11. 15). In this variety the course of the interradial
ridges is indistinctly visible on the dorsal surface of the plate (fig. 20), which is slightly
hollowed in the centre. The floor of the central cavity is, however, very thick and solid,
and its middle portion presents no trace whatever of any perforations for the canals of
~ing cirrhi, though those proceeding to the three marginal cirrhus-sockets are just
jle under the projecting lip (fig. 21, u), which conceals several others. These indicate
that more cirrhi either have been or would have been developed had the animal lived

_ . their external openings having been obliterated in the former case (the more
probable one) and not yet formed in the latter.

"• In Act. Solaris (PI. V. fie. 2) the ventral surface of the plate-like eentro-
dorsal piece is very nearly flat, as in. Ant. rosacea (PI. IV. fig. 15) and in Act. polymorpha,
var. 1 (PI. VI. fie. 15), rising but slightly from the circumference towards the centre,
and marked by five interradial elevations, along the top of each of which runs a long and
narrow basal gi Its width is tolerably uniform from its base until near its

end, where its sides suddenly approach one another, and meet at a little distance within
the margin of the plate, where the ridge formed by their union also ceases without
extending outwards beyond the general surface of the plate. The same is the case in

HE. P. H. CAEPEXTEE ON THE GENTS ACTEXOMETEA. 67

the large Act. robusta (PL Y. fig. 14), in which three of the hasal furrows are widest at
their central ends, and consequently triangular, while the others are somewhat irregular
in shape.

In both these specimens numerous small openings are visible on the floor of the
central cavity of the centrodorsal piece, but the principal ones leading to the marginal
cirrhus-sockets are concealed under its projecting lip. In the small centrodorsal piece of
Act. pectinate/., however, these last are very distinct (PI. Y. fig. 7, ?() and correspond
in number to the eleven external cirrhus-sockets (PI. Y. fig. 6, T7), so that the five
principal cirrhus-vessels leaving the quinquelocular organ would seem to divide at once
within the cavity of the centrodorsal piece, and not within the substance of its walls, as
is the case in Act. polynwrplta. The ventral surface of the centrodorsal plate of Act.
pectinata (PL Y. fig. 7) is by no means so flattened as in the closely allied Act. Solaris
(fig. 2), but rises considerably from the circumference towards the centre, and the inter-
radial ridges are well marked. The basal grooves {b.g) are narrow and parallel-sided, and
terminate within the margins of the plate, beyond which the interradial ridges are not
continued, so that there are none of the small processes extending outwards from the
angles as in some forms of Act. polymorpha. The median line of each of the radial areas
is occupied by a deep depression, which is particularly distinct at its central end. A
similar depression, though developed to a less extent, exists also in Act. polymorpha,Yax. '2
(PL VI. fig. 17, r.ar).

(§ 46) Xearly all the observers who have studied Comatula have regarded the " Knopf,"
or centrodorsal piece, as of essentially the same nature as the stem of the stalked Crinoids.
The first author who put forward this opinion was Schweigger 1 ; and Miller's views 2 , pub-
lished two years later, were fundamentally the same, though somewhat modified in
form ; for the centrodorsal piece was regarded by Miller as composed of two separate
pieces, one forming the floor of the cavity and the other its sides and rim. The former
was described by him as a pentagonal imperforated plate, " analogous in situation to the
first columnar joint of the Crinoidea ; but as it is not required to transmit the passage
to the alimentary canal s (no prolongation of the column existing in this animal), it is
without central perforation."

The other or ventral half of the centrodorsal piece was regarded by Miller as an
annular or basin-shaped plate, representing the " pelvis" or basal circlet of JPentaerinus,
though he described it as marked externally by numerous sockets for the attachment of
the cirrhi, which in Pentacrinus are borne by the stem-segments only, and never by
the basals.

Goldfuss, who in most points followed Miller's views, differed from him considerably
with regard to the nature of the centrodorsal piece of Comatula mediterranean, which
they had both studied ; and his conclusions, though not absolutely correct, are much
nearer the truth than those of Miller. Pinding most specimens to bear three rows of

1 Op. tit. p. 64. - Op. tit. pp. 129,

:i It must be remembered that the canal which occupies the centre of the Crinoidal stem was originally sup; - . I
to be a continuation of the alimentary canal, and not, as we now know it to be, of the general perivisceral cavity
or coelom.

SECOND SERIES. — ZOOLOGY, VOL. II. 10

68 ME. P. H. CAEPENTEE ON THE GENES ACTINOMETEA.

dorsal cirrki, he described this species as having a column of three segments, and gave
a sectional figure in support of his statements L , which shows three segments below the
circlet of first radials, each bearing a row of cirrhi. It is doubtful how far this figure can
be relied on as accurate, though I have occasionally met with somewhat similar appear-
ances myself. Goldfuss, like Miller, was unacquainted with the remarkable condition
of the basals in this type; and as the "pelvis" described by Miller in Comatula was
rightly regarded by him as representing apart of the stem of Penlacrinus, he was led
to believe in the absence of basals in Com. mediterranea, though he found them in the
Com. multiradiata [Comaster), in which he described the rudimentary column as con-
sisting of only a single segment. Miiller was led, by his comparison of the component
pieces of the calyx of Comatula with those of the calyx of Pentacrinus aster/a (Captd-
medasce), to recognize the very close general correspondence between tbem ; and he
pointed out 2 that the presence of cirrhi at the upper end of the stem of the Penta-
crinoid larva on the one hand, and on the centrodorsal plate bearing the first radials
of the young Comatula on the other, indicate that the latter is comparable to the stem
of PentacriuHs, which bears the cirrhi in verticils separated by longer or shorter in-
tervals. This view of Midler's was pretty generally recognized as the true one, and it
was adopted and greatly strengthened by Wyville Thomson and Dr. Carpenter, who
came to precisely the same conclusions upon developmental grounds. The former
defined it as representing a " coalesced scries of the nodal stem-joints in the stalked
Crinoids," namely, of those joints which bear whorls of cirrhi, so that " the centro-
dorsal plate with its dorsal cirrhi in Antedon is the homologuc of the stem with its
cirrhi in the stalked Crinoids." Ludwig 3 also, while referring to the development of
the centrodorsal as the enlarged uppermost stem-segment, speaks of it as " ein zusarn-
mengedrangter, oberer Stengelabschnitt, in welchem das verkalkte Gewebe keine Sonde-
rung in untereinander^eleo'ene Glieder erfahren hat."

(§47) The first rudiment of the stem of the Pentacrinoid larva as described by
Wyville Thomson 4 consists of a series of delicate calcareous rings forming a curved
line, which passes backwards from beneath the centre of the lower ring of plates, the
embryonic basals. Within each of these is formed a hollow sheaf of parallel calcareous
rods, united together by short anastomosing lateral branches ; the upper one of these, on
which the lower edges of the basal plates rest, soon becomes considerably wider and
thicker than the rest. " Daring the earlier stages of the growth of the Pentacrinoid it
is simply a circular band of the ordinary calcified areolar tissue, enclosing a sheaf of the
peculiar fasciculated tissue of the stem, gradually enlarging, with a central aperture con-
tinuous with the bore of the tube-like stem-joints."

This ring is subsccjuently developed into the permanent centrodorsal piece ; but the rudi-
ments of the first dorsal cirrhi do not appear around its lower contour until very much
later. New rings are developed immediately beneath it, until there are fifteen or sixteen

1 Petref. German, p. - " Lau des Pentacrinus, p. 10.

3 "Zur Anatomic ties BJiizo rinus lofot, nsis, "Zuitschr. fiir wiss. Zool. Bd. xxix. p. 127.

4 '• On the Embryogcny of Antedon rosaceus (Lin ck) (Comatula rosacea of Lamarck)," Phil. Trans. 1805, pp. 53G,
5o7.

ME. P. II. CAEPENTEE OX THE GENUS ACTINOMETEA. G9

ill all, the length and diameter of which are gradually increased by the deposition of new
calcareous material at their extremities and upon their outer cylindrical surface.

Dr. Carpenter 1 has shown that, at or about the period at which the suppression and
metamorphosis of the embryonic oral and basal plates begins, " the production of new
calcareous segments in the stem appears to cease, and a remarkable change begins to
show itself in the one on which the calyx rests. Instead of increasing in length, its
original annular disk augments in diameter, becoming convex on its lower surface and
concave on its upper, and it extends itself over the bottom of the calyx in such a manner
as to receive in its concavity the apices of the basal plates ;" and that portion of its
under surface which extends itself beyond the segments whereon it rests begins to be
marked by small tubercles, which are the origins of the dorsal cirrhi, while it also
" augments not only in absolute but in relative diameter, extending itself over the dorsal
or outer surface of the basal plates, which at the time of the detachment of the body

from the stem are almost entirely concealed by it A second whorl of cirrhi is now

developed, after the same manner as the first, between the latter (with which it alternates
in position) and the base of the calyx (pi. xlii. fig. 3), and a third whorl generally makes
its appearance before the detachment of the Pentacrinoid, so that the young Antedon
possesses ten cirrhi in different stages of advanced development, and from one to five still
rudimentary."

After the detachment of the young Antedon from its stem a minute five-rayed per-
foration is visible for a short time in the somewhat depressed central portion of the
inferior surface of the centrodorsal piece. It is the remains of the original " communi-
cation between the cavity of the basin-shaped plate and the central canal that is still
left in the upper segments (at least) of the stem. This perforation, however, is very
soon closed up by an extension of the calcareous network, so that no trace of it remains
visible either internally or externally."

We have thus seen that the centrodorsal piece " first presents itself in a form which
nowise differentiates it from the other joints of the cylindrical stem, but begins to take
on an extraordinary increase in a peripheral direction at the time when the dorsal cirrhi
first sprout forth, aud thenceforward remains in closer connexion with the calyx than
with the rest of the stem, from which it separates itself so soon as the dorsal cirrhi are
sufficiently developed to serve for the attachment of the animal." Each of these cirrhi
receives a " sarcodic thread, which proceeds from the sarcodic axis contained within the
cavity of the basin, and runs along the central canal of the cirrhus to its termination."

New cirrhi gradually appear between those previously formed and the base of the
calyx, and each receives a peduncle of sarcodic substance from the central axis ; and
" since the arrangement of the whole aggregate of such peduncles is distinctly verti-
cillate, the want of a definite plan in the grouping of the cirrhi on the external surface
of the centrodorsal plate seems attributable to their very close apposition."

During the whole period of the growth of the centrodorsal basin there is a " progressive
exuviation of the first-formed cirrhi from within outwards, concurrently with the develop-
ment of new ones near the margin, those cirrhi which surrounded the summit of the

1 Phil. Trans, he. clt. p. 732.

10*

70 ME, P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

stem being first shed and their sockets filled up by new deposit, and the space thus
formed being gradually widened by the progressive exuviation of the cirrhi that bound
it, and the filling up of their sockets." Thus the flattened central portion of the
dorsal surface of the plate by which it was originally attached to the joint of the stem
next beneath it increases very much in extent, and finally comes to bear a considerable
proportion to its diameter (PL IV. fig. 1). In Act. polymorpha (PI. VI. figs. 2, 7, 14,
16, 20), as we have seen, it extends over the whole of the dorsal surface of the plate, and
to a certain extent also in Act. robusta (PI. V. fig. 15). In Ant. Hschrichtii, however, it
does not reach any great extent, for most of the first-formed cirrhi do not appear to be
cast off as in Ant. rosacea and Ant. ecltica, or, if they are lost, their sockets are not obli-
terated, but they seem to be replaced by others, for I have frequently found young and
rudimentary cirrhi among the larger and perfectly developed ones around the central
portion of the large hemispherical " Knopf" of this species l .

(§ IS) In most pedunculate Crinoids, in which the calyx rests upon the uppermost
segment of the stem, this segment, instead of being the largest, is the smallest, being
the latest formed, while the base of the calyx is formed by the thickened and expanded
basals. Hence, as Dr. Carpenter remarks 2 , "it seems clear that the extraordinary
development of the highest segment of the stem into the centrodorsal basin, which is
characteristic of the mature Antedon, is connected with the multiplication of the pre-
hensile cirrhi which extend themselves from its dorsal surface."

At the base of the quinquelocular organ, and lying on the bottom of the centrodorsal
basin, but enclosed, together with the five chambers, in the above-mentioned fibrillar
envelope (JV), winch is probably of a nervous nature, there is, both in Antedon and in
Actinometra, a succession of verticils of five triangular leaflets 3 . As already shown by
Dr. Carpenter, there can be little doubt but that the lower ones of these mark the
origins of the earlier cirrhal cords from the crinoidal axis. They increase in size from
below upwards, and from the extremities of some of the upper leaflets there issue groups
of three diverging cords that proceed to the cirrhi which are developed at a later period
around the periphery of the centrodorsal piece.

Greeff i has found the older cirrhus-cords still in connexion with these leaflets.
Apparently unaware of the original existence and subsequent removal of the cirrhi cor-
responding to them, he drew a distinction between the vessels which they enclose, and
which end close under the dorsal surface of the plate, and the vessels enclosed in the

1 I cannot altogether confirm Mailer's statement (' Gattung Comatula,' p. 239 (3) ) that the central apical portion of
the centrodorsal in Ant. Eschrichtii, where it was formerly united to the stem, may be covered with cirrhi. In all
the individuals of this species which I have examined (and they are many) there is always a small apical space
quite free from cirrhi ; it may not bo wider than the diameter of a large cirrhus-sockct, but it is always to be found.
I imagine that by the expression " Da es Antedon-Arten giebt, bei denen auch der mittlere Theil des C'entrodorsale
Cirrhen tragt (Antedon Eschrichtii v.. \\.) " (Crinoideen, p. 69, note), Ludwig does not mean any thing more than
that the centrodorsal is covered with cirrhi to a much greater extent than is usual in most Comatulce, where there is
generally a central space of considerable extent entirely free from cirrhi. Schliiter lias also expressed his doubts
respecting the accuracy of Midler's statement.

■ Proceedings 11. S., No. 100, 1870, p. 218. 3 These are seen in section in PI. VIII. figs. 3, 7.

4 Marburg Sifzungsberichte, No. 5, 1S76, p. 91.

ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 71

more peripherally placed cords proceeding from the upper leaflets, which enter the later-
developed cirrhi. Ludwig 1 regarded them as rudimentary structures in Ant. rosacea,
because he found them in Ant. Eschrichtii to enter the more centrally placed cirrhi, which
are not removed, hut persist throughout life, as already mentioned. It will be evident,
however, from the facts stated above, that these cirrhus-cords, which end on the dorsal
surface of the plate, are not rudimentary structures, but the proximal ends of more
complete cords that have undergone a retrogressive metamorphosis, which in Ant.
Eschrichtii is not carried so far as in Ant. rosacea.

These facts all tend to strengthen the view first expressed by Wyville Thomson, that
the centrodorsal piece represents a coalesced series of the nodal stem-joints of the stalked
Crinoids. In Ant. Eschrichtii six or even more rows of cirrhus-sockets may be traced on
the hemispherical surface of the " Knopf," each row corresponding to a node in the
stem of Pentacrinus. Even in those Actmometrce in which only one row of sockets is
visible externally, the composite character of the centrodorsal piece is indicated by the
verticils of degenerate cirrhus- vessels at the base of the chambered organ (PI. VIII.
figs. 3, 7), and by the partially obliterated openings on the central part of the floor of
the centrodorsal cavity (PI. V. figs. 2, 7, II; PI. VI. figs. 3, S, 10, 15, 17, it).

Sometimes, indeed, the " Knopf " may actually assume a more or less columnar
form, as in the specimen of Ant. macrocnema mentioned in sect. II, and in the genus
Solanocrinus ; in both of which three or four alternating rows of cirrhi are visible.
In these forms we may reasonably suppose that the columnar centrodorsal was de-
veloped by the enlargement of the uppermost stem-segment on which alternating whorls
of cirrhi successively appeared, just as in Ant. celtica (PI. IV. figs. 1, 6, 8), Ant. rosacea
(fig. II), and Act. robusta (PL V. fig. 15), but not in such numbers as to obscure the
alternate arrangement (p. 69).

(§ 19) Gotte-, to whom we owe a series of most beautiful observations on the de-
velopment of the water-vascular system and perivisceral cavity of Comatula, has recently
questioned the accuracy of those observations of Wyville Thomson and Dr. Carpenter,
according to which the uppermost of the embryonic stem-segments developes into the
centrodorsal piece, and lias also attacked the view that it may possibly in some
cases arise from the fusion of two or more stem-segments as represented in Goldfuss's
figure.

His description of its origin is as follows : — " Die Anlagen der Centrodorsalplatte sind
schmale, aber doch netzformige Skeletstreifen, welche gleichzeitig mit den Basalia an
deren unteren Biindern entstehen und die obersten, noch eng zusammengedriingten
Stielgliederanlagen umgeben (fig. 13). Es ist daher spater, wenn diese Stelle sich
verschmachtigt, nicht immer ganz leicht, jene Anlagen der Centrodorsalplatte von den
obersten Stielgliedern zu unterscheiden. Beachtet man jedoch, dass sie anfangs das 5.-8.
Stielglied, und nachdem diese abwarts gcriickt sind, das 9., 10., 11., 12. u. s. w. umschlies-
sen, was Thomson iiberhaupt nicht erwahnt, so kann man sich der Ueberzeugung

1 Beitriige, loc. cit. p. 69.

2 " Vergleioh. Entwickelungsgesoli. d. Comatula mediterranca." Archiv, f. rnikrosk. Anat. Bd. xii. 1876, p. ij'.tl.

72 MR, P. H. CARPENTER ON THE GENUS ACTINOMETRA.

nicht verschliessen, das die Skeletzone, aus welclier die Centrodorsalplatte hervorgeht,
unabhangig von den eigentliclien Stielgliedern, mehr in Anschluss an die Basalia und
wohl als rudimentare Wiederholungen dersclben sich entwickelt. Besonders lehrreich
fiir diese Auifassung sind die gar nicht seltenen stiellosen Misshildungen der Comatula-
larven welche ich beobachtet babe. An solchen finden sich in der hinteren Korpcrhalfte,
welcbe ihre urspriinglichen Dimensionen behalt, statt der Stielglieder grosse netzformige
Platten welche den B,auin zwischen den Basalia und dem Endknopf ausfiillen (pi. xxviii.
fig. 50). Vergleicht man sie niit den viel schwachcren Anlagen der Centrodorsalplatte,
so spricht die Darstellung sehr an, dass sie durch die Stielbildung in ihrer Entwickelung
gehemmt und im umgekehrten Falle gefordert werdeu."

The only normal figure given by Gotte in support of his views represents a ciliated
larva, very much younger than the pentacrinoid stage, and with only eight stem-seg-
ments, over parts of the four uppermost of which are traces of a calcareous network
connected with the lower end of one of the embryonic basal plates. This network, which
reaches a more extensive development in the malformation represented in Gotte's other
figure, does not appear in any one of Wyville Thomson's figures of Coma tula larva?,
either in the free-swirnming or in the pentacrinoid condition. As his observations
were carried on for four years, in each of which he followed out the development of
several broods of embryos, it is impossible to suppose that he can have overlooked it had
it been present in the larva? of the British variety investigated by him. It is possible
that the early-formed irregular calcareous ring, " considerably wider and broader than the
ordinary rings of the stem, which lies immediately beneath the basal plates, and subse-
quently develops into the permanent centrodorsal plate," may represent the network
figured and described by Gotte. But then, as the latter says, Thomson makes no mention
of its extending downwards around the other stem-segments ; he gives, however, a series
of figures which, taken in connexion Avith those of the later stage given by Dr. Carpenter,
demonstrate conclusively that the above-mentioned ring does develope into the per-
manent centrodorsal piece. Gotte gives no figures whatever of the pentacrinoid stage.
If, as I believe to be the case, the network described by him as the rudiment of the
centrodorsal piece really does represent the primitive centrodorsal ring of Wyville
Thomson, commencing, be it remembered, as a network of small curved hollow spicules,
then his observations are in complete accordance with the views of "Wyville Thomson and
Dr. Carpenter. Gotte offers no explanation of its downward extension over the remain-
ing stem-segments as described by him in the Mediterranean variety ; and nothing of the
kind is described by the two above-mentioned observers as occurring in the British
variety, unless, indeed, it be the deposit of calcareous material upon the outer cylindrical
surface of each stem-segment by which its diameter is increased.

It is possible that this deposit might commence to be formed at an earlier period in the
Mediterranean variety than in the British one ; but it is difficult to understand its down-
ward extension from the rudiment of the centrodorsal plate as described by Gotte.

(§ 50) The condition of the centrodorsal piece in Ant. rosacea and in Acthwmelra
gives us, I believe, the means of understanding a problematical Cretaceous fossil, first

ME, P. H. CAEPENTER ON THE GENUS ACTINOMETRA. 73

described by Goldfuss \ of which neither be nor any subsequent observer bas given a
satisfactory explanation 2 .

Glenotremites was at first placed by Goldfuss among the Echinoidea, and was sup-
posed by him to have some relationship with the Cidaridcs. It is a somewhat hemi-
spherical body, in the centre of the flattened upper surface of which is a large round
opening, called by Goldfuss the mouth. " Um den Mund liegen fiinf grosse ovale Locher
and zwischen diesen fiinf fiaehe Einnen, die sich bis zum llande erstrecken, wo ihre

Vertiefung nicht auslauft, sondern durch einen erhabenen Saum begriinzt wird

Die Locher gehen trichterformig in die Tiefe ; die Einnen sind die Felder der Fiihler-
gange." These grooves were supposed by Goldfuss to be perforated by minute pores for
the passage of tentacles.

The convex dorsal side of the body bears numerous sockets for the attachment of
cirrhi ; but Goldfuss compared these at first to the large tubercles of the Cidaridte. At
the apex are five smaller apertures ; and Goldfuss suggested that these might be respi-
ratory and the others genital, or, more probably, that both, like the cirrhus-sockets,
marked the points of attachment of various kinds of spines. Subsequently, however, in
his description of a second species, G. conoideus, he spoke of the larger apertures as
ovarian openings, and recognized the resemblance between the sockets on the convex
surface and the similar ones on the dorsal surface of the centrodorsal piece of Comatula
to which the cirrhi are articulated ; and he suggested that Glenotremites might be more
nearly related to the Comatulidee than to the JEchinidce, as he had at first thought.
Agassiz 3 adopted this view, and placed Glenotremites among the Crinoids, and near to
Comatula. Like Goldfuss, he regarded the central aperture as a mouth; but the five
punctated grooves radiating from it, which were supposed by Goldfuss to be provided
with tentacles, were regarded by Agassiz as the points of insertion of the radii. He did
not attempt to explain the five large openings on the ventral surface and the five smaller
apical ones. Reenter 4 , who, like all subsequent writers, accepted the view that
Glenotremites is the centrodorsal piece of a Crinoid allied to Comatula, regarded the
former as " trichterformigen Arm-Anfangen oder Mund-Winkeln," but did not under-
stand those of the dorsal surface.

D'Orbigny 5 , who confused Glenotremites with Comaster and Solanocrinus under one
name, Comatula, and Pictet 6 , who retained it as a separate genus, did not attempt to
offer any further explanation of its peculiarities, and, so far as I know, Agassiz and
R center's views have been generally accepted.

1 Petref. German, i. p. 151), Taf. xlix. fig. 9, Taf. li. fig. 1, and ii. p. ISO, Taf. clx. fig. 18.

2 The following section was written early in 1877, and was in the hands of the Secretary of the Linnean Society
in June of that year. The substance of a portion of it was referred to in my paper on Pentacrintts and Rhizocrirms
('Journal of Anatomy and Physiology,' Oct. 1S77, p. 45). I am therefore exceedingly glad to find, from a paper
published early in 1878 (" TJeber einigc astylide Crinoiden," Zeitschrift der deutsehen geologischeu Gesellschaf t, Jahr-
gang 187S, p. 33), that Schluter has independently given the same explanation of Glenotremites as had occurred to
myself. I learn from his paper that even as late as 1871 Goldfuss's original views were still held by Geinitz (Elb-
thalgebirge, i. 1871, p. 91).

3 Prodrome, Joe. cit. p. 289. 4 Lethaja Geognostica, v. p. 177.

6 Cours elementaire, ii. p. 138. c Traite de Pale'ontologie, iv. p. 290.

71 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

That Glenotremites is the centrodorsal piece of a Com a tula there can, I think, he little
donht ; hnt I see no reason to regard the central opening as a mouth, any more than in
any other of the centrodorsal pieces represented in Plates IV., V., and VI. In all these
cases the centrodorsal cavity, as we know from Gotte's observations, is derived from the
posterior part of the right peritoneal diverticulum of the larval alimentary canal, and
is therefore a part of the general body-cavity or enteroccel. It is occupied by the dorsal
half of the quinquelocular organ which rises through its central opening, the so-called
mouth of Glenotremites, and is continued as the "axial prolongation" (PI. VIII.
fig. 3, a.p) through the central aperture of the rosette upwards into the middle of the
visceral mass.

In Ant. rosacea this central opening is surrounded by five depressions (PL IV. fig. 15, q),
which are the dorsal terminations of the five radial diverticula of the body-cavity en-
closed between the radial spout-like processes of the rosette and the internal faces of the
first radials. These diverticula exist both in Antedon and in Actinometra (PL VIII.
fig. 3, a.r.c), but do not always reach the ventral surface of the centrodorsal piece as in
Ant. rosacea (PL IV. fig. 15). If we suppose the above-mentioned depressions (5) placed
radially around the centre of this surface to be deepened sufficiently to become openings
leading into the centrodorsal cavity, they would occupy precisely the same position as
the so-called genital openings J of Glenotremites ; and simply effect a more open com-
munication between the two parts of the coelom contained in the centrodorsal piece on
the one hand, and the general cavity of the calyx on the other, than when the ventral
surface of the former presents only a single central opening, as in Antedon and Acti-
nometra.

If the view advanced above be correct, it follows that the peripheral part of the areas
around the " genital openings " of Glenotremites are the representatives of the radial

1 These so-called " genital openings " were described by Goldfuss as " Locher." Schliiter, however, merely speaks
of them as " Gruben " (pp. 33, 42), and uses the same term for the whole cavity of the centrodorsal piece, " welche das
Herz (??) oder gekammerte Organ aufnimmt," and is therefore spoken of by him as the " Herzgrube." But from
his expressions, " centralo Herzgrube fiinfseitig " or " zehnseitig," he obviously intends "Herzgrube" to mean nothing
more than the central opening of the ventral surface of the centrodorsal, which he elsewhere calls the " Nahrungs-
canal " (!), although he evidently understands its real meaning.

I cannot therefore clearly make out from Schluter's paper whether the " Eadialgruben " are real perforations or
mere depressions, as in Ant. rosacea, which, by-the-bye, is the same species as the Antedon europceus of Greeff, and
not different from it as Schliiter seems to think. His figures (pi. i. figs. 1, 4, & 10, and pi. ii. figs. 1 & 3) appear to
represent ventral openings in the centrodorsal of some fossil Anteclons, just as described by Goldfuss in Glenotremites ;
but then he refers (p. 33) to Ludwig and Greeif (!) as describing the radial pits of Ant. rosacea as " sackforrnige, in
den Kalkseheitel eindringende blindgeschhssene Erweiterungcn der Leibeshohle."

His use of the word " blindgcschlossene " would seem to indicate that the " Eadialgruben " of his specimens are
really pits, closed below as in Ant. rosacea, and not actual openings; for in the latter case these extensions of the
ccelom contained within the radial axial canals would have opened into the centrodorsal cavity (also a part of the
coelom), i.e. into that part of it which was not filled up by the chambered organ, and they could not then be accu-
rately described as " blind.'' The " Eadialgruben " seem, however, to have been actual perforations in Ant. semi-
ghoosus (Sehl. pi. i. fig. 10); for Schliiter speaks of them (p. 42) as " mit der Centralgrube verschmolzen (reiehen
aber tiefer hinab)," though he suggests the possibility of this being due to an accidental fracture of their central bony
border. The point is one of some interest ; for in no recent Comatula yet known are the "Eadialgruben" more than
simple pits, such as arc generally found in Ant. rosacea.

ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 75

areas on the ventral surface of the centrodorsal piece of Ant. rosacea (PL IV. fig. 15, r.ar),
in each of which lies the convex dorsal surface of a single first radial.

What, then, are the five radiating punctated grooves which Agassiz and Poemer
regarded as the articular surfaces for the attachment of the five arms of Glenotremites?
I befieve them to be the representatives of the basal grooves on the ventral surface of
the centrodorsal piece of Actinometra (PL V. figs. 2, 7, 11, PL VI. figs. 3, 8, 10, 15, 17,
21, b.g). They are sometimes slightly developed in Ant. rosacea, one lying between every
two of the depressions (q) mentioned above (PL IV. fig. 15), in precisely the same manner
as the grooves and " genital openings " alternate on the ventral surface of Glenotremites.

We thus see that the peculiarities of the ventral surface of Glenotremites may be
readily explained by what we know of the corresponding parts in Antedon and Actino-
metra. The apertures in the centre of the dorsal surface admit of an equally simple
explanation.

The quinquelocular organ forming the dorsal termination of the axial prolonga-
tion of the adult Qomatula consists of five chambers, arranged around a central axis
which contains numerous vessels. In Pentacrin/is there is no centrodorsal piece, but the
quinquelocular organ is contained in a cavity, the sides of which are formed by the first
radials above and by the basals below. Its five chambers are not closed below, but
narrow considerably, and are continued down the stem as five long vessels arranged
symmetrically around a central axis. The same appears to be the case in the stem of
the Pentacrinoid larva of Qomatula ; for, as already mentioned, Dr. Carpenter has
described a minute five-rayed perforation occupying the central portion of the dorsal
surface of the recently detached Antedon. I regard this perforation as homologous with
the five small apertures arranged around a single central one on the dorsal surface of
the centrodorsal piece of Glenotremites, and with the similar openings on the underside
of the calyx of the other stalked Crinoids — for example, of Ciqrressocrinus. The fact
that the young Antedon rosacea has only three rows of dorsal cirrhi when liberated
from its stem, while there are four or six rows on the dorsal surface of Glenotremites,
does not at all tell against this view. Indeed Sars 1 has shown that the pentacrinoid stage
persists in Antedon Sarsii very much longer than in Ant. rosacea, and he has found
specimens with nearly thirty cirrhi still in a pedunculate condition, the cirrhi being
placed in such close proximity to one another that any trace of a distinct order in their
arrangement was entirely obliterated. The exterior of the centrodorsal piece of Ant.
Sarsii, therefore, immediately after its liberation from the stem, would present (its size,
of course, excepted) a very similar appearance to the convex dorsal surface of Gleno-
tremites, viz. a central five-rayed opening, or possibly even a single opening with five
others round it, the rest of the surface being covered with sockets for the articulation of
the dorsal cirrhi.

(iv.) The Pentagonal Base of the Calyx.

(§ 51) In all the Actinometra? with which I am acquainted the Pentagonal Base of the
calyx formed by the close mutual adhesion of the five first radials, together with the

1 ' Criuoides vivants,' p. 57.
SECOND SERIES. — ZOOLOGY, VOL. II. 11

76 ME. P. H. CAEPENTEE ON THE GEM US ACTINOMETEA.

rosette or metamorphosed basals, differs in many points from that of Ant. rosacea and
of all the other species of Antedon which I have examined.

In all of these the external or distal faces of the first radials slope at a considerahle
angle from above and within downwards and outwards, so that a view of the upper or
ventral aspect of the radial pentagon formed by their union (PI. IV. figs. 4, 17) shows,
not only their small superior or ventral faces around the central funnel-shaped space (F),
but also the greater part of their inclined external faces (PI. IV. figs. 6, 8, 14), viz. the
fossae (/) for the attachment of the muscles between the first and second radials (PI. IV.
fig. 5, r.m) and the smaller ones (/*) which lodge the interarticular ligaments, the distal
opening of the central canal (c.c), and the large transverse articular ridge (i), together
with more or less of the large fossa (/) which lodges the elastic ligament. The amount
of this fossa which is visible on the ventral aspect of the radial pentagon varies in
different cases.

In correspondence with this inclination of the distal faces of the first radials of Antedon
to the vertical or dorsoventral axis of the calyx, their ventral faces are much reduced and
are very small in comparison with the dorsal ones.

When the ventral surface of an isolated first radial of Ant. rosacea is examined (PI. IV.
fig. 12 a), it is seen to be divided into a central and a peripheral portion by two curved
ridges, bending towards each other along the median line, and there separated by a
furrow (fj). The central portion only is the true ventral face of the radial. It slopes
inwards, so as to contribute to the formation of the central funnel-shaped space (PI. IV.
fig. 17, F) occupying the centre of the radial pentagon, and partially filled up by a cal-
careous network formed by the inosculation of processes which proceed from the internal
and ventral faces of the surrounding radials (PI. IV. fig. 12 a, en). The peripheral
portion, on the other hand, slopes outwards, and is, in fact, the upper or ventral half of
the inclined external face (fig. 14), namely, the large vertical lamella? in which the mus-
cular fossa? are excavated. The upper and inner edges of these lamella? form the curved
ridges above mentioned, each of which has two limbs, one superior (fig. 12 a, g^), which is
horizontal, or nearly so, and forms the external boundary of the ventral face, and one
inferior (g 2 ), which descends along the median line of the inclined external face towards
the great transverse articular ridge (i) ; it is separated from the corresponding inner
edge of the other muscular fossa by the furrow (/J, which may therefore be called
" intermuscular."

These curved ridges thus produce great inequalities in the ventral aspect of the
radial pentagon (PL IV. fig. 17). The walls of its central funnel present an alternation
of radiating ridges and furrows, of each of which there are ten. Five of the furrows
(v.i.f) correspond Avith the divisions between the component pieces, and are therefore
interradial, while the ridges which bound them are the superior limbs (#,) of the curved
ridges above mentioned, one belonging to each of the two contiguous radials. Of the
other five furrows, one passes along the middle of each of the five radials, and both the
ridges which bound it belong to the same piece — being, in fact, the median or descending
margins (g 2 ) of the large vertical lamella? in which the muscular fossa? (f) are excavated.
These furrows, therefore, are simply the intermuscular furrows of the distal faces (PL IV.

MR. P. H. CARPENTER OX THE GENUS ACTINOMETRA. 77

figs. 12 a, 14, fj), and they only appear on the ventral aspect of the radial pentagon
because of the inclination of these faces to the vertical axis of the calyx. They do not
appear to exist in Ant. celtica (PI. IV. figs. 4, 6, 8), in which the radial muscles are
far larger than in Ant. rosacea, so that the vertical lamellae to which they are attached
attain a ninch greater size. These are placed at such an angle to the dorsal portions of the
distal faces of the ra dials that they stand up nearly vertically, and form the outer wall
of the central funnel-shaped space (fig. 4, F) which leads downwards into the cavity of
the centrodorsal piece. Its pentagonal rim is formed, as in Ant. rosacea (fig. 17), hy
their superior margins, two of which, belonging to contiguous radials, bound each of the
five ventral interradial furrows (v.i.f) which mark the angles of the pentagon. In the
centre of each of the five sides of this pentagonal rim is a deep notch (fig. 6, f 2 ) ; but it
does not descend on to the distal face of the corresponding radial so as to form an inter-
muscular furrow bounded by the median descending margins of the muscular fossa?, as
in Ant. rosacea (figs. 14, 17, /I) ; for these fossae are so large, and extend so far towards
the median line, that their inner margins unite and form a prominent vertical ridge
(figs. 4, 6, //;;), which passes below into the elevated rim around the opening of the central
canal (c.c).

(§ 52) These five notches in the sides of the pentagonal rim of the central funnel of
the radial pentagon in Ant. celtica (PI. IV. fig. 6, f 2 ) represent the points at which in
Ant. rosacea the superior or central end of each intermuscular furrow (f_) passes at a
slight angle, due to the inclination of the distal face, into a shallow depression (figs. 12 a,
17, v-r.f) occupying the centre of the small ventral face of each first radial. This depression,
which is much better developed in Actinometra, is far more distinct in some specimens of
Ant. rosacea and Ant. celtica than in others, and in the dry state is barely visible.
When, however, the interior of the calyx is viewed from above after the visceral mass
has been removed, so as to lay open the circumvisceral ccelom, and expose the ventral
aspect of the radial pentagon, the position of the ventral radial and interradial furrows is
indicated by dark lines converging towards the centre (fig. 5). These are due to the fact
that the parietal layer of the peritoneum which lines the interior of the calyx descends
into these depressions, so that its pigment is here more thickly aggregated than on the
rest of the ventral surface. A similar slight depression lined by the pigmented peri-
toneum exists on the median line of the ventral face of the second and third radials and
of the basal brachial segments, and it lodges the dorsal portion of the cceliac canal, which,
in the intervals between the segments, sends down diverticula between the muscles
connecting them, so that its course is readily traceable by the greater intensity of the
pigment along the median bine of the segments and between the two muscles connecting
every pair (fig. 5, v.r.f). At the base of the arms the cceliac canal becomes broken up
by connective-tissue septa into a number of intercommunicating spaces, which open
freely into the general cavity of the calyx or circumvisceral ccelom. The dorsal part of
the canal, however, retains its primitive relation to the skeleton and muscles, and is
lodged in the furrows on the ventral faces of the radials (fig. 5).

We have already seen that the inner wall of the funnel-shaped space (F) occupying the
centre of the radial pentagon is formed by the inclined ventral faces of the five first

11*

78 MR. P. H. CARPEXTER OX THE GEXES ACTIXOMETRA.

radials (figs. 4, 12 a, 17). These are not simply plane, but are usually more or less
divided up by delicate calcareous processes which extend to meet the ventral face of the
rosette, and collectively form a complicated network (c.n), filling up the central funnel,
and often partially bridging over the ventral radial furrow, so as to convert it into an
incomplete canal.

At the inner margin of the ventral face this furrow turns downwards, and passes
directly into a nearly vertical furrow occupying the median line of the proximal or
internal face (PL IV. fig. 12 c, a.r.f), and more or less completely converted into a canal
by the union of irregular processes, which extend themselves from its sides to meet the
rosette. As it descends towards the dorsal face and passes between the inner raised edges
of the two apertures {/, y) of the central canal, this axial radial farrow becomes a complete
canal, for its edges are closely applied to the inflected margins of. one of the five radial
spout-like processes of the rosette (PI. IV. figs. 13, 16, p).

The five canals thus formed may hence be regarded as enclosing cavities directly con-
tinuous with the cceliac canals of the arms, in the direction of which they lie ; and they
thus enclose portions of the body-cavity, which I will call the radial ccelom 1 .

They open on the dorsal surface of the radial pentagon by five large openings (PI. IV.
fig. 16, Q), that correspond with five more or less distinctly marked circular depressions,
which are placed radially on the ventral surface of the centrodorsal piece around the
margins of its central cavity (fig. 15, q), and the cauals end blindly in these depressions.
"Where these canals are enclosed by the spout -like processes of the rosette, they are com-
pletely shut off both from one another and from the dorsal extension of the ccelom which
occupies the central funnel-shaped space within the radial pentagon (figs. 1, 17, F), and
passes down into the cavity of the centrodorsal piece through the central opening of the
rosette (fig. 16, r.o). On the ventral side of the rosette, however, these radial axial canals
are only partially complete, and are in free communication with the numerous plexiform
spaces into which the funnel-shaped space is broken up by the above-mentioned cal-
careous network. The central portion of this system is very irregular; but peripherally
the plexus becomes more regular, and five axial interradial canals are traceable between
the five radial ones, with which, as with the centre of the plexus, they are in free
communication.

These interradial canals are continuous with the interradial furrows which are visible
on the ventral aspect of the radial pentagon (PL IV. figs. 4, 17, i'-i-f), and they enclose
diverticula of the circumvisceral ccelom to which the name interradial ccelom may be
given. They do not descend so far towards the dorsal surface as the axial radial canals,
and are not, like the latter, enclosed (normally, at any rate) by spout-like processes of
the rosette ; for their course towards the dorsal surface is terminated by the five short
triangular processes of the rosette (figs. 3, 7, 13, 16, o), which are directed towards the
sutures between the five radials.

(§ 53) This is well seen in Ludwig's schematic vertical section through the body of
Ant. rosacea-, in which the radial ccelom (Lr) is rightly represented as both longer and

1 The general relation of these axial radial canals is precisely the same in Act inometra as in Antedon. See PI. VIII.
figs. 3, 6, a.r.c. 2 Beitrage &c. Taf. xix. fig. 74.

ME. P. H. CAEPEXTEE ON THE GENITS ACTEXOMETEA. 79

larger than the interradial coelom (Id). It is also seen in Taf. xv. figs. 25, 26, on a larger
scale ; and in Taf. xiv. figs. 20-24 hoth the radial and the interradial diverticula of the
body-cavity are seen in transverse section, separate from one another towards the dorsal
side, but communicating freely nearer the ventral surface, both with one another and
with the centre of the plexus. In figs. 20-24, Ludwig has accidentally lettered them
L' and L" respectively. This is unfortunate, as these letters are employed by him in
his other figures to designate the circumvisceral and axial body-cavity ; while in fig. 26
he uses the same letter L to designate the system of plexiform spaces occupying the
central funnel of the pentagonal base, as he employs in his other figures for the inter-
visceral division of the body-cavity.

This hardly agrees with his text ; for on p. 43 he says : — " Ueber den ersten Eadialien
lost sich die axiale Leibeshohle in eine Sumnie von mit einander allseitig communiciren-
den Maschen raumen auf, welche zwischen die ersten Eadialien eindringen, bier das
dorsale Organ [i. e. axial prolongation] umgeben und endlich mit zehn blmdgeschlos-
senen Fortsetzungen endigen, von denen fiinf radiar gerichtet sind (Lr), funf interradiar
(Id). Der Dorsalcanal [ = cceliac canal] des Amies giebt seine Lage dicht fiber den
Kalkgliedern und zwischen und fiber deren Aluskelpaaren nicht auf bis er fiber dem
ersten Eadiale angekommen ist, wo er sich gleichfalls in die schon erwahnten Maschen-
raume auflost. Letztere stehen also in Verbindung mit der axialen Leibeshohle und mit
den Dorsalcanalen der Arine, aber sie dehnen sich auch ferner nach oben und seitlich aus,
und erfiillen bier den Raum der rings um die axiale Leibeshohle zwischen dem Yentral-
canal und dem Dorsalcanal in der radiaren Halfte, zwischen Ventralperistom und Dorsal-
peristom in der interradiaren Halfte der Scheibe fibrig bleibt." This space, the general
perivisceral cavity, falls naturally, as Ludwig has pointed out, into two divisions — one
external or circumvisceral, between the visceral mass and the body-wall, and corresponding
to the " cceloni " of Dr. Carpenter ; and one internal or intervisceral, surrounding the axial
body-cavity (or axial canal of Dr. Carpenter), and occupying the spaces between the
various coils of the alimentary canal within the visceral mass. This last corresponds
to the intramural spaces and mesenteric sinuses of Dr. Carpenter ', and not to the former
only, as Ludwig appears to think (p. 55).

Of all the divisions of the body-cavity this intervisceral cceloni is the one which is
least directly connected with the plexiform network between the first radials (Ludwig,
figs. 26, 74, L) and with the cceliac canals of the arms ; for it is completely separated from
the latter by the visceral layer of the peritoneum, except at the minute aperture in the
under surface of the visceral mass, where the axial prolongation, coming up from the quin-
quelocular organ through the central vacuity of the pentagonal base, enters the inter-
visceral cceloni contained within the visceral mass. "When the latter is turned out of
the calyx the intervisceral ccelom contained within it is, of course, removed at the same
time, while the plexiform system of spaces between the first radials, and the continuations
of the cceliac canals of the arms which terminate in it, are laid open ; both of these, there-
fore, are manifestly portions of the general c/;'c^//iviseeral cavity surrounding the visceral
mass. Ludwig, however, makes the following statement (p. 90) : — " Die Hauptab-

1 Proc. E. S. no. 166, 1S76, pp. 216, 217, 225.

80 MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

sclmitte der Leibeshohle in Scheibe und Arnien stelien niiteiuander paarweise in engerer
Bezielnrng, indem sick dieaxiale Leibeshole fortsetztin die Ventralcanale [subtentacular]
der Arnie und Pinnulse, die interviscerale in die Dorsalcanale [cceliac] und die eircum-
viscerale in die Genitalcanale."

I cannot corroborate this statement except with regard to the axial body-cavity, the
connexion of which with the subtentacular canals of the arms was first shown by
Dr. Carpenter.

The ventral portion of the circumvisceral body-cavity, viz. the limited and much
divided space between the parietal and visceral layers of the ventral peritoneum, certainly
does stand in direct connexion with the genital canals of the arms ; but its dorsal
portion, viz. the space between the visceral mass and the skeleton of the calyx, is, as
already mentioned, far more directly a continuation of the cceliac than of the genital
canals. The former gradually increase in size as they approach the disk, becoming very
large in the second and first arm-segments, and traversed by numerous connective-tissue
bands, which are directly continuous with those of the circumvisceral space ; while the-
genital canal remains relatively small, and is nothing more than a space in the horizontal
septum separating the subtentacular and cceliac canals.

The beautiful investigations of Gotte x have shown that the primitive ccelom of the
pentacrinoid larva of Antedon consists of two parts : (1) an oral or ventral one, de-
veloped from the left peritoneal diverticulum of the primitive alimentary canal ; and
(2) a dorsal one, which sends a prolongation backwards into the stem, and is developed
from the corresponding right peritoneal diverticulum. These divisions of the primitive
coelom had been previously described by Dr. Carpenter -, Metschnikoff 3 3 and Greeff 4 ,
to all of whom, however, their origin was unknown. The last observer regarded the
ventral division as " den vom Wassergefasssystem und der hinteren Leibeshohle geschie-
denen urspriinglichen Blutsinus ;" for he supposed it to be continuous with the cavity
of the axial prolongation, which he called the " dorsovcntral Gefassstrang." Dr. Car-
penter has found, however, that this structure breaks up into five branches, one of which
goes to each of the primitive rays, and developes into the so-called " genital rachis " of
the arms, while the oral ccelom of the pentacrinoid larva (the "Bloodsinus " of Greeff),
sends off an extension into each of the arms as its subtentacular canal. In the direction
of the radii it forms, of course, the subtentacular canals of the disk ; but elsewhere, or
interradially, it becomes gradually limited by the enlargement of the visceral mass, and
by the formation of adhesions between its upper surface (visceral layer of the peritoneum)
and the parietal layer lining the under surface of the ventral perisoine ; so that the
ventral portion of the circumvisceral ccelom enclosed between these layers, to which the
primitive oral ccelom gives rise, becomes v r ery much reduced in extent. We do not yet
know the precise origin of the genital canals of the arms ; but it seems most probable
that, like the ventral portion of the circumvisceral coelom with which they are connected
in the disk, they are developed out of the lower or dorsal half of the extension into the

1 Op. eit. p. 591, Taf. xxvi. fig. 19. 2 Phil. Trans, he. eit. p. 728 ; Proc. R. S. no. 166, p. 228.

3 " Beitr. z. Entwickelungsgesch. einiger niederen Thiere," Bull, de l'Acad.. Imp. des Sciences de St. Petersb. torn. sv.
1871, pp. 502-509. 4 Marburg Sitzungsbericbte, 1876, No. 5, p.

ME. P. H. CAEPENTEE OX THE GENUS ACTINOMETRA. 81

arms of tlie primitive oral coelom — the upper or ventral portion of which gives rise to
the suhtentacular canals of both arms and disk.

The dorsal or aboral coelom of the pentacrinoid larva lies beneath the annular
mesentery, and forms the dorsal half of the primitive circumvisceral coeloin, long before the
alimentary canal is sufficiently convoluted to give rise to a distinct intervisceral coelom.
Like the oral coelom it sends off radial extensions into the developing arms, but beneath
the horizontal partition extended from the mesenteric bands, and these become the cceliac
canals. Its dorsal prolongation gives rise to the cavity of the centrodorsal piece, and
ultimately to the central canals of the calcareous segments. Both of these, together with
the plexiform space between the first radials and the cceliac canals converging to it, are
therefore, like the greater part of the circumvisceral coelom, derived from the right peri-
toneal diverticulum of the primitive digestive cavity ; while the left one gives rise to the
suhtentacular canals of the disk and arms, and to the ventral portion of the circum-
visceral coelom. The primitive distinction between the oral and the aboral ccelom of the
larva, indicated by the mesenteric fold, becomes, however, gradually obliterated by the
development of numerous similar septa of connective tissue, and by the growth of the
alimentary canal and its consequent winding.

The axial canal, continuous above with the oral, and below with the aboral ccelom, is
produced by the limitation of the central space left by the coiling of the intestine around
the stomach ; while the remainder of the spaces between the coils become the inter-
visceral ccelom, which is therefore not developed to any extent until after the cceliac
canals of the rays have been extended from the primitive aboral coelom.

(§ 51) All the species of Antedon do not agree with Ant. rosacea and Ant. celtica in
the great inclination of the distal faces of the first radials to the vertical axis of the calyx,
so that these faces enter into the composition of the ventral aspect of the radial pentagon.
In a new and undescribed Antedon from the Philippines this inclination is very slight;
and in a view of the pentagonal base from above but little more is seen than the proper
ventral faces of the component radials. In this respect, therefore, forms such as these
present an approximation to Actinometra (PI. V. fig. 1, PI. VI. figs. 5, 12, 23), in which
the distal faces of the first radials are nearly or quite vertical, and do not at all enter
into the composition of the ventral aspect of the pentagonal base, which consists simply
and entirely of the five adjacent ventral faces of the component radials.

In Ant. rosacea (PL IV. fig. 17) these form a five-pointed star, the central surface of
which slopes rapidly downwards and inwards as the inner wall of the central funnel (F) ;
while its five rays correspond with the divisions between the component radials, and are
bounded by the large ridges which form the upper and outer margins of the two adjacent
muscular fossa? (/) of every pair of contiguous radials (PI. IV. fig. 12 a, g x ). The sutures
between the radials are marked by slight depressions of their ventral surfaces, and these
are completed by the ridges at their sides into the ventral interradial furrows already
described (PI. IV. figs. 1, 17, v.i.f) ; they occupy the five rays of the star, and alternate
with the five shallow depressions (v.r.f) lying in the centre of the ventral faces of the
first radials. In Actinometra, as will be seen further on, these depressions become very
marked ; but in Ant. rosacea they are hardly deep enough to be called furrows, and are

82 ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

generally more or less concealed by the calcareous network occupying the opening of the
central funnel. They are the ventral continuations of the five canals enclosed by the
radial spout-like processes of the rosette, and they pass downwards and outwards in the
reentering angles of the star into the intermuscular furrows on the distal faces of the
component radial s (PI. IV. figs. 12 a, 1-1, 17, f). These reentering angles, which are
bounded by the superior margins of the two muscular fossse of each radial, are more open
in Ant. celtica (fig. 1) than in Ant rosacea (fig. 17), so that the rim of the central funnel
becomes more nearly pentagonal, having somewhat the shape of a Goniaster. This is
still more marked in Ant. Esckrichtii, while in Actinometra it becomes a regular pentagon
(PI. V. fig. 4, PI. VI. figs. 5, 12, 23).

In correspondence with the nearly vertical position of the distal faces of the radials in
Actinometra, their ventral faces, which in Ant. rosacea and Ant. celtica have a very
steep inward slope, occupy a nearly horizontal position, sloping but very gently inwards
towards the central space, so that the opening of the funnel becomes widely expanded.
Its inner walls, formed by the adjacent ventral faces of the contiguous radials, which are
relatively much larger than in. Antedon, are generally more or less sculptured out into
a series of radiating ridges and furrows, the number and distribution of which vary in
different species.

(§ 55) In Act. pectinata the ventral surface of each first radial (PI. V. fig. 9 a) is
nearly as even and regular as that of Ant. celtica (PL IV. fig. 4) or Ant. rosacea (PI. IV.
figs. 12 a, 17), and in some cases it may be even more so. It is, however, both absolutely
and relatively larger, as it is not encroached upon by the distal face, which stands nearly
at right angles to it, and the furrow (v.r.f) occupiug its median line is far more distinct
than in either of the two species of Antedon. These points are clearly seen in a com-
parison of figs. 5 on Plates IV. and V., which represent the ventral aspect of the calyx,
as seen after removal of the visceral mass, in Ant. celtica and Act. pectinata respectively.
In the former (PI. IV. fig. 5) the second and third radials and the bases of the arms are
at a higher level than the pentagon of the first radials, owing to the inclination of the
distal articular faces of the latter ; but in Actinometra (PI. V. fig. 5) the whole ventral
surface of the calyx is iu one horizontal plane, as the opposed articular faces of the first
and second radials are parallel to the vertical axis of the calyx, and not more or less
inclined to it, as in Antedon.

We have seen that in Ant. rosacea the lateral margins of the ventral faces of the first
radials (PL IV. figs. 12 a, 17) are somewhat depressed, so that when two pieces are in
contact a shallow interradial groove marks their line of union on the ventral side. It
is deepened into a furrow (v.i.f) by the elevation at its sides of the ridges forming the
superior margins of the muscular fosste of the inclined distal face. This interradial
depression also occurs in Act. pectinata (PL V. figs. 5, 9 a &c.) ; but as the vertical
lamella; (fig. 9 c, g) in which the muscular fossae are excavated are very small, and do
not extend inwards so as to encroach upon the ventral face, there are no ridges at the
sides of this interradial depression (fig. 5, v.i.f) converting it into a deep furrow as in
Ant. rosacea (PL IV. fig. 17), so that it is far less conspicuous than the corresponding
radial furrow (PL V. figs. 5, 9 a, v.r.f).

MR. P. H. CARPENTER OX THE GENUS ACTLXOMETRA. 83

The ventral aspect of the radial pentagon of Act. Solaris (PI. V. fig. 4) consists almost
entirely of the conjoint ventral faces of its component pieces ; the distal faces are very
slightly inclined to the vertical axis of the calyx, so that portions of the fossae lodging
the radial muscles and the interarticular ligaments hecome visihle (PL Y. fig. 4?,f, h).
The ridges (g{) which bound the muscular fossa? superiorly form by their apposition the
outer margin of the ventral surface of the pentagonal base, which is interrupted at ten
points, five being radial and five interradial. The former, which lie between the two
muscular lamellae of each radial, indicate the union of the intermuscular furrows of the
distal face with the ventral radial furrow occupying the median line of the superior face
(PL V. fig. 1, v.r.f) ; while the latter, which are at the angles of the pentagon, are the
outer ends of the ventral interradial furrows (v.i.f) corresponding with the sutures
between every two contiguous radials, the superolateral edges of which are slightly cut
away, so that by the apposition of every two pieces an interradial furrow is formed.

These interradial furrows, like the radial ones, slope gently inwards towards the centre.
The two sets, as soon as they pass into the axial furrows on the internal faces, become
respectively converted into five radial and five interradial axial canals by the union with
one another in successive pairs of small processes extending from the intervals between
them towards the central calcareous network (PL V. fig. 1, c.n). These processes are
the central ends of ridges which are developed on the two halves of the ventral surface of
each first radial, between its median furrow {v.r.f) and its lateral margins. The small
and irregular furrows between them usually converge towards the radial or interradial
furrows, where they begin to descend into the corresponding axial canals; but in two
cases (PL V. fig. 4, x ) they are also converted into canals by the small bridge-like pro-
cesses above mentioned. These intermediate canals, like the normal radial and inter-
radial ones, are in free communication with the rest of the spaces in the calcareous net-
work, just as in Ant. rosacea; but the radial ones do not extend so far towards the
dorsal surface of the pentagonal base as in this species, as will be seen when we come to
study its dorsal aspect (PL Y. fig. 3).

(§ 56) In Act. robusta (PL A*, fig. 11) this sculpturing or development of ridges and
furrows on the ventral faces of the first radials is carried much further than in Act.
Solaris. The muscular fossae (f) are also somewhat deeper, and the median radial
furrows which proceed inwards from the intervals between their superior margins (g^)
along the ventral faces of the radials are broken up very soon into a number of small
irregular furrows ; all converge, however, towards the centre, by the development of
numerous ridges of a similar nature to those rising from the lateral halves of the ventral
faces in Act. Solaris.

These ridges completely obliterate all traces of any regularity in the passage of the
radial furrows into the central calcareous network (en), as was so marked in Act. Solaris
(fig. 4). The interradial furrows, too, are not particularly distinct, as the ventral surfaces
of the radials fall away but little towards their lateral margins. Towards the centre,
however, they become more distinct, and arc bridged over by processes extended from
the above-mentioned ridges, so that they pass downwards as canals into the system of
plexiform spaces occupying the central funnel of the pentagonal base. The position of

SECOND SERIES. — ZOOLOGY, VOL. II. 12

84 MR. P. H. CAEPEXTEE OX THE GENUS ACTINOMETRA.

one of these axial interradial canals is indicated, in PI. V. fig-. 11, by a brown bristle (II)
which has been passed along it.

In the type of Act. polymorpha the distal faces of the radial pentagon are placed
somewhat more vertically than in Act. Solaris, so that scarcely any trace of the muscular
fossae is to be seen on its ventral aspect (PL VI. fig. 5). This is still more the case in
var. 4 (fig. 23), in which the ventral aspect of the radial pentagon exhibits nothing but
the extremely sculptured and inclined ventral faces of its component pieces ; it is divided
into a very large number of ridges and farrows, nearly every one of the latter having a
canalicular ojiening into the central network (c.n). The radial furrows arc thus entirely
obliterated; and as there is no corresponding intermuscular furrow on the distal face
(as in Ant. celtica, PI. IV. figs. 4, 6), there is nothing to indicate their position on the
outer margin of the radial pentagon. But the interradial furrows (PI. VI. fig. 12, v.i.f)
are readily distinguishable by their being somewhat deeper and straighter than the
secondary radial furrows. This is also the case, but to a less extent, in the type (fig. 5)
and in varieties 2 and 3 ; but var. 1 is somewhat different, and in this respect approaches
Antedon rosacea more than any other Actinometra with which I am acquainted. The
distal faces of the radial pentagon (PI. VI. fig. 12) are perceptibly inclined to the
vertical axis of the calyx, so that even the opening of the central canal (c.c) appears on
its ventral aspect. The muscular fossae (/) are deep, so that their superior margins
project inwards and encroach somewhat on the ventral faces ; and the median furrows
of the latter are tolerably deep, their outer extremities passing over into the inter-
muscular furrows (f) of the distal faces. The interradial furrows between the elevated
lateral halves of the ventral faces (PI. VI. fig. 12, v.i.f) are also deep, but the ventral
faces are plain and scarcely at all sculptured, so that both radial and interradial furrows
pass down with tolerable regularity into the peripheral axial canals of the central cal-
careous network (c.n).

(§ 57) In Antedon rosacea, as we have already seen, the five radial diverticula of the
ccelom terminate blindly on the ventral surface of the centrodorsal piece in five depres-
sions (PI. IV. fig. 15, q), which are disposed around the opening of its central cavity (ccl.c).
In correspondence wdth these depressions the dorsal surface of the pentagonal base pre-
sents five large openings (PI. IV. fig. 16, Q), disposed in like manner around the margins
of its central space. These openings are the dorsal terminations of the five radial axial
canals, and are formed by the application of the five radial spout-like processes of the
rosette (figs. 13, 16, p) to the inflected margins of the two openings (.(•', y) on the internal
face of each first radial (fig. 12 c), through which the secondary basal cords (X, 1\) pass
on their course from the fibrous mass enveloping the quinquelocular organ to the
circular commissure contained within the radial pentagon (compare PI. VIII. fig. 2).

The existence of these five large openings (PL IV. fig. 16, Q) is due to the fact that the
dorsal face of each first radial presents a deep notch in the centre of its inner margin
(fig. 12 b, Q') ; this notch indicates the continuation towards the dorsal surface of the
radial axial furrow on the internal face (fig. 12 c, a.r.f) ; and when this furrow becomes
converted into a canal by the application to its inflected edges of one of the spout-like
processes of the rosette, the notch on the dorsal face also becomes converted into a cir-

ME. P. H. CARPENTER OX THE GENUS ACTINOMETRA. 85

cular opening. So far as I know, these openings are tolerably constant on the dorsal
surface of the radial pentagon of Ant. rosacea ; but the five depressions corresponding to
them on the ventral surface of the centrodorsal piece are very variable in the distinct-
ness of their development ; and Dr. Carpenter has found that in some cases they may be
absent altogether \

This last condition, in which there are no radial depressions (q) on the ventral surface
of the centrodorsal piece, appears to be the normal one in Ant. celtica, in which I have
rarely found any traces of such depressions (PI. IV. fig. 2). The margin of the central
opening is usually almost circular, though sometimes bluntly stellate as in Ant. rosa
(fig. 15) ; at the same time, the five openings ( Q) upon the dorsal surface of the radial
pentagon are but little developed or even entirely absent. In PI. IV. figs. 3, 7, they are
represented as present in the small variety and absent in the large one ; but I have
sometimes found exactly the reverse to be the case.

"We shall find the same variability in the presence or absence of these openings in
Actinometra, not only in different individuals of the same species, but in the same indi-
vidual. This fact shows that too much reliance must not be placed on the presence,
absence, or difference in size of similar openings in the calyx of the fossil Crinoids
(the interradial " Li'icken " in Cupressocrinus, for example) as characters of any
systematic value. The absence or slight development of these openings in Ant. celtica is
principally due to the fact that the inner margin of the dorsal surface of the first radials
is not notched but straight, the radial axial furrow not being continued so far towards
the dorsal surface as in Ant. rosacea ; and also that processes grow inwards from the two
sides of the dorsal end of each of the five spout-like rays of the rosette, so that the
lumen of the canal it encloses becomes much diminished ; while in some cases similar
processes are put forward from the margin of the first radial, which unite with the
others so completely as eutirely to obliterate the lumen of the radial axial canal, and
thus form its dorsal boundary.

(§ 58) The dorsal aspect of the pentagonal base of the calyx of Actinometra is by no

1 Schliitcr (op. cit. p. 37) has proposed a division of the (fossil) Comatulce into two groups, characterized as fol-
lows : — (1) Centrodorsal with no radial pits and a round " Nahrungscanal " (central opening) ; and (2) Centrodorsal
with radial pits and a quinquepartite opening.

These characters, however, are far too uncertain to be of any systematic value. For example, Schliiter himself
notes the absence of the " Eadialgruben " in his own specimen of Solanocrinus scrobiculatus, while they were present
in one examined by Quenstedt (Echinodermen, p. 179). I have some specimens of Ant. celtica answering to the
first, and others to the second of the above definitions: and although most specimens at Ant. rosacea would be classed
in group (2), yet individuals with a pentagonal or even quinquepartite opening, ?»tZ no radial jiits, are not uncommon.
This last condition is very common among the ' Challenger ' Antedons. In fact, the radial pits of Ant. rosacea and
Ant. celtica arc peculiar to these species, and not always present even in them. In no other recent Comatula have
I found any thing exactly like them. They are not parts of the generally concave surface of each radial area, but
have distinct peripheral borders marking them off from these surfaces, and corresponding to the openings of the radial
axial canals enclosed within the spout-like processes of the rosette. In Act, pectinata, Act. polymorpTia, var. 2, and
in a new ' Ch.illeii' ' \edon there is a distinct pit at the central end of each radial area, which is merely the

deepened termination of a depression occupying its median line. Its nature (sects. 45 & (!1 ) is essentially the same as
that of the radial pits of Ant. rosacea, but its appearance is very different. Hence I cannot corroborate Sehluter's
statement that " Slanche Arten" possess '• Eadialgruben." He ouly describes them in ~> out of his 11 fossil species,
i besides the Solaru crinns scrooieidatus examined by Quenstedt.

12*

86 ME. P. H. CAEPENTEE OX THE GENUS ACTINOMETRA.

means of such a simple nature as it is in Ant. celttca (PL IV. figs. 3, 7), where it con-
sists solely of the five adjacent dorsal surfaces of the component radials. These are
somewhat elevated in the centre hut fall away towards the sides, where they are sepa-
rated from one another by slight furrows, corresponding in position with the five inter-
radial ridges on the ventral surface of the subjacent centrodorsal piece (PL IV. fig. 2, i.e).
In Actinometra, however, these dorsal interradial furrows are very marked ; but they are
not usually visible on the dorsal aspect of the radial pentagon, as they are occupied by
five long processes which radiate outwards from the angles of the central vacuity in
which the rosette lies (PI. V. fig. 3 ; PL VI. figs. 4, 9, 13, 21, S).

The presence of these rays of the basal star introduces an element of considerable
complexity into the dorsal aspect of the pentagonal base of Actinometra ; and its nature
will be best understood if we commence with the study of its component pieces in
the large Act. robusta. The dorsal face of each first radial of this species (PL V.
fig. 12) is slightly convex, so as to fit into the somewhat depressed radial area correspond-
ing to it on the ventral surface of the centrodorsal piece (PL V. fig. 14, r.ar). The
centre of its inner margin is, as in Ant. rosacea, marked by a deep notch, which indicates
the position of the axial radial furrow occupying the median line of the internal face
(fig. 10, a.r.f.). The latter is converted into a canal by the union of its inflected edges
with those of one of the radial spout-like processes of the rosette (figs. 12, 13, p), in the
manner already described by Dr. Carpenter for Ant. rosacea.

The central notch on the inner margin of the dorsal face thus becomes a round opening
(fig. 12, Q), similar to that seen in Ant. rosacea (PL IV. fig. 1G, Q). A bristle passed
through this opening towards the ventral side, therefore, will follow the course of one of
the axial radial canals, in which its lower end is concealed by the spout-like radial
process of the rosette (PL V. figs. 12, 13, I). On the ventral side of the rosette the
radial axial canal is incomplete, as the furrow on the internal face is only partially
bridged over by the calcareous processes which extend themselves from its sides to meet
the rosette ; the bristle which lies in the furrow is therefore visible here and there through
the openings in the network {cm) formed by the inosculation of these processes (figs.
10, 13, 1). This is best seen in fig. 13, which is a view of two radials from within, together
with that portion of the rosette which corresponds to and is united with them; and also
in fig. 10, which represents the internal face of a single radial, from which the portions
of the rosette that are normally united with it have been removed, so that the whole
of the internal face is exposed. The bristle I is seen to lie in the deep furrow between the
two raised edges of the apertures {x 1 , y) of the central canal, and to pass upwards under
the network extending from the ventral half of the internal face, where it follows the
course of the axial radial furrow and emerges on the ventral aspect of the radial. The
furrow in which it lies is here continued into the numerous irregular furrows of the
ventral face which converge towards the centre of its inner margin (PL V. fig. 11).

Just above the dorsal surface of the radial, the axial furrow occupying the median
line of its internal face gives off a large horizontal diverticulum into the substance of its
calcareous tissue, which extends outwards for some distance between the central canal
and the dorsal surface of the radial (fig. 10, r.c') ; and, like the axial furrow or canal as

ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 87

it is in the natural coudition when the rosette is in situ, encloses a dorsal extension of the
body-cavity or ccelom. I have seen no trace of these diverticula in any other species of
Comatula that I have yet examined ; but they are very large and well marked in each of
the five first radials in Act. robusta.

(§ 59) The furrows which occupy the median line of the ventral and internal faces of
the first radials thus terminate in Act. robusta (PI. V. fig. 12) precisely as in Ant. rosacea
(PL IV. fig. 16), by five large openings (Q) on the dorsal aspect of the radial pentagon,
which are closed in the natural condition by the ventral surface of the centrodorsal plate
on which the radial pentagon rests.

In Ant. rosacea the course of the slightly marked interradial farrows which pass down
from the ventral aspect of the radial pentagon into the peripheral portion of the central
calcareous network is terminated interiorly by the five triangular interradial processes of
the rosette ; for the apices of these processes unite with the two members of every pair of
contiguous radials, just between the two adjacent apertures of their central canals (PI. IV.
figs. 3, 7, 10, o).

In Actinometra, however, the interradial furrows both are more marked on the ventral
surface of the radial pentagon, and, like the radial ones, become converted into canals,
terminating by five openings upon its dorsal aspect.

In Ant. rosacea the edge which separates the lateral and dorsal faces of each first
radial is tolerably sharp and straight (PL IV. fig. 12 b, c); but in Act. robusta it is
somewhat truncated (PL V. figs. 10, 12, 13), so that when the lateral faces of two radials
are in apposition a deep interradial furrow appears along the line of union of their dorsal
surfaces (fig. 12, a.i.f). In the middle of the inner margin of the floor of this furrow is
a notch similar to that marking the centre of the inner margin of the dorsal face of each
single radial, both in this species and in Ant. rosacea (PL IV. fig. 12 b, Q'), except that
two radials take part in its formation instead of only one. This notch marks the con-
tinuation towards the dorsal surface of an interradial furrow from the ventral aspect of
the pentagonal base.

The edges between the internal and lateral faces of each first radial are truncated in
the same way as those between the dorsal and lateral faces (PL V. figs. 10, 13). In the
natural condition, therefore, when the lateral faces of all the radials are in apposition
with one another in pairs, there are five axial interradial furrows alternating with the
radial ones, which occupy the median lines of the internal faces. The ventral portions of
these, as of the axial radial furrows, are partially bridged over by the inosculating
calcareous processes which extend themselves towards the ventral aspect of the rosette
from the internal faces of the five first radials, so that a bristle passed along their course
is only partially visible (PL V. fig. 13, II).

These superior portions of the axial interradial furrows are in free communication,
both laterally, with the radial furrows occupying the intervals between them, and cen-
trally with the remaining spaces of the calcareous network, of which system these two
sets of furrows form the peripheral part. Interiorly, i. e. towards the dorsal surface, each
of these axial interradial furrows passes between the two outer lips of the adjacent aper-
tures (x, x') of the central canals of two contiguous radials along the line of union of

8S MS. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

which the interradial furrow is situated. The outer lips of these apertures, like the inner
ones (PI. V. fig-. 10), are raised and applied to the similarly inflected edges of the five
spout-like interradial processes of the rosette, so that the furrow lying hetween the aper-
tures becomes converted into a complete canal. A bristle, therefore, which lies in the course
of this furrow (PI. V. figs. 12, 13, II) is concealed by the interradial process (o) of the
rosette. The dorsal end of the latter unites with the margins of the notch described above
at the central end of the dorsal interradial furrow, so as to produce a roundish interradial
opening on the dorsal aspect of the pentagonal base, through which the bristle passed
along the axial interradial furrow emerges from its concealment beneath the interradial
process of the rosette (figs. 12, 13, II). The manner in which these openings are closed
in the natural condition by the central ends of the rays of the basal star will be best
described further on.

(§ 60) In the condition and relative inclination of their dorsal and internal faces, the
first radials of Act. robusta are more like those of Ant. rosacea than those of any other
of tin- various Actinometree which I have examined. In Ant. rosacea the ventral sm'face
of the eentrodorsal piece (PL IV. fig. 15) is almost flat, as the five radial areas into which
it is divided lie nearly in a horizontal plane; and I he corresponding dorsal surfaces of
the five first radials are likewise horizontal, and form an angle of but little more than 90°
with the internal faces (PI. IV. fig. 12 b, c). In Act. robusta this angle becomes more
obtuse, so that the dorsal surfaces of the radials are somewhat inclined to the horizontal
plane (PL V. figs. 10, 13) ; and, in correspondence with this, the radial areas on the
ventral surface of the eentrodorsal (fig. 11, r.ar.) have a slight downward and outward
slope between their central and peripheral margins, so that the whole surface rises very
gradually from the circumference towards the centre.

This is also the case in Act. Solaris, in which the dorsal surface of the radial pentagon
slopes slightly downwards from its margin towards the opening of the central vacuity
in which the rosette is situated (PL V. fig. 3), so as to correspond with the gradual eleva-
tion between the circumference and centre of the ventral surface of the eentrodorsal on
which it rests (fig. 2).

Act. Solaris also agrees with Act. robusta in the fact that the sides of the dorsal inter-
radial furrow (fig. 3, a.i.f) which is produced by the truncation of the adjacent supero-
lateral edges of two contiguous radials are simple and straight, and not raised into
leaf-like folds, as in Act. pectinata (fig. 9 b) and Act. polymorpha (PL VI. figs. 9, 13, 21 ;
PL VII. figs. 1 d, 1 d, b.g).

In Act. Solaris there are none of the apertures which occur in Ant. rosacea and Act.
robusta, by which the axial radial canals open upon the dorsal surface of the radial
pentagon (PL IV. fig. 10, and PL V. fig. 12, Q). We have already seen (sect. 57) that they
may be absent in Ant. celtica (PL IV. fig. 3) ; and their absence in Act. Solaris is due to
the same cause as in this case, viz. to the want of a central notch on the inner margin
of the dorsal face of each first radial, and to the obliteration of the lumen of each canal
by the ingrowth of calcareous tissue from its sides.

In the closely allied Act. pectinata, however, these openings may be present (and not
improbably also in other specimens of Act. Solaris than the one which I have examined);

ME. P. II. CAEPENTEH ON THE GENUS ACTINOMETEA. 89

for the inner margin of the dorsal face of each first radial exhibits a slight median notch
(PL V. fig. 9 b, Q), though it is by no means so distinct as in Ant. rosacea (PI. IV.
figs. 12 b, 16) and Act. robusta (PL V. fig. 12).

In this species too the ventral surface of the centrodorsal plate (PL V. fig. 7) rises very
perceptibly from its circumference towards its centre ; and the dorsal face of each first
radial is very considerably inclined to the vertical internal face, the angle between the
two almost reaching 135° (PL V. fig. 9 b, c). Consequently, when the radial is viewed
from its dorsal side, the large projecting lips of the two apertures of its central canal are
seen below the central or inner edge of the inclined dorsal face (fig. 9 b, x, y). These
are not seen in a similar view of a first radial of Ant. rosacea, in which the inclination
of the dorsal to the internal face is very little over 90° (PL IV. fig. 12 b, c).

(§ 61) In Ant. rosacea and Act. robusta the slight convexities of the dorsal surfaces
of the first radials fit into the correspondingly slight concavities in the centre of the
radial areas on the ventral surface of the centrodorsal piece (PL IV. figs. 2, 3, 15, 16;
PL V. figs. 12, 11). In Act. pectmata, however, these areas are occupied by median
depressions, increasing somewhat in depth from their peripheral to their central ends
(PL V. fig. 7, r.ar) ■ but the dorsal faces of the first radials do not exhibit corresponding
ridges, for they have similar median depressions, which are also deepest at their central
ends (PL V. fig. 9 b, c, d.r.f).

When, therefore, the dorsal surface of the radial pentagon and the ventral surface of
the centrodorsal piece are in their normal state of apposition, they are separated from
one another along the median lines of the five radials by five cavities or radial spaces ;
these are largest at their central ends, and extend in a peripheral direction to open
externally by five small openings situated on the margin of the small centrodorsal
piece, beneath the radial pentagon which rests upon and extends considerably beyond it.
These "radial spaces " are seen in section in PL VIII. figs. 5-8, which represent parts
of foiu- sections selected out of a series that was cut through a decalcified calyx of
Act. pectmata.

The section represented in fig. 5 was cut across the angle of two radials {A, B) near
the edge of the centrodorsal piece, and the open outer ends of the radial spaces are cut
somewhat obliquely (r.s). Pig. 6 represents a section much nearer the centre, and the
closed inner ends of the radial spaces are seen just beneath the lower ends of the axial
radial canals (a.r.c), but not communicating with them. In fig. 7 two other spaces are
seen, cut almost longitudinally, as the section is one from the other side of the centre,
through the radii, C & E, almost in the direction of their axial nervous cords (»), beneath
which are the radial spaces (r.s) between the dorsal surfaces of the first radials and the
ventral surface of tbe centrodorsal piece. Lastly, in fig. 8, which represents a section
still further from the centre, and cut transversely to the direction of radius 1), the closed
central end of the corresponding radial space is seen, as in fig. 7, on the dorsal side of the
axial nervous cord (n); at either side of it (s) is the expanded lower end of one of the
axial interradial canals seen in fig. 7 (a.i.c).

The external medium which occupies these radial spaces between the radial pentagon
and the centrodorsal piece is only shut off from the dorsal portion of the coeloui en-

90 MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

closed within the radial pentagon, and from the centrodorsal coeloni, by the small
bony bars at their expanded central ends, which form the thickened inner or central
edges of the dorsal faces of the first radials (PL V. fig. 9 b, c.) They are slightly
developed in one of the varieties of Act. polymorpha, in which the radial areas of the
small centrodorsal piece (PI. VI. fig. 17, r.ar) and the corresponding dorsal faces of the
first radials (PI. VII. fig. I a, d) both exhibit median depressions, which gradually
increase in depth from their peripheral to their central ends. In both these cases the
centrodorsal piece is relatively very small and by no means conceals the first radials,
while its ventral surface rises very considerably from the circumference towards the
centre (PI. V. fig. 7, and PI. VI. fig. 17). The meaning of these radial spaces is to me
quite obscure. [n no other Comatula have I found any thing at all comparable to them
except in Act. robusta, where the axial radial canals give off horizontal diverticula (PL V.
fig. 10, r.c'), which extend outwards in a peripheral direction in the substance of each
first radial just beneath its dorsal surface. These, enclosing diverticula of the dorsal
portion of the ccelom, are, of course, in indirect communication with the external
medium, while the radial spaces in Act. pectinata are altogether outside the substance of
the first radials, communicate directly with the exterior, and are completely shut oil' from
the dorsal ccelom. There is, therefore, scarcely any resemblance between the two sets of
cavities, although they occupy very nearly similar positions, i. e. between the centrodorsal
piece and the whole or the greater part of the mass of the first radials.

In some species of the fossil Apiocrmus, however, cavities similar to the radial spaces
in Act. pectinata appear to exist between every pair of contiguous basals and the first
radial, which rests upon them and alternates with them in position. As the basal circlet
is generally regarded as comparable to a stem-segment, it is evident that the posi-
tions of these cavities in Act. pectinata and in Apiocrmus respectively are homologous
with one another. In Apiocrinus rotundas five lateral openings were discovered by
Miller ' on the circumference of the body, " in or between the lateral surfaces of the
joints of the pelvis (basals) and the insertion of the first costal (radial) joints," which in
one case he thought he was able to trace as a canal or perforation " passing through the
joint of the pelvis into the space between it and the costal joints, extending perhaps
thence into the perivisceral cavity" (i.e. into the dorsal division of the body-cavity).

Miller supposed these to be the openings of oviducts leading to an ovary situated in
this dorsal ccelom, just in the same manner as the five openings on the ventral surface
of the centrodorsal piece of Glenotremitcs were (till lately) regarded as genital openings,
although the genital glands of all the recent Crinoids with which we are acquainted are
situated in the arms and pinnules.

Similar openings to those seen by Miller in Ap. rotundas have been described in Ap.
obconicus by Goldfuss 2 , who also supposed them to lead into the body-cavity. This is,
however, certainly not the case with the homologous openings in Act. pectinata.

The interarticular pores in the upper part of the stem of Peiitacrinus are also homolo-
gous with the external openings of the radial spaces in Act. pectinata. They are the

1 Op. tit. p. 31. 2 Petref. German, p. Is7, Taf. vii. fig. 5, a, b, c.

ME. P. H. CAEPEXTEE ON THE GENUS ACTINOMETEA. 91

openings of spaces between the successive segments, which are similarly situated, with
regard to the radial symmetry of the animal, to the radial spaces in Act.pectluala, viz.,
in the direction of the radii ; and they are produced in the same way, by the apposition
of two grooves radiating outwards from the centre of each stem-segment, which are
largest at their central ends and shallowest towards the periphery.

(§ 62) In Ant. rosacea and celtica (PI. IV. figs. 3, 7, 16), and in Act. rohusta and Solaris
(PI. V. figs. 3, 10, 12, 13), the sides of the interradial furrows (d.i.f) on the dorsal surface
of the radial pentagon are simple and straight ; but in Act. pectinata that portion of the
dorsal surface of each first radial which is next to its truncated lateral edge is raised
into a sort of curved ridge or fold (PI. V. fig. 12 b, b.f), so that in the natural con-
dition of mutual apposition of the five first radials the dorsal interradial furrows become
somewhat lancet-shaped. They correspond in position with the basal grooves on the
ventral surface of the subjacent centrodorsal piece (PI. V. fig. 7, b.g), and in the cavity
formed by the apposition of the edges of these two grooves lie, as will be subsequently
seen, the five rays of the basal star.

The first radials of Act. polymorpha are very similar to those of Act. pectinata. Those
of variety 1 (PI. VI. fig. 12) are like those of the type (PI. VII. tig. 1), except in the
simpler condition of their ventral surface, which is far less marked by secondary ridges
and furrows than is the case in the type (PI. VI. fig. 5). In the other three varieties,
the first radials of which resemble one another very closely, this sculpturing of the
ventral surface is even more marked than in the type (PI. VI. fig. 23 ; PI. VII. fig. 1 c).
The angle between the dorsal and internal faces is considerably less in the type
(PL VII. fig. 1 a,d) and in var. 1 than in varieties 2, 3, and 1, the first of which
resembles Act. pectinata in the presence of a median depression of the dorsal face
(PL VII. fig. 1 a,d, d.r.f), which corresponds with a similar depression along the
median line of the radial areas of the small centrodorsal piece (PL VI. fig. 17, r.ar).
This dorsal interradial furrow does not exist in varieties 3 and 1, nor in var. 1 (fig. 13),
while there is a trace of it in some specimens of the type, but not in others (figs. 1, 9).
In like manner the development of the openings of the radial axial canals on the
dorsal surface of the pentagonal base, which are so large in Ant. rosacea and in Act.
rohusta (PL IV. fig. 16 ; PL V. fig. 12, Q), is in Act. polymorpha extremely variable. In
two specimens of the type (PL VI. figs. 1, 11) they are entirely absent, as in Act.
Solaris (PL V. fig. 3) ; in another the inner margin of the dorsal face of each first radial
exhibits a slight median notch (PL VI. fig. 9 ; PL VII. fig. la, Q'), which would be
completed into an opening by the apposition to it of the end of one of the radial
spout-like processes of the rosette.

In variety 1 this notch is fairly marked, and five small openings are consequently
visible around the central vacuity, on the dorsal surface of the pentagonal base (PL VI.
fig. 10, Q). In varieties 2 (PL VII. fig. la) and 3 it is somewhat more distinct; and
in var. 1 it exists iu three of the first radials, but not in the other two, so that there
arc only three openings on the dorsal surface of the pentagonal base (PL VI. fig. 21, Q).

The extent to which the basal folds are developed at the sides of the dorsal inter-
radial furrows is also very variable in Act. polymorpha. We have seen that, although

SECOND SERIES. — ZOOLOGV, VOL. II. 13

92 ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

absent in Act. Solaris (PL V. fig. 3), they are well marked in the closely allied Act. pec-
tinata (PI. V. fig. 9 h, h.f). In two specimens of the type of Act. polymorpha (PI. VI.
figs. 4, 11) they are absent altogether, while in a third they are very well marked (fig. 9,
b.f), as also in each of the varieties, three of which are represented in PI. VI. figs.
13, 21, and PI. VII. fig. 4 a. In all these cases the borders of the interradial furrows
on the dorsal surface of the pentagonal base, which are produced by the apposition of
the truncated superolateral edges of every pair of contiguous radials, assume a leaf-like
appearance, owing to the presence of the folds at their sides. The precise shape of these
leaves, which is different in the type and in all the four varieties, corresponds very
closely with the shape of the basal grooves on the ventral surface of the centrodorsal
piece, with which they also correspond in position. They further resemble them in
the fact that they are entirely devoid of the pigment which is so abundant on the other
parts of the surfaces of the radial pentagon and centrodorsal piece ; so that when these
last are separated from one another, the dorsal interradial furrows on the pentagonal
base, like the basal furrows on the centrodorsal piece, stand out sharp and distinct as
five white leaflets on a dark-brown background. They are best marked invar. 4 (PL VI.
fig. 24, b.f), in which the basal folds of every pair of contiguous radials are rather widely
separated from each other about the middle of their length. This is also the case,
though to a less extent, in var. 1 (tig. 13), where, as in var. 4, the dorsal interradial
furrows correspond very closely in shape with the basal grooves on the centrodorsal
piece (fig. 15).

This is particularly distinct in the specimen of the type represented in PI. VI.
figs. 8, 9, in which one of the basal grooves is much shorter than the rest, and does
not reach the margin of the centrodorsal piece. The basal folds at the sides of the
dorsal interradial furrow corresponding to this aborted groove are also imperfectly
developed, so that the borders of its outer end are simple and straight ; as they are
throughout the whole course of the furrows in Act. solans (PL V. fig. 3) and Act. robusta
(PL V. fig. 12).

This last condition may also occur in the type of Act. polymorpha (PL VI. figs. 4, 11) ;
and in correspondence with it the basal grooves on the centrodorsal piece are simple and
almost parallel-sided (PL VI. figs. 3, 10, b.g), just as in Act. Solaris and Act. robusta
(PL V. figs. 2, 14).

This correspondence in the appearance of the dorsal interradial furrows and basal
grooves which is also seen in Act. polymorpha, var. 2 (PL VI. fig. 17; PL VII. fig. 4 a),
is not, however, an invariable one; for in Act. pectinata the basal folds are very well
marked (PL V. fig. 9 b, b.f), and the dorsal interradial furrows, therefore, leaf-like in
appearance, as in most specimens of Act. polymorpha (PL VI. figs. 9, 13, 24). The basal
grooves, however, on the ventral surface of the centrodorsal piece are narrow and
parallel-sided (PL V. fig. 7, b.g), just as in the allied species Act. Solaris and Act. robusta
(PL V. figs. 2, 14).

The external or distal faces of the first radials of Act. polymorpha differ not a little
from the corresponding faces in Ant. rosacea and Ant. celtica (PL IV. figs. 4, 6, 8 , 12 a,
11, 17) ; in which, especially in the latter, the fossse (f) for the attachment of the muscles

ME. P. H. CAEPEXTEE OX THE GENUS ACTIXOMETEA. 93

are very large, and considerably more extensive than those which lodge the interarticular
ligaments (I/).

In Act. polymorpha (PI. VI. fig. 1; PI. VII. figs, lb, 4b), however, the muscular
fossae are very small, being best developed in var. 1 (PL VI. fig. 12,/); while the fossae
(h) lodging the interarticular ligaments are very extensive, and separated by the down-
ward continuation of the intermuscular furrow (f t ), which reaches the dorsal margin of
the opening of the central canal (c.c). The external faces of the first radials of varieties
2 (PI. VII. fig. 4 b), 3, and 4 resemble one another, but differ from the corresponding
faces in the type (PI. VII. fig. lb) in being somewhat higher in proportion to their
width, and in the fact that the fossae (j) lodging the elastic ligaments are relatively
smaller, not extending so far into the lower or dorsal angles of the face as is the case in
the type.

(v.) The Basals.

(§ G3) We have already seen that all the older observers regarded Comatula as devoid
of those five pieces resting upon the top segment of the stem to which, in the other
Crinoids, MuTler gave the name of " basals ; " and it was not until Dr. Carpenter l dis-
covered the extraordinary metamorphosis undergone by the embryonic basals of Coma-
tula (Antedon) rosacea and then transformation into the "rosette," that the existence
of basals, although internal and concealed in the adult animal, was recognized.

The rosette of Ant. rosacea and Ant. celtlca (PL IV. figs. 3, 7, 16, H) is a peculiarly
shaped circular plate, occupying the dorsal half of the central cavity in the pentagonal
base of the calyx, which lies much nearer to the dorsal surface of the pentagonal base
in the latter species than in the former.

A normal rosette consists of a disk perforated in the centre with ten rays proceeding
from it. Pive of these rays (PL IV. tig. 13, o') are short, triangular in form, and nearly
flat, and their position is interradial, as they are directed to the sutures between the five
radials, their apices joining the contiguous pairs of these just between the two adjacent
apertures (x, x') of their central canals.

Alternating with these five interradial processes of the rosette are five radial ones
(fig. 13, p), each of which has parallel margins inflected on its ventral aspect in such
a manner as to form a groove ; while the process itself is so curved towards its dorsal
aspect that this groove reaches the periphery of the rosette, and then terminates
abruptly as if truncated.

The inflected margins of each of these five radial or, as Dr. Carpenter has called them,
" spout-like " processes of the rosette are applied to the similarly inflected margins of the
dorsal half of the axial radial furrow, lying between the two apertures of the central
canal on the internal face of each first radial (fig. 12 c, x', //). In this manner a com-
plete radial axial canal is formed, which, as we have already seen, terminates on the
dorsal surface of the radial pentagon by the opening Q (fig. 10), or becomes closed
before it reaches the dorsal surface by the union of ingrowths developed from its walls.

Besides this very intimate union between the peripheral portion of the rosette and the

1 Phil Trans. 1865, pp. 744, 745.

13*

9i MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

internal faces of the first radials, its central portion is also frequently connected with
the radial pentagon by delicate processes, which sometimes sprout forth irregularly
from the inner margins of the component pieces of the latter ; but sometimes form a
more regular ingrowth, which considerably contracts the central space on the ventral
aspect of the disk, and becomes continuous with an annular projection from the ventral
face of the rosette.

(§ 64) Before attempting to understand the complicated condition of the basals in
Actinometra, it Avill be well to study the mode in which the embryonic basals of Ant.
rosacea become metamorphosed into the rosette, as described by Dr. Carpenter, from
Avhose memoir the following account is principally taken.

In the young animal each basal is a flattened irregularly pentagonal plate, the apex of
which lies between a pair of radials that partially rest upon it. On the ventral surfaces
of the basal plates lie the five primary basal cords 7\ IV, X, Y, Z, proceeding from the
angles of the quinquelocular organ. Each of these divides into two branches, V x , V 2 ,
. * . . Z u Z 2 , the secondary basal cords, which pass on to the ventral faces of each pair of
contiguous radial plates, c. <j. X, and 1\ to one radial, T 2 and X x to the next, and so on.

Both basals and radials gradually become much thickened by an endogenous extension
of the calcareous network, which takes place in such a manner that the basal cords
come to lie in furrows channelled out on the ventral surfaces of the plates. By a further
endogenous growth of the radial plates these furrows are converted into canals (the
" central canals " of Johannes Mtiller), which at first lie close under the ventral surfaces
of the plates, but come gradually, by a continuation of the same process, to lie in then-
central axis.

In the basals, however, this process of endogenous growth is followed by one of
absorption; for the cribriform film of which each basal is originally composed, and
which still forms its external layer, now undergoes absorption, especially in its central
portion, where it covers in the dorsal side of the primary basal cord ; so that the central
space left by the incomplete union of the proximal ends of the five embryonic basal
plates is extended on its dorsal aspect into five broad rays, though on its ventral aspect,
where it is bounded by the last-formed portion of the endogenous reticulation, it shows
no corresponding increase. It is this last-formed ventral portion which persists in the
adult as the five triangular intcrradial processes of the rosette (PI. IV. figs. 3, 7, 13,
16, d).

The formation of the five radial or spout-like processes is somewhat more complicated.
The removal of the external layer in the centre of the dorsal aspect is carried so far as
to leave nothing but a kind of thickened margin along those sides of the plate which are
received between the first radials ; and by an extension of the same process along the
median dorsal line of each plate as far as its salient angle, so as completely to remove
the terminal portion of its inferior or dorsal layer, its two lateral portions become sepa-
rated from each other at their distal ends, and remain as small curved processes extend-
ing outwards. Those of every two contiguous basals now unite to form a sort of ray
curving towards the dorsal aspect; and this is the rudiment of one of the five radial or
spout-like processes of the rosette, the shape of which becomes much more strongly pro-

ME. P. H. CAEPENTEB ON THE GENUS ACTINOMETEA. 93

nounced with the subsequent increase of its size (PI. IV. fig. 13, p). The rosette is thus
essentially formed at the expense of the secondary or ventral layer of the original basals,
the ends of the curved spout-like processes being- the sole residue of their primary or
dorsal layer ; and since, by the removal of the median portion of that layer in each plate,
the primary basal cords are left bare upon their dorsal aspect, they now pass from the
angles of the quinquelocular organ into the central canals of the first radials, on the
inferior or dorsal side of the calcareous skeleton which occupies the base of the calyx ;
instead of reaching them by passing, as they did in the first instance, along its superior
or ventral face or, as at a later period, through the middle of its substance.

Each of these primary basal cords, X, Y, &c, which are thus interradial in position,
divides into two branches, X u X,, Y x , Y 2 , &c, towards the periphery of the rosette, on
the dorsal surface of which it rests. These branches lie in the shallow channels which
mark the union of the base of each interradial triangular process (fig. 13 1), </) with the
two curved lateral processes above mentioned, each of which unites with a corresponding
process from the adjacent basal to form one of the five spout-like rays Qj) of the rosette.
The apex or peripheral end of each triangular process is directed to the suture between
two contiguous radials (figs. 3, 7, 16, d), to which it is attached just between the two
adjacent apertures (a; .?') of their central canals. Into these canals pass the secondary
basal cords X u X 2 , one into each of the two contiguous radials, so that one lies on each
side of the interradial process of the rosette.

(§ 65) As a general rule, this process, both in Ant. rosacea and in Ant. celtica, is short,
triangular, and slightly curved towards the ventral side. It is not always so, however,
for I have frequently met with specimens of Ant. rosacea in which one or more of the
interradial processes of the rosette, after bending for a short distance towards the ventral
side, turns suddenly downwards, and extends towards the dorsal surface of the radial
pentagon. At the same time the parallel margins of each of these abnormally developed
processes are so inflected towards the dorsal surface as to form a narrow interradial
spout-like process. This is so applied to the projecting and similarly inflected outer
edges of the adjacent openings of the central canals {x, x 1 ) in two contiguous radials
as to convert the interradial furrow lying between them into a complete axial inter-
radial canal, precisely similar in character to the radial axial canals already described
(§§52,57).

In one case which I have met with, four out of the five interradial processes of the
rosette were of this character. In the rosette represented in PL IV. fig. 13, only two of
the interradial processes (o) are long and spout-like, the other three (o') are short and
triangular, like those of a normal rosette.

This abnormal condition of the interradial process of the rosette of Ant. rosacea is of
considerable interest, as it is the normal one in Actlnometra and in many species of
Antedon.

Not only the interradial, but also the radial processes of the rosette of Ant. rosacea
may exhibit departures from their usual shape ; for the removal of the primary or
dorsal layer at the salient angle of one or more of the five embryonic basals may be
incomplete, so that the ends of the curved rays of the rosette exhibit lateral processes,

96 MR. P. H. CAEPENTER ON THE GENUS ACTINOMETRA.

which are the remains of the upper margins of the primitive hasal plates on which
the first radials rested. Occasionally the apex of the original hasal is left unabsorbed,
so that the two lateral curved processes which remain after the removal of the primary
external layer along the median line of each "plate remain in connexion with one another;
as is seen in the bottom part of the rosette represented in PL IV. fig. 13. The trian-
gular iuterradial process (o'), which is developed from a secondary calcareous deposit
on the ventral side of the original basal, has here become more or less completely united
with these primary bars connecting the two lateral portions of the basal. The latter
retain their primitive relation to the first radials, for they remain united with them
along the inner margins of their dorsal faces (fig. 10, b.b) ; and as they partially cover
in the secondary basal cords, X u X 2 , &c., on their dorsal aspect before they enter the
central canals of the first radials, I will call them the " basal bridge " (PL IV. fig. 13,
b.b).

This basal bridge is well seen in situ in the specimen of Ant. rosacea represented in
PL IV. fig. 10. It is remarkably well developed, being nearly as distinct as in Actino-
melra (PL V. figs. 3, 12, and PL VI. figs. 1, 13, 21, b.b), in which its presence is normal,
and not al (normal, as in Ant. rosacea. It is also slightly developed in the specimen of
Ant. celtica represented in PL IV. fig. 3 ; but in fig. 7 no trace of it is visible.

(§ G6) This tendency to an incomplete metamorphosis of the embryonic basals of Ant.
rosacea, and consequently to the abnormal persistence of a more embryonic condition
than usual, is of considerable interest, because in Actmometra and in many Anledons
a basal bridge, representing the apex and unabsorbed margins of the embryonic basal
plates, is normally present (PL V. figs. 3, 8, 12, and PL VI. figs. 4, G, 13, 18, 19, 22, 24,
b.b). While at the same time, as already mentioned, the iuterradial processes of the
rosette (o), which are developed from a secondary or ventral layer, are large and spout-like,
as is abnormally the case in Ant. rosacea, and acquire a connexion with the remains of the
primary or dorsal layer which forms the basal bridge. The complicated rosette thus
constituted also becomes united with the large, more or less spindle-shaped, rays of the
basal star (S), the origin of which, as will be subsequently seen, is totally different from
that of the rosette.

A single " compound basal," as it may be called, of Actinometra, thus consists of two
distinct elements — (i) the incompletely metamorphosed embryonic basal, and (ii) a single
ray of the basal star. Its position is iuterradial, as it occupies the space enclosed between
the apposed edges of the basal furrows on the ventral surface of the centrodorsal piece,
and of the iuterradial furrows ou the corresponding dorsal surface of the radial pentagon
(PL VI. figs. 4, 13, 24,#).

An isolated compound basal which is thus constituted, when seen from its dorsal side
(PL V. fig. 8 b, PL VI. fig. 22 b), shows : — (i) more or less of the calcareous network {cm)
which unites the ventral surface of the rosette to the internal faces of the first radials ;
(ii) a large iuterradial spout-shaped process (o) ; (iii) two small, radial, curved processes
(//), extending outwards from the base of the iuterradial process, and representing the
unabsorbed lateral portions of the primary layer forming the embryonic basal plate.

ME. P. H. CARPENTER ON THE GEXUS ACTINOMETRA. 97

(iv) The basal bridge (b.b), consisting of two calcareous bars, that represent the
unabsorbecl peripheral margins of the embryonic basal on which two first radials rested.
They extend towards one another from the outer ends of the small radial processes, until
they meet at a point that represents the apex of the embryonic basal, and is situated
on the dorsal side of the peripheral end of the interradial process (o), developed from the
secondary or ventral layer, which becomes united with the basal bridge.

(v) The ray (S) of the basal star, which is joined to the interradial process and to the
basal bridge, along the line of union of the two primary bars constituting the latter with
one another and with the secondary interradial process, i. e. at the apex of the embryonic
basal. The development of this ray is quite different from that of either the primary or
the secondary portions of the compound basal. It is really a tertiary structure, being
nothing more than a deposition of calcareous material in the substance of the connective
tissue of the synostosis between the centrodorsal piece and the radial pentagon.

(vi) At the sides of the interradial process (o), bounded laterally by the radial pro-
cesses (p), and externally by the bars of the basal bridge {b.b), are two large apertures,
*u #3j >/\, 1/-j, &c., in each compound basal. Through these apertures pass the secondary
basal cords, X x , X,, Y l} Y,, &c. (PI. VIII. fig. 3), which result from the bifurcation of the
primary cords, X, Y, Z, proceeding from the angles of the quinquelocular organ. The
two secondary cords lie in the depressions on the dorsal surface of the compound basal,
between the central ends of its radial and interradial processes. They then pass outwards
through the apertures (a\, a\, &c.) beneath the bars of the basal bridge, and enter the
adjacent openings (%, a/, &c.) on the internal faces of the two contiguous first radials 1 ,
which contribute to form the dorsal interradial furrow occupied by the single fusiform
ray (S) of the corresponding basal.

The ventral surface of each of these rays of a compound basal (PI. V. fig. 8a; PI. VI.
figs. 6, 18, 22 a) is not flat, like the dorsal surface, but occupied by a prominent median
ridge, so that the ray is triangular in section. This ridge does not extend quite to the
central end of the ray, which is occupied by a considerable depression (s), forming the
peripheral end of the groove contained in the spout-like interradial process (o). In the
natural condition, when the basals are in situ and in connexion with the radial pen-
tagon, the inflected edges of this process unite with those of the axial interradial furrow
to form an axial interradial canal. This terminates on the dorsal surface of the radial
pentagon by a small opening situated at the central end of the dorsal interradial furrow
(PI. V. fig. 12, II), in which furrow the tertiary element (&') of the corresponding com-
pound basal is received. The depression (s) at the central end of the ray (PI. V. fig. 8 a ;
PI. VI. figs. G, 18, 22 a) lies over this opening, and thus forms a blind end to the axial
interradial canal (PI. VIII. figs. 3, 5, 7, a.i.c ; fig. 8, s)— precisely in the same manner
as the depressions (q) on the ventral surface of the centrodorsal piece of Ant. rosacea
(PI. IV. fig. 15) receive the blind ends of the axial radial canals which open on the
dorsal surface of the radial pentagon by the five large openings, Q (PI. IV. fig. 16).

A view of a single compound basal does not, of course, show one of the large and

1 In PL V. fig. 13, four of these openings are seen on the internal faces of the two contiguous first radials,
viz. x', y, y\ z.

98 MR. P. II. CARPENTER ON THE GENUS ACTINOMETRA.

spout-like radial processes of the rosette ; for each of these is a composite structure,
formed by the apposition of two of the small curved lateral processes of contiguous basals
(PL VI. rig-. 22, p'). This is seen in PI. VI. rigs. 6, 18, 19, particularly in the last two ; for
the union of the adjacent lateral processes (//) of the two contiguous basals which are
there represented, is seen to be incomplete, so that a slight fissure is visible along the
median line of the dorsal surface of the composite radial process (fig. 19, p). The peri-
pheral end of this radial process is united to those of the interradial processes (o) at its
sides by the bars of the basal bridge (b.b). Their central ends arc also united around the
opening of the rosette (PI. VI. figs. 4, 19, 24, II) ; but their median portions are separated
by the two apertures (a? 2 , y x ) by which the two adjacent secondary basal cords ( X>, li) pass
out under the bars of the basal bridge, to enter the two openings of the central canal on
either side of the axial radial furrow on the internal face of the first radial (PL V.
figs. 9 c, 10 ; PL VII. figs. 1 a, 4 a, a/, y), with the inflected inner edges of which the
radial spout-like process (p) unites.

The openings (% u y._) by which the other branches (X u Y 2 ) of the two primary basal
cords pass outwards to reach the central canals of the other two radials corresponding to
these two basals are best seen in a dorsal view as shown in PL VI. fig. 19. This also
shows the two outer lateral processes (//) of these united basals, which would naturally
unite with those of the two next contiguous basals, one on each side, to form two more
radial spout-like processes.

(§ G7) The tertiary elements which form the rays of the basal star vary very con-
siderably in their shape and in the completeness with which they are developed, just as
do the interradial furrows on the dorsal surface of the radial pentagon in which they lie.
In Act. Solaris (PL V. fig. 3, d.i.f) these have no curved folds at their sides, and the rays
of the basal star (*S') are only imperfectly calcified rods, long and narrow, like the basal
grooves on the ventral surface of the centrodorsal piece into which they are received
(PL V. fig. 2, b.g). In Act. pectinata, however, although the basal grooves are long and
narrow as in Act. Solaris (PL V. fig. 7), yet the dorsal interradial furrows are widened by
the presence of large curved basal folds at their sides (fig. 9 b, b.f) ; and in correspond-
ence with these the tertiary basal elements (fig. 8, S) are much wider, and also far
more perfectly calcified (being solid throughout), than is the case in Act. Solaris.

In Act. robusta the central ends of these basal rays are wide and stout, and com-
pletely calcified as in Act. pectinata ; but their peripheral ends are much thinner, and
consist of a simple curved plate, which forms a sort of bridge over the dorsal interradial
furrow (PL V. fig. 12, S), the borders of which are straight, as in Act. Solaris (fig. 3, d.i.f),
and not marked by any lateral folds. The basal grooves on the ventral surface of the
centrodorsal piece are also simple and nearly parallel-sided (fig. 14, b.g).

In Act. polymorpha and its varieties the condition of the basal star varies extremely,
like that of the basal folds and basal grooves, the development and shape of which
exhibit a very close correspondence with the appearance of the rays of the basal star.

Thus, in that specimen of the type of Act. polymorpha in which the basal grooves are
narrow and parallel-sided, and all terminate well within the margin of the centrodorsal
piece (PL VI. fig. 3, b.[/), Avhile no basal folds are present at the sides of the dorsal inter-

ME. P. H. CARPENTER ON THE GENUS ACTIXOMETRA. 99

radial furrows (fig. 4, d.i.f), the rays of the hasal star are short and flattened, and do not
hy any means reach the angles of the radial pentagon (fig. 4, S). Their dorsal surface is
somewhat depressed along the median line, while the depression (fig. 6, s) at the central
end of the ventral surface which receives the Mind end of the axial interradial canal is
continued outwards in a peripheral direction somewhat further than in Act. pectlnata
(PI. V. fig. 8 a, s); and the median ridge which runs from its end to the apes of the ray
is less marked than in this species.

In another specimen of the type, however, in which hoth hasal grooves (PL VI.
fig. 8, b.g) and hasal folds (fig. 9, bf) are wide and well marked, the basal rays are
stout and thick, with a fairly distinct median ridge on their ventral surface. In fig. 8,
three of them are seen occupying their normal position in the hasal furrows on the
ventral surface of the centrodorsal piece, with which they are closely connected, while
the other two rays have remained in connexion with the rosette and radial pentagon.

In a third specimen of the type the ray S of the hasal star are very imperfectly
developed ; two only extend for any distance towards the angles of the radial pentagon
(PL VI. fig. 11, <$'), while of the other three little or nothing is to he seen. In this
specimen, as in the one first described, there are no basal folds, and the basal grooves
are parallel-sided and only imperfectly developed (fig. 10, b.g). It is also remarkable
for the fact that the absorption of the apex and outer margins of each embryonic basal
plate seems to have been very incomplete ; for the bars of the basal bridge are so wide,
and extend so far towards the centre from the inner margins of the dorsal surfaces of the
first radials, with which they are closely united, that they entirely conceal the apertures
in the compound basals (PL V. figs. 3, 8; PL VI. figs. 1, 6, 13, 18, 19, 22, 21, x u x,, y x ,
y 2 , &c.) through which the secondary basal cords pass in order to reach the central
canals of the first radials. Consequently nothing is seen of the rosette in a dorsal view
of the pentagonal base but its central opening surrounded by a raised rim (PL VI.
fig. 11, r.o). In all the other figures, however (PL V. fig. 3 ; PL VI. figs. 1, 13, 21), these
apertures are large and distinct, every one being situated between a radial (p) and an
interradial process (o) of the rosette.

In Act. polymorpha, var. 1, both basal grooves (PL VI. fig. 15, b.g) and basal folds
(fig. 13, b.f) are well marked and somewhat lancet-shaped in form; the rays of the
basal star wdrich occupy the former are much flattened dorsally (fig. 22 b, S), as in one
of the specimens of the type (fig. 1, S). They are not, however, so short as in this case,
but, like the basal folds at their sides, reach the outer angles of the radial pentagon.

This is also the case in the other three varieties. In var. 4 the basal folds diverge
considerably at about the middle of their course (fig. 21, b.f), so that the dorsal inter-
radial furrow is here very wide, and then rapidly narrows towards its peripheral end.
In correspondence with this, the basal rays also widen somewhat from their narrow
central ends, and then begin to decrease in width as they approach the angles of the
radial pentagon (fig. 24, S) ; they are also marked by a slight median furrow along their
dorsal surface, as is the case in one specimen of the type (fig. 4, S). In varieties 2 and 3,
as in the type, and in var. 1, the basal rays are widest at their central ends (figs. 18,
19, 22, S). In both cases the basal grooves (figs. 17, 21, b.g) and basal folds (PL VII.

SECOND SERIES, — ZOOLOGY VOL. II. 11

100 ME. P. H. CARPENTER ON THE GENUS ACTINOMETEA.

fig. 4(1, b.f) are well developed and somewhat lancet-shaped in form, as in var. 1
(PI. YI. fig. 13, b.f, fig. 15, b.g).

In var. 2, and still more in var. 3, the mode of union of the hars of the basal bridge
(figs. 19, 22 b, b.b) with one another and with the basal rays (S), is seen very distinctly
at the central end of the latter; much more so than in Act.pectinata (PL V. fig. 8 b), in
which, as in the other specimens figured (PI. V. fig. 3 ; PI. VI. figs. 4, 11, 13), the
various elements of each compound basal are so completely united, that the lines of
unction between them become almost indistinguishable.

(§ 68) The complicated condition of the basals described above in Actinometra is not
altogether peculiar to this genus, as I was first inclined to believe ; for in Ant. Eschricldii
a basal star may be developed to a greater or less extent. In his paper on JPhanogcuia
Loven gives a diagram 1 of the dorsal aspect of the pentagonal base of the calyx of this
species for comparison with that of Thanogenia. It shows five large rays extending
from the periphery of the rosette to the outer angles of the radial pentagon, with the
constituent elements of which they alternate in position ; and in the text he speaks of
them as belonging to the rosette. Unfortunately, his paper is written in Swedish, so
that I have been unable to ascertain precisely what his views were with respect to the
homologies of these rays.

His figure also shows five radial openings on the dorsal surface of the pentagonal
base, which correspond with the dorsal openings of the radial axial canals in Aid. rosacea
(PL IV. fig. 1G, Q).

In neither of the two specimens of Aid. Eschrichtii which I have been able to
examine do these openings exist, so that they are probably somewhat uncertain in their
occurrence, as in Ant. celtica (PL IV. figs. 3, 7) ; and in neither of these specimens is
the basal star developed to any thing like the extent that it is in the specimen figured by
Loven. In one (PL IV. fig. 10, S) the rays are excessively small and inconspicuous,
only extending for a very short distance along the dorsal interradial furrows (d.i.f),
while the corresponding basal grooves on the interradial elevations of the centrodorsal
piece are also very slightly developed (fig. 11, b.g). The basal bridge also, connecting
two successive rays of the basal star, is barely traceable around the inner margin of the
radial pentagon (fig. 10, b.b).

In the other specimen which I examined the basal rays were somewhat better
developed, occupying a larger portion of the dorsal interradial furrows, and extending
further outwards towards the margin of the radial pentagon ; although still remarkably
slender and delicate, somewhat as in Act. Solaris (PL V. fig. 3, S), and by no means so
large as in the specimen figured by Loven 2 .

The interior of the calyx of Ant. Eschrichtii is much simpler than that of Actinometra,

1 Loc. cit. p. 230, m.

2 Sinco the above lines were written, I have examined several other specimens of Ant. Eschricldii. None of
them have the radial openings above mentioned, nor arc the rays of the basal star so large as in the specimen figured
by Loven ; but they arc always present, though less regularly developed than in Actinometra. Further, ray work on
the 'Challenger' Comatulce has brought out the fact that a basal star is nearly always present in Antedon as well as
in Actinometra, so that the British species (Ant. rosacea, Ant. celtica) are remarkable for its absence, rather thun
Ant. Eschrichtii for its presence.

ME. P. H. CAEPEXTEE (XN THE GENUS ACTINOMETEA. 101

as Avill be seen by comparing PL IV. fig. 9 with PI. V. fig. 13, both of which represent
the internal aspect of two united first radials. In Actinometra (PI. V. fig. 13) there is
an abundant calcareous network (c.n) in connexion with the internal faces of the radials,
which are marked by well-developed radial and interradial furrows. In Ant. JEschrichtii,
however (PI. IV. fig. 9), the processes forming tins network are but little developed;
there is no axial interradial furrow, and even the radial one is indistinct, except near
the dorsal surface, where it passes between the raised edges of the two apertures
(x', y, y', z) of the central canal which unite with the inflected edges of a radial spout-
like process (p) of the rosette.

The interradial process of the rosette (o') is short and broad, but without the spout-
like character which it has in Actinometra — being simply directed, as in the normal
condition of Ant. rosacea (PI. IV. figs. 3, 7, 16), to the line of suture between the two
contiguous radials, to which it is attached between the two adjacent apertures (fig. 9, y, y')
of their central canals.

(§ G9) The remarkable variation in the extent to which the rays of the basal star may
be developed, as described above in Actinometra aud in Aniedon JEschrichtii, is due to the
fact that they are not calcifications in a nucleated protoplasmic network like the ordinary
elements of the skeleton. They are the result of a calcareous deposition, of a more
or less regular character, around the connective-tissue fibres which effect the synostosis
with the centrodorsal piece of every pair of contiguous radials along the line of
contact of the latter ; so that their position is, as we have already seen, interradial.

In PI. III. fig. 5 is seen the lower end of a vertical section cut transversely to the plane
of synostosis of two decalcified first radials (;•,) olAct. polymorpha, close to their peripheral
margin wdiere they are not concealed by the centrodorsal piece, so that the fibres of the
elastic ligaments connecting them with the second radials are cut somewhat obliquely (/J.

The threads of the protoplasmic network of which the organic basis of the radials is
composed pass somewhat rapidly at then surfaces into the connective-tissue fibrils (L)
which run horizontally between them and elfect the synostosis. These fibrils being very
closely set, the superficial portions of the calcareous reticulation forming the skeleton
of the radials, which are ossified around their ends, are very much more dense than the
central portion produced by calcification of the protoplasmic network.

PI. VIII. fig. 4 represents a section, from the same series as the previous one, across
the line of union of the same two radials (A, B), rather nearer to the centre of the calyx,
so that their dorsal surface appears no longer free, but partially covered by the centro-
dorsal piece (ccl). The lower portion of this section, more highly magnified, is seen in
PL III. fig. 6. The synostosis of the radial areas of the centrodorsal piece with the
dorsal surfaces of the first radials is effected by simple and not specially numerous con-
nective-tissue fibres (/) ; these pass directly from the protoplasmic basis of the one piece
into that of the other, in a direction vertical to the plane of the opposed surfaces, just
as in an ordinary synostosis. But in the direction of the interradii the course of the fibres
is different, and they have a deeper origin in the substance of the centrodorsal piece
than those which occupy the radial areas (PL III. fig. 6).

There are three principal masses of these longer inten adially placed fibres : — two smaller

14*

102 ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

ateral ones (S 3 ), in which the fibres have the same direction as those occupying the radial
areas ; and a large median mass, in which the fibres ascend vertically for some distance
(S-,) and then diverge to the two sides (S { ), where they pass into the protoplasmic basis
of the radials ; and the horizontal fibres (L) which pass between the radials fill up the
open angle caused by the divergence of the ascending fibres. There is thus a much greater
development of connective-tissue fibres, effecting the synostosis of the centrodorsal piece
with the radial pentagon, in the interradial than in the radial planes. This is well seen
in PI. VIII. fig. 3, which represents a longitudinal section through the calyx of Act.poly-
morpha. On the right side it is interradial, passing through the synostosis of the first
and second radials of the two radii, A, B ; and the connective-tissue fibres (S 3 ) connecting
the centrodorsal piece with the edges of these two radii are longer and more abundant
than those on the left side (/), passing between the centrodorsal piece and the first radial
of radius B, which is cut longitudinally.

This is also seen in PI. VIII. figs. 5-8, which represent portions of four out of a series
of sections through a decalcified calyx of Act. pectinata. These are in the same plane as
the section of the calyx of Act, polymorpha represented in fig. 4, i.e. transverse to the synos-
tosis of the radii A, B on the one side of the centre, and to the radius B on the other.

.Pig. 5 represents a section, rather nearer the centre than fig. 4, passing vertically
along the axial interradial canal (a.i.c) ; beneath the dorsal end are seen the vertical
ascending fibres (S 2 ), which have a much deeper origin in the substance of the centro-
dorsal piece of this species than in Act. polymorpha. The diverging fibres are not seen,
as they give rise by their calcification to the long basal ray (PI. V. fig. 8, S); and this
section passes through the depression at the central end of its ventral surface (PL V.
fig. 8 a, s) in which the axial interradial canal terminates.

Pig. 6 is somewhat nearer the centre, but still shows the long vertical fibres (S-,) in the
interradial plane, together with a portion of the central calcareous network aud the
axial radial canals (a.r.c) corresponding to the two radii A, B.

Pig. 7 is just beyond the centre, i. e. across the inner end of the first radial of B, so
that no vertical fibres arc visible, as they are only interradial in position. Two sets of
them, however, are seen in fig. 8, which shows the first radial of B cut transversely
rather further from the centre, so that the fibres (I) effecting its synostosis with the
adjacent radials of C and E are cut obliquely; beneath these are seen the interradial
ascending fibres (S 2 ), which diverge slightly at their upper extremities (S\).

These diverging fibres and the upper ends of the vertical ones are the basis around
which the calcareous material forming the rays of the basal star is deposited. As the ven-
tral surface of the centrodorsal piece on which these rays rest is much higher at the centre
than at the circumference, it is impossible to obtain horizontal sections in which these
five rays are seen at all complete. Oblique sections, however, may be obtained in which
one or more of them are cut along the greater part of their length. Two such sections,
seen from their dorsal side, are represented in PL VIII. figs. 1 & 2 ; their lower left-hand
portions lie nearer the dorsal surface of the calyx than the upper right-hand portions.

The centre of fig. 1 is occupied by the fibrous envelope N of the quinquelocular organ,
from the dorsal portion of which cords (n.c) proceed to the cirrhi (cir). At the top and

MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 103

right of the figure are seen portions of three first radials (t\), the central parts of which
are lighter than the more peripheral parts, as the section here passes through the slightly
developed unpigmented fibrous tissue (I) connecting the radials with the centrodorsal
piece (cd), the peripheral portion of which consists of the same pigmented protoplasmic
network as the substance of the radials.

Two of the synostoses between the latter are seen at the top of the figure ; but they
do not quite reach to the centre, where their place is occupied by the central ends of two
of the rays of the basal star (#1), the remaining three rays of which are visible in the
lower part of the figure for the greater portion of their length. Two of them are also seen
in the left or more dorsal portion of the next section (fig. 2), which also shows three
synostoses (L) between the radials in the right or more ventral part of the figure.

The sections of these rays of the basal star appear very dark, not from the presence
of pigment, which is entirely wanting in their fibrous basis, but because of the abund-
ance and very close approximation of the fibres of this basis ; these are the diverging
and vertical fibres seen in PI. VIII. figs. 3, 4, and more highly magnified in PI. III.
fig. 6 (S u S 3 ). Just in the same way, in PI. VIII. fig. 3, the section appears much darker
on the right-hand side, where it cuts the closely approximated connective-tissue fibres
effecting the synostosis between two first radials transversely, than on the left-hand side,
where it passes through the more open protoplasmic network of the individual segment
of a single radial, I).

(§ 70) This great development of fibrous tissue along the interradial portions of the
centrodorsal piece and of the pentagonal base of the calyx accounts for the fact, often
mentioned already, that there is no pigment in the substance of the rays of the basal
star (which is a more or less complete calcification of the central portions of these inter-
radial fibrous masses), nor in the walls of the basal grooves on the centrodorsal piece, nor
in those of the dorsal interradial furrows on the inferior surface of the pentagonal base,
which are calcifications of the smaller lateral masses of long fibres running directly from
the organic basis of the centrodorsal piece into that of the first radials (PL III. fig. 6, 6' 3 ).
These lateral fibres have a common point of origin in the substance of the centrodorsal
with the vertical and diverging fibres (#„ S 2 ), around which the calcareous tissue of the
basal rays is deposited. It is therefore easy to understand that the calcification may
in some cases be so complete that the basal rays formed around the median fibres (S 1} S 2 )
may become completely united with the walls of the basal grooves formed around the
lower ends of the two lateral fibrous masses (*$'.,) ; as is the case in the specimen of Act.
polymorpha represented in PL VI. fig. 8, where two of the rays of the basal star (S) are
so completely united with the floor and sides of the basal grooves in which they lie that
the line of junction between them becomes indistinguishable.

The fact that the rays of the basal star are calcifications in connective tissue and not
in the ordinary nuclear tissue which forms the organic basis of the other parts of the
skeleton, also affords an explanation of the great variations in the extent to which the
rays are developed. The general arrangement of the fibres constituting the interradial
portions of the synostosis between the centrodorsal piece and the radial pentagon is essen-
tially the same in Antedon as in Actinomctra. In Ant, rosacea and Ant. celtlca they

104 ME. P. H. CAEPEXTEE OX THE GENUS ACTINOMETEA.

never seem to undergo calcification, though this may take place in Antedon JEschrichti%
either only very slightly, as in the two specimens examined by myself (PI. IV. fig. 10, S),
or to a considerable extent, as in that figured by Loven. In Actinometra also the extent
of calcification of the rays of the basal star is very variable. In one specimen of Act.
polymorpha scarcely any trace of them is visible (PI. VI. fig. 11, S) ; in another they are
short, but otherwise well developed (fig. 4, S) ; while in others they may extend very
nearly to the outer angles of the radial pentagon (figs. 13, 18, 19, 22, 24, S).

In Act. pectmata, again, we have found them to be large and thick (PL V. fig. 8, S),
while in the closely allied Act. Solaris (PI. V. fig. 3, S) they were slender rods, only imper-
fectly calcified here and there — the intervening portions of the dorsal interradial furrow
seen in the prepared skeleton (d.i.f) being occupied in the fresh state by masses of fibrous
tissue, which are removed by the action of the alkali used in preparation.

(§ 71) We have seen that the basal circlet of Actinometra is somewhat complicated
in its nature, and consists of two entirely distinct elements, viz. a central rosette, which
we may fairly suppose to be the result of the metamorphosis of the embryonic basal
plates, as in Antedon rosacea, and five more or less completely ossified rays extending
from it in a peripheral direction.

The rosette is situated on the ventral side of the quinquelocular organ, from the fibrous
envelope of which proceed the primary basal cords. These are very short, and soon
bifurcate, so as to give rise to ten secondary cords, which pass through the ten apertures
(i\, v 2 , . . • z u 2 a ) m the peripheral portion of the rosette in order to reach the central
canals of the first radials.

As already remarked by Ludwig , we may fairly regard those elements of the skeleton
in which the bifurcation of the primary basal cords occurs as homologous throughout
the different genera of the Crinoids. This leads us to the conclusion that the rosette
of Antedon and Actinometra is homologous with the united central ends (at least) of the
basals of Pentacrmus, which are perforated by canals that lodge the five bifurcating
fibrillar cords proceeding from the dorsal angles of the quinquelocular organ, and not
from its ventral angles as in Comatula.

The question now arises, Where are we to seek for the homologues of the five rays of
the basal star in most Comatulce? Unfortunately the only type in which we find
a condition any thing like that described above is a fossil one, the Solanocrinus of
Goldfuss ; so that it is difficult to ascertain the precise relation of its basals to the canals
proceeding from the quinquelocular organ that was undoubtedly contained in the cavity
of its deep centrodorsal piece.

The upper surface of the latter, according to Goldfuss 2 , presents " five radiating
elevations on which the pelvis articulates." They correspond to the interradial elevations
on the ventral surface of the centrodorsal piece of Antedon and Actinometra.

The basals themselves vary in appearance in the different species. In S. costatus and
S. scroblcidatus they are only " fiinf schmale Strahlen die sich zwischen die Nahte der
Bippenglieder einsenken;" but in S. Jcegeri they are much wider, " so dass sie auf der
ganzen Gelenkfiaeke zusammenstossen und bier ftinf ausstrahlende Furchen zur Aufnahme

1 Beitriige &c. he. cit. p. 07. a Petrof. Germ. Joe. cit. p. L66.

ME. P. H. CAEPEXTEE OX THE GEXTJS ACTIXOMETEA. 105

der Saule bildeii " (p. 168). This species, however, is possibly not a Solanocrmus at all,
but the head of a stalked Crinoid.

In both these cases the peripheral ends of the basals appear on the external surface of
the calyx, between the centrodorsal piece and the radial pentagon, although the extent
to which they are visible is very different.

The long and narrow prismatic-shaped basals of S. costatus evidently represent
the five rays of the basal star of Actlnomctra. The iuterradial elevations on the
ventral surface of the centrodorsal piece are marked by five median grooves for the
reception of the basals, just like the basal grooves of Actinometra ; and these iuterradial
elevations are continued beyond the margins of the radial areas, just like the small
processes (t) in some species of Actinometra (PL VI. figs. 14, 15). They correspond with
five longitudinal ridges on the outer surface of the columnar centrodorsal which separate
the rows of cirrhus -sockets.

We do not, of course, know whether there was a rosette in S. cuslalns. I am
inclined to think that this was not the case, as the central ends of the five basals are in
: contact with one another laterally for a short distance, instead of being united by narrow
bars forming a basal bridge, as in Actinometra ; and their internal or proximal faces
were probably perforated by the opening of a short bifurcating canal lodging the fibrous
cords on their way to the central canals of the first radials, as in the closely similar basals
iof Pentacrinus asteria. Hence these five basals as a whole would represent the circlet of
compound basals in Actinometra, viz. the rosette together with the rays of the basal star.
Whether, however, only the united central ends of the basals of S. costatus represent the
embryonic basal plates, like the rosette of Antedon and Actinometra, or whether the whole
star results from a metamorphosis of the embryonic basals, is a question which must
remain in doubt, though the latter is by far the more probable supposition. Apart from
the analogy of Pentacrinus, as the peripheral ends of the basal rays extend beyond the
margin of the radial pentagon, it is hardly likely that they can be the result of cal-
cification in the iuterradial portions of the synostosis between the radial pentagon and
centrodorsal piece, as in Actinometra and Antedon Eschrichtii.

The calyx of the doubtful S. Jvgeri presents a great advance upon that of S. costatus
with respect to the development of the basals, which led Pictet L to propose the erection
of this species into a separate genus. Instead of being long and narrow, and in contact
only by their central ends, as in S. costatus and 5. scrobiculaius, they are broad and wedge-
shaped, and in contact along their whole sides, so as to form a complete calcareous disk
entirely separating the radial pentagon from the centrodorsal piece.

This is occasionally their position in Pentacrinus, though there are but few species of that
genus in which the basals are relatively so large and complete as in Solanocriuus Jcegeri.
In P. asteria, and in the two fossil species P. briareus and P. subangularis, they are
small and cuneiform and only in contact by their central ends, just as in S. costatus, so
that the greater portion of the radial pentagon is in contact with the top stem-seg-
ment. In P. Mulleri they are in contact for about half their length, and then diverge,
while in P. Wycille-Thomsoni they are completely united with one another along the

1 Op. cit. p. 2SS.

106 MR. P. H. CAEPENTEE ON THE GENUS ACTINOMETRA.

whole length of their sides, so as entirely to cut off the radial pentagon from the top
stem-segment, just as in S. Jcegeri.

There can therefore he little douht that the hasals of Pentacrinus are homologous with
those of Solanocrinus, and therefore analogous to the compound hasals of Actinometra,
which, as we have seen, are not entirely developed out of the embryonic basal plates.
It would seem, in fact, as if in Pentacrinus and Solanocrinus the embryonic basal
plates became directly transformed into the basals of the adult ; while in Comatula
they undergo metamorphosis into the central rosette by the absorption of the greater
portion of their dorsal or primary tissue, and the development of a secondary ossification
on the ventral side of the original plates.

In Ant. rosacea the metamorphosis is much more complete than in most Antedons
and in Actinometra, in which new skeletal elements are developed by a more or less
complete tertiary ossification in masses of connective tissue, that correspond precisely in
position, and to a certain extent also in shape, with the basals of Solanocrinus and Penta-
crinus. The latter being most probably direct products of the growth of the embryonic
basals are therefore strictly homologous only with the rosette of Actinometra, although
analogous in position to the whole circlet of compound basals in this genus, viz. to
the rosette and basal star taken together.

(§ 72) It is interesting to observe the different position of the basals with regard to
the chambered organ in Comatula and in the various species of the stalked Crinoids.

In Comatula this organ is situated in the cavity of the centrodorsal piece (PL VIII.
fig. 3) which is on the dorsal side, not only of the radial pentagon, but also of the rosette
or metamorphosed basals; and the nervous cords proceeding from its fibrillar envelope
to enter the central canals of the first radials come off from its ventral angles.

The large centrodorsal piece of Comatula is developed by the growth of the top stem-
segment of the Pentacrinoid larva. In Pentacrinus, which remains pedunculate
throughout life, the top stem-segment is the youngest and smallest. Its central
cavity is far too small to contain the quinquelocular organ forming the upper end of the
central axis of the stem, which contains five longitudinal chambers expanding slightly
at every nodal segment, where each of them gives off a single cirrhus-vessel 1 .

There is no special increase in the diameter of these chambers in the top stem-seg-
ments, and they do not expand into the large chambers of the quinquelocular organ
until near the level of the ventral surface of the basal circlet which surrounds the
dorsal half of the chambered organ. The ventral portion of the latter is contained in the
lower part of the central funnel-shaped space enclosed within the radial pentagon, where
it is surrounded by a very dense calcareous network, through which the axial prolongation
containing the superior continuations of the five chambers of the quinquelocular organ
ascends, on its way to enter the visceral mass, just as in Comatula. In consequence of
this relatively higher position of the chambered organ in Pentacrinus than in Comatula,
the nervous cords which enter the central canals of the first radials come off from its
dorsal angles, and not from the ventral ones as in Comatula.

Iu Comatula, therefore, the walls and floor of the cavity enclosing the chambered organ

1 Pentacrinus and Bhizocrinus, loe. eit. pp. 43-40.

ME, P. H. CAEPEXTEE ON THE GENUS ACTIXOMETEA. 107

are formed almost entirely by what was once a stem-segment ; while in Pentacrinus this
cavity is a part of the central space enclosed within the radial and basal pentagons, which
respectively form the ventral and dorsal portions of its side walls. Among the fossil
Apiocrinklce we find an intermediate condition between these two extremes. Thus in Ap.
mespiUformis, as seen in Goldfuss's figure 1 , the first radials are small, but the basals are
very large and curved outwards, so as to enclose a large central cavity ; this we may
fairly suppose to have lodged a chambered organ, as Ludwig 2 has found that in Rhizo-
crmus, the modern representative of this family, the axis of the stem expands into a
chambered organ just as described above in Pentacrinus. I have found the same to be
the case in Bathijcrinus.

This organ is contained in H. 1 of of easts in a large and apparently simple segment,
described by Sars 3 as the expanded uppermost stem-segment. Pourtales 4 , however, for
reasons which will be discussed further on, regards it as composite and as representing
the five basals. In this case the relative position of the chambered organ is precisely
identical with that which we may suppose it to have occupied in Ap. mespiUformis,
namely, on the dorsal side of the radial pentagon, but not within the uppermost stem-
segment, as in Comidula. In Ap. rosacens 5 the relative position of the chambered organ
must have been very much as in Pentacrinus, though slightly higher; for the cavity in
which it was contained was almost entirely enclosed between the enlarged first radials,
while the basals only form its floor and the very lowest portion of its side walls. Lastly,
in Ap. Ililleri 6 the chambered organ must have lain altogether on the ventral side of the
basals; the superior surfaces of which form by their apposition the floor of a cavity whose
side walls are entirely composed of the adjacent inner faces of the contiguous first radials.

This condition is thus precisely the opposite of that which we find in Comatula, where
the cavity containing the chambered organ is not only altogether outside the radial
pentagon, but also on the dorsal side of the rosette or metamorphosed basals.

(§ 73) In the works both of Goldfuss and Miller may be found incidental suggestions
that the " basis," or circlet of basals, and, indeed, the whole of the lower part of the calyx
of the stalked Crinoids, may be regarded as representing expanded stem-segments, each
broken up into five parts. Muller, however, was the first to put this idea into a definite
form. He described the basals of Pentacrinus as a metamorphosed stem-segment 7 , or as
" zerfallene Theile eines obersten Stengelgliedes ;" for they correspond in position with
the five leaf-like figures on the articular surfaces of the stem-segments which mark the
positions of the five longitudinal tendons.

The fibrous bundles composing these tendinous cords are separated from one another
by a very regular calcareous network, which is deposited around and between them,
somewhat as in the rays of the basal star of Actinometra. They are attached (by their

1 Petref. Germ., Taf. lvii. fig. 1, n.

' " Zur Anatomie des li. lofotensis," Zeitschr. f. wiss. Zool. Ed. xxix. p. 1:2:2.

3 Crinoides vivants, loc. tit. pp. -4, 12.

4 " On a new Species of Bhizocrinus from Barbadoes,'' Zoological Results of the Hassler Expedition, p. 28. Cam-
bridge, U. S., 1374. 5 Petref. Germ., Taf. lvi. fig. 3, e.

6 Petref. Germ., Taf. lvii. fig. 2, d. 7 " Bau des Pentacrinus" loc. eit. pp. 10, 25.

SECOND SERIES. — ZOOLOGY, VOL. II. 15

108 MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

upper ends) to the lower surfaces of the five basals in the same way as the ligaments of
the arms are attached to the brachial segments, the fibres of the one passing gradually
into the protoplasmic basis of the other. The basals cannot, however, be regarded as
simply ossifications in this fibrous tissue of the same nature as the basal rays of Actlno-
metra ; for, as shown above, there is every reason to believe that they are developed, like
the other elements of the skeleton, out of the embryonic basal plates ; although, so far as
position is concerned, they are precisely homologous with the calcareous deposits within
the tendinous cords of the stem.

On the other hand, the basal rays of Actinometra, which are similar in position, though
not in origin, to the peripheral portions of the basals of Pentacr'mus, are of the same
nature as the calcareous tissue of the leaf-like areas of the stem-segments, being simply
the result of the deposition of calcareous material around and between connective-tissue
fibres.

In many of the fossil Articulate Crinoids the lateral union of the basals with one
another is so very complete that the lines of junction between them are not always
visible, and the "basis" has therefore been described either as entirely absent or as
replaced by the uppermost stem-segment, which, according to Midler's view, it is sup-
posed to represent.

This is particularly the case in the Apiocrlnidce and in Eugeniacr'mus. Miller, who
was the first to describe the latter type l , mistook the first radials of E. caryophyllatus
for the basals, and described them as firmly anchylosed to what he supposed to be the
"superior columnar joint." Goldfuss 2 , however, rightly determined this last to be a
part of the first radials, which are very much prolonged downwards, while, at the same
time, he described the basals as replaced by the enlarged uppermost stem-segment,
which articulates with the inferior surface of the elongated first radials. E-cemer 3 did
not accept this view of Goldfuss's, although he recognized that the " superior columnar
joint " of Miller was simply a dorsal prolongation of the first radials ; but, like Miller,
he described these last as the basals.

It is most probable that Goldfuss's view is the truer one, as in Hagenow's figure 4 of
Eugeniacrinus Eagenoicii, in which the first radials are not prolonged downwards as in
E. caryophyllatus, the piece on which they rest, representing that which Goldfuss called
the enlarged uppermost stem-segment of E. caryophyllatus, is seen to be distinctly com-
posite ; for its external surface is marked by five sutural lines, alternating in position
with those between the first radials, and evidently indicating the lines of union of five
basals.

The "•eoloaical collection of the British Museum, which I have been able to examine,
thanks to the kindness of Mr. Henry Woodward, contains a very interesting series of
specimens from the Chalk which are labelled Apiocrinus elliptlcus.

In some of them the basals form a complete ring, separating the radials from the upper
stem-joint, which is very much enlarged. But in other specimens the basals appear

1 O+h cit. p. 111. - Petref. Germ. torn. cit. p. 102. 3 Lethoea Gcoguostioa, ii. Tbeil 4, p. 115.

4 Min. Jabrb. he. cit. ix. p. 13.

ATE. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 109

externally merely as small triangular pieces, not meeting laterally ; so that they exhibit
the same differences as the basals of Pentacrinus and of Solanocrinus. In many specimens
the sutures between the basals, radials, and top stem-joint are clear and distinct ; but in
others there is no trace of them at all, just as in some examples of JEugeniacrinus and
Bhizoci'inus ; but this is hardly a satisfactory reason for supposing the basals to be
internal and concealed, as has been done in the case of the last-named genus.

Even in some species of Pentacrinus the basals appear to be very closely united to
one another, and to assume the form of an uppermost stem-segment. Thus in P. sca-
laris, Goldfuss 1 , there is no appearance whatever of small wedge-shaped basals, such as
are found in P. briareus and in P. asterla ; but, as remarked by the Messrs. Austin 2 ,
they appear to be united into a single plate, which resembles an " enlarged columnar
joint." The same was probably the case in the Jurassic genus Isocrinus, described by
Von Meyer 3 , though it is, of course, possible that in both these cases the basals may
have been internal and concealed, as in Comatula.

I have endeavoured to show elsewhere 4 that in the recent Phizocrimis we find a
strikingly similar case to that presented by JEugeniacrinus, viz., the sutures between the
basals, visible externally in one species and not in another, or, rather, not invariably in
another. In M. lofotcnsis the first radials rest upon a large and expanded apparently
simple segment, which was described by Sars 5 as the expanded uppermost stem-segment;
and a small circular plate situated in the central vacuity between the first radials, with
which, as well as with the enlarged uppermost stem-segment, it is closely connected, was
regarded by him as representing the metamorphosed embryonic basals of Comatula.

Pourtales's observations 6 , as well as my own subsequent ones, have led me to believe
that the piece called the enlarged uppermost stem-joint of P. loftensis by Sars and
Ludwig 7 is composed (if not entirely, at any rate in great part) of five closely anchylosed
basals. Schluter s is evidently not acquainted with the evidence on which this view rests,
or he would scarcely suggest that B. Rawsonii might not be a Rhizocrinus at all, because
its basals differ from those described in P. loftensis by Sars and Ludwig ; although
these two observers are not themselves in accordance as to which parts of the interior
of the calyx are to be regarded as concealed basals.

Sir Wyville Thomson 9 takes the same view as Pourtales and myself ; for he describes
how " in Rhizocrinus the funnel-shaped piece formed by the coalescence of the basals with
the fused first radials above and the dilated upper joint of the coalesced upper joints of
the stem beneath, makes up a large part of the cup ; " and his descriptions of the
calices of Hyocrinus and Bathycrinus, both genera allied to Rhizocrinus, together with

1 Petref. Germ. torn. cit. p. 173, Taf. Is. fig. 10, b.

2 'A Monograph on Kecent and Fossil Crinoidea,' p. 121 (Bristol, 1845).

3 " Isocrinus imd Chelocrinus," Museum Senkenbcrgianum, p. 251 (Frankfurt, 1837).

4 " Pentacrinus and Rhizocrinus," loe. cit. pp. 47-53. 5 Crinoides vivants, loc. cit. p. 4.

6 Hassler Expedition, Joe. cit. pp. 28, 29.

7 Rhizocrinus lofotensis, loc. cit. pp. 121, 122.

8 0/(. cit. p. 29. Schluter was unfortunately unable to make himself acquainted with Pourtales's memoir.

9 " Notice of new living Crinoids belonging to the Apiocrinidse," Journ. Linn. Soe. Zool. vol. xiii. p. 48.

15*

110 ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

the analogies of Apiocrinus and Eugenia crinus, strengthen Pourtales's view still more
completely.

The occasional fusion of the upper stem-joints with the lower part of the calyx, as
described ahove in the Apiocrinklce, is an excellent illustration of Midler's idea respecting
the correspondence between the basis and the stem-joints.

This correspondence, however, is by no means entitled to rank as a serial homology.
The earliest condition of the basals shows them to be five separate plates developed in a
spiral around the aboral ccelom of the Crinoid embryo 1 . They have distinct homologies
in the apical system of the other Echinoderms 2 ; while the stem-segments, surrounding
the aboral coelom much in the same way as the basal circlet, arc simple undivided pieces
from the first, and seem to be almost or quite unrepresented in the other Echinoderms.

(vi.) The Second and Third Radlals.

(§ 74) The second radial oi Act. poly morpha (PI. VII. fig. 2), like that of Ant. rosacea,
is an oval, somewhat discoidal plate, having two nearly parallel faces — one internal or
proximal, articulating with the first radial, the other external or distal, articulating with
the third radial. The internal face (fig. 2 a) closely resembles the external face of the
first radial (PI. VII. fig. 1 b), with which it articulates, being divided transversely by a
large articular ridge (*) into a dorsal and a ventral portion ; the former is entirely occupied
by the fossa lodging the clastic ligament (/), which is particularly deep just below the
opening of the central canal (c.c). Prom the ventral margin of this opening arise
the two ridges which bound the intermuscular furrow (f x ), and are joined near their
upper extremities by the transverse secondary ridges separating the large fossae (h)
that lodge the interarticular ligaments from those (f) lodging the flexor muscles of the
ray ; the latter are excavated in a pair of thin lamelke, which extend upwards from
the proper ventral margin of the plate, as is seen in a view of the distal face (fig. 2 b,
g). Besides the above-mentioned ridges and fossa?, which correspond to similar ones
on the distal face of the first radial, the proximal face of the second radial shows two
lateral processes, in which shallow fossa? (k) are excavated. These processes represent the
outer portions of the distal face, which is somewhat wider than the proximal one, as the
lateral faces are not set at right angles to the two terminal ones, but form an oblique
angle with the proximal face, so that the outline of the radial, when seen from the dorsal
or ventral side, is trapezoidal in form (fig. 2 c, d). The shallow fossae which are exca-
vated in these lateral faces lodge the ligamentous substance by which the second radials
are united with one another in pairs : the extent of this union is, as above remarked,
very variable in different specimens, being generally greatest where the number of arms
is largest (PL II. figs. 9, 11).

The external or distal face (fig. 2 b) is much simpler in character than the proximal
one, as no muscles are attached to the vertical lamella? which rise from its ventral
margin above the articular face proper. This last is divided by a vertical ridge (*) that

' Gotte, he. at. pp. 595, G20.

2 P. H. Carpenter, " On the Ural and Apical Systems of the Echinoderms." Quart. Journ. Alicr. Sci. xviii. (1878)
pp. 371, 382.

ME. P. H. CAEPEXTEE OX THE GENUS ACTING3IETEA. Ill

passes round the opening of the central canal (c.c) into a pair of lateral fossae (It), which
give attachment to the large interarticular ligament connecting the second with the third
radial. The proximal face is not quite vertical, but slightly inclined towards the distal
one, so that the ventral face is not much more than an edge. When the piece is viewed
from the ventral side, therefore (fig. 2 c), little else is visible but the fossa? for the
muscles (f) and interarticular ligaments (//) of the proximal face and the intermuscular
furrow (f) descending along its median line.

The second radials of var. 1 are very similar to those of the type, except that, as in
the first radials, the muscular fossae are relatively somewhat larger. In var. 2, however,
they are very much smaller (PL VII. fig. 5 a,/); and there are no vertical lamella; pro-
jecting from the ventral margin of the distal face (fig. 5 b), as is the case in the type.
The lateral fossae (k) lodging the ligamentous substance which connects the second radials
with one another are somewhat more marked, as the union of the second radials in pairs
is more complete than in the type, though not so complete as in varieties 1, 3, and 4.

The two latter also agree with var. 2 in the fact that the proximal and distal faces of
the second radials are nearly parallel, and less inclined to one another than in the type
and in var. 1 ; so that the fossae for the muscles and interarticular ligaments are barely
visible when the piece is seen from the ventral side (fig. 5 c), as there is a proper ventral
face. Its median line is occupied by a continuation of the furrow on the ventral surface
of the first radial (figs. 4 c, 5 c, v.r.f), while its lateral portions are divided up, in the
same way as those of the first radial, into secondary ridges and furrows.

(§ 75) The third or axillary radial of Act. polymorpha, which gives attachment to
two primary arms, presents three articular surfaces — an internal one corresponding to the
distal face of the second radial, and two external ones, inclined to one another, with which
the bases of the arms articulate.

The proximal face (PI. VII. fig. 3 a) is precisely similar in character to the distal face
of the second radial, being divided, like it, by a vertical ridge into two lateral fossae (//)
which lodge the interarticular ligaments. Its articular margin, when viewed from the dorsal
side (fig. 3 d), is perfectly straight, and does not project in the middle as in Ant. rosacea ;
so that the possible amount of lateral movement between the second and third radials
must be extremely slight. Two vertical lamellae (g) project from the uppermost margin
of the internal face ; but they do not form part of the surface of articulation with the
second radial, as they are excavated into fossae on their outer side for the attachment
of the proximal ends of the outer muscular bundles passing between the axillary radial
and the lowest segments of the primary arms. The two inner muscular bundles are
attached to the two sides of a projecting wedge-shaped process (cl) on the external or
distal face, the " clavicular" of Schultze 1 , which occupies the angle between the two ob-
liquely placed articular faces for the basal arm-segments. These are of precisely the
same character as the external faces of the first radials (fig. 3 b), consisting, besides
the muscular fossae (/) just mentioned, of two others for the interarticular ligaments (//),
and of a large dorsal fossa (j) lodging the elastic ligament, and separated from the other
two by a transverse articular ridge (i) , in the centre of which is the opening of the central
canal (c.c).

1 Loc. cit. p. 5.

112 ME. P. H. CARPENTER ON THE GENUS ACTINOMETEA.

The median line of the ventral face (fig. 3 c) is occupied hy a ventral radial furrow
continuous with that on the ventral face of the first radial ; it divides iuto two branches,
one of which passes on either side of the clavicular, in order to be continued on to the
basal arm-segments. The proximal end of this furrow is indicated by the deep notch
separating the two vertical lamella3 which project upwards from the proper internal
face (fig. 3 a, g), and through which the base of the clavicular is seen.

In the type and in var. 1 the lateral portions of the ventral face of the third radial
are plain, and not sculptured ; but in varieties 2-1 they are divided up by secondary
ridges and furrows (fig. 6 c), just like the ventral faces of the first and second radials
(figs. 1 c, 5 c). In these varieties also there are no vertical lamellae projecting upwards
from the ventral margin of the internal face (fig. 6 a), which is also the case in the ex-
ternal face of the second radial (PI. VII. fig. 5 b), as the muscular bundles passing between
the first and second radials, and between the third radials and basal arm-segments, are
smaller than in the type. In var. 2 there would appear to be more power of lateral
movement between the second and third radials than is the case in the type ; for
although, as in the type, there is no projection in the middle of the proximal articular
margin of the third radial, yet the distal articular margin of the second radial shows
a slight indication of such a median prominence (fig. 5 d), which is absent in the type.
It would seem though, to be replaced to a certain extent by the greater thickness of
the vertical articular ridge (*) around the opening of the central canal, which is seen, in
PI. VII. fig. 2 d, to project a little beyond the level of the dorsal surface of the radial ;
so that when the opposed ridges of the second and third radials are in contact with
each other, the third may possibly have a very slight power of lateral movement upon
the second, though by no means so great as in Ant. rosacea, in which the median pro-
minence on the internal articular margin of the third radial is very marked.

The second and third radials of var. 2 differ from those of the type of Act. polymorpha
and of all the other varieties in the very marked convexity of their dorsal surfaces, which
renders them considerably higher than the first radials ; so that when the whole calyx is
viewed from the exterior, the inner circle of first radials, which are only very little con-
cealed by the small centrodorsal piece, seems somewhat sunk within the outer circle
formed by the second and third radials.

This marked convexity is well seen in PI. VII. fig. 5, a, d, and fig. 6, b, d, especially
when these figures are compared with those of the corresponding parts in the type
(fig. 2, a, d, fig. 3, b, d).

List of Works referred to.

1. Adams, J., " Description of some Marine Animals found on the coast of Wales," Trans. Linn. Soc.
vol. v. p. 7 (London, 1800).

2. Agassiz, L., " Prodrome d'une Monographie des Radiaires ou Echinoderines," Annales d. Sci. Nat.
2 mc se'r. Zool. vii. p. 288.

3. Agassiz, AL, " Revision of the Echini," Illust. Cat. Mus. oC Comp. Zool., No. vii. (Cambridge, U.S.
1872-71).

4. Austin, T. and T., Jun., 'A Monograph on Recent and Fossil Crinoidca' (Bristol, 1815).

ME. P, H. CARPENTER ON TIRE GENUS ACTINOMETRA. 113

5. Barrelieri, Jac, 'Plantrc per Galliarn, Hispauiam et Italiam observatae ' (Paris, 1714).

G. Beyrieh, E., " Ueber die Crinoideen des Muschelkalks," Abliaudl. d. k. Akad. zu Berlin, 1857.

7. de Blainville, H. M. D., 'Manuel d'Actinologie ' (Paris, 1834).

8. Bohlschc, AY., "Ueber Actinometra Bennettii und eine neue Coniatula-Art [Antedon Dubenii),"
Wiegm. Arcbiv fur Naturgesch. 1866, i. p. 92.

9. Broun, H. G., ' Die Klasseu und Ordnungen des Thierreichs,' Band ii. Aktinozoen (Leipzig,
1860) .

10. Carpenter, P. Herbert, "Remarks on the Anatomy of the Arms of the Crinoids, Part I.," Journ.
of Anat. aud Physiol, vol. x. April, 1876, pp. 571-585.

11. Carpenter, P. Herbert, The same, Part II., Journ. of Auat. and Physiol, vol. xi. Oct, 1876,
pp. 87-95.

12. Carpenter, W. B., " Researches ou the Structure, Physiology, and Development of Antedon rosa-
ceus, Part I.," Phil. Trans, vol. 156, pp. 671-756, jus. xxxi.-xlii.

13. Carpenter, W. B., "On the Structure, Physiology, and Development of Antedon rosaceus," Proc.
Roy. Soc, No. 160, 1876, pp. 211-231, pis. 8, 9.

14. Carpenter, W.B., " Supplemental Note to the above Paper," P. R. S., No. 169, 1876.

15. Columna, Fabius, ' Phytobasanus, sive Plantarurn aliquot Historia ' (Neapoli, 1592).

16. Dujardin and Dupe, ' Histoire Naturelle des Zoophytes Echinodermes' (Paris, 1862).

17. 'Encyclopedic Methodique,' " Histoire Naturelle des Vers, des Mollusques, des Coquillages et
Zoophytes" (Paris, 1789-1832), Partie des Vers.

18. De Freminvillc, "Memoire sur un uouveau Genre de Zoophytes de TOrdre des Radiaires," Nouv.
Bull. d. Sci. par la Soc. Philom. torn. ii. p. 349 (Paris, 1811).

19. Gegenbaur, C, ' Grundriss der verglcich. Anat/ (Leipzig, 1874).

20. Goldfuss, G. A., 'Petrefacta Germanise' (Dusseldorf, 1826-35).

21. Goldfuss, G. A., "Beitr. zur Petrefactenkunde," N. Acta Acad. Leop.-Carol. Nat.-Cur. xix. a.
1839, p. 318.

22. Goldfuss, G. A., "Beitr. zur Petrefactenkunde," Neues Jahrb. f. Mineral. (1841), p. 86.

23. Gotte, Al., " Yergleich. Entwickelungsgesch. der Comatula mediterranea," Archiv f. mikros. Anat.
Band xii. 1876, pp. 58 3-648, Taf. xxv.-xxviii.

24. Gray, J. E., " Notice on the Digestive Organs of the Genus Comatula, and on the Criiwidea of
Miller," Annals of Philosophy, N. S. vol. xii. p. 592 (1826).

25. Greeff, R., " Ueber den Bau der Echinodcrmen, 3. Mittheilung," Sitzungsb. d. Gesellsch. z. Be-
ford. der gesam. Naturwiss. zu Marburg, No. 11, 1872, pp. 155-169.

26. Greeff, R., The same, 4. Mitthciluug, " Ueber den Bau der Crinoideen," Marburg. Sitzungsb.
No. 1, 1876, pp. 16-29.

27. Greeff, R., The same, 5. Mittheilung, " Ueber das Herz der Crinoideen," Marburg. Sitzungsb.
No. 5, 1876, pp. 83-95.

28. Yon Hagenow, Fr., "Monographic der Riigen'schen Kreide-Versteinerungen, II. Abtheilung,
Radiarien und Annulaten," Neues Jahrb. fur Mineralogie, 1840, p. 664.

29. Heusinger, C. P., " Bemerkungen iiber den Verdauungskanal der Comatulen," Meckel's Archiv
f. Physiol. 1S26, p. 317.

30. Heusinger, C. P., " Anatomischc Untersuchungen der Comatula mediterranea," Zeitsch. f. organ.
Physik, Band iii. 1833, pp. 367-374, Taf. x., xi.

31. De Koninck, L., and H. Le Hon, ' Recherches sur les Crinoi'des du Terrain Carbonif ere de la Bel-
gique ' (Bruxclles, 1854).

32. Lamarck, J., ' Systeme d'Animaux saus Vertebrcs/ 2 mo edit., Paris, 1816, torn. ii. p. 532.

33. Lange, W., " Beitr. z. Anat. und Histiologie der Asterien und Ophiuren," Morphol. Jahrb. Band
ii., 1876, pp- 211-286, Taf. xv.-xviii.

114 ME. P. H. CAEPENTEE ON THE GENES ACTINOMETRA.

34. Leach, W. E., ' Zoological Miscellanies/ vol. ii. p. 61 (London, 1815).

35. Leuckart, F. S., " Einiges iiber das Asteroiden-Geschlccht Comatula, Lam., iiberlianpt, und liber
C. mediterranea iusbesondere," Zeitsch. f. organ. Physik, Band iii. 1833, p. 385.

36. Linckii, Joliannis Henrici, Lipsiensis, ' De Stcllis Marinis liber singularis ' (Lipsiae, 1733).

37. Linnaeus, K., ' Systema Naturae/ editio decima tertia, pars vi. p. 3166 (Lipsiae, 1788).

38. Loven, S., " Phanogcnia, ett hittills okiindt slagte af fria Crinoideer," Ofvers. af Kongl. Ve-
tensk.-Akad. Forhandl. 1866, No. 9, pp. 223-233.

39. Ludwig, Hubert, "Beitr. zur Anat. der Crinoideen," Nachrichten von der Konigl. Gesellscb.
der Wisscnscb. und der G. A. Universit'at zu Gottingen, No. 5, 1876, pp. 105-114.

40. Ludwig, Hubert, ' Morpliologiscbe Studien an Echinodermen, I. Beitr. zur Anat. der Crinoideen'
(Leipzig, 1877) ; Separat-Abdruck aus der Zeitsch. f. wisscnscb. Zool. Band xxviii.

41. Ludwig, Hubert, "Zur Anatomie des Rhizocrinus lofotensis, M. Sars," Zeitschr. f. wiss. Zool.
Band xxix. pp. 101-130, Taf. v., vi.

42. Luidi, Eduardi, ' Lithophylacii Britanniei Iclmograpliia ' (Londini, 1699).

43. Luidi, Eduardi, ' Praelectio de Stcllis marinis Oceani Britanniei, nee non de Asteriarum, Entro-
chorum, et Encrinorum origine' (Oxford, 1703).

1 1. Meckel, J. F., " Ucbcr die Oeffnungen des Speisekanals bci den Comatulen," Meckel's Archiv f.
Physiol. Band iii. 1823, p. 470.

45. Metschnikoff, E., "Beitr. zur Entwickelungsgesch. einiger niederen Thiere/ J Bull, de l'Acad. Imp.
des Sci. de St. Petersb. torn. xv. 1871, pp. 502-509.

46. Meyer, C. E. H., " Isocrinus und Ckelocrinus" Museum Scnkenbergianum (Frankfurt, 1837).

47. Miller, J. S., < A Natural History of the Crinoidea' (Bristol, 1821).

48. Miiller, J., " Ueber den Bau des Pentacrinus caput-Medusa," Abhandl. d. k. Akad. zu Berlin,
1843; Abst. inWiegm. Archiv f. Naturgesch. 1840, i.

49. Miiller, J., "Ueber die Gattungen und Arten der Comatulen," TViegm. Archiv f. Naturgesch. 1841, i.

50. Miiller, J., " Neue Beitr. zur Kenntniss der Arten der Comatulen/ 5 Wiegm. Archiv f. Naturgesch.

1843, i.

51. Miiller, J., " Nachtrag zu der Abhandlung iiber die Comatulen," Monatsb. d. Berlin. Akad. 1846,

p. 177.

52. Miiller, J., " Ueber die ^Gattung Comatula, Lam., und ihre Arten," Abhandl. d. k. Akad. z.

Berlin, 1849.

53. Miiller, J., " Ueber den Bau der Echinodermen," Abhandl. d. k. Akad. z. Berlin, 1853.

54. Norman, A. M., " On the Genera and Species of the British Echinodermata," Annals and Mag.
of Nat. Hist. ser. iii. vol. xv. p. 98.

55. d'Orbigny, A. D., ' Cours elementaire de Paleontologie et de Geologic stratigraphiquc ' (Paris,

1850-52).

56. Petiveri, .Tacobi, ' Gazophylacium Naturae et Artis' (Londini, 1711).

57. Petiveri, Jacobi, ' Aquatilium animalium Amboinensium Icones et Nomina' (Londini, 1713).

58. Philippi, R. A., " Alecto alticeps, n.^sp., eine [tertiare Comatula-Art von Palermo," Neues Jahrb.
f . Mineral. 18-14, p. 540.

59. Pictct, F. G., ' Traite de Paleontologie ' (Paris, 1857) .

60. de Pourtales L. F., " Contributions to the Fauna of the Gulf Stream at Great Depths," Bull,
of the Mus. of Comp. Zool. vol. i. no. 6.

61 de Pourtales L. F. " List of the Crinoids obtained on the coasts of Florida and Cuba by the
United States Coast Survey Gulf-Stream Expeditions, in 1867, 1868, 1869," Bull, of the Mus. of Comp.
Zool. vol. i. no. 11, pp. 355-358.

62 de Pourtales L. F., " On a new Species of Rhizocrinus from Barbadoes/' Zool. Results of the
Hassler Expedition, Must. Cat. of the Mus. of Comp. Zool. no. viii. pp. 27-31, pi. v.

ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 115

63. Retzius, A. J., ' Nova Acta/ Stockholm, 1783, no. 12, p. 234.

04. Eetzius, A. J., ' Dissertatio sistcns species cognitas Asteriarum' (Lundae, 1805).

65. Eceraer, F., In ' Lethcea Geognostica/ drittc Auflage, 1851, Theile iv. v.

66. Rosinus, M. R,, ' Tentaminis de Lithozois ac Lithophytis olim marinis, jam vcro subterraneis,
prodromus; sivc dc stellis marinis quondam, nunc fossilibus, disquisitio' (Hamburg, 1719).

67. Sars, M., ' Memoires pour servir a la connaissance des CriuoVdcs vivauts ' (Christiania, 1868).

68. von Scldottlieim, S., ' Nachtrage zur Petrefactenkunde,' Abtli. ii. (Gotha, 1823), p. 48.

69. Schultze, L., ' Monog. dcr Ecliinodcrmen des Eiflerkalkes ' (Wien, 1866). Besonders Abgedriickt
aus dem xxvi. Bande der Denkschr. der mathemat.-natu.rwissensc.haft. Classe d. kaiserl. Akad. d.
Wissensch.

70. Schweigger, A. P.,' Beobacht. auf naturhistorischen Reisen ' (Berlin, 1819).

71. Semper, C, " Kurze anatomische Bemerkungen iiber Comatula," Arbeiten aus dem zoolog.-
zootom. Institut zu TVurzburg, Band i. 1874, pp. 259-263.

72. Simroth, H., " Anatomic und Schizogonie der Ophiactis wrens, Sars, Theil I." Zeitsch. f. wiss.
Zool. Band xxvii. Heft 4.

73. Teuscher, R., " Beitr. zur Anatomie der Echinodermen, I. Comatula mediterranea" Jenaische
Zeitsch. f. Naturvrissensch., Band x. 1876, pp. 243-262, Taf. vii.

74. Teuscher, R., " Beitr. zur Anat. der Echinodermen, II. Ophiuridse," Jenais. Zeitsch. Band x. 1876,
pp. 263-280, Taf. viii.

75. Teuscher, R., "Beitr. zur Anat. der Echinodermen, III. Asteridre," Jenais. Zeitsch. Band x. 1876,
pp. 493-514, Taf. xviii., xix.

76. Thomson, Wyv., " On the Embryogeny of Antedon rosaceus, Linck {Comatula rosacea of
Lamarck)," Phil. Trans, vol. 155, 1865, pp. 513-544, pis. xxiii.-xxvii.

Addenda.

77. Baudelot, E., " Etudes generates sur le systeme ncrveux. Contribution a l'histoire du systeme
nerveux des Echinodermes," Arch, de Zool. experim. et g^nerale, tomei. pp. 177-210.

78. Carpenter, P. Herbert, "On some points in the Anatomy of Pentacrinus and Rhizocrinus," Joum.
of Anat. and Physiol, vol. xii. Oct. 1877, pp. 35-53.

79. Carpenter, P. Herbert, "On the Oral and Apical Systems of the Echinoderms, Part i.," Quart.
Journ. Micr. Sci. vol. xviii. new series, Oct. 1878, pp. 351-383.

80. Gegenbaur, Carl, ( Grundriss der vergleichenden Anatomic,' zweite Auflage (Leipzig, 1878).

81. Grube, Descriptions of three new Comatula (C. Icevissima, C. Mertensi, Act. borneensis) in Jah-
Iresber. d. schles. Gesellsch. 1875, Nat, Hist. Sect. pp. 54, 55*.

82. Ludwig, H., " Bcitrage zur Anatomie der Asteridcn," Zeitschr. f. wissensch. Zool. Band xxx.
pp. 150-212, Taf. v.-vii.

83. Schluter, Clemens, " Ueber cinige astylidc Crinoiden," Zeitschr. d. dcutsch. gcol. Gesellsch. Jahrg.
j 1878, pp. 28-66, Taf. i.-iv.

84. Thomson, Sir C. Wyvillc, "Notice of new living Crinoids belonging to the Apiocrinidae," Journ.
Linn. Soc, Zool. vol. xiii. pp. 47-55.

* I have, unfortunately, been unable to get a .sight of this paper, and only know of it from the reference to it in
Leuckart's ' Jahresbericht.'

SECOND SERIES. — ZOOLOGY, VOL. II.

10

116 MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

DESCRIPTION OF THE PLATES.

The following letters denote the same parts throughout all the Plates.

A, B, C, D, E. The five Radii or Amhulacra.

A u A 2 , B u B„, C„ Co, D u Z> 2 , E v E t . The ten Primary Arms.

a.c. Axial ccelom. a.p. Axial prolongation.

a.i.c. Axial interradial canal. a.r.c. Axial radial canal.

a.i.f. Axial interradial furrow. a.r.f. Axial radial furrow.

An. Anal tube.

b u b 2 , b 3 , b e , U) . First, second, third, sixth, and tenth brachials. h.h. Basal bridge.

b.f. Basal fold. b.g. Basal groove.

b.m v Muscles between the radial axillary and the first brachials.

b.ni . Muscles between the second and third brachials.

c.c. Central canal of the calcareous segments of the rays and arms.

c.c'. Central canal of the eirrhus-segments. ch. Chambers of the quinquelocular organ.

rd. Centrodorsal piece. dr. Cirrhus.

cd.r. Centrodorsal coelom. el. Clavicular.

c.n. Calcareous network in the central vacuity of the pentagonal base of the calyx.

co.c. Commissural canals in the first radials. d.i.f. Dorsal interradial furrow.

cv.c. Circumvisceral coelom. d.r.f. Dorsal radial furrow.

d l} d.„ d.a. First, second, and axillary distichals. ep. Epithelial wall of the alimentary canal.

F. Central funnel-shaped space enclosed within the pentagonal base.

/. Muscular fossae. <]■ Vertical lamellae of the calcareous segments.

/j. Intermuscular furrow. g x . Their superior margins.

/ 2 . Notch representing it in Ant. celtica. g r Their inner lateral margins.

g s . Ridge formed by the union of these in the first radials of Ant. celtica.

h. Fossae lodging the intcrarticular ligaments.

i. Transverse articular ridge. i.eo. Interradial commissure.

i.e. Interradial elevations on the centrodorsal piece.

ir.c. Intcrvisceral coelom. j. Fossre lodging the clastic ligaments.

k. Fossae lodging the ligamentous substance between the sides of the second radials.

L. Ligamentous substance between the sides of the first radials.

/. Ligamentous substance between the first radials and the centrodorsal piece.

L'. Incompletely decalcified portions of the skeleton.

L r Interarticular and "1 ligaments between the first and second radials, or second and third

/,. Elastic J brachials.

L-,. Interarticular ligaments between the second and third radials. M. Mouth.

N. Fibrillar nervous envelope of the quinquelocular organ.

n. Axial nervous cords of the rays and arms. o. Interradial spout-like processes of the rosette.

n Their branches. o'. Interradial triangular processes.

v. r. Axial nervous cords of the cirrhi. o.b. Ovoid bodies. P. Peristome.

Pe. Uncalcified pcrisome between the radii. p. Radial processes of the rosette.

p. Small curved processes at the sides of a single basal.

px,p v p.a. First, second, and axillary palmars.

Q. Radial openings on the dorsal surface of the pentagonal base.

ME. P. H. CARPENTER OX THE GENUS ACTLXOMETEA. 11

Q'. Notches on the inner margins of the dorsal faces of the first radials.

q. Radial depressions on the centrodorsal piece. R. Rosette.

''ij r si r - a - First, second, and axillary radials. r.ar. Radial areas on the centrodorsal piece.

r.c. Diverticulum from the radial ccelom into the substance of the first radial.

r.co. Intraradial commissure. r.o. Central opening of the rosette.

r.m. Muscles between the first and second radials. r.s. Radial space. S. Rays of the basal star.

S 1} S& S 3 . Diverging, vertical, and lateral fibres effecting the interradial portions of the synostosis

between the centrodorsal piece and the radial pentagon.
s. Depressions at the central ends of the rays of the basal star.
tpi, sp.a. First and axillary suprapalmars. Sy. Syzygium.

t. Short processes at the angles of the centrodorsal piece.
U. Sockets for the attachment of the dorsal cirrhi.
u. Inner openings of the cirrhus-canals in the centrodorsal piece.
v.i.f. Ventral interradial furrow. v.r.f. Ventral radial furrow.

V, W, X, Y, Z. The five primary basal cords proceeding from the angles of the quinqueloeular organ.
Vi, V 2 , W x , W v X u X 2 , Y x , Y 2 , Z u Z v The ten secondary basal cords produced by the bifurcation of

the primary ones.
i\, r 2 , ?i'„ W 3 , x v x 2 , y v y v g lt Z v The apertures in the basals through which the secondary cords pass.
v, v', iv, 10, x, at, y, y', z, z . The corresponding apertures of the central canals on the internal faces

of the first radials.

In PI. II. figs. 9-11 indicate the position of the mouth (ventral) relative to the radial skeleton

(dorsal) .
* In PL V. fig. 4 indicates the passage of the ventral radial canal into the central calcareous network

within the radial pentagon by two openings, instead of by only one as usual.

Plate I.

Diagrams of the distribution of the ambulacra on the disks of different species of Comatula. The red

lines mark the interradial intervals. Figs. 1-4 copied from Muller.
Fig. 1. Antedon rosacea. Fig. 3. Act. Wahlberghii.

2. Actinometra solans. 4- Act. multiradiata.

In figs. 5-16 the tentaculiferous grooves are marked by dark lines, and the non-tentaeuliferous
grooves by fainter lines.

Fig. 5. Act. Solaris. Proportion of non-tentaculiferous arms, ^
6. Act. polymorpha : Type. » i's

7. „ „ » 18

„ < ;

8. „ „ » 20

9. „ Var. 1. „ H

10. „ Type. „ 25

11. „ „ » ^6

12. „ „ » 28

13. ,, „ » 2S

14. „ Var. 2

15. „ Type.
1G. Var. 4. All the arms tentaculiferous

2 a

15. „ Type. » ;i

16*

118 ME. P. H. CAEPENTEE OX THE GENUS ACTINOMETEA.

Plate II.

Fig. 1. Diagram of the distribution of the ambulacra in a new Actlnometra from the Philippines.

2. Superior or ventral aspect of the disk of Act. polymorpha, var. 2, the oral pinnules having been

cut away near their bases, x 3.

3. Piece of an ordinary tentaculiferous arm of Act. polymorpha, from about the middle of its

length, seen from above, x 4.

4. Terminal portion of the same arm. x 4.

3. Piece of the; middle portion of a non-tentaculiferous arm, borne upon the same axillary as that

represented in figs. 3 & 4. Ventral view, x 4.
G. Termination of the same arm. x 4.

7. Distichal, palmar, and lower brachial segments of one of the radii of Act. polymorpha, var. 4,

showing the white line which occupies the middle of the dorsal surface of the skeleton, x 4,

8. CeDtrodorsal piece and one radius of a monstrous specimen of Act. polymorpha, showing the very

irregular form of the centrodorsal piece [cd] and the imperfect condition of one of the
palmar series, which consists simply of one axillary segment bearing brachials upon one of
its distal faces, and two suprapalmars [sp u sp.a) upon the other, x 4.

9. Diagram of the calyx of a small thirteen-armed specimen of Act. polymorpha : Type, x 4.

10. A similar diagram of a larger specimen with 2G arms. x 3.

11. A similar diagram of a specimen of var. 3, with 39 arms. x 2.

The * in these three figures ('.), 10, 11) indicates the position of the mouth on the ventral side of the
disk relatively to the radial skeleton of the dorsal side.

Plate III.

Fig. 1. Terminal comb of an oral pinnule of Act. pectinata. x 20.

2. Oral pinnule of Act. polymorpha, Type, x 10.

3. Oral pinnule of Act. polymorpha, Var. 4. x 10.

4. Portion of a horizontal section through the synostosis of two first radials of Pentacrinus Wyville-

Thomsoni. x 110.

5. Lower portion of a vertical section through the peripheral end of the synostosis of two first radials

of Act. polymorpha. x 110.

6. Lower portion of a similar section, taken rather nearer the centre of the radial pentagon, showing

the disposition of the fibres which effect the synostosis of the first radials with the centrodorsal
piece, both in the radial (/) and the interradial planes [S it S 2 , S 3 ). x 110.

7. Longitudinal section through one of the muscles (b.m._) and interarticular ligaments (L^ con-

necting the second and third brachials, x 110.
8-10. Cirrhi of Act. polymorpha. Type, and vara. 1 & 2.

8. Type, x G. a, adult; b, very young; c, nearly mature.

9. Isolated cirrhus-segments of the Type, x 40. a, a basal segment ; It, a terminal segment.

10. Cirrhus of var. 2. X 6.

11. Cirrhus of Tar. 1. x 6.

Plate IV.

Figs. 1-8 of Ant. celtica. All x 7.
Figs. 1 & 2. Centrodorsal piece, after removal of the cirrhi, as seen from its dorsal (fig. 1) and ventral
(fig. 2) sides.
3 & 4. Pentagonal bass of the calys, as seen from its dorsal (fig. 3) and ventral (fig. 4) sides.

ME. P. H. CAEPENTER ON THE GENUS ACTINOMETEA. 119

Fig. 5. Interior of the calyx, as seen after removal of the visceral mass.

6. Lateral view of the base of the calyx with the centrodorsal piece in situ.

7. Pentagonal base of the calyx of a smaller variety, as seen from its dorsal side.

8. Lateral view of the base of the calyx of the same variety, with the centrodorsal piece in situ.

Figs. 9-11 of Ant. Eschrichtii.

Fig. 9. Two united first radials, together with the portion of the rosette which is in connexion with
them, as seen from within, x 7.

10. Pentagonal base of the calyx as seen from its dorsal side after removal of the rosette occu-

pying its central cavity, x ■>\.

11. Centrodorsal piece seen from its ventral side. x 3£.

Figs. 12-17 of Ant. rosacea, all x 7, except fig. 13, which is x 15.

Fig. 12. Isolated first radial, a. ventral, /;. dorsal, c. internal aspect.

13. Abnormally developed rosette, with two spout-like interradial processes (o) and a basal bridge
(b.b.) connecting the ends of two of the radial processes (p). x 15. a. ventral, b. dorsal
aspect.
1 1. Lateral view of the base of the calyx with the centrodorsal piece in situ.
15. Centrodorsal piece seen from its venti'al side.
1G & 17. Pentagonal base of the calyx as seen from its dorsal (fig. 16) and ventral (fig. 17) sides.

Plate V.

The figures all x 7, except fig. S, which is x 15.

Figs. 1-4 of Act. Solaris.

Figs. 1 & 2. Centrodorsal piece as seen from its dorsal (fig. 1) and ventral (fig. 2) sides.

3 & 4. Pentagonal base of the calyx as seen from its dorsal (fig. 3) and ventral (fig. 4) sides.

Figs. 5-9 of Act. pectinata.

Fig. 5. Interior of the calyx as seen after removal of the visceral mass.

6 & 7. Centrodorsal piece as seen from its dorsal (fig. G) and ventral (fig. 7) sides.

8. An isolated compound basal, x 15. a. ventral, b. dorsal aspect.

9. An isolated first radial, a. ventral, b. dorsal, c. internal aspect.

Figs. 10-15 of Act. robusta.

Fig. 10. Internal aspect of an isolated first radial.

11-13. Two united first radials, together with those portions of the rosette which arc in con-
nexion with them, as seen from above (tig. 11), below (fig. 12), and within (fig. 13).
In these four figures (10-13) I. indicates a bristle passed along the axial radial canal; II. another
passed along the axial interradial canal; and III. a third, entering the central canal by one of the aper-
tures on the internal face (z 1 ), and coming out through the aperture of the commissural canal (co.c.) on
the lateral face.
Figs. 14 & 15. Centrodorsal piece as seen from its dorsal (fig. 15) and ventral (fig. 14) sides.

Plate VI.
All these figures are x 7, except figs. 0, 18, 19, 22, which are all x 15.
Figs. 1-11 of Act. polymorpha, Type. Figs. 1-6, from one specimen.
Fig. 1. Lateral view of the base of the calyx, with the centrodorsal piece in situ.
2. The same parts seen from the dorsal side.

120 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETKA.

Fig. 3. Ventral aspect of the centre-dorsal piece.

4 & 5. Pentagonal base of the calyx as seen from its dorsal (fig. 4) and ventral (fig. 5) sides.
C. Two united compound basals as seen from their ventral side. x 15.
7, 8, & 9, from a second specimen.

7 & 8. Centrodorsal piece as seen from its dorsal (fig. 7) and ventral (fig. 8) sides. In fig. 8
three of the rays of the basal star are seen occupying the basal grooves (b. y), their proper
connexion with the rosette having been broken.

9. Pentagonal base of the calyx as seen from its dorsal side after removal of the rosette and

basal star.
10 & 11, from a third and abnormally developed specimen.

10. Centrodorsal piece seen from its ventral side.

11. Dorsal aspect of the pentagonal base of the calyx.

Figs. 12-15 of Act. polymorpha, var. 1.

Figs. 12 & 13. Pentagonal base of the calyx as seen from its ventral (fig. 12) and dorsal (fig. 13)
sides.
14 & 15. Centrodorsal piece as seen from its dorsal (fig. 14) and ventral (fig. 15) sides.

Figs. 16-19 of Act. polymorpha, var. 2.

Figs. 16 & 17. Centrodorsal piece as seen from its dorsal (fig. 16) and ventral (fig. 17) sides.

18 & 19. Two united compound basals as seen from their ventral (fig. 18) and dorsal (fig. 19)
sides.

Figs. 20-22 of Act. polymorpha, var. 3.

Figs. 20 & 21. Centrodorsal piece as seen from its dorsal (fig. 20) and ventral (fig. 21) sides.
22. An isolated compound basal as seen from its ventral (a) and dorsal (b) sides.

Figs. 23 & 24 of Act. polymorpha, var. 4.
Pentagonal base of the calyx as seen from its ventral (fig. 23) and dorsal (fig. 24) sides.

Plate VII.

In this Plate are shown the first, second, and third radials of the type of Act. polymorpha
(figs. 1-3) and of var. 2 (figs. 4-6).
The different aspects shown are designated as follows: — a. Internal or proximal face ; b. External
or distal face ; c. Ventral or superior face; d. Dorsal or inferior face.

Act. polymorpha, Type. Act. polymorpha, var. 2.

Fig. 1. First radial. Fig. 4. First radial.

2. Second radial. 5. Second radial.

3. Third or axillary radial. 6. Third or axillary radial.

Plate VIII.

All the figures are x 18. Figs. 1 & 2 of Act. polymorpha, Type.

Figs. 1 & 2. Two successive oblique sections through the base of a decalcified calyx, viewed from their
dorsal side.

Fig. 1 is the more inferior, i.e. nearer the dorsal surface. Its left-hand lower portion shows the centro-
dorsal piece only, with its marginal cirrhi (cir.) which receive fibrillar cords (n.c.) from the

ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 121

central mass (A 7 ) enveloping the quinqiielocular organ. Proceeding outwards from the
centre are seen five dark rays (Si), which represent the closely fibrillar organic basis of the
five rays of the basal star. In the upper part of the figure two of them are very short, only
their central ends being visible, as the section has here passed above the level of their outer
ends through the substance of three first radials (;•]).

Fig. 2. In this section only two of the basal rays (SJ are visible, as the greater part of it has passed
above the level of the synostosis (/) between the first radials and the centrodorsal piece. In
the centre are seen the chambers of the quinquelocular organ (c/i), with their ventral openings
into the vessels contained within the axial prolongation. At the right of the figure are seen
the lower ends of the axial canals, both radial (a:r.c.) and interradial (a.i.c.) ; their cavities
are generally crossed by transverse septa, which divide them up into two or three intercom-
municating smaller ones. The interval between every two of these canals is occupied by one
of the secondary basal cords (Y. 2 , Z u Z 2 , &c.), produced by the bifurcation of the short primary
cords (V, Y, &c.) proceeding from the angles of the quinquelocular organ. They are connected
with one another laterally by interradial and intraradial commissures (Leo. & r.co.) and
enter the central canals of the first radials (r x ) in successive pairs, so that the axial nervous
cord (») of each radius is composed of fibres derived from two primary basal cords (Y„,Z 1)
&c), just as in Antedon.

Fig. 3. A vertical longitudinal section through a decalcified calyx, passing on the right through the
synostosis of two first radials (A, B) and the fibrillar basis (<S 2 ) of one of the basal rays, and
on the left through the segments (j-j, r 2 , r.a) of radius D. The first of these is united to
the centrodorsal piece by connective-tissue fibrils (/) similar to, but less abundant than,
those in the interradial portion of the section, around which, in the natural condition, the cal-
careous material forming one of the basal rays is deposited. The passage of the ventral
furrows (cr.f.,v.i.f.) into the axial canals (a.r.c, a.i.c.) is also well seen in this section.
Its centre is occupied by the quinquelocular organ, from the ventral portion of which the
axial prolongation (a.p.) rises into the circumvisceral ccelom (c.v.c), which, together with
the lower end of the wide axial ccelom (a.c), occupies the space between the ventral surface
of the skeleton and the lower or dorsal wall of the convoluted alimentary canal. It is
traversed by numerous connective-tissue septa, which divide it up into a system of spaces,
communicating freely with those both of the intervisceral and of the axial ccelom (ice, a.c).

Fig. 4. Transverse section through the synostosis of two first radials (A, B) near the peripheral
margin of the centrodorsal piece (cd.) , showing the radial (/) and the interradial (S„) fibres
which effect the synostosis between it and the united first radials. The latter form the
organic basis of one of the rays of the basal star.

Figs. 5-8. Four vertical sections, selected from a series, through a decalcified calyx of Act. pectinata.

5. Section through the adjacent inner ends of two first radials (A,B), showing the axial inter-
radial canal (a.i.c.) between them, and the open outer ends of the radial spaces (r.s.)
between their dorsal surfaces and the ventral surface of the centrodorsal piece. The central
ends of their axial nervous cords (//.) are cut very obliquely.

6. A section rather nearer the centre, showing the closed central ends of the radial spaces (r.s.)
of the same two radii (A, B) and their axial canals (a.r.c) ; also the four secondary basal
cords which unite in successive pairs (Z. 2> V\ and V 2 , W-j) to form their axial cords, cut
obliquely.

7. A section from a little the other side of the centre, through the outer end of one of the
chambers (ch) of the quinquelocular organ, corresponding to radius D. The radial spaces
(r.s) of C and E arc cut almost longitudinally ; and above them, in the interior of the
radials, are seen the axial nervous cords, with one of the two secondary basal cords

122 ME. P. H. CAEPEXTEE ON THE GENUS ACTINOMETEA.

(X u Y,) by which each is connected with the central nervous envelope of the quinquelocular
organ. The other branches (X 2 , Yj) of the two primary cords (X& Y) combine to form the
axial cord of the radius D. The inner end of its first radial is seen in the centre of the upper
part of the figure (r{), separated from those of G'&_Eby the axial interradial canals (a.i.c).
Fig. 8. A section rather further from the centre of the calyx, showing the first radial of D cut trans-
versely, with the closed central end of its radial space [r.s.) . At the sides of the latter are
the expanded dorsal ends of the axial interradial canals seen in fig. 7 ; they are received in
depressions (s) at the central ends of the rays of the basal star, which are ossified around
the vertical fibres (So) only, and not, like the stouter more peripheral portions of the rays,
arouud both vertical and diverging fibres, as is seen in Plate III. fig. 6, S u S 2 .

[ 123 ]

II. On some New Species of Nudibranchiate Mollmcafrom the Eastern Seas.
By Cuthbert Collingwood, M.A., M.B., F.Z.S., &c.

(Plates IX. & X.)

Read March 7th, 1878.

THE very considerable numbers of naked-gilled Mollnsca which have been found
upon our own shores would lead one to suppose that on other coasts, in which climatal
conditions were more favourable, they would be very commonly met with. Having
collected not fewer than twenty-eight species upon a very small section of our northerly

Wes'tof^r , H " T 1 '^ 1 SGarCh UP ° U tr ° piCal Sh ° reS W ° m ^ ld - a ^ant
haivest of these highly mterestmg and beautifully-tinted animals. In any such research

however, seasonal changes must not be overlooked. There can be no doubt in the mini

of any one who has ransacked tropical localities that in them the highest development

o colour and he most curious vagaries of form are to be found ; but I am persuaded

ha a zoologist who pays but a brief visit to a number of coasts in succession is far less

likely to make a considerable collection of species of Nudibranchiata than is anyone

who, confined to a single favourable locality for a considerable period, is thus able to

pursue h* examination of the same spot through all the different seasons of the year

Hence it is that the collections of Sir W. Elliot on the Madras coast (32 species), of

1 1. Kelaar on the Ceylon coast (42 species), and of Mr. Angas in New Smith Wales

hat'ofT^T S ° laV§ V^!f Q the V °^ e ° f thG ' Asteolabe ' y^d but 18 new species,
that of the Samarang half a dozen species, and my own researches (the fruits of eve™
opportunity within my reach during a period of rather more than twelve months) resulted
only m the discovery of the 16 new species described in the following paper

The greater number of species of this group may be found upon stony shores near low-
water mark, and especially at low spring-tide. They adhere to the under side of stones

labW l ^ tlU ' Uing tLem 0Ver " Alth ° U - h SOme of them a PP<*r to be

capable of swimming, and most of them can float, branchiae downwards, upon the surface

ol the water I never saw them in this position in the sea, however calm. In fact the Y

have invariably been found by me in a passive condition, like little shapeless masses of

oft, coloured matter, in depressions and crevices of stones, where they have probably

retired at that particular juncture, when the wash of low tide has disturbed the water

and rendered it both turbulent and turbid; from both of which disadvantages the same

spot would be free at all times, except at dead low-water. I have, however, dredged up

a very beautifully-eolourecl and delicate species from a depth of 20 fathoms, which

notwithstanding the rough handling of the dredge and the company of shells, corallines'

and sponges, seemed in no way incommoded or less lively. Again, some species of

JNudibranchiata, as Glaucus, appear to be free ocean-swimmers, and such I have taken

the towmg-net m the Formosa channel; while another species {Scyllcea pelagica) is

SECOND SERIES.— ZOOLOGY, VOL. II. 17

124 DR. CUTHBERT COLLING-WOOD ON SOME NEW SPECIES OF

abundant in the open ocean of the Atlantic, but always adherent by its clasping foot to
the fronds of the gulf-weed.

Nor do the exquisite colours of this group depend entirely upon geographical position.
On our own northerly coasts richly tinted species occur, such as Doris flammea, D. coc-
cmea, and many brilliant species of Eolis, &c. ; while Doris sordida is a Red-Sea species,
D. tristis is from the Madras coast, and D. exanthema from Ceylon ; and the same
remark also applies equally to other genera and families. Nevertheless, as a rule, the
more brightly coloured species are more commonly met with on tropical shores ; and of
the genus Ghromodoris, remarkable for their brilliancy, no species are found upon our
own shores, but, although some occur on the Mediterranean coasts, the majority are
characteristic of the hotter regions.

A very remarkable eiicunistance in the history of these delicate animals is their
extraordinary geographical distribution. Mr. Abraham remarks {I.e. jwstea) that the
well-known and almost first-described British species, Doris tuberculata, has also been
found in New Zealand on the one hand, and at Vancouver's Island on the other ; and
my own experience has remarkably confirmed the widespread habitats of certain species.
Interesting in this respect was the fact of my meeting with the same species, within a
few days' or weeks' interval, on different sides of the China Sea. Thus a Chromodoris
which I had already found in the Pescadores was the first thing I picked up, some time
later, on the reef of Labuan (Borneo). Veiy soon after I met with a second species at
Labuan, which I at once recognized as one I had already captured and figured in the
Haitan Straits, a little south of the river Min, coast of China. A Doris which I found
on a submerged reef in the centre of the China Sea, I afterwards found again on another
island off the coast of Borneo, west of Sarawak river. A minute species also (probably
an immature Trevelycma) I found on two occasions in localities separated by 150 miles
of sea.

A circumstance due, I think, to what I have already mentioned about seasonal changes
is worth recording. I had searched in vain upon some rocks in the harbour of Hong Kong,
and having mentioned this to a resident gentleman, who occasionally himself made the
same researches, he kindly offered to go with me and show me where he found them.
But we were equally unsuccessful, and could obtain none. The same thing occurred at
Labuan. Having showed my drawings to a gentleman interested in natural history, he
at once recognized them as animals of which he assured me there were many beautiful
species to be found ; and he also kindly conducted me to his hunting-ground, but equally
in vain, and to his surprise he could show me none where he had been in the habit of
finding them. In fact, although I met at Labuan with species I had already found
elsewhere, I only added one new species to my list at that place. In both these instances
I imagine I was at the localities mentioned at unfavourable seasons of the year, and not
at those seasons when my informants assured me they had met with numerous species.

I should mention, however, one other Labuan species which unfortunately I was
unable to record, owing to the following curious circumstance. It was a large tuber-
culated Doris, of a mottled grey colour, l< inches long, with capacious tentacles and
expansive gill-tufts, of which I found two or three specimens upon a reef in one of these

NEDIBEANCHIATE MOLLUSCA FEOM THE EASTEEN SEAS. 125

expeditions. I brought them home, and placed them in water till the next day, intending
by daylight to draw and describe them. But on visiting them in the morning, I found
that they had performed a spontaneous amputation of the mantle close to the body all
round. It was done as cleanly as if by a pair of scissors. A large Pyrula in the same
vessel was at first credited with this act ; but other specimens having been placed separate
in clean water, not being able to attend to them immediately, I found on visiting them
next day that they also had amputated their mantles and were destroyed. It appeared
to be a suicidal act, or " happy despatch," similar to the self-evisceration of Holothuriae
and breaking up of Comatula3 under the influence of the gradual fouling of the water.

These delicate and beautifully tinted animals are so entirely altered as to their form
and colour by spirit, that I can scarcely understand how new species can be satisfactorily
described from spirit-specimens. Alcohol bleaches their colours, and contracts to shape-
lessness the most beautiful elements of their form, the mantle, but more especially the
tentacles and branchiae, so that they bear no resemblance whatever to the living animal.
It keeps them, however, fit for dissection. The spicula, odontophores, buccal collars, and
other important classificatory characters are thus well preserved. Glycerine, while it keeps
for a considerable time their colour and form, renders them soft and comparatively useless
for dissection. The only way to retain a correct idea of their living character is to make
careful drawings of them in their active condition ; and whatever value the present
illustrations may have is founded upon this circumstance, for they are faithful to the life.

With regard to the two species of the Polybranchiate family of Phyllidiadae (the only
ones I met with, and found side by side in a rock-pool on the coast of Borneo), I have
consulted Bergh's elaborate paper in the ' Naturhistorisk Tidsskrift ' for 1868-69, and
also that in Heft 10 of Semper's ' Beisen im Archipel der Philippinen,' in both of which
papers species are described and figured ; but although there is some resemblance, I
cannot assure myself that the Phyllidiella pustulosa figured by him is the same as my
Phyllidia spcctabilis.

Nothing more need be added in these introductory remarks concerning the species to
be described in this paper. As to the works to which I have been indebted in the inves-
tigation, I would chiefly mention two, viz. the well-known and invaluable ' Monograph '
of Messrs. Alder and Hancock, published by the Bay Society, which has been to me as
a companion ever since its publication ; and, secondly, the " Bevision of the Anthobran-
chiate Nudi branchiate Mollusca," by Mr. P. S. Abraham, in the Proc. Zool. Soc. 1877 :
other memoirs are referred to in the text.

List of the Genera and Species enumerated in the present Paper.

Albania formosa.
Triopa Principis-Wallise.
Trevclyana felis.
Doridopsis arboresmis.
— rubra.
Phyllidia spectabilis.
Fryeria variabilis.
Scylhea pelagica.
Bornella nianuorata.

Doris pecten.

cresceutica.

Cbromodoris iris.

Bullockii.

aurco-purpurea.

tumulifcra.

tennis.

funcrea.

Alderi.

17*

126 DR. CUTHBERT COLLINGWOOD ON SOME NEW SPECIES OF

Order NUDIBBANCMIATA, Cuvier (1817).
Family DORIDIDiE, Alder and Hancock (1855).

Genus Doris, Linnaeus (1758).

Doris pecten, Coll., n. sp. (Plate IX. figs. 1-5.)

Length nearly -^ inch. Body oval, of a deep greenish-blue colour all over, studded
with minute papillae of a darker tint. Mantle large, covering the foot entirely. Dorsal
tentacles short, lamellated, dark-coloured, paler at the bases. Branchial, consisting of
seven or eight simple leaflets, are ranged in a crescentic form, the horns of the crescent
pointing forward and embracing a small crescentic area of a paler tint. Under surface
of the mantle paler than the upper, and spotted ; foot brownish below.

Two specimens of this little Boris were found in a rock-pool above low-water mark on
Bush Island, entrance to the harbour of Ke-lung, North Formosa, May 29. When at
rest, the posterior part of the mantle is drawn in, and the branchiae are seen in profile,
looking like a comb stuck in behind, whence its specific name.

Doris crescentica, Coll., n. sp. (Plate IX. figs. 0-8.)

Length 3 inches ; breadth 2 inches. Body broad, flat, tubcrculated. Mantle capa-
cious, covering the whole body, and largely projecting beyond the posterior extremity of
the foot ; very broad and round anteriorly, but about the middle of the body constricted
on either side to about one half its diameter ; edge very thin and flat, and puckered all
round the margin with numerous large and small folds. Plentifully covered with large
warty excrescences arranged crescentically and concentrically around the anterior and
posterior margins, where they are but slightly elevated : an elongated irregular excrescence
runs along the centre of the dorsum, commencing anterior to and between the tentacles,
and terminating at the branchiae,on cither side of which are arranged large irregular bosses
a quarter of an inch high. On either side of the branchiae there is a plain and thinner
irregular portion, larger on the left than on the right side. Dorsal tentacles large, cluh-
shaped, acuminated, and arising from a projecting eye-like sheath with irregular opening;
the club-shaped portions laminated, the peduncles smooth. Branchiae, of six compound
leaflets, much branched, and arranged in a wide, round, anal orifice ; the plumes nearly
equal in size. Mead with two small oral tentacles, concealed beneath the mantle.

Colour and general appearance. — General colour a brownish olive, the elevations and
bosses paler. Upon the large bosses is a tinge of pink, surmounted by a whitish apex ;
a similar pink tinge upon the thinner part of the mantle round the margin. The thin
non-tuberculated parts of the mantle on cither side of the branchiae straw-colour. Ten- |
tacle-sheaths same as the body-colour on which they arc situated, the tentacles themselves
somewhat darker. Branchiae pale brown, the stems darker, and the edges of the leaflets
whitish. Anal orifice Avhite. Under surface — anterior half of the mantle pale brown- i
pink; posterior half yellowish, a broad irregular reddish band immediately surrounding
the whole body.

NTJDIBEANCHIATE MOLLUSCA EEOM THE EASTEEN SEAS. 127

Of this remarkable species I found one specimen upon a coral block in a shallow patch
about 3 feet deep on the Fiery Cross Reef, China Sea, August 4th. The animal crawled
freely, but did not float on its back while under observation. When turned over on its
back it at once regained its natural position. The posterior portion of the mantle varied
in form from various spontaneous degrees of constriction ; sometimes it appeared to be
nearly of an oval form, at others it resembled a Tetlujs in shape.

Subsequently I obtained a specimen of this species on the island of Barundum west
coast of Borneo, of a gigantic size for Nudibranchiata, October 8th, upon coral blocks
between tide-marks. Of the two specimens found here, one was 6* inches Ion- and ^
wide. In all respects they resembled the one figured, even, to a great extent, in the
arrangement of the tubercles or bosses.

From a careful comparison of the description given by Kelaart* of his Boris exanthemata
with my drawings and descriptions of this species, I have come to the conclusion that
they are not identical, although evidently nearly allied. The form of the dorsal tentacles
the crescentic arrangement of the anterior tubercles, as well as several points in the
colour, to say nothing of the constriction of the mantle posteriorly (which might be
accidental), all point to different species; nor is the foot of the present species deeply
grooved and notched in front, as is the species described by Kelaart. I have named my
species from the crescentic arrangement of the tubercles on the anterior border of the
notseuni.

I am by no means certain, though, that the animal named Boris cerebralis, Gould f
may not be identical. The Nudibranch referred to, 5 inches long by 2f broad is stated
to have been taken from a reef in Sandalwood Bay, Feejee Islands. Comparing
Br. Couthouy's coloured sketch with my own, the two not being drawn exactly in a
similar position, considerable difference is manifest ; but if taken along with Dr Gould's
description in the text, it seems quite possible they may be the same species. Should
further research prove this to be the case, B. crescentica must necessarily be regarded as
a synonym.

Genus Chromodoeis, Alder & Hancock (1855).
Cheomodoris iris, Coll., n. sp. (Plate IX. figs. 9-11.)

Length 2J; inches. Body elongated, slightly tuberculated in profile, presenting two
rounded elevations, with depression between. Mantle scanty, disclosing the foot on either
side and posteriorly ; much waved at the margin. Anteriorly there is a thin-lobed fim-
briated veil, and posteriorly it is also divided into three lobes immediately behind the
branchial tufts. Borsal tentacles slender, conical, and finely lamellated, situated upon
slight elevations of the mantle. Bronchia composed of eight or ten simple leaflets
arranged in a cup-shaped form like the petals of a flower ; the leaflets delicately pinnate
Foot large, projecting about f inch behind the mantle, and tapering to a point posteriorly.

* Ann. & Mag. Nat. Hist, 3rd ser. vol. iii. p. 300 (1859).

t U.S. Exploring Expedition, Mollusca and Shells, Text, 1852, p. 208, and Atlas, 1856, pi. xxiii. figs 393 a-e
and Abraham, P. Z. S. 1877, p. 212. '

128 DR. CUTHBERT COLLINGWOOD ON SOME NEW SPECIES OF

Colour and general appearance. — Mantle deep blue, with a narrow edging of bright
yellow, and an irregular yellow stripe on either side of the median line, or broken up into
yellow spots. Large black roundish spots are scattered irregularly over the surface. The
foot is of a lighter blue colour, spotted irregularly with yellow and black spots ; the black
ones roundish, the yellow ones forming an irregular line. Branchiae and tentacles rich
vermilion, the latter arising from tricoloured bases.

Spawn, a spiral ribbon of a pale straw-colour, deposited under observation.

Two specimens of this splendid species I obtained in a basaltic rock-pool under a large
stone, at Makung, Pescadores Islands, May 12th, about midway between high- and
low-water marks. In August of the same year I obtained one specimen of the same
species from a reef on the shore of Labuan Island. Prom the beautiful combination
of the three primary colours presented in this species, I have given it the name of the
rainbow-goddess Iris.

Gould's Doris smaragdma* bears a distant resemblance, but is a smaller animal, more
greenish, with an indigo-blue margin and mantle, tentacles and branchiae yellow.

Under Ghromodoris rimcinataf and C. Semper 1 1 Dr. Rudolph Bergh has described
two new species from the Philippines ; but though in some points as to coloration and
markings there is an approach to C. iris, the distinction, in absence of dorsal and mar-
ginal yellow stripes, with greyish tentacles and branchiae and other particulars besides
geographical distribution, warrants specific separation.

Chromodobis Btjllockii, Coll., n. sp. (Plate IX. figs. 15-17.)

Length 2f inches. Body rather compressed, translucent. Mantle broad and square
in front, narrowing behind, and having the sides and posterior portion of the foot
uncovered. Dorsal tentacles ^o mcu l° n g> slender, consisting of a cylindrical pedicle,
smooth, and supporting a club-shaped, spirally laminated head. JBranchice consisting of
seven simple leaflets arranged in three sets, and arising from a thick retractile peduncle
situated in a cylindrical sheath of the mantle, a little more than halfway from the anterior
edge of the mantle to the posterior point of the foot. This peduncle gives off one leaflet
in front, and two lateral branches, each of which gives rise to three leaflets. The leaflets
are angular posteriorly, and edged with delicate papillae upon the anterior aspect.
Head concealed by the mantle, and bearing two small oral tentacles. Foot long and
fleshy, extending nearly three fourths of an inch beyond the posterior edge of the
mantle.

Colour and general appearance. — Body semitransparent. Head of a deep amethystine
tint, shading behind the dorsal tentacles through paler amethyst to reddish upon the
back ; an opaque white edging all round the mantle. Peduncles of the tentacles and
branchiae deep amethyst ; laminated portion of the tentacles and leaflets of the branchiae
deep orange-yellow. Foot pale amethyst, becoming deeper at the posterior portion, where
it is as deep as at the anterior part of the mantle.

* U. S. Exploring Expedition, vol. xii. p. 290 ; Atlas, pi. xxii. figs. 390 a-c.

t Semper's ' Reisen ini Arehipel der Philippine^' 1S77, Band ii. Heft 11, p. 479, and Hct't 10, pi. 53. figs. 5-12.

J Bergh, op. tit. supra, Exit 11, p. 482. pi. 55. figs. 2-7.

NUDIBEANCHJATE MOLLUSCA FEOM THE EASTERN SEAS. 129

Of this magnificent species I dredged one specimen in GO fathoms, off Recruit Island,
North Pacific, about 150 miles N.E. of Formosa. It was quite lively, and lived several
days, moving freely about, and floating foot uppermost, but never showing any tendency
to leave the surface of the water. I have named it after Captain Charles Bullock, R.N.,
a gentleman greatly interested in these studies, and from whom I met with much
assistance and courtesy.

Chromodoris aureo-purpurea, Coll., n. sp. (Plate IX. figs. 18-22.)

Length 1| inch. Bod;/ entirely covered by the mantle, except a small portion of pos-
terior end of the foot. Mantle ample, entire, smooth, broad anteriorly. In a second
specimen the edge of the mantle was somewhat indented in one or two places. Dorsal
tentacles small, club-shaped, upon cylindrical footstalk, curving backwards and outwards ;
the club-shaped portion finely laminated. Branch'ue of ten leaflets, forming a small
double ring ; the leaflets conical, pinnate, and diminishing in size from before backwards.
Foot slightly projecting behind mantle. Head rounded in front, with two minute oral
tentacles at the sides.

Colour and general appearance. — Upper surface with a general yellow tinge, and
covered over with small, irregular blotches of bright yellow, of a roundish or elliptical
form. Mantle edged with faint violet, and an irregular row of deep violet-shaded spots
running all round upon the faint edging, both being equally distinct upon the upper and
under side. The laminated portions of the tentacles dark violet, shading off at the foot-
stalk to the colour of the mantle. Branchiae, leaflets, and midribs deep violet at their
distal ends, becoming paler below, where they merge into the colour of the mantle.
Underside yellowish white, with a bright yellow spot invisible from above.

Two specimens were obtained under moderate-sized rough stones upon the inner shore
of Slut Island, Haitan Straits, on the Chinese coast, near low-water mark, June 30th.

These animals were very active, moving rapidly along with a gliding motion, at which
time their mantles were broad and flowing over the sides, the anterior portion somewhat
square. The tentacles were in constant and graceful motion, and they also frequently
floated, foot uppermost, on the surface. When at rest, however, the animal assumed a
nearly round form.

Mr. Andrew Garrett* has described and figured a new species (his Goniodoris Tryoni),
71 millims. long, from the Society Islands (Mus. Godeffroy, Hamburgh), which suggests
likeness to my Chromodoris aiirco-purpurea. His example is mentioned as creamy white,
margined with violet, and umber or tawny flesh-coloured branchial plumes and tentacles
tipped with violet. The body-spots, however, are deep black ocelli surrounded with
white. The branchial plume has 24 divisions, the posterior shorter than those in front,
and each terminally divided. Dr. R. Bergh later t refers to Garrett's species under the
designation Chromodoris Tryoni, and further describes and gives the anatomy of other
specimens collected by Prof. Semper, these differing slightly in tint and markings. The

* Troc. Acad. Nat. Sci. Philad. 1873, p. 232, pi. iv.

t In ' Ttoi.scn ira Archip. d. Philipp.' 1 S77, Band ii. Heft 11,1'. MO, and in Journ. d. Mus. Godeffroy, 1S77, Heft 14.
pi. iv. figs. 12-23.

130 DE. CUTHBEET COLLINGWOOD ON SOME NEW SPECIES OF

clearly defined black spots of Chromodoris Tryoni, as compared with the yellow blotches
in C. aureo-purpurea and other particulars noted by Bergh, give me reason to regard
that now described from the China coast as entitled to specific rank.

Chromodoris tumulifera, Coll., n. sp. (Plate IX. figs. 23-26.)

Length f inch. Body oblong, depressed, obtuse at either end. Mantle ample, covering
the whole body, except the posterior portion of the foot ; entire, tuberculated. Dorsal
tentacles smooth and delicate, club-shaped, the club-shaped extremities very finely lami-
nated, and twice as long as the cylindrical pedicle. Branchice small, consisting of nine
simple leaflets, the anterior largest, and diminishing in size posteriorly, the two hindmost
being rudimentary. Head crescentic, with two acute angles forming small tentacles,
one on either side.

Colour and general appearance. — Mantle translucent yellowish white, the upper surface
irregularly strewed with large, roundish, well-defined tubercles of a rich carmine colour
and tumuliform profile. Round the mantle runs a broadish band of chrome-yellow,
defined exteriorly, but somewhat fimbriated interiorly, leaving a narrow edging of the
mantle tint all round the outside : upon this edging are two carmine spots on the anterior
and two on the posterior angles of the mantle. The tentacles and branchial are of the
same tint as the mantle, the latter delicately formed and difficult of observation.

One specimen of this liandsome species found under a moderate-sized rough stone on
the south side of Slut Island, Haitan Straits, coast of China, in June. It was a somewhat
inactive animal, moving but slowly, but swimming occasionally foot uppermost on the
surface.

In August of the same year, being on the island of Labuan, on the very opposite side
of the China Sea, I met with this species more than once, at Pulo Pappan and Pulo Daat,
two islets between Labuan and the mainland of Borneo. These specimens were about
the same size as the Chinese one, but differed in that the carmine tubercles were more
numerous and encroached upon the chrome border ; the branchiae also were more deve-
loped, and I was led to imagine that the specimen figured was a young individual.

In colouring, this animal bears some resemblance to the Doris petechialis, Gould * ;
but this latter is 2 \ inches long, \\ inch broad, has vermilion-coloured tentacles, pinkish
branchiae, a more lemon-coloured margin and partially slate-coloured dorsum, and its
habitat is Honololu, Sandwich Islands. Dr. Gould admits his drawings are somewhat
imperfect, but sufficient for identification of the species.

The deep carmine spots recall the black ones of Garrett's Chromodoris Tryoni, I. c, but
in other respects the two cannot well be confounded.

Curomodoris tenuis. Coll., n. sp. (Plate IX. figs. 27-29.)

Length f inch. Body long and slender, very attenuated when in motion. Mantle
entire, covering the whole body, excepting the posterior portion of the foot ; broad and
squarish in front, and narrower from behind the tentacles backwards, bluntly pointed
posteriorly. Dorsal tentacles short and club-shaped, laminated, the suture anterior.

* U. S. Explor. Exped. Moll. vol. xii. p. 296 : Atlas, pi. 22. fig. 391.

NUDIBRANCHIATE MOLLUSCA FEOM THE EASTEEN SEAS. 131

Branchiae small, consisting of seven small and simple leaflets arranged in a circle, the
anterior leaflet somewhat larger than the others, and the posterior pair smallest. Foot
long and narrow, slightly tubular, projecting beyond the mantle posteriorly.

Colour and general appearance. — Mantle opaque white with a slight tinge of yellow,
especially on the anterior portion, edged with chrome-yellow, slightly shading off inte-
riorly. The whole mantle is covered with minute roundish spots of carmine, irregularly
distributed, absent only from the most anterior portion, the spots varying in size from
mere specks to roundish definite spots. Tentacles yellowish, but not so bright as the
border of the mantle; the bases whitish. Branchiae pale yellow. Foot edged with
chrome posteriorly, like the mantle. Under surface yellowish, foot and mantle with a
faint edging of chrome-yellow, the carmine spots showing through at the sides of the
head.

Two specimens were found on the under surface of a block of coral in a shoal patch of
reef in the midst of the China Sea, named Eiery Cross Reef. It is possible the spot
might be uncovered occasionally at low spring tides, but was now 3 feet under water.
Notwithstanding this, however, tbese little creatures when captured were continually
crawling out of the water and resting upon the dry edge of the vessel in which they were
contained, under which circumstances they had a short and stumpy aspect. When placed
in the water they were very lively and at once commenced crawling, having first stretched
themselves to double their previous length, with a proportionate tenuity. While crawling
they had a remarkably slender and linear appearance. They also swam freely on the
surface, foot uppermost. (August.)

In most respects the Chromodoris tenuis agrees with the Doris aspersa, Gould*, save
size, his specimen being 1^ inch long by \ an inch broad, and its habitat Vincennes Island,
Paumotu group. It may be questionable in this case whether we have specific distinction,
or whether size and other slight variation may not be attributable to difference of age,
sex, or geographical range. The D. aspersa, D. cerebralis, &c, given by Gould, were
drawn from nature by Mr. Joseph P. Couthouy, Naturalist to the U. S. Explor. Exped.,
I and therefore may be deemed reliably correct.

• Chromodoris eunerea, Coll., n. sp. (Plate IX. figs. 30-33.)

Length If inch. Body simple, stout, except when actively in motion, when it becomes
I attenuated, obtuse in front. Mantle smooth and entire, having the posterior portion of
i the foot exposed during progression, well rounded anteriorly. Dorsal tentacles small,
i arising from a small simple sheath, laminated, with scarcely any pedicle. Branchiae
composed of seventeen or eighteen leaflets arranged in a convoluted form, the larger
i leaflets in front, the smaller behind, leaflet irregularly branched. Head concealed by
the mantle and supporting a pair of oral tentacles.

Colour and general appearance. — Border of mantle narrowly edged with orange;

general aspect of the upper surface a rich dark brown, with yellowish-white or white

lines, following the direction of the border of the mantle, but in some places slightly

ramifying and sometimes anastomosing. The posterior and exposed portion of the foot

* U. S. Explor. Exped. Mollusca and Shells, p. 304 ; and Atlas, pi. 25. figs. 399 a-c.

SECOND SERIES. — ZOOLOGY, VOL. II. 18

132 DE. CUTHBERT COLLLNGWOOD ON SOME NEW SPECIES OF

is bordered and streaked like the mantle. The tentacles have a white ring, sometimes
two rings, round the sheaths ; they are reddish above, deepening to dark brown below ;
rows of opaque white spots are arranged nearly regularly, parallel with the lamina? ; at
the apex is a small whitish ring. The branchiae are reddish, the inner side of the stems
of the leaflets marbled below with brown and white ; the whole branchial tuft studded
like the tentacles with minute spots of opaque white, having a very beautiful appearance.
The under surface of the foot is white, with an orange border, like the mantle ; the sides
of the foot striped as above, only the white lines are whiter, but less distinct, and the dark
body-colour paler.

This is a remarkably handsome species. The longitudinal white striae become widened
when the animal is at rest, and very much attenuated when in progression. The branchiae
and tentacles look as though studded with little pearls. It is somewhat remarkable
that these animals never floated upon the surface of the water while under observation,
and if placed upon their backs always immediately turned over.

These specimens w r ere taken upon a reef east of the island of Labuan, Borneo, and one
on the adjacent islet, Pulo Pappan, in August.

Chbomodoris Alderi, Coll., n. sp. (Plate IX. figs. 34-37.)

Length 2 inches. Mantle capacious, covering the head, squarish in front, slightly
emarginated. Dorsal tentacles very small, and placed upon very short pedicles, finely
laminated, having the commissure anterior ; they have precisely the appearance of small
cochineal insects. Branchial of ten simple four-sided angular leaflets, some of them
bifurcating near the apex, arranged in an imperfect circle, curving outwards and sur-
rounding the anal orifice. Head with two small white oral tentacles. Foot somewhat
tubular posteriorly, and extending about \ inch beyond the mantle.

Colour and general appearance. — General body-colour an opaque yellowish white or
cream-colour, a border of bright orange running all round the edge of the mantle and
projecting portion of the foot. Back beautifully marbled with reddish brown, an irregular
row of carmine spots placed all round the marbled portion, between it and the orange
border. Tentacles laminated alternately with crimson and white. Branchiae reddish,
the angles crimson. Under surface of a delicate transparent white.

A most beautiful species, slow in its movements, which were confined to crawling and
floating upon its back. One specimen only, found between tide-marks in a sandstone
rock-basin in Ke-lung Harbour, North Formosa, May 31st. I have named it after the
late Mr. Joshua Alder, whose name will always be connected with the history of this
elegant order of Mollusca.

Genus Albania, Collingwood, nov. gen. (1878).

Corpus depressum, niolle, semipellucidum. Notseuin amplissimum, undulatum et inversum. Rkino-
plioria flexibilia, sine vaginulis. Caput velo bilobato obsituru. Branchiae e circa 7 foliolis,
separation retractilibus, compositse.

Body depressed, soft, semitransparent. Notaeurn (mantle) ample, undulated, and
turned up at the sides. Dorsal tentacles (rhinophoria) flexible, without sheaths.

NTTDIBEANCHIATE MOLLUSCA FEOM THE EASTERN SEAS. 133

Branchiae consisting of about seven compound leaflets, each separately retractile. Head
with a bilobated crenated veil.

I have named this genus after the late If r. Albany Hancock, a gentleman so well
known in connexion with the history of these animals. For euphony's sake I have so
adapted the name as to make it agree with Formosa, the native island of this elegant
species. The genus must be placed near BZexabranchus.

Albania Formosa, Coll., n. sp. (Plate X. figs. 1-5.)

Length 2 inches. Body extremely delicate, almost semitransparent. Mantle broad
and capacious, forming, as it were, wings or fins on either side ; the edges turned over the
back when at rest. Dorsal tentacles large and rather thick, consisting of a conical,
bluntly pointed, finely laminated portion, with the commissure anterior, mounted upon
a cylindrical pedicle of equal length, without a sheath, non-retractile. Branchial con-
sisting of seven compound leaflets, each having three or four or five branches, and arising
from a common thick pedicle, but separately retractile, the whole forming a ring near the
posterior extremity of the mantle ; very delicate and almost transparent. Head with a
bicrcscentic veil, and studded round with a fringe of minute papilla?. Foot extending
j inch beyond the mantle posteriorly, somewhat rounded.

Colour and general appearance. — General tint a pale rose, darker and richer on the
back, forming a Vandyke pattern nearly regular on either side. Edge of the mantle
opaque white, with a wide inner border of crimson, the whole studded with minute
whitish translucent points, the greater or lesser abundance of which effects the gradation
of colour. Laminated portion of the tentacles crimson ; pedicles pale rose. Fringe of
the veil orange ; veil and posterior portion of the foot yellowish ; under surface pale rose.

One specimen of this singular and beautiful Nudibranch was taken in a red sandstone
rock-pool in Ke-lung Harbour, North Formosa, May 31st. When placed in a vessel of
sea-water it commenced swimming freely with a vertical vermicular movement and
extreme grace. The mantle was spread out wide, the tentacles thrown back, like ears,
ind the anterior and posterior extremities of the body thrown upward till they met above,
lien partially thrown back, the mantle waving in a vermicular manner from anterior to
posterior edge. It continued swimming like this for a considerable time. It did not
:rawl about like other Xudibranchs, but when not swimming remained in a more or less
contracted form, the mantle constantly changing its aspect. When, however, I turned
It over, it floated on its back like its congeners. I have called the species Jbrmosa, both
I roni its beauty and the island of which it is a native.

Family POLYCERID^], Alder & Hancock (1855).
Genus Triopa, Johnston (1838).

tsiopA Princlpis-Walll^, Coll., n. sp. (Plate X. figs. 6-11.)

Length f inch. Body slender, narrow, rounded in front, obtusely pointed behind.
'lantle scanty, just covering the body, smooth, and furnished with papilla; round the
nterior portion and along the sides. The anterior papilla?, eight in number, arranged

18*

134. DE. CUTHBEBT COLLINGWOOD ON SOME NEW SPECIES OE

in a crescentic form as a veil ; the lateral papillae five on either side. These papillae
consist each of a conical stem with pinnae, the lateral papillae being about twice the
length of those round the head. Dorsal tentacles club-shaped, the upper half swollen
and finely laminated, upon a cylindrical pedicle. Branch he of five leaflets arranged
round the anal orifice upon an elevated portion of the body ; the anterior leaflet much
the largest, and the two posterior minute. Each leaflet simply pinnatifid, and feather-
like in general aspect.

Colour and general appearance. — Body of a general pale orange-yellow, darker between
the tentacles and along the median line, spotted irregularly with minute dots of orange.
Upper half of the papillae with larger spots of orange. Pinnae translucent yellowish.
Branchiae pale orange-yellow. Tentacles spotted with orange about the central parts of
the laminated portion.

One specimen found beneath stones near low-water mark, on Slut Island, Uaitan Straits,
coast of China, June 30th. It was rather sluggish in its habits, but swam on the surface,
foot uppermost. I have named the species from the resemblance of the branchiae to the
well-known crest of the Prince of Wales.

Genus Trevelyana, Kelaart (1858)*.
Trevelyana felis, Coll., n. sp. (Plate X. figs. 12-14, immature.)

Length % inch. Body simple, smooth, of a uniform scarlet colour, the intestines
showing darker upon the dorsal surface. Mantle indistinct, covering the entire body.
Dorsal tentacles two, perfectly simple, conical, scarlet. Branchice, none visible.

Several specimens of this little animal occurred upon a stone in a tide-pool on the
basaltic shore of Makung, Island of Pong-hou, Pescadores, in May ; and in June, six
weeks later, I also found it on Slut Island, Haitan Straits. It was very active and flexible,
assuming at different times the most singular forms, resembling in turn a fox, a rabbit,
a cat, according to its different attitudes. It swims like the other nudibranchs on the
surface of the water, foot uppermost. I cannot speak with great certainty of it, but
believe it to be an immature species of Trevelyana, the trivial name being added to call
other observers' attention to it. Qu. In what respects does it stand to the T. (Stenodoris)
rubra, Peascf , from the Pacific ?

Family DOBIDOPSID^E, Alder and Hancock (1864).

Genus Doridopsis $, Alder and Hancock (1864).

Doridopsis arborescens, Coll., n. sp. (Plate X. figs. 15-17.)

Length If inch. Mantle capacious, enveloping the body, and overlapping the foot
laterally and anteriorly; the edges deeply cut and puckered all round; smooth and

* Trevelyana, Kel. Journ. Asiatic Soc, and Aim. & Mag. Nat. Hist. 3rd ser. (1858) vol. i. p. 257. Messrs. Alder
and Hancock, Trans. Zool. Soc. vol. v. p. 132 (footnote) suggest that "this genus may possibly bo synonymous with
the Cfymnodoris of Stimpson, Proceed. Philad. Acad. Nat. Sc. 1855."

t Amer. Journ. Conch, vol. ii. p. 206, pi. 4. fig. 2.

J See " Indian Nudibranchiato Mollusca," Trans. Zool. Soc. vol. v. p. 124.

XUDIBRANCHIATE MOLLTJSCA FROM THE EASTERN SEAS. 135

velvety. Dorsal tentacles rather large, consisting of a thickish footstalk pointing
forwards and a laminated club-shaped portion curving backwards; the whole seated
within the rim of a narrow retractile sheath. Branchiae very large, consisting of about
seven compound ramified leaflets, radiating from around the anal aperture, and situated
near the posterior edge of the mantle. The anterior leaflets (pointing forward), when
fully expanded, reach nearly halfway over the back; the posterior leaflets smaller. Each
leaflet finely pinnatifid, and the whole forming a beautiful star nearly an inch in dia-
meter, concealing the posterior portion of the animal. Foot large, occasionally visible
beyond the posterior edge of the mantle ; deeply cleft posteriorly.

Colour and general appearance. — The whole of the mantle a rich deep velvety blackish
brown, edged with fight chestnut. The peduncles of the tentacles blackish, translucent ;
the laminated portion like the mantle, and tipped with chestnut ; branchial leaflets
dark brown, tipped with light greyish at the edges. Underside of the mantle blackish,
translucent ; and foot light brown, shading to chestnut along the margin.

Two specimens of this handsome Nudibranch were found among rough stones on Slut
Island, Haitan Straits, coast of China, near low-water mark, June 30th.

The animals were sluggish, not moving much, nor fast; they floated readily, foot
uppermost, on the surface. The mantle was so capacious that in some attitudes they
appeared as broad as long; but when ordinarily crawling, the large and beautiful branchiae,
were very conspicuous, occupying nearly the posterior half of the body, and concealing
the posterior margin of the mantle and end of the foot, which at that time projects
beyond the mantle.

Messrs. Alder and Hancock * have shown that in the case of Doridopsis nigra there are at
least three or more varieties with gradation of tints and markings, and inhabiting the
coasts of the Loochoo Islands, Madras, and Ceylon. Our species, D. arborescens, appears
quite distinct ; but nevertheless, with such tendency to variation in a form closely related,
careful comparison with a series might cause a different view to be taken as to its
separation or identity with species already described by other authors.

Doridopsis rubra, Kelaart. (Plate X. fig. 18.)

Length 1| inch ; breadth 1 inch. Body thick, sluggish, opalescent. Mantle capacious,
thin, covering the whole body, except the posterior portion of the foot ; smooth, semi-
transparent, folded posteriorly during progression. Dorsal tentacles pyramidal, short
and thick, curved laterally, the footstalks smooth, as long as the upper portions, which
are swelled, pointed at the extremities, and laminated. Branchiae consisting of six
compound leaflets, like feathers, somewhat unequal in size, surrounding the anus, retrac-
tile. Read small, covered with the mantle. Foot left uncovered posteriorly during
progression.

Colour and general appearance. — Mantle of a rich rose-colour, darker in the thicker
parts of the back, and paler upon the thinner portions of the sides, marbled on the back
Avhen contracted, from corrugation of the surface. Branchiae of the same rose-tint as

* Ind. Nudib. Moll., Trans. Zool. Soc. vol. v. p. 128.

136 DE. CUTHBEET COLLINGWOOD OX SOME NEW SPECIES OF

the mantle. Dorsal tentacles of a rich rose-colour, the footstalks paler. Underside of
mantle pale rose, and underside of the foot inclining- to yellowish.

One specimen, found under a stone hetween tide-marks in Singapore harbour, immedi-
ately west of the town, in December. It was sluggish in habit, and crawled slowly, but
floated upon the surface foot uppermost.

This specimen died in the vessel of water, discolouring the fluid of a pink tinge/which,
however, did not appear to injure two Planariae in the same water.

Two figures of this species occur in Sir "VV. Elliot's 'Madras Nudibranchs '*. One of
these, with black markings on the mantle, is supposed to be the typical specimen, and one
of a more pure rose-colour the variety. I am disposed to think, however, that the
present description will be found typical, and that those diverging into markings are
varieties.

Family PHYLLIDIADvE, Lamarck (1809).
Genus Phyllidia, Cuvier (1708).

Phyllidia spectabilis, Coll., n. sp. (Plate X. figs. 19-23.)

Length 2 inches ; greatest breadth f inch. Body oval, tuberculated, covered with the
mantle. Tentacles two, short, placed near together, rather more than | inch from the
anterior extremity of the mantle, tapering, slightly curved, laminated, black, retractile
within a simple fixed sheath, which is situated on one of the tubercles of the mantle.

Upper surface covered with numerous irregularly-shaped tubercles, arranged in groups
of from one to ten or twelve, these clusters each perfectly distinct and similarly coloured.
The ground of the mantle is jet-black and smooth, forming a network which ramifies
among the groups of tubercles ; the bases of these groups polygonal, of a pale emerald-
green colour, the most elevated kuobs being whitish. A narrow, black, irregular edging
surrounds the dorsal surface, enclosing all the groups of tubercles, outside which is a
smoother and paler, irregular, and very slightly tuberculated margin. Under surface —
Foot greyish, oval ; a small bitentaculated head in front, the tentacles immovable. A
narrow cleft bisects the posterior half-inch of the foot. Mantle ample, surrounding the
head and foot on all sides. At the junction of the mantle with the foot and on the under
eck-e of the mantle is a close row of lamelliform branchiae, small in size, the series sur-
rounding the whole body, except the head.

This beautiful Phyllidia, in captivity, deposited a long, irregular, and narrow ribbon of
spawn, of a whitish colour, from an aperture in the side of the body. One specimen,
found under a block of coral between tide-marks on Pulo Barundum (or Marundum),
west coast of Borneo.

Doubts might be expressed whether this species may not come rxnder that named
Phyllidia ptistulosa, Cuv., P. verruculosa, Cuv., and Phillidiella pustulosa, Berghf. If

* Alder and Hancock, Trans. Zool. Soc. vol. v. p. 12G, pi. xxxi. figs. 1 and 2.

t See Cuvier, Ann. du Mus. (1804) vol. v. p. 268, pi. sviii. A. fig. 8 ; Mem. p. 3. fig. S ; also Bergh, Monog-
p. 511, and in Sernper's Eeisen im Arehip. d. Philipp. Band ii. Heft 10, p. 3^2.

NUDIBRANCHIATE MOLLUSCA FROM THE EASTEEN SEAS. 137

such should be the case, this Nudibranch has a wide distribution. As I cauuot, however,
myself regard theni as identical, I prefer to give that now described a separate specific
designation .

Genus Fryeria, Gray (1853).

Fryeria variabilis, Coll., n. sp. (Plate X. figs. 24-28.)

Length 2f inches ; breadth 1 J inch. Body convex, elongated, thickly tuberculated.
Tentacles two, situated f of an inch from the anterior margin of the body, arising from
simple sheaths, in which they are retractile ; the sheaths and tentacles black, the latter
being also finely laminated. Mai/ tie covering the head and sides of the foot, but leaving
tbe posterior point of the foot exposed during progression.

Colour. — The smooth ground-portion black, but supporting a number of irregularly-
scattered roundish tubercles, mostly single, a few double or treble ; the tubercles largest
and most loosely scattered over the central dorsal region, smaller and more closely
clustered along the sides. The colour of these tubercles was in one specimen pinkish, in
others of a pale emerald-green. In all respects, however, they are evidently the same
species. Foot smooth, black above where visible. Under surface. — Foot dark grey
beneath, blackish at the sides. Head small, with two small black tentacles concealed by
the mantle. Underside of the mantle blackish. A row of lamelliform branchige occu-
pying the junction of the foot and mantle beneath all round the body, excepting only
the head. Anus situated under the mantle on the right side, about f inch from the
anterior margin.

. Several specimens were found under blocks of coral on the reef of Palo Barundum,
west coast of Borneo. Animal very sluggish in its movements.

Family SCYLLJEIVJE, Alder & Hancock (1855).

Genus Scyll.ea, Linna3us (1758).

Scyll^a pelagicAj Linn. (Plate X. figs. 29-33.)

Length 1^ inch (average). Body smooth, opalescent, narrow, compressed, with two
broad tentacle-sheaths and two pairs of broad branchial lohes ; the posterior portion
raised to a crest, wedge-shaped, and notched like a cock's comb. Dorsal tentacles two,
lamellated, in broad and somewhat clavate compressed sheaths, straight in front and
finely serrated and crenatcd behind, with smooth and entire broad footstalks, the apex
somewhat flattened and depressed. Branchial lobes broad, flat, in two pairs, situated on
broad entire peduncles of nearly equal size ; externally smooth, internally having a
number of very delicate tree-like tufts irregularly scattered, of varying size, the large
below and the smaller above. Scad with a projecting crenated veil. Foot narrow, the
edges folding over inwardly, but when the animal crawls on glass a small portion of the
foot appears flattened out for the purpose.

Colour and general appearance. — General colour yellowish brown, darker along the
edges of the papillary prominences and the tentacles and their connecting ridges, inter-

138 DR. CUTHBERT COLLINGWOOD ON SOME NEW SPECIES OP

spersed here and there with minute white spots. Tufts on papillae [light brown. On
either side of the body is a row of opaque white projecting tubercles ; and between them
and tbe papillary prolongations are some minute turquoise spots, three or four in number.
The body generally is opalescent, with faint brown markings.

Spawn, a loose straw-coloured coil, entwining the leaves and berries of Sargassum
bacciferum, and imbedded in a mass of transparent jelly.

Portion of a branch of Sargasswm bearing a floating bladder, and with ($p) a coil of spawn of Scyllcea fdagica attached.

Nat. size.

Considerable numbers of this pelagic species were found upon the Sargassum floating
in lat. 25 c X., long. 37° W., most pieces of the weed having one or more specimens. The
animals were in a constant movement of contraction and writhing. In the water they
swam freely, moving the head and tail from side to side alternately, so as nearly to touch
one another ; and when thus swimming were always, owing to the weight of the papillary
prolongations and tentacles, back downward, and bore a grotesque resemblance to a
four-legged animal with long ears (such as a Skye terrier). They would also attach
themselves to the surface of the vessel by a sort of sucker formed by a small cylindrical
portion of the foot (fig. 33, PI. X.). There is no figure of this species in Alder and
Hancock's work, nor do I know of any good figure. The present ones are from life, and
fig. 30, PL X. represents its peculiar falling aspect, as mentioned above.

Alder and Hancock (T. Z. S. v. p. 136) have described a Scyllcea marmorata and S. viridis
as new species, the former of which, save in size, differs little, if at all, from the Linnean
species S. pelagica ; but they admit that it is difficult to decide as to what constitute
specific differences in the genus. In Dr. Bergh's Monograph of Scyllcea* there is an
excellent resume of the subject, and a number of varieties of S. pelagica described and
referred to, as well as a good account of the anatomy of the genus given.

Family DENDRONOTLD^, Alder & Hancock (1855).

Genus Borxella, Gray (1817).

Bornella marmorata, Coll., n. sp. (Plate X. figs. 31-38 )

Body long and slender, with seven pairs of papillae. Head small, with a veil of five

* Semper's Reisen, 7. c, Band ii. Heft 8, p. 315 el sc</.

NTTDIBRANCHIATE MOLLUSCA FROM THE EASTEEN SEAS. 139

or six radiating papillae on either side, deeply cleft in the middle. Dorsal tentacles
situated upon thickish footstalks, from which spring one large hranch (the uppermost)
and four shorter branches, forming altogether a radiating sheath, in the midst of which
is a slender laminated tentacle. Marginal processes in seven pairs, the anterior longest
and most compressed, the posterior pair small and rudimentary. Each pair of papilla?
varies in size and development : the first pair consists of three elongated bodies of the
size of the largest branch in the radiating sheath of the tentacles ; the second pair has
two such ; the third and fourth also, but smaller ; the last pair has but one. On the
inner side of each of these bundles is a finely ramified dendriform branchia, not so large
as the processes, the processes and branchia all arising from a common stem. Foot long,
and pointed behind. Anus situated on the right side, between the dorsal tentacles and
first pair of papillae.

Colour and general appearance. — Body beautifully marbled all over with streaks of
vermilion, strongest on the back, and more delicate and faint on the sides of the foot.
Papillae tipped with vermilion, and also the radiating projections of the veil, but neither
are marbled like the body. Dorsal tentacle-branches also tipped with vermilion, and the
largest branch marbled.

Three specimens of this beautiful species rewarded a brief search among the rocks near
the landing-place at Aden in March. They were under stones, not far from high-water
mark. The animals were extremely active, often swimming with a lateral vermicular
movement, at which time the body was laterally compressed. They were seldom quite
at rest.

I here specifically distinguish this form, although previously * I inclined to regard it
as the B. digitata, Adams t ; but a more careful examination since discloses points of
difference which had previously escaped my notice. B. digitata was first found at Sunda
on floating fuci, but it has since been got on the Madras coast by Sir Walter Elliot J, and
in the Philippines by Professor Semper §. Bergh § enumerates six species belonging to
the genus, and he adds very considerably to the account of the anatomy furnished by
our countrymen Messrs. Alder and Hancock ||.

DESCRIPTION OF THE PLATES.

The line beside certain of the figures denotes the natural size of the animal. The separate figures of
tentacles, branchial tufts, and other parts are all considerably enlarged.

Plate IX.

Figs. 1-5. Doris pecten, Coll., n. sp. : 1, general view of animal, its dorsum; 2, under surface of same ;
3, a dorsal tentacle (rhinophore) ; 4, a branchial tuft ; 5, outline of the animal at rest.

* " Observations on the Distribution of some Species of Nudibranchiate Mollusca in the China Sea," Ann. & Mag.
Nat. Hist. 1868, ser. 4, vol. i. p. 91 ; and ' A Naturalist's Rambles in the China Seas' (London, 1868), p. 9.

t Voyage of the ' Samarang," p. 07, pi. six. fig. 1. X Trans. Zool. Soc. vol. v. p. 131.

§ Iteisen im Archip. d. Thilipp. Band ii. Heft 7, p. 301. || Loc. cit. T. Z. S. swpra.

SECOND SEKTES. — ZOOLOGY, VOL, II. 19

140 ON SOME NEW SPECIES OF NUDIBRANCHIATE MOLLUSCA.

Pigs. 6-8. Doris crescentica, Coll., n. sp. : 6, general view of animal, its dorsum (nat. size) ; 7, a dorsal

tentacle and its sheath, anterior view ; 8, a branchial tuft (or branch). Sketch taken from

specimen obtained at the island of Barundum, west coast of Borneo.
Figs. 9-14. Chromodoris iris, Coll., n. sp. : 9, lateral view of animal in progression, the branchiae partly

closed ; 10, another view of the same, with the branchiae fully expanded ; 11, a side view of the

animal contracted and at rest ; 12, a lamina of the branchiae ; 13, a dorsal tentacle ; 14, the

spawn.
Figs. 15-17. Chromodoris Bullockii, Coll., n. sp. : 15, the animal in progression; 16, the underside of

the head; 17, plan in outline of the brancbial laminae.
Figs. 18-22. Chromodoris aureo-purpurea, Coll., n. sp. : 18, dorsal surface of the animal in progression;

19, outline from above of the animal at rest ; 20, diagram of the mode of arrangement of the

branchiae ; 21, a branchial lamina; 22, a dorsal tentacle.
Figs. 23-26. Chromodoris tumuli/era, Coll., n. sp. : 23, dorsal surface of the animal in progression; 24,

a tubercle of the dorsum seen in profile; 25, a branchial lamina; 26, a dorsal tentacle.
Figs. 27-29. Chromodoris tenuis, Coll., n. sp. ; 27, upper surface of the animal in progression; 28, the

under surface of the animal as seen swimming with foot ; 29, anterior view of a dorsal tentacle.
Figs. 30-33. Chromodoris funerea, Coll., n. sp. : 30, animal in progression as seen from above; 31, a

dorsal tentacle; 32, a branchial tuft ; 33, an outline plan of the branchiae.
Figs. 34-37. Chromodoris Alderi, Coll., n. sp. : 31, dorsal surface of the animal in progression ; 35, larnime

of the branchiae ; 36, a dorsal tentacle ; 37, the suctorial mouth.

Plate X.

Figs. 1-5. Albania formosa, Coll., nov. gen. et sp. : 1, animal in movement as seen swimming, its upper

surface ; 2, the veil ; 3, a tuft of the branehue ; 4, a dorsal tentacle ; 5, an outline of the

under surface.
Figs. 6-11. Triopa Principis-JVallue , Coll., n. sp. : 6, animal in progression, upper surface, the branchiae

expanded ; 7, side view of the animal at rest ; 8, a papilla of the dorsum ; 9, a papilla of the

head ; 10, a dorsal tentacle ; 11, outline of a branchial twig.
Figs. 12-14. Trevelyana felis , Coll., n. sp., immature : 12, animal in progression ; 13, animal as contracted

at rest; 14, natural size.
Figs. 15-17. Doridopsis arborescens, Coll., n. sp. : 15, the animal in progression ; 16, outline of the under

surface of the body ; 17, a branchial tuft.
Fig. 18. Doridopsis rubra, Kel. : diagram of the branchiae.
Figs. 19-23. Phyllidia spectabilis, Coll., n. sp. : 19, upper surface of the animal ; 20, outline of the under

surface of the same; 21, a dorsal tentacle and sheath; 2.2, diagram portion of branchiie ; 23,

the spawn.
Figs. 24-28. Freyeria variabilis, Coll., n. sp. : 24, animal, its upper surface; 25, diagram of the under

and lateral surfaces, showing position of branchiae ; 26, an under view of the head ; 27, a dorsal

tentacle ; 28, sketch portion of the branchiae.
Figs. 29-33. Scylhea pelayica, Linn. : 29, dorsal surface of the animal; 30, position which the animal

assumes in falling through the water; 31, a dorsal tentacle sheath; 32, branchial tuft covering

the inner surface of the body ; 33, the foot from below.
Figs. 34-38. Bornella marmorata, Coll., n. sp. : 34, outline sketch of the head; 35, a dorsal teutacle;

36, the veil ; 37, a branchial lobe ; 38, markings on the mantle.

[ 141 1

111, On the Anatomy of Ants. By Sir John Lubbock, Bart, MB F.B S F.L S
B.CL LL.B Vice- Chancellor of the University of London, President of 'the
Entomological Society. J

(Plates XI. & XII.)

Bead February 6th, 1879.

Introductory Bemarks.

IN conjunction with the observations on the habits of Ants, which the Society has done

me the honour of publishing from time to time in the Journal, I have also been studying

their anatomy especially with reference to the muscular system. Of the anatomy of the

Society '1877* **** ^ ^^^ ™ ^ ' Tnm8action8 of the Microscopical

The present paper is devoted to the thorax, with special reference to Basins favus.
Though it is founded on numerous dissections, and on more than 1000 sections beautifully
prepared for me by Mr. Newton, of the School of Mines, and Mr. Robertson, of Oxford it
is still very imperfect ; and I am only induced to bring it before the Society in its present
incomplete state because, while I hope it will be found to add somewhat to our knowledge
I see little prospect of being able to work out the subject as thoroughly as I could wish '
As a general rule, the thorax of insects is considered to consist of three more or less
well-marked segments, usually known by the names suggested by Nitzsch-prothorax
mesothorax, and metathorax. r '

Dr. Ratzeburg, however, published in 1832 a memoir (< Ueber Entwickelung der fuss-
losen Hymenopteren-Larven, mit besondercr Riicksicht auf die Gattung Formica ') in
which he maintained that the fifth segment of the larva forms, not the so-called « scale >'
or firstabdommal segment, but the hinder part of the thorax. This view has also been
maintained by Audouin and Latreille; while, on the contrary, others, as, for instance,
Kirby and Spence and MacLeay, consider the thorax of these insects to be composed
ot three segments, as usual.

Burmeister, indeed, roundly observes (< Manual of Entomology,' Shuckard's transl.
p. b5J that Audouin s assertion is unfounded.

Lepelletier cle St.-Fargeau, in his < Histoire Naturelie des Hvmenopteres ' (183G)
expresses the same opinion. « II me parait," he says, "plus simple parler comme voient

7? y6 T • n, m n alS °' ^ MS eXCdlent ' Introd ^ tio » to the Modern Classification

of Insects (1810, p. 227), adopts the same view. It may, he admits, « be asserted that,
as the body of the imago possesses two or three segments fewer than exist in the body of
larva, we may suppose that the loss of one of these segments takes place at least in this
manner, and in this part of the body. This, however, can only be done by admitting that
the head and three thoracic segments of the imago are composed of five larvta-se.-
ments instead of four, an admission negatived by all analogy with pedate larva,."

* See Quekett Lecture, Monthly Microscopical Journal (1877). vol. xviii. p. 121.

19

*

142 Sffi JOHN LUBBOCK ON THE ANATOMY OF ANTS.

Newport, in his article " Insecta," in the ' Cyclopaedia of Anatomy and Physiology,'
p. 920, says : — " at first we were inclined to Audouin's opinion, more especially on
account of what we shall presently find in Lepidoptera, in which the fifth segment, in its
atrophied condition, is as much connected with the thorax as with the ahdomen. On
further examination, however, we are satisfied that that portion of the metathorax
which is posterior to the incisure belongs to the third segment of the thorax."

Fenger, in his " Allgemeine Orismologie der Ameisen " (Arch, fur Naturg. 1862, p. 315),
treats the thorax as consisting of three segments, and does not even hint at any dif-
ference of opinion on the subject. Mayr also, in his excellent ' Die europaischen For-
miciden,' p. 4, and Schenck, in his " Bcschreibung nassauischer Ameisen- Arten " (Jahrb.
des Ver. fur Naturkunde im Herz. Nassau, 1852), adopt the same view. Lastly, Forel
(' Fourmis de la Suisse,' p. 5) says that the thorax " se divise en trois segments, comme
chez tous les insectes : prothorax, mesothorax et metathorax."

It would perhaps hardly be appropriate to refer to more general or condensed works
in which the thorax is stated to consist of three segments, as, for instance, by Owen
(' Lectures on Invertebrate Animals,' p. 193), Claus (' Grundziige der Anatomie,' p. 557),
Bolleston (' Forms of Animal Life,' p. cix), &c. ; for these eminent authors, thougb expres-
sing no qualification, perhaps only meant to describe a general, and not necessarily an
invariable, rule.

Huxley, in the ' Introduction to the Classification of Animals,' p. 58, observes, with his
usual care and accuracy, that " three, or perhaps, in some cases, more, somites unite, and
become specially modified to form the thorax."

Notwithstanding the high authorities who have adopted the opposite opinion, and
although the first appearance of the tborax seems to support their view, for my own part
I cannot but think that Batzeburg's opinion was correct. Packard (' Guide to the Study
of Insects,' p. 66) has given figures of the metamorphoses of Bombus, from which it seems
clear that the fifth segment of the larva forms the posterior portion of the thorax of the
perfect insect. Lacaze-Duthiers (Ann. des Sc. Nat. 1853, p. 231), Palmen (Zur Morph. des
Tracheensystems), and Beinhard (Berl. ent. Zcits. 1865) also advocate the same view.

The position of the spiracles affords also strong evidence in support of the same opinion.
It is generally stated in works on the anatomy of insects that there are on the thorax
two pairs of spiracles, the first between the pro- and mesothorax, the second between the
meso- and metathorax.

According, indeed, to Burmeister (' Handbook of Entomology,' p. 164), this is also the
case with the Hymenoptera, which " all possess four (spiracles) in the thorax, two of which
are upon the limits of the prothorax, between it and the mesothorax, and the other two
lie between the meso- and metathorax. In the Hymenoptera, in which the thorax con-
sists of a hard, horny case, and the segments are closely united together, the posterior
pair of spiracles lie upon the metathorax itself, whereby they distinguish themselves from
all the other orders." In fact, however, as may be seen from the accompanying
figures (PI. XL figs. 2, 4, 5), the thorax of Ants possesses, not two, but three, pairs of
spiracles.

Tbe two first pairs are situated between the pro- and mesothorax and the meso- and

SIR JOHN LUBBOCK ON THE ANATOMY OF ANTS. 143

metathorax, as usual, and evidently correspond with the two pairs of thoracic spiracles of
other insects. The third pair is situated at the side of the so-called metathorax ; but in
no case whatever do we find among insects two pairs of spiracles on one segment. Such
an arrangement would be contrary to the whole plan of organization of the Arthropoda.
It is obvious, therefore, that the third pair of spiracles corresponds to that which in other
insects lies between the thorax and the first abdominal segment. Burmeister, as we
have seen, remarks that certain Hymenoptera " distinguish themselves from all other
orders " in having a pair of spiracles " on the metathorax itself;" but he supposes that
these correspond to the spiracles which are ordinarily situated between the meso- and
metathorax, overlooking the fact that these spiracles also exist as usual. It seems clear,
therefore, that the portion of the body posterior to the third pair of spiracles really cor-
responds to the first abdominal segment in ordinary insects.

Nor are the respiratory organs alone in pointing to this conclusion. The internal
chitinous appendages clearly divide the thorax into four portions ; and I think it may
be said that the thorax contains four ganglia, though the last (PI. XL fig. 2, G 4 , PL XII.
fig. 2) is certainly not large.

The Prothorax.

The upper part of the prothorax, or pronotum, is formed in Lasius jlacus by a single
arched chitinous plate (PL XL figs. 1, 1, & 6, B), which slopes downwards from its posterior
border towards the head, where it forms a sort of keel (PL XL figs. 1 & 2). Seen exter-
nally and from the side, its lower border appears to join the upper edges of the propectus ;
but a transverse section (PL XII. fig. 4) shows that this is not so, but that the propectus
is continued for some distance beyond the lower margin of the pronotum, and is then
connected with it by a membrane which passes from the upper margin of the propectus to
the lower one of the pronotum: The propectus tapers in front (PL XL fig. 5), terminating
on each side in two teeth, which lock into two corresponding teeth (PL XL figs. 6 & 7,
and PL XII. fig. 1, X) or processes at the back of the head. The propectus is divided
into two plates (Plate XL fig. 2, C & T), one anterior and one posterior, which, moreover,
are divided into lateral regions by a central ridge. The anterior plate of the propectus
has in front a deep bay or depression, at the two horns of which are the above-mentioned
teeth or processes. Each region of the anterior division of the propectus has therefore
roughly the form of a triangle with arched sides. The posterior division of the pro-
pectus is elliptic in form, and not so large as the anterior division, to which it is firmly
attached.

The propectus is therefore attached to all the surrounding chitinous plates by flexible,
though tough, membranes. It hangs, indeed, something like the under body of a carriage;
xnd from the fact that the anterior horns of the prothorax interlock with the posterior
arocesses of the head, if the propectus is turned round it carries the head with it. On
;he other hand, if the head be retracted, the posterior processes of the head, from their
position with reference to the anterior horns of the prothorax, prevent the head of the
nsect from being turned round against its will.

144 SIE JOHN LUBBOCK ON THE ANATOMY OF ANTS.

The posterior surface of the propectus is connected with the anterior edge of the meso-
pectus hy a tough, hut flexible, membrane.

I have found it difficult to understand the descriptions given of the interior skeleton
of the thorax hy previous writers, nor do their figures give much assistance. In the
normal insect-thorax there appear to he seven principal processes — four springing from
the back, and called by Kirby and Spence the phragma, prophragma, mesophragrna, and
metaphragma ; and three from the sternum, named by the same authors profurca,
mesofurca, and postfurca. In the worker Ants the four superior processes are not
developed, but the furca, mesofurca, and postfurca are very important ; they give
attachment to various muscles, and serve also to protect the nervous system. Kirby and
Spence, however, dismiss them very summarily, and, as regards the processes of the endo-
sternum, state that they "are not sufficiently remarkable to require particular notice " *.
Burmeisterf says that in the prothorax (of the Hymenoptera) "there are two strong
pointed processes, each of which has a double root. The exterior one comes from the
margin of the prosternum, and the interior one from the central ridge of the same part.
Between these roots the muscles of the coxae pass, and between the processes themselves
run the pharynx and the nervous cord ; and it is to these processes that the connecting
muscles of the pronotum and prosternum are attached. In the mesothorax we first rind
the prophragma, a small, not very high, horny partition, which descends from the anterior
margin of the mesonotum ; and we next find a delicate ridge, which encompasses the
whole distinctly separated mesonotum. The mesosternum and scapuke are closely joined
in a half-ring, and from the central carina of this ring springs a broad strong ledge, which
at its upper margin is furnished on each side with a strong process ; they form with the
ledge a rectangular cross, and serve as points of insertion for the muscles of the coxa- of
the middle legs, lying on each side contiguously to the central ridge." As regards the
metafurca, all he tells us is, " between the metanotum and metaphragma the two large side
pieces and their auxiliaries lie, separated from each other by furrows, from which, inter-
nally, strong ridges spring, and to which the muscles of the posterior legs are attached."

Graber, in his admirable ' Die Insekten,' truly observes that the endoskeleton has been
almost entirely neglected by recent entomological writers. I trust, however, that the fol-
lowing description and the accompanying figures may give some idea of the endoskeleton
as it exists in the workers of Lasiusjlaciis.

The hinder plate of the propectus turns upwards at approximately a right angle, and
is produced into the antefurca (PI. XI. figs. 1, 2, 5, & 6; PI. XII. fig. 8), a chitinous pro-
cess which extends more than halfway up the dorsum, leaving, however, a central orifice
(PL XII. fig. 4) through which the nervous chords penetrate, while the oesophagus and
the heart pass between the upper edge of the antefurca and the dorsum.

As seen from behind (PI. XII. fig. 4) it has the form of a cross with four arms. In the
middle of the centre piece is an oval orifice, the wider end below, through which the
nervous system passes. The centre of the upper part sends out a process both anteriorly
and posteriorly, as shown in PL XL fig. 2 ; in fact it forms a sort of case for the protection
of the ganglia.

• ' Introduction to Entomology,' vol. iii. p. 587. t Op. cit.

SIE JOHN LUBBOCK ON THE ANATOMY OF ANTS. 145

The medifurca (PI. XI. fig. 2 ; PL XII. fig. 5, Med) rises from the medipectus. It is
much more elongated and slender than the antcfurca, and has the form of a Y, the upper
arms of which, however, are connected by a cross bar, thus leaving a triangular orifice
with rounded angles, through which runs the nervous chord. To a process of the cross
har is attached the muscle which elevates the prothorax.

The postfurca (PI. XL fig. 2 and PL XII. fig. 6) also has somewhat the form of
a Y. The stem, however, is much shorter, the branches are curved, and the cross bar
is absent. The postfurca arches forwards, so that the upper part of the arms approach
those of the medifurca, with which they are connected by tendinous fibres. Between
the medifurca and the postfurca lies the third thoracic ganglia.

Muscles of the Mead.

There are two elevators of the head on each side (a & a 1 ). The first (PL XL figs.
1, 2, & 5, a) is a thin muscle, which rises from the back near the middle line, at the
junction of the pro- and mesothorax, and, passing forwards, is inserted at the upper margin
of the occipital foramen, where the posterior margin of the head joins the intersegmental
membrane. The second is more powerful. It (PI. XL figs. 1, 2, a 1 ) rises from the anterior
surface of the upper part of the antefurca, and, passing forwards and slightly upwards, is
inserted close to the preceding. The heads of attachment of this muscle reach almost
across the segment.

The first depressor of the head (PI. XL figs. 1, 2, b), like the second elevator, is attached
to the anterior face of the antefurca, but at a lower level, and, passing over the prothoracic
ganglion, is attached to the inferior margin of the occipital foramen.

The second depressor of the head (PL XL fig. 1, PL XII. fig. 1, b x ) is attached to the
central and hinder part of the propectus, and, passing directly forwards, is also attached
co the lower edge of the occipital foramen.

The rotators of the head are five in number on each side. The first (PL XL figs. 1, 2, &

t,c) rises from the middle of the lateral wall of the pronotum, and, passing downwards

ind inwards, is attached to the anterior toothed process of the propectus. The second

'otator passes from the middle of the lateral wall of the propectus (PL XL figs. 4, 6, &7, c l ),

md is attached to the outer anterior toothed process of the prosternum. The third

■otator (PL XL figs. 4, 6, 7, c' 2 ) lies rather nearer the middle of the segment. In front it

s attached to the inner toothed process, and posteriorly to the lateral and posterior

vail of the propectus, a little behind the preceding. The fourth rotator (PL XL

igs. 1, 2, 6, d) commences at the anterior process of the propectus, close to the preceding,

md, passing backwards and slightly inwards, is attached to the anterior central process

>f the antefurca. The fifth rotator (PL XL figs. 1, 2, & 6, d l ), rises with the preceding,

mt passes diagonally across the segment to be attached to the lateral edge of the antefurca.

Although the muscles of the head of Coleoptera, as described by Straus-Durckheim in

Melolontha, and as given generally by Burmeister in his ' Handbook of Entomology,' are

aore complex than those which are found in Ants, yet neither of these authors describe

ny muscle exactly comparable to the following.

This muscle (PL XL figs. 1, 5, e) differs from the preceding in that, while they taper

146 SIE JOHN LUBBOCK ON THE ANATOMY OF ANTS.

as they pass forward, it, on the contrary, rises from the anterior surface of the pronotum
by several, somewhat diverging heads, and, passing backwards and slightly downwards,
is attached to the upper part of the antefurca. It would therefore seem to draw the
propectus, and consequently to push the head, forwards. It is obvious that if the head
is projected forwards, and the propectus then retracted, so that the head could move
freely towards each side, it would be easily turned by the rotators above described. On
the contrary, if it be retracted, or if the propectus be thrown forward, so that the
posterior process of the head interlocks with the anterior processes of the propectus,
the head would be so situated as to retain its position even against a considerable force.

The next muscles to be mentioned are the elevators of the antepectus; these are
two in number. The first (PI. XL figs. 1, 2, 4, 5, and PL XII. fig. 4,/) rises from
near the middle of the pronotum, and, passing downwards, is attached to the anterior
process of the antepectus. The second is weaker; it is attached to the side of the
pronotum, and, passing downwards (PL XL fig. 5,/ 1 ) and slightly inwards, is also
attached to the anterior process of the antepectus, close to the preceding. The attach-
ment of the first large rotator of the head (e) lies between those of these two muscles,
as may be seen in PL XL fig. 5, where f and f l represent the heads of these two
muscles, which, when they contract together, would tend to elevate the antepectus.

The depressor of the antepectus is smaller. It commences (PL XL figs. 5, 6, 7, g) at
the lower edge of the pronotum, and, passing upwards, is attached to the upper edge of
the antepectus, which therefore, on contracting, it draws downwards.

Front Legs and their Muscles.

The legs consist of the following segments : — 1, coxa ; 2, trochanter ; 3, femur ; 4, tibia ;
and 5, tarsus, this latter being composed of five segments.

The description given by Straus-Durckheim of the muscles by which the legs are
moved has been adopted by most subsequent writers. According to him, the anterior
le°'s have five muscles, four flexors and one extensor. The first flexor rises from the
superior lateral and anterior surface of the prothorax, and is attached to the posterior
border of the coxa. The second and third flexors rise from the superior and posterior
surface of the prothorax, and are attached to the coxa just outside the preceding. The
fourth flexor rises from the external portion of the posterior surface of the " rotule,"
and is attached to the posterior edge of the coxa. Lastly, the extensor rises from the
pronotum, near the first flexor, and acts immediately in opposition to the preceding.

The number of muscles in the Ant appears to be greater than in Melolontha, audi
the disposition is in many respects dissimilar.

The first muscle of the leg (PL XL figs. 4, 7, and PL XII. figs. 1, 4, h) rises from the I
anterior lateral wall of the prothorax, and, passing downwards and backwards, is
attached to the upper anterior angle of the condyle of the coxa, which, therefore, it
would tend to draw forwards and inwards.

The second (*, PL XL figs. 1, 2, 4 ; PL XII. figs. 1, 3, & 4) lies transversely in the lower
and posterior portion of the antepectus. In PL XL figs. 1 & 2 it is seen in section. In
PL XL fis?. 4 it is severed close to its attachment. It rises from the central ridge

o

SIK JOHN LUBBOCK ON THE ANATOMY OF ANTS. 147

of the antepectus, and, passing trausversely across the segment, is attached to the
posterior and outer edge of the leg, at the summit of the projecting head or condyle.
It would tend to extend the leg laterally.

The third (i 1 , PL XII. fig. 4) is attached to the antefurca, and, passing downwards
and outwards, is attached close to the preceding.

The fourth and fifth muscles of the fore legs are of a different character, penetrating
into the coxa. The fourth rises from the upper edge of the antepectus in front of the
antefurca (PL XL fig. 4, and PL XII. fig. 4, k), and passes downwards into the coxa.

The fifth rises partly from the hinder wall of the antefurca, partly from its posterior
spur (PL XI. figs. 1, 2, 4, & 7, /), and, like the preceding, passes down into the coxa.
The upper part of the muscle is joined by some fihres, which pass round the posterior
process of the antefurca and are attached to the pronotum.

The seventh is attached to the outer and posterior edge of the coxa, and, passing
backwards and inwards, is attached to the anterior surface of the medifurea. It is not,
however, well shown in any of my sections.

In addition to these muscles, the coxa contains two others, one of which rises from the
upper and outer wall and passes downwards and inwards, while the other, rising from the
upper and inner wall, passes downwards and outwards.

The small trochanter (PL XII. fig. 1, tr.), in addition to the above-mentioned fibres of
the flexor of the femur, contains only a short single muscle, which at its lower end is
ittached to the thigh.

The femur (PL XII. fig. 1, fm) contains two muscles. The extensor is attached to
she upper surface of the segment, the fibres being attached to one side of a long tendon,
svhich at its lower end is attached to a chitinous piece at the upperside of the head of
;he tibia. The flexor is situated rather on the lower side of the segment ; but the fibres
liverge from both sides of the tendon, and some of them cross those of the extensor
nuscles. Some of the central fibres pass into the trochanter, and are attached to its
nner margin. The lower end of the tendon of the flexor is attached to a chitinous
irocess.

The tibia presents some very remarkable points, with reference to which I may perhaps
)e permitted to quote a passage from a paper of mine published in the ' Microscopical
Journal,' 1877.

Remarks on the Tibial Organ.

In the year 1844 Von Sicbold* described a remarkable organ which he had discovered
n the tibiae of the front legs of Gryllus, and which he considered to serve for the purpose
|)f hearing. These organs have been also studied by Burmeister, Brunner, Hensen,
^eydig, and others, and have recently been the subject of a monograph by Dr. V. Graberf,
vho commences his memoir by observing that they are organs of an entirely unique
haracter, and that nothing corresponding to them occurs in any other insects or, indeed,
n any other Arthropods.

* " Ueber das Stinim- und Gehor-Organ der Orthopteren," Wicgmann's Arch. f. Natur. 1S44.
t Die tympanalen Sinnes-Apparate dcr Orthopteren, von Dr. Titus Graber, 1875.
SECOND SERIES. — ZOOLOGY, VOL. II. 20

148 SIR JOHN LUBBOCK ON THE ANATOMY OF ANTS.

I have therefore been very much interested by discovering in Ants a structure which
seems in some remarkable points to resemble that of the Orthoptera. As will be seen
from a glance at Dr. Graber's memoir, and the plates which accompany it, the large
trachea of the leg is considerably swollen in the tibia, and sends off, shortly after entering
the tibia, a branch, which, after running for some time parallel to the principal trunk,
joins it again. See, for instance, in his Monograph, pi. ii. fig. 43, pi. vi. fig. 09, pi. vii.
fig. 77, &c. Now I have observed that in many other insects the tracheae of the tibia
are dilated, sometimes with a recurrent branch. The same is the case even in some mites.

I will, however, reserve what I have to say on this subject, with reference to other
insects, for another occasion, and will at present confine myself to the Ants. If we
examine the tibia, say of Lasius Jlaviis, we shall see that the trachea presents a remark-
able arrangement, which at once reminds us of that which occurs in Gryllus and other
Orthoptera. In the femur it has a diameter of about 3<fo5 °f an mcn > as soon, however,
as it enters the tibia it swells to a diameter of about - 5 -^ of an inch, then contracts again
to -g^o, and then again, at the apical extremity of the tibia, once more expands to -^>q.
Moreover, as in Gryllus, so also in Formica, a small branch rises from the upper sac,
runs almost straight down to the tibia, and falls again into the main trachea just above
the lower sac. The remarkable sacs at the two extremities of the trachea in the tibia
may also be well seen in other transparent species, such, for instance, as Myrmica rugi-
nodis and JPheidole megacephala.

At the place wdiere the upper tracheal sac contracts there is, moreover, a conical
striated organ (x), which is situated at the back of the leg, just at the apical end of the
upper tracheal sac. The broad base lies against the external wall of the leg, and the
fibres converge inwards. In some cases I thought I could perceive indications of bright
rods, but I was never able to make them out very clearly. This also reminds us of a
curious structure which is found in the tibia of Locustidaa, between the trachea, the
nerve, and the outer wall, and which is well shown in some of Dr. Graber's figures.

L b'

Other Organs of the Prothorax.

The anterior pair of spiracles, as already mentioned, lie (PL XL figs. 4, 5, Sp l ) between
the pro- and mesothorax. The tracheal tube immediately behind the spiracle is provided
with a short muscle, as already described in other insects by MM. Landois and Thelen*.
The ganglion (PI. XL figs. 2, 6, & 7, G l ) is of considerable size, and is connected ante-
riorly with that of the head, and posteriorly with that of the mesothorax, by a double
commissure. In the latter case the commissures pass through an orifice in the antefurca,
which thus not only serves as a support to the muscles, but also as a protection to the
nervous system.

The oesophagus passes straight through the prothorax, and, indeed, does not enlarge
into the crop until it reaches the enlarged part of the abdomen. In the upper part of the
prothorax lie the large thoracic salivary glands (PI. XL fig. 2, gl).

A considerable part of the upper and anterior portion of the prothorax is occupied by
the thoracic salivary glands, which I have already described in the ' Microscopical Journal.

* Zeitsehr. f. wiss. Zool. 1867, p. 1S7.

SIR JOHN LUBBOCK ON THE ANATOMY OF ANTS. 149

They consist of a number of branched and twisted tubules which gradually unite in a
single duct. This duct then swells into a capacious receptacle, after which it contracts
again, and after joining the corresponding duct from the other side, passes through the
neck into the head, and then, after a meandering course, opens at the upperside of the
under lip. The duct consists of an epithelial layer of cells, within which is a structureless
membrane, strengthened, as is so often the case with the ducts of glands, by chitinous
ridges, which give it very much the appearance of a trachea. Fig. 3, PI. XI., represents
a glandular organ situated in the lower part of the thorax of Myrmica ruginodis imme-
diately above the base of the anterior leg.

Mesotliorax and Middle Legs.

The mesotliorax is much more closely connected with the metathorax than with the
prothorax (PL XI. fig. 2). Like the prothorax it consists of an upper and lower more or
less arched plate. The upper plate or mesonotum (PI. XL figs. 2, 5, 6, lies) is oblong,
somewhat emarginate behind, the spiracles (PI. XL figs. 2, 4, Sj) 2 ) being situated at the
posterior angles. In front the mesonotum projects some way over the sides of the pro-
thorax ; and as the middle legs are attached quite at the posterior end of the metapectus,
they, as well as the posterior legs, lie under the metanotum, and seem at first sight as if
they belonged to the hinder division of the thorax.

The depressor of the prothorax (PI. XL figs. 2, 4, 5, in) arises from the junction of the
meso- and metathorax, beneath the spiracle, and passing down and forwards is attached
to the lower posterior edge of the prothorax, which therefore it